INFORMATION to USERS the Most Advanced Technology Has Been Used to Photo Graph and Reproduce This Manuscript from the Microfilm Master

INFORMATION TO USERS The most advanced technology has been used to photo graph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are re produced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information Company 300 North Z eeb Road, Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600 Order Number 9011227

Cell culture models for ueural trauma

Murphy, Eric James, Ph.D. The Ohio State University, 1989

UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106 Cell Culture Models for Neural Trauma

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University By Eric James Murphy, BA

*******

The Ohio State University

1989

Dissertation Committee: Approved By

Lloyd A. Horrocks

Gerald P. Brierley

Richard P. Swenson

Adviser Department of Physiological Chemistry Copyright by Eric James Murphy 1989 This work is dedicated to my son Cameron and to my wife Cindy, and to my parents

Roger and Sandy

ii ACKNOWLEDGMENTS

In the course of five years there are many people to thank for helping me in my graduate studies. These people include my coworkers

Marilyn, Marianne, Kim, Carolyn, Allen, Royal, Paula and Paul. I reserve special thanks for Yutaka "Ruskie" Hirashima and Carl Bates.

They have been there over the past year to make the timely completion of this work possible. I also thank Laura Dugan for her instruction in HPLC techniques. I thank Lynda, Barb and Patty for providing large quantities of cell cultures on an almost impossible schedule.

I thank Che Maxwell for her help in the first and last year of my studies.

I would especially like to thank my wife, Cindy and son Cameron for their patience during the past year. Cindy did a great job preparing the manuscript and worked long hours to get the job done.

I thank my parents for their support and for taking care of Cam this summer. I thank my Dad for instilling in me a work ethic that enabled me to continue to work hard in order to complete the task at hand.

Of course I have to thank Dr. Dan Funk for his superb surgical work on my knee and my physical therapists Big Dave and Robyn. I also thank my brother, Brian for his support during my injury and recovery as well as my friends Steve Sutherland, Steve Myers, Mark,

Shirley, and Veronica.

Lastly, I have to thank Dr. Lloyd Horrocks for his support and understanding. He has prepared me well for a career in science and I am sincerely thankful for this preparation. VITA

13 March 1963 Born - Bucyrus, Ohio

1984 BA (Biology and History) Hastings College Hastings, Nebraska

1985-Present Graduate Research Assistant Department of Physiological Chemi stry The Ohio State University Columbus, Ohio

PUBLICATIIONS

E.J. Murphy, L. Joseph, R. Stephens, and L.A. Horrocks (1986) Phospholipid Composition of Human Endothelial Cultures of Vascular Tissue. Presented at American Society for Neurochemistry Meeting, Montreal, Canada. Trans. Am. Soc. Neurochem. 17: B139.

M. Jurkowitz, L. Dugan, E.J. Murphy, M. Waugh, and L.A. Horrocks (1986) Inhibition of Lysoplasmalogenase by 2-acylglycerophospho- ethanolamine. Presented at American Society for Neurochemistry S atellite Symposium on 11 Molecular Mechanisms of Neuronal Responsivity" in Burlington, Vermont.

E.J. Murphy, D.K. Anderson, E.D. Means, and L.A. Horrocks (1987) Pressure Induced Trauma in Neuronal Cell Cultures: A Novel Model for Spinal Cord Injury. J. Neurochem., 48: 550A.

E.J. Murphy, R. Stephens, M.S. Jurkowitz, and L.A. Horrocks (1987) Plasmalogen Quantitation in Cell Culture Lipids by Acid Hydrolysis Followed by High Pressure Liquid Chromatography. J. Neurochem., 48: S60B.

E.J. Murphy, D.K. Anderson, E.D. Means, and L.A. Horrocks (1988) Pressure Induced Trauma in Neuronal Cell Cultures. Presented at American Society for Neurochemistry Meeting, New Orleans, Louisiana, Trans. Am. Soc. Neurochem., 19: BUB.

v E.J. Murphy, E. Roberts, and L.A. Horrocks (1989) Aluminum Silicate Toxicity in Cell Cultures. American Society for Neurochemistry Meeting, Chicago, Illin o is. Trans. Am. Soc. Neurochem., 20: B221.

E.J. Murphy, D.K. Anderson, E.D. Means, and L.A. Horrocks (1989) Pressure Induced Trauma in ROC-1 Oligodendroglial Cell Cultures. American Society for Neurochemistry Meeting, Chicago, Illin o is. Trans. Am. Soc. Neurochem., 20: A141.

E.J. Murphy and L.A. Horrocks (1989) Cell Culture Model for Mechanical Trauma Induced by Pressure. 7th Annual Neurotrauma Meeting, Phoenix, Arizona.

E.J. Murphy and L.A. Horrocks (1989) Pressure Induced Trauma in Cell Cultures. Society for Neuroscience, Phoenix, Arizona.

Y. Hirashima, E.J. Murphy, and L.A. Horrocks (1989) Purification of Plasmalogens with Rhizopus delemar Lipase and Naja naja naja Phospholipase Ap. American Society for Neurochemistry, March 5-10, 1989, Chicago, Illin o is. Trans. Am. Soc. Neurochem., 20: A265.

M. Jurkowitz-Alexander, H. Ebata, J.S. Mills, E.J. Murphy, and L.A. Horrocks (1989) Solubilization, Purification, and Characterization of Lysoplasmalogen alkenylhydrolase (lysoplasmalogenase) from rat liver microsomes. Biochim. Biophys. Acta, 1002: 203-212.

E.J. Murphy and L.A. Horrocks (1989) Mechanisms of Action of CDPcholine and CDPethanolamine on Fatty Acid Release during Ischemia of Brain. In Lipid Mediators in Ischemic Brain Damage and Experimental Epilepsy, N.G. Bazan (Ed.). S. Karger AG, Basel, Switzerland. TABLE OF CONTENTS

page

DEDICATION...... ii ACKNOWLEDGMENTS...... iii

VITA...... v

LIST OF TABLES...... ix

LIST OF FIGURES...... xi

LIST OF PLATES...... xv

CHAPTER

I. INTRODUCTION 1

II. COMPOSITION

Introduction ...... 7 Materials and Methods ...... 12 Results...... 30 Discussion ...... 39

III. ALUMINUM SILICATES

Introduction ...... 48 Materials and Methods ...... 59 Results...... 62 Discussion ...... 154

IV. PRESSURE

Introduction ...... 168 Materials and Methods ...... 180 R esults...... 182 Discussion ...... 231

v ii CHAPTER

V. CONCLUSION...... 241

APPENDICIES

A...... 244

B...... 251

BIBLIOGRAPHY...... 270

v i i i LIST OF TABLES

Table page

1. Long phospholipid separation ...... 21

2. Short phospholipid separation ...... 25

3. Phospholipid composition of cell cultures ...... 32

4. Plasmalogen content of cell cultures ...... 33

5. FA composition of the major phospholipids in primary neurons c e lls ...... 34

6. FA composition of the major phospholipids in N1E-115 c e lls ...... 35

7. FA composition of the major phospholipids in ROC-1 c e lls ...... 36

8. FA composition of the major phospholipids in HUVE c e lls ...... 37

9. Comparison of phospholipid compositions of ROC-1, C-6 glioma, and glial cells ...... 42

10. Clay induced LDH release in ROC-1, N1E-115 and HUVE c e lls ...... 138

11. Clay induced cell death in N1E-115 andHUVE c e lls 139

12. Clay induced FA release in ROC-1 c e lls ...... 245

13. Clay induced FA release in N1E-115 c e lls ...... 247

14. Clay induced FA release in HUVE c e lls ...... 249

15. Pressure induced FA release in ROC-1 cells 1 min recovery...... 252

16. Pressure induced FA release in ROC-1 cells 10 min recovery ...... 253

ix Table page

17. Pressure induced LDH release in ROC-1 cells 1 min recovery ...... 254

18. Pressure induced LDH release in ROC-1 cells 10 min recovery ...... 255

19. Pressure induced FA release in ROC-1 cells function of recovery time ...... 256

20. Pressure induced LDH release in ROC-1 cells function of recovery time ...... 257

21. Pressure induced FA release in N1E-115 cells 1 min recovery ...... 258

22. Pressure induced FA release in N1E-115 cells 10 min recovery ...... 259

23. Pressure induced LDH release in N1E-115 cells 1 min recovery ...... 260

24. Pressure induced LDH release in N1E-115 cells 10 min recovery ...... 261

25. Pressure induced FA release in N1E-115 cells function of recovery time ...... 262

26. Pressure induced LDH release in N1E-115 cells function of recovery time ...... 263

27. Pressure induced FA release in HUVE cells 1 min recovery ...... 264

28. Pressure induced FA release in HUVE cells 10 min recovery ...... 265

29. Pressure induced LDH release in HUVE cells 1 min recovery ...... 266

30. Pressure induced LDH release in HUVE cells 10 min recovery ...... 267

31. Pressure induced FA release in HUVE cells function of recovery time ...... 268

32. Pressure induced LDH release in HUVE cells function of recovery time ...... 269

x LIST OF FIGURES

Figure page

1. TLC separation of lipids ...... 18

2. HPLC long separation of phospholipids ...... 23

3. HPLC short separation of phospholipids ...... 27

4. Molecular structures of phospholipid subclasses 28

5. FA release in ROC-1 cells by bentonite ...... 73

6. FA release in ROC-1 cells by kaolinite ...... 75

7. FA release in ROC-1 cells by erionite ...... 77

8. FA release in ROC-1 cells by montmorilIonite ...... 79

9. FA release in ROC-1 cells Comparison of clay type 1 h incubation ...... 81

10. FA release in ROC-1 cells Comparison of clay type 6 h incubation ...... 83

11. FA release in ROC-1 cells Comparison of clay type 24 h incubation ...... 85

12. FA release in N1E-115 cells by bentonite ...... 87

13. FA release in N1E-115 cells by kaolinite ...... 89

14. FA release in N1E-115 cells by erionite ...... 91

15. FA release in N1E-115 cells by montmorilIonite 93

16. FA release in N1E-115 cells Comparison of clay type 1 h incubation ...... 95

17. FA release in N1E-115 cells Comparison of clay type 6 h incubation ...... 97

xi Figure page

18. FA release in N1E-115 cells Comparison of clay type 24 h incubation ...... 99

19. FA release in HUVE cells by bentonite ...... 101

20. FA release in HUVE cells by kaolinite ...... 103

21. FA release in HUVE cells by erionite ...... 105

22. FA release in HUVE cells by montmoril Ionite ...... 107

23. FA release in HUVE cells Comparison of clay type 1 h incubation...... 109

24. FA release in HUVE cells Comparison of clay type 6 h incubation ...... Ill

25. FA release in HUVE cells Comparison of clay type 24 h incubation ...... 113

26. Bentonite toxicity in cells Comparison of cell type 1 h incubation ...... 115

27. Bentonite toxicity in cells Comparison of cell type 6 h incubation ...... 117

28. Bentonite toxicity in cells Comparison of cell type 24 h incubation ...... 119

29. Kaolinite toxicity in cells Comparison of cell type 1 h incubation ...... 121

30. Kaolinite toxicity in cells Comparison of cell type 6 h incflbation ...... 123

31. Kaolinite toxicity in cells Comparison of cell type 24 h incubation ...... 125

32. Erionite toxicity in cells Comparison of cell type 1 h incubation ...... 127

33. Erionite toxicity in cells Comparison of cell type 6 h incubation ...... 129

34. Erionite toxicity in cells Comparison of cell type 24 h incubation ...... 131

x ii Figure page

35. MontmonlIonite toxicity in cells Comparison of cell type 1 h incubation ...... 133

36. MontmorilIonite toxicity in cells Comparison of cell type 6 h incubation ...... 135

37. Montmorillonite toxicity in cells Comparison of cell type 24 h incubation ...... 137

38. Pressure induced FA release in ROC-1 cells 1 min duration ...... 190

39. Pressure induced FA release in ROC-1 cells 3 min duration ...... 192

40. Pressure induced FA release in ROC-1 cells 5 min duration ...... 194

41. Pressure induced FA release in ROC-1 cells 10 min duration ...... 196

42. Composite of pressure induced FA release in ROC-1 cells 10 min recovery ...... 198

43. Composite of pressure induced FA release in ROC-1 cells 1 min recovery ...... 200

44. Pressure induced FA release in R0C-1 cells at various recovery times ...... 202

45. Pressure induced FA release in N1E-115 cells 1 min duration ...... 204

46. Pressure induced FA release in N1E-115 cells 3 min duration ...... 206

47. Pressure induced FA release in N1E-115 cells 5 min duration ...... 208

48. Pressure induced FA release in N1E-115 cells 10 min duration ...... 210

49. Composite of pressure induced FA release in N1E-115 cells 10 min recovery ...... 212

50. Composite of pressure induced FA release in N1E-115 cells 1 min recovery ...... 214

x ii i Figure page

51. Pressure induced FA release in N1E-115 cells at various recovery times 3 min duration 15 atm pressure 216

52. Pressure induced FA release in HUVE cells 1 min duration ...... 218

53. Pressure induced FA release in HUVE cells 3 min duration ...... 220

54. Pressure induced FA release in HUVE cells 5 min duration ...... 222

55. Pressure induced FA release in HUVE cells 10 min duration ...... 224

56. Composite of pressure induced FA release in HUVE cells 1 min recovery ...... 226

57. Composite of pressure induced FA release in HUVE cells 10 min recovery...... 228

58. Pressure induced FA release in HUVE cells at various recovery times 10 minduration 15 atm pressure 230 LIST OF PLATES

PI ate page

I. ROC-1 cells incubated with montmoril Ionite ...... 141

II. ROC-1 cells incubated with buffer ...... 143

III. HUVE cells incubated with bentonite ...... 145

IV. HUVE cells incubated with kaolinite ...... 147

V. HUVE cells incubated with erionite ...... 149

VI. HUVE cells incubated with montmoril Ionite ...... 151

VII. HUVE cells incubated with buffer ...... 153

xv CHAPTER I

INTRODUCTION

Today's political environment for scientists is treacherous. As

the animal rights movement gains momentum, scientists have been put

to the test defending the basic premise of science, the scientific

method. The quest for knowledge results in hypotheses, experi

mentation and the generation of data which are analyzed and

conclusions drawn. Scientists have utilized animals for the purpose

of fulfilling the need to answer these complex questions. Yet, today

there is mounting pressure to develop new ways to test our

hypotheses. This pressure has resulted in the dramatic increase in

the use of cell cultures for experimentation. However, the question

arises, do cells in vitro represent the situation in vivo? While

agonist-receptor interactions are easily studied, more complex

questions involving many disease or pathophysiological states are

not.

Cell cultures are effective to use in studying numerous

biochemical problems. These uses range from receptor studies to trauma models. Cells in culture are grown in a chemically defined medium. This medium can be easily manipulated to meet the desired experimental conditions. Toxicological and pharmacological studies can be performed using a large number of cells. This permits rapid screening in numerous cell types. Studies in vivo do not permit the

flexibility that cell cultures do.

The development of an in vitro model for spinal cord injury is

essential to understand the pathophysiology. Cells comprising the

spinal cord can be examined individually to determine their reaction

to mechanical trauma and hypoxia. Pharmacological agents can be

tested which inhibit or decrease the undesirable effects of trauma

and enhance the desirable effects of trauma. A model has been

developed over the past several years which utilizes an increase in

atmospheric pressure to cause a mechanical trauma in cell cultures mimicing that seen in vivo.

Cell cultures can be used to examine thecellular toxicology of

various compounds. Aluminum silic ate containing clays are an

environmental toxin which are extensively used in industry and

consumer goods. Evidence indicates aluminum silicates may play a

role in the etiology of Alzheimer's disease. Cell cultures were used to examine the possible toxicology of aluminum silicate containing clays to cells of neuronal origin.

Cell culture models are being reported in the literature with an

increasing frequency. A model for cerebral ischemia and the resulting postischemic edema has been established using C-6 glioma cells (Jakubovicz et a l., 1987). Cellular hypoxia, an important component in myocardial infarction, spinal cord injury and cerebral

ischemia, has been simulated in astrocytes by exchanging the normal cell incubation atmosphere of 95% air, 5% C02 with 95% N2, 5%

C02 (Yu et a l., 1989). Cell death begins to occur after 18 h of hypoxia and is nearly complete after 24 h. Deenergization has been produced using potassium cyanide to block oxidative phosphorylation

(Huang and Gibson, 1989). This model results in an increase in intracellular Ca^+ and an alteration of polyphosphoinositide metabolism which is time dependent. Lipid peroxidation has been produced in vitro using FeCl2 (Means et a l., 1986). This resulted in only minor changes in neurons after 16 h unless polymorphonuclear leukocytes (PMNL) were included in the culture. The addition of PMNL caused the onset of cellular necrosis by 8 h and the demise of the neurons after 16 h. Spinal cord injury has been simulated in vitro using fetal mouse spinal cord explants grown in culture and trauma induced by dropping the blunt end of a dressmaker pin onto a cell

(Balentine et al., 1988). Morphologically the cells showed signs of trauma comparable to those seen in vivo.

Shear stress has been studied using endothelial cultures. Shear stress induces an increase in prostacyclin synthesis (Bhagyalakshmi and Frangos, 1989). These levels may represent more closely those found in vivo because of the constant mechanical stress on the cells in vivo. Shear stress causes platelet aggregation (Heliums et al.,

1987). Agents which increase cAMP levels decrease this aggregation but also increased platelet lysis. Fluid mechanical stress increases prostaglandin and leukotriene synthesis in platelets, PMNL and human umbilical vein endothelial cells (Mclntire et a l., 1987). Cells have been cocultured to more closely simulate the in vivo state which includes cell-cell interactions (Johns et al., 1988; Rudge et al.,

1989). Cell cultures are becoming increasingly used to simulate in vivo environments and stresses to study the reaction of cells to these situations. All of these present a new method to examine complex problems. One advantage of all cell culture models is the ease of manipulating cell conditions.

Fatty acid release is one index of cell trauma. During cerebral ischemia fatty acids and diacylglycerols are released from several different phospholipids (Ideda et al., 1986; Horrocks et al., 1984;

Yoshida et a l., 1980, 1986). Following compression trauma to the spinal cord, fatty acids and diacylglycerols are released (Demediuk et a l., 1985b,d). During fluid mechanical stress, arachidonate metabolism through the cyclooxygenase and lipoxygenase pathways increases indicating increased substrate availability (Mclntire et a l., 1987). Pressure induced trauma causes an increase fatty acid release by cultured cells (Murphy et a l., 1987, 1988, 1989a).

Lactate dehydrogenase, a ubiquitous cytosolic enzyme, is another index of cell trauma (Yu et a l., 1989; Koh and Choi, 1987). Release of lactate dehydrogenase by the cell indicates a compromised membrane which has lost its integrity and is unable to maintain the intracellular environment.

Free fatty acids and diacylglycerols have a detrimental effect on cellular metabolism. Lysophospholipids, formed by the deacylation of phospholipids, exert detergent-like effects on cell membranes

(D'Amato et al., 1975; Weltzien et al., 1979a,b). These detergent effects cause an increase in fluidity by increasing the intermolecular space between membrane components (Lee, 1983). Fatty acids and diacylglycerols intercalate into the membrane further increasing the fluidization of the membrane (Michel 1 et al., 1975).

Both free fatty acids and diacylglycerols inhibit ATPases, primarily

Na+,K+-ATPase and Mg2+-ATPase, during cerebral ischemia and spinal cord injury (Clendenon et a l., 1985; Goldberg et a l., 1985;

Rhoads et a l., 1982). Polyunsaturated fatty acids induce brain edema

(Chan and Fishman, 1978), cause neurotransmitter amino acid release from synaptosomes (Rhoads et a l., 1983; Drejer et a l., 1985), and inhibit reuptake of released neurotransmitter amino acids (Rhoads et al., 1982; Chan et al., 1983). Polyunsaturated fatty acids also cause an induction of superoxide radical formation in primary astrocytes (Chan et a l., 1988). Fatty acids uncouple oxidative phosphorylation in mitochondria reducing the overall energy charge in the cell (Majewska et a l., 1978; Hillered and Chan, 1988). Thus, fatty acids and diacylglycerols released following cell trauma have a detrimental effect on cell metabolism as well as an increased fluidization of the membrane.

Increasing the membrane fluidity causes an increase in permeability of the cell to Ca2+ (Boonstra et a l., 1982). As cellular integrity is lost, an unregulated influx of Ca2+ occurs.

Ca2+ has been implicated in cell death (Piper, 1988; Starke et al.,

1986; Trump and Berezesky, 1983). As intracellular levels of Ca2+ begin to increase, there is an increased rate of phospholipid hydrolysis (Shier and DuBourdieu, 1982). This increased hydrolysis further exacerbates the cell's situation. An increase in Ca2+ may also disrupt the cytoskeleton which results in membrane blebbing

(Laiho et a l., 1983). Ca2+ causes the extensive degradation of neurofilament and microtubule proteins (Banik et a l., 1987). This would cause the disruption of the cytoskeleton aiding the loss of membrane integrity and eventual cell lysis. Furthermore, as integral membrane proteins are degraded, there is an increase in molecular motion in the bulk membrane lipid component corresponding to an increase in fluidity (Meier et a l., 1987).

In summary, it is essential for scientists to begin to use cell cultures to examine their hypotheses. In many areas of biomedical research this has already begun to occur. Cell cultures are useful tools in studying the effects of various pharmacological, toxins, agents, and traumas. Fatty acid release is a good indicator of membrane perturbation and damage. This damage increases membrane flu idity and permeabilizes the membrane. As the membrane is compromised, intracellular Ca^+ concentrations increase causing inevitable cell death. As a result of membrane breakdown and cell death, lactate dehydrogenase (LDH) is released, thereby making LDH a good marker for cell death. CHAPTER II

COMPOSITION

Introduction

Phospholipids play a major role in maintaining cell membrane structure and integrity. The perception of the importance of phospholipids has changed as their role in receptor-mediated events becomes evident. However, the basic phospholipid composition for many of the cells currently used in cell culture is unknown. Thus, it was important to determine the phospholipid composition for primary murine neuronal cells, human umbilical vein endothelial cells, ROC-1 oligodendroglia cells and N1E-115 neuroblastoma cells.

The fatty acid compositions of the major phospholipid classes and cholesterol levels were also determined for each cell type.

Endothelial cells are no longer regarded as a metabolically passive cell whose only function is to line vascular tissue. It has been known that endothelial cells produce prostacyclin, a potent platelet anti-aggregant (Weksler et al., 1977). However, the endothelium possesses numerous receptors which regulate the formation of various metabolites (Hammerson and Hammerson, 1985). Bradykinin stimulates increased prostacyclin (Clark et al., 1986; Crutchley et al., 1983; Hong and Deykin, 1982) and thromboxane (Selivonchick and

7 Roots, 1976) production. Bradykinin stimulates polyphosphoinositide hydrolysis and increases prostacyclin synthesis through binding to the bradykinin 2 receptor (Derian and Moskowitz, 1985, 1986).

Histamine induces polyphosphoinositide metabolism in HUVE cells resulting in increased thromboxane and prostacyclin synthesis (Resink et a l., 1987). The histamine H-l receptor is linked to histamine-stimulated inositol phosphate accumulation in the HUVE cell

(Lo and Fan, 1987). The decrease in polyphosphoinositides linked to the stimulation of histamine receptors produces diacylglycerols which stimulate an increase in protein kinase C activity (Halldorsson et a l ., 1988).

Phospholipids have several other roles in endothelial cells.

Endothelial cells produce platelet-activating factor through a Ca^+ influx or a protein kinase C linked mechanism which causes the hydrolysis of the acyl group from the s/7-2 position of l-0-alkyl-2- acyl-s/?-glycero-3-phosphocholine (Whatley et al., 1989).

Phosphatidylcholine is a major contributor of arachidonic acid following thrombin stimulation of endothelial cells (Thomas et al.,

1984). The calcium ionophore A 23187 also stimulates the release of arachidonic acid from phosphatidylcholine (Martin and Wysolmerski,

1987). These releases of arachidonic acid provide substrate for the synthesis of prostacyclin, thromboxane and hydroxyeicosatetraenoic acids (Buchanan et a l., 1985; Gorman et a l., 1985; Yamaja Setty et a l., 1985; Ingerman-Wojenski et a l., 1981; Nawroth et a l., 1985;

Marcus et a l., 1978). All of these arachidonic acid metabolites may have an important role in the interactions between platelets and the endothelium (Moncada and Vane, 1979). Thus, phospholipids have an

integral role in signal transduction in endothelial cells as well as

being a source of arachidonic acid for eicosanoid synthesis (Schror,

1985).

ROC-1 oligodendroglia cells are a cell hybrid between a C-6 rat

glioma and a calf oligodendrocyte. The morphological and biochemical

characteristics of the ROC-1 closely resemble those of primary

oligodendrocyte (McMorris et a l., 1981). One inherent problem with

growing primary oligodendroglial cells in cell culture is the

difficult procedure in isolating the cells from tissue (Pleasure et

al., 1977; Gonatas et al., 1982). Once isolated the cells are

difficult to grow in serial culture in large quantities (McCarthy and deVellis, 1980). One major advantage of the ROC-1 cell is its

ability to be serially cultured and grown in large quantities (McMorris et a l., 1981).

In vivo the oligodendroglia are responsible for the synthesis and processing of lipids prior to their deposition into myelin membranes

(Deshmukh et a l., 1988; Polak and Szuchet, 1988). The cell also acts as a support cell for the neurons providing nutrients and maintaining the myelin sheath (Deshmukh et a l., 1988). No phospholipid composition of ROC-1 cells has been reported in the literature.

However, the partial phospholipid composition of glial cells has been reported (Witter and Debuch, 1982; Deshmukh et a l., 1988) as well as the composition of the C-6 glioma cells (Robert et a l., 1983). Thus, this is the first report on the lipid composition of the ROC-1 cells. Neuroblastoma cells have been used extensively in studying receptor function. Neurotensin stimulates polyphosphoinositide hydrolysis through a G protein linked receptor in N1E-115 neuro blastoma cells (Amar et a l., 1987). Thrombin causes a marked increase in cGMP in N1E-115 cells (Snider and Richelson, 1983). This increase appears to be linked to an increase in phospholipase A2 activity and production of lipoxygenase products. The lipoxygenase products stimulate the cGMP increase following thrombin stimulation and act as an intracellular messenger (Snider and Richelson, 1983). Muscarinic acetylcholine receptors also mediate cGMP synthesis in N1E-115 cells

(El-Fakahany and Richelson, 1980). Bradykinin causes an increase in cGMP. This increase is Ca^+ dependent and appears to be linked to a phospholipase A2 with subsequent formation of lipoxygenase products (Snider and Richelson, 1984). Like the histamine Hj and muscarinic receptors, the bradykinin receptor may be linked to a phospholipase A2 which produces arachidonic acid for the synthesis of lipoxygenase products. These products mediate the levels of cGMP

(Snider et a l., 1984). Thus, many of the receptors studied in

N1E-115 neuroblastoma cells act through a phospholipase A2 to release substrate for lipoxygenase pathway. The lipoxygenase pathway metabolites mediate cGMP increases in the cells.

Primary neuronal cells, like primary oligodendroglia cells, are difficult to grow in culture. Several protocols have been established for growing primary neuronal cells in culture (Mersel et al., 1987; Ransom et al., 1977). The cultures are difficult to produce in large quantities and take 18-21 days to reach confluency (Demediuk et a l., 1985c). This does not permit th eir use in projects which demand large quantities of cells over short periods of time.

Therefore, like ROC-1 cells, the N1E-115 neuroblastoma cells are used to replace primary neuronal cells due to their ab ility to be grown in large quantities and serially.

The phospholipid composition of N1E-115 and primary neuronal cells has not been reported, so this is the first report of these compositions.

Phospholipids have several roles within the cellular membrane.

After receptor stimulation by various agonists, the phosphatidyl- inositol 4,5-6/sphosphate is hydrolyzed by a phospholipase C. The resulting diacylglycerol and inositol trisphosphate act as second messengers. Diacylglycerols stimulate protein kinase C and inositol trisphosphate releases intracellular Ca^+ (Sekar and Hokin, 1986;

Nishizuka, 1984; Majerus et a l., 1985). Recently phosphatidylcholine has been postulated to act as a signal transducer (Pelech and Vance,

1989; Loffelholz, 1989). Like polyphosphoinositide hydrolysis, a diacyl glycerol is formed through a phospholipase C or D mechanism which is activated upon receptor stimulation. Unlike the polyphos phoinositides, there is no increase in intracellular Ca^+ levels.

However, the diacylglycerol produced does stimulate protein kinase C.

The hydrolysis of choline plasmalogens has been implicated in several receptor responses which result in the release of arachidonic acid (Horrocks et a l., 1986b). Choline plasmalogens may play a role in bradykinin receptor activation and subsequent signal transduction

(Horrocks et a l., 1986a). The choline plasmalogens have also been 12 implicated in regulating protein kinase C following a phospholipase C hydrolysis of the choline moiety. The resulting 1-alkenyl-2-acyl -sn- glycerol may inhibit protein kinase C activity (Murphy and Horrocks,

1989b).

Phosphatidyl inositol also can act as a protein anchor (Phelps et al., 1988; Low et al., 1986). Phosphatidyl inositol binds to a protein thereby reducing its lateral motion within the membrane.

Membrane proteins often exhibit a need for specific phospholipids bound either covalently or by hydrophobic interactions for optimal activity (Sanderman, 1978). Thus, phospholipids have a dynamic role in cellular metabolism and are no longer thought of as merely molecules which form the selectively permeable lipid bilayer.

The purpose was to determine the basal cholesterol, phospholipid and phospholipid FA composition in neuronal, N1E-115, ROC-1, and HUVE cell cultures.

Materials and Methods

Human Endothelial Cells

Cells were supplied by Dr. Ralph Stephens (Department of

Pathology and Ohio State University Cell Culture Service, Columbus,

Ohio). Cells were isolated from human umbilical veins by treating the luminal surface with 0.1% type II collagenase (Gimbrone et al.,

1974). The cells were plated upon human fibronectin coated glass tissue culture flasks (23 cm^) containing M-199 media (Gibco) supplemented with 10% fetal calf serum, 150 /ig/ml endothelial cell growth factor (Maciag et a l., 1979), 90 /ig/ml Na-heparin (Thornton et a l., 1983), and antibiotics. The endothelial cells were incubated in a 95% humidified air, 5% C0 2 atmosphere at 36°C. When the cells in a flask reached confluency the media was removed and the cells rinsed with PBS to remove any detached cells. Trypsin was added to remove attached cells which were washed with PBS buffer, sp lit and plated on

100 mm glass cell culture plates coated with human fibronectin.

Cells were used from passages 5 through 12. Endothelial origin was confirmed by the presence of factor VIII antigen (Jaffe et a l.,

1973).

N1E-115 neuroblastoma Cells

These neuroblastoma cells were derived from murine neuroblasts and are known to exhibit many properties of normal differentiated neuronal cells (Richelson, 1979). Cells were obtained from

Dr. E llio tt Richelson (Department of Psychiatry and Pharmacology,

Mayo Foundation, Rochester, MN). The cells were initially plated in

T-75 cm2 plastic tissue culture flasks in Dulbecco modified Eagle's medium high glucose (Gibco, Grand Island, NY) with 10% fetal calf serum, 10 mM HEPES and Penn/strep antibiotics and Fungizone. The cells were incubated in a 90% humidified air, 10% C0 2 atmosphere at a constant temperature of 37°C. Upon reaching confluency (70-85%) the medium was removed and the cells washed with Puck's buffer.

Puck's buffer was added to the cells and the flask was placed in the incubator for 10 minutes or until the cells detached. The cell suspension was split and replated on 100 mm plates. Cells were used for experiments upon reaching confluency. Passages 13 through 30 were used. 14

Murine Neuronal Cells

Primary neuronal cells were derived from the ventral portion of

11-13 day old mouse embryos (Ranson et a l., 1977). The dorsal root ganglion and meninges were removed and ventral portions of 20 spinal cords were minced in 100 ml of media. The minced cords were transferred to 323 mM HEPES buffer containing 0.25% trypsin and incubated in 90% humidified air, 10% CO 2 atmosphere at a constant temperature of 35°C. After 30 minutes the minced cords were transferred to a medium comprised of 78% modified Eagle's medium

(Gibco, Grand Island, NY), 10% heat inactivated horse serum, 10% fetal calf serum, 1% L-glutamine and 1% DNase. The cells were plated onto collagen-coated glass 100 mm tissue culture plates. After two days in culture, the medium was changed to 93% modified Eagle's medium containing 5% horse serum, 1% 1-glutamine and 1% azide supplement. The azide supplement contained Dulbecco's phosphate buffered saline (PBS), NaN 0 (10 /fg/ml), transferrin T (100 mg/ml), progesterone (40 nM), putrescine (200 /zM), Na selenite (60 nM), cortocosterone (2 /zg/ml), triiodothyronine (20 ng/ml) and bovine serum albumin (BSA 200 /zg/ml). The cells were fed 2 to 3 time per week by removing half of the medium and replacing i t with fresh medium. After 18 to 21 days in culture, the cells were ready for use (Demediuk et a l., 1985a).

ROC-1 Oligodendroglia Cells

ROC-1 cells were formed through the hybridization of a rat C -6 glioma and calf brain oligodendrocytes (McMorris et a l., 1981). The cells proliferate well in culture and exhibit biochemical characteristics of both C-6 glioma cells and normal differentiated oligodendrocytes (McMorris et a l., 1981). Cells were obtained from

F.A. McMorris (Wistar Institute of Anatomy and Biology, Philadelphia,

PA). The cells were maintained in modified Eagle's medium supple mented with 10% fetal calf serum, and HAT (containing 100 mM hypoxanthine, 0.4 lift aminopterin and 16 /xM thymidine). The cells were fed by removing half of the medium and replacing i t with fresh medium. The cells were maintained in a 95% humidified a ir, 5% C0 2 atmosphere at a constant temperature of 37°C. Upon reaching confluency, the medium was removed and cells washed with PBS to remove any detached cells. Trypsin was added to remove the attached cells. The cell suspensions were split and plated upon 100 mm plastic culture plates. Upon reaching confluency the cells are used for experimentation.

Lipid Extraction

Confluent cells were extracted using n-hexane: 2-propanol (3:2 v/v) (Hara and Radin, 1978). For cells grown on glass culture plates the cellular lipids were extracted using n-hexane: 2-propanol (3:2 v/v) added directly to the plates. For cells plated upon polystyrene culture plates a modified extraction method was used. Only

F.A. McMorris (Wistar Institute of Anatomy and Biology, Philadelphia,

PA). The cells were maintained in modified Eagle's medium supple mented with 10% fetal calf serum, and HAT (containing 100 /iM hypoxanthine, 0.4 mM aminopterin and 16 jiM thymidine). The cells were fed by removing half of the medium and replacing i t with fresh medium. The cells were maintained in a 95% humidified a ir, 5% CO£ atmosphere at a constant temperature of 37°C. Upon reaching confluency, the medium was removed and cells washed with PBS to remove any detached cells. Trypsin was added to remove the attached cells. The cell suspensions were split and plated upon 100 mm plastic culture plates. Upon reaching confluency the cells are used for experimentation.

Lipid Extraction

2-propanol was added directly to the plates. Prior to extraction the medium was removed and the cells washed with two 3 ml portions of cold PBS buffer to remove traces of medium. The plates were immediately placed upon dry ice and frozen to minimize damage during extraction and cell removal (Demediuk et a l., 1985a). Two 3 ml aliquots of n-hexane: 2-propanol (3:2 v/v) for glass plates or two 2 ml aliquots of 2-propanol for plastic plates were used to extract the lipids from the cells. The initial aliquot was added to the frozen cells which were subsequently removed from the plate by using a Teflon cell scraper. The second aliquot was used to rinse the plate. For cells extracted with the 2-propanol system, the aliquots were added to tubes containing 6 ml of n-hexane to yield a final ratio of n-hexane: 2-propanol (3:2 v/v). Major cellular debris was removed by filtration through glass wool or by centrifugation in a table top centrifuge. Lipid extracts were stored in a large amount of solvent under a nitrogen atmosphere to lim it autooxidation. These extracts were maintained at -20°C for up to two months with no effect on phospholipid composition.

Thin Layer Chromatography

Lipid extracts were dried under a stream of nitrogen and the tubes rinsed with 1-2 ml of chloroform. The chloroform was evaporated and approximately 100 p1 of chloroform was added prior to spotting on thin layer chromatography (TLC) plates.

Silica G plates (Analtech) were activated overnight at 110°C.

Three samples were spotted per plate using oleic acid as a standard in the fourth lane. The 100 jxl sample was spotted on the plate and two 100 p] rinses were subsequently spotted by micropipette. The neutral lipids were separated using a solvent system containing petroleum ether: diethyl ether: acetic acid (110:90:3.8 by volume)

(Demediuk et a l., 1985a). This solvent system separated in ascending order: phospholipids (origin), cholesterol, diacylglycerols, free fatty acids and triacylglycerols (Figure 1). The various lipid bands 17 were visualized after spraying with 1 mM 6-p-toluidino- 2-naphthalenesulfonic acid (TNS) in 50 mM Tris HC1 (pH 7.4) (Jones et al., 1982).

Ultraviolet light was used to cause the TNS to fluoresce. The desired bands are marked and scraped into test tubes.

Diacylglycerol and triacylglycerol bands were scraped into 16 x

125 mm te st tubes, 2 ml methanol added and stored under nitrogen.

The bands with free fatty acids were scraped into screw top test tubes and 2 ml of toluene: methanol (1:1 v/v) added. The samples were placed under nitrogen and capped. Cholesterol was stored in 1 ml ethanol. All samples were stored at -20°C until use.

Fatty Acid Methyl Esters

Free fatty acids are converted to methyl esters in order to increase the v o latility of the long chain fatty acids. Toluene: methanol: sulfuric acid (2 ml, 100:100:4 by volume) was added to the free fatty acids stored in toluene: methanol (1:1 v/v). This produced a final sulfuric acid concentration of 0.38 M. Nitrogen was placed over the reaction mixture, and the tube was tightly capped and vortexed. The tubes were incubated in a Dubnoff shaking water bath for 4 hours at 65°C (Akesson et a l., 1970). The tubes were removed and permitted to cool to room temperature. Water (1 ml) was added and the tube vortexed to stop the reaction. Petroleum ether (3 ml) was added and the tube vortexed to extract the fatty acid methyl esters (FAME). The upper phase containing FAME was pipetted into a test tube, covered with nitrogen and stored at -20°C until gas liquid chromatography ( 6LC) analysis. 18

Solvent Front ^ > ’ Triacyl glycerol

r— ------/ - — } 11 Free Fatty Acid C— ——■—^

1,3 Diacyl glycerol /---______---> {____ ^—-----

1,2 Diacyl glycerol 11 Cholesterol

Phospholipid (origin)

Figure 1 Thin layer chromatography separation of neutral lipids. Solvents were petroleum ether: diethyl ether: acetic acid (110:90:3.8 by volume). Silica Gel G plates from Analtech were activated overnight at 110°C. Bands were visualized with TNS under uv light. 19

The phospholipid acyl chains were converted to FAME through a base catalyzed transesterification reaction. The sample was placed into a test tube and dried under nitrogen. Methanol (1 ml) was placed into the tube and the reaction was initiated by the addition of 1 ml of 1M methanolic KOH at room temperature (Brockerhoff,

1975). The sample was mixed. After 2 minutes the reaction was terminated by the addition of 400 p1 of ethyl formate. FAME were extracted with 2 ml of hexane. The upper phase was pipetted off and the remaining MeOH washed with 2 ml hexane which was then removed.

The FAME were stored in a nitrogen atmosphere at -20°C until GLC analysis.

High Performance Liquid Chromatography

Prior to phospholipid separation, the cell extracts were filtered using a Rainin 0.2 pm Nylon f ilte r . The sample was dried under nitrogen and redissolved in a known volume of HPLC grade n-hexane:

2-propanol: water (3:2 + 5.5% by volume) prior to chromatography.

The cell extracts were separated into major phospholipid classes by HPLC. Solvents used were HPLC grade n-hexane and 2-propanol from

E.M. Science (Cherry H ill, NJ). Solvents were filtered through a

0.5 pm Mi Hi pore FH-type Nylon f ilte r and degassed. Solvent A was n-hexane: 2-propanol (3:2 v/v) and solvent B was n-hexane:

2-propanol: water (3:2 + 5.5% by volume). Water was purified using a

Millipore water purification system. The HPLC system consisted of two Altex 100A dual piston pumps, an Altex 420/421 controller and an

Altex 210 injection port. The Dupont Zorbax Silica column (4.6 mm x

250 mm 5-6 pm silica) was maintained at a constant temperature of 20

34°C with a Jones Chromatography heating block (Columbus, OH). An

Isco V4 UV variable wavelength detector was used to detect peaks at

205 nm.

The chromatographic procedure permitted the separation of all major phospholipids including the separation of the acidic phospho lipids phosphatidyl serine (PtdSer) and phosphatidyl inositol (Ptdlns) as well as the resolution of lysophosphatidylethanolamine

(lysoPtdEtn) (Dugan et al., 1986a). The polyphosphoinositides were also separated (Dugan et al., 1986b). Retention times for each phospholipid class were relative to the polarity. The elution order was from least polar to most polar (Figure 2). The HPLC program for the long phospholipids separation is found in Table 1. The lower lim it of sensitivity for optimal separation was 100 nmol of injected lipid phosphorus.

Plasmalogens were separated by HPLC after mild acid hydrolysis

(Murphy et al., 1987) (Figure 3). The ethanol amine glycerophospholipid and choline glycerophospholipids were fractionated by HPLC.

The collected peaks were dried in a tube under nitrogen and the tubes inverted over 5 drops of concentrated HC1 in a test tube cap for 5 minutes. This caused the hydrolysis of the alkenyl ether bond of the plasmalogens while the alkyl ether, acyl ester and acyl ester, acyl ester fractions remained intact (Figure 4). The HC1 vapors were removed with nitrogen and the sample was re-extracted in n-hexane:

2-propanol (3:2 v/v) and re-chromatographed using the same HPLC solvent system as previously described, with a different HPLC program (Table 2). 21

Table 1. Long phospholipid separation used to separate phospholipids

Time Function Value Duration

0.00 min. FIowrate 1.8 ml/min 0.00 min. 0.00 Chart speed 0.4 cm/min 0.00 0.00 % Solvent B 45% 0.00 40.00 % Solvent B 66% 4.00 47.00 % Solvent B 76% 4.00 65.00 % Solvent B 100% 4.00 75.00 Flowrate 2.5 ml/min 0.70 110.00 % Solvent B 45% 2.00 110.00 Chart speed 0 cm/min 0.00 110.00 FIowrate 1.8 ml/min 0.70

Initial conditions are: column temperature 35°C, flowrate 1.8 ml/min and solvent B 45%. The injector has a 20 n1 loop. The column was a 5 fim Dupont Zorbax Silica (4.6 mm x 25 cm) column. Solvent A is n-hexane: 2-propanol (3:2 v/v). Solvent B is n-hexane: 2-propanol: water (3:2 + 5.5% by volume). 22

Figure 2 HPLC separation of phospholipids as described in

Materials and Methods and Table 1. Abbreviations are:

NL, neutral lipids; CL, cardiolipin; PE, ethanolamine

glycerophospholipid; PG, phosphatidylglycerol; PA,

phosphatidic acid; PI, phosphatidyl inositol; LPE,

lysophosphatidylethanol amine; PS, phosphatidyl serine;

PCplas, choline piasmalogen; SM, sphingomyelin; DPI,

phosphatidyl inositol 4-phosphate, LPC, lysophospha-

tidylcholine; and TPI, phosphatidyl inositol

4,5- b/sphosphate NL

PG PA

ro to L....« i . i .. I , 10 20 30 40 50 60 70 80 90

Figure 2 Time (min) 24

Gas Liquid Chromatography

FAMEs were separated and quantitated by gas liquid chromatography

(GLC). A set of standard curves was done to establish relative retention times and relative correction factors (RCF) for each FAME.

The internal standard was 17:0. This standard was used for the standard curves and for each sample run. Each FAME was quantitated using the following equation:

[nmol/sample]=[(peak area/17:0 peak area)(RCF/ig//il)(6/il)(1000nmol//tmol)]/M.W.

The GLC system consists of a Shimadzu GC- 8A, a Supelco SP-2330 capillary column (30 m long) and a flame ionization detector.

Detector linearity was determined using commercial standards (NU

Check Prep, Elysian, MN and Supelco, Bellfonte, PA). Column temperature was maintained at 190°C with nitrogen as the carrier gas at a pressure of 0.5 kg/cm2. Detector and injector temperatures were maintained at 220°C. A sp lit ratio of 100:1 was used. Peak area data were collected with a Nelson Analytical 760 series intelligent interface and computed with Nelson model 2600 software.

Assays

Phospholipids were quantitated by assay of lipid phosphorus

(Rouser et a l., 1969). The tubes were taken to dryness in an oven.

To each tube, 0.5 ml distilled water and 0.65 ml perchloric acid was added. The samples were placed in a heating block for one hour at

180°C in a fume hood. For added precaution, a funnel connected to a water aspirator was placed over the samples to remove dangerous 25

Table 2. Short Phospholipid Separation

Time Function Val ue Duration

0.00 min. Chartspeed 0.4 cm/min 0.00 min 0.00 Flowrate 1.8 ml/min 0.00 0.00 % Solvent B 76% 0.00 8.00 % Solvent B 100% 3.00 28.00 Flowrate 2.5 ml/min 0.70 35.00 % Solvent B 45% 1.00 40.00 Chart speed 0 cm/min 0.00 40.00 FIowrate 1.8 ml/min 0.70

This method was used to separate EtnGpl, lysoPtdEtn, ChoGpl and lysoPtdCho. Initial conditions are: column temperature is 35°C, flowrate is 1.8 ml/min and Solvent B 45%. The injector has a 20 p 1 loop. The column was a 5 pm Dupont Zorbax Silica (4.6 mm x 25 cm) column. Solvent A is n-hexane: 2-propanol (3:2 v/v). Solvent B is n-hexane: 2-propanol: water (3:2 + 5.5% by volume). 26

Figure 3 HPLC separation of ethanol amine glycerophospholipid and

choline glycerophospholipids following acid

hydrolysis. Abbreviations are PtdEtn, phosphatidyl-

ethanolamine; lysoPtdEtn, lysophosphatidylethanolamine;

PtdCho, phosphatidylcholine; and lysoPtdCho, lysophos-

phatidyl choline. HPLC Chromatograph

Neutral Lipids

PtdEtn

lysoPtdEtn PtdCho

lysoPtdCho

20 Figure 3 28

1-alkenyl- 2-acyl phospholipid

Ho Acid hydrolysis I 0 C- 0— C—C - R II I R-C-C-0-C-H 0 I II _ C -O -P -O head group I I H2 0 "

1-alkyl- 2-acyl phospholipid

0 C-0-C-C-R

R-C-C-O-C-H O' I I C- 0—P -0 head group I II H2 0

1,2-diacyl phospholipid

Ho 0 I I' 0 C - O - C - C - R II I R-C-C-O-C-H 0 - I I C - o - P - 0 " head group I II H2 0

Figure 4 Molecular structures of phospholipid subclasses 29

fumes. After one hour the samples were removed and allowed to cool

to room temparature. To each tube 3.0 ml water, 0.5 ml 10% ascorbic

acid and 0.5 ml ammonium molybdate was added. After vortexing, the

samples were placed in boiling water for 5 minutes to fully develop

the chromophore. After cooling to room temperature, the samples were

read at 797 nm in a Beckman DU-65 spectrophotometer equipped with a

sipper.

Proteins were assayed using a modification of a dye-binding

procedure (Bradford, 1976). Each lipid extract was centrifuged in a

Model HN-5 International Equipment Company table top centrifuge to

pellet all non-lipid cellular debris. The lipid extract was removed

and stored at -20°Cunder a nitrogen atmosphere. The remaining

residue dried under a gentle stream of nitrogen. Two m illiliters of

water and 200 pi of 2M K0H was added to each sample. The samples

were sonicated using a Branson 185 sonifier set on 7 for two short 5

second bursts. This was sufficient to solubilize the denatured

protein p e lle t. Bovine serum albumin (BSA) was used to establish a

standard curve. The BSA standard and samples contained equimolar

concentration of KOH. Addition of K0H had no undesirable effect upon

the assay. This was determined comparing the equation of the line

for the BSA/KOH standard curve and the BSA alone standard curve. The

samples were read at a wavelength of 595 nm in a Beckman DU-65

spectrophotometer equipped with a sipper.

For cholesterol assays (Bowman and Wolf, 1962), 3 ml ethanoland

3 ml of iron working reagent was added slowly to each sample and mixed well using a vortexer. The iron reagent contained 30

2.5 g FeCl 3*6H20 in 100 ml of 85% phosphoric acid. Eight m illiliters of stock reagent diluted to 100 ml with concentrated

H2SO4 made up the working reagent. After 30 minutes the samples were read at 550 nm against a reagent blank in a Beckman DU-65 spectrophotometer equipped with a sipper. A standard curve was constructed using serial dilution of a standard containing 1 mg/ml of cholesterol.

Results

Primary Neuronal Cells

The choline glycerophospholipids (ChoGpl) and ethanolamine glycerophospholipids (EtnGpl) make up the majority of the total phospholipids of primary murine neuronal cells (Table 3). The lysophospholipids comprised a high mole percent of the total lipid phosphorus when compared to the abnormally low content of all the other phospholipids excluding ChoGpl and EtnGpl. Nearly 18% of the total Gpl are plasmologen (Table 4).

The majority of the arachidonic acid (20:4 n- 6) and docosa- hexaenoic acid (22:6 n-3) were found in the EtnGpl (Table 5). In

ChoGpl the major fatty acids were in decreasing order 16:0, 18:1 and

16:1. In Ptdlns and PtdSer the major fatty acids were in decreasing order 18:0, 16:0 and 18:1. The major fatty acids in the EtnGpl were in decreasing order 18:0, 18:1, 20:4 n -6 and 16:0. The majority of all PUFA were found in the EtnGpl although moderate amounts were present in Ptdlns and PtdSer. The cholesterol content in primary neuronal cells was 224 ± 38 nmol/mg protein. 31

N1E-115 Neuroblastoma

The ChoGpl make up the largest mole % of the total phospholipid in N1E-115 neuroblastoma cells (Table 3). The order of the major phospholipids in decreasing order was ChoGpl > EtnGpl » CerPCho >

Ptdlns = PtdSer. The lysophospholipids were found in relatively smaller amounts when compared to the total composition than in primary neuronal cells. The choline and ethanol amine plasmalogen contents comprised nearly 10% of the total Gpl (Table 4).

In the N1E-115 cells, only the EtnGpl contained a large amount of arachidonic acid (20:4 n- 6) although this amount was variable (Table

6). The phospholipids were comprised of primarily saturated and monounsaturated fatty acids. ChoGpl consistently contained 18:3 n -6 and n-3 series. The major fatty acids in Ptdlns and PtdSer were

16:0, 18:0 and 18:1. The ChoGpl contained large amounts of 18:1,

16:0 and moderate amounts of 18:0. The major FA in EtnGpl were 18:1 and 18:0.

The cholesterol content in N1E-115 cells was 28 ± 19 nmol/mg protein.

R0C-1 01igodendroglia

The ChoGpl make up the largest mole % of the total phospholipid in R0C-1 oligodendroglia cells (Table 3). The order of the major phospholipids in decreasing mole % was ChoGpl > EtnGpl > CerPCho >

Ptdlns = PtdSer. The lysophospholipids make up a small portion of the total phospholipids. The total plasmalogen content was nearly 18% of the total Gpl (Table 4). Table 3 Phospholipid Composition of Cell Cultures

N1E-115 Neuronal ROC-1 HUVE

PtdoGro 1.0 ± 0.4 nd 1.0 ± 0.4 2.6 ± 2.9 PtdGro 1.5 ± 0.1 nd 1.2 ± 0.7 nd EtnGpl 23.9 ± 1.3 44.8 ± 2.1 27.6 ± 3.3 17.6 ± 2.6 Ptdlns 2.9 ± 0.4 1.0 ± 0.7 5.8 ± 1.4 1.9 ± 3.1 PtdOH nd 1.7 ± 0.9 nd nd lysoPtdEtn 0.9 ± 0.2 1.5 ± 0.3 0.8 ± 0.8 3.1 ± 3.4 PtdSer 2.0 ± 0.7 2.5 ± 0.6 5.6 ± 1.5 5.4 ± 2.7 ChoGpi 59.4 ± 1.4 43.5 ± 1.7 43.7 ± 5.1 42.8 ± 6.3 CerPCho 5.8 ± 1.3 2.9 ± 0.9 13.7 ± 2.8 8.9 ± 2.7 PtdIns4P 0.5 ± 0.2 1.0 ± 0.4 0.8 ± 0.2 1.6 ± 1.7 lysoPtdCho 1.3 ± 0.2 1.1 ± 0.9 0.6 ± 0.2 7.5 ± 5.0 PtdIns4,5P 2 nd nd nd 1.2 ± 0.9 (n=4) (n=4) (n=5) (n=12)

The values are the mole % of each phospholipid class. Values are means ± standard deviation, nd means not determined. The abbreviations are as follows: PtdgGro, cardiolipin; PtdGro, phosphatidyl glycerol; EtnGpl, ethanol amine glycerophospholipid; Ptdlns, phosphatidyl inositol; PtdOH, phosphatidic acid; lysoPtdEtn, lysophosphatidyl ethanol amine; PtdSer, phosphatidyl serine ChoGpi, choline glycerophospholipid; CerPCho, sphingomyelin; PtdIns4P, phosphatidyl inositol 4-phosphate; lysoPtdCho, lysophosphatidylcholine; PtdIns4,5P2, phosphatidyl inositol 4,5-bisphosphate Table 4 Plasmalogen Content

N1E-115 Neuronal ROC-1 HUVE

PlsEtn % EtnGpl 26.1 ± 4.8 28.2 ± 2.6 53.5 ± 10.5 42.8 ± 0.4

PlsEtn % Gpl 6.2 12.5 14.8 7.0

PlsCho % ChoGpi 6.1 ± 0.2 11.2 ± 3.8 8.1 ± 3.6 9.3 ± 1.9

PlsCho % Gpl 3.6 4.9 3.5 3.8

The above abbreviations are as follows: PlsEtn, ethanol amine plasmalogen; PlsCho, choline plasmalogen; EtnGpl, ethanolamine glycerophospholipid; ChoGpi, choline glycerophospholipid; Gpl, glycerophospholipid. Values are means ± standard deviation. 34

Table 5 Fatty acid composition of the major phospholipid class in primary neuronal cells

EtnGpl ChoGpi Ptdlns PtdSer

16:0 11 ± 1 45.7 ± 1 25 ± 4 21 ± 7 16:1 4 ± <1 10 ± 1 6 ± 6 11 ± 4 18:0 25 ± 1 5 ± <1 31 ± 11 32 ± 11 18:1 24 ± 1 32 ± 3 16 ± 4 20 ± 6 18:2 2 ± <1 2 ± <1 3 ± 2 4 ± 1 18:3 n -6 1 ± <1 <1 ± <1 nd nd 18:3 n-3 <1 ± <1 <1 ± <1 nd nd 20:3 n-9 7 ± 1 1 + 1 13 + 11 4 ± 3 20:3 n -6 1 ± <1 <1 ± <1 nd nd 20:4 n -6 18 ± 1 2 ± <1 4 ± 5 2 ± 2 22:6 n-3 6 ± <1 nd nd nd

n=4 n=4 n=4 n=4

Values are mole % for each FA detected in the phospholipid class. Values are means ± standard deviation. The nd indicates not detected. Abbreviations are EtnGpl, ethanolamine glycerophos pholipid; ChoGpi, choline glycerophospholipid; Ptdlns, phospha tidyl inositol ; PtdSer, phosphatidyl serine. 35

Table 6 Fatty acid composition of the major phospholipid class in N1E-115 cells

EtnGpl ChoGpi Ptdlns PtdSer

16:0 11 ± 4 2 1 + 8 30 ± 14 29 ± 15 16:1 7 ± 5 8 ± 5 6 ± 7 3 ± 3 18:0 22 ± 2 14 ± 4 30 ± 9 35 ± 11 18:1 49 ± 6 51 ± 7 30 ± 11 25 ± 7 18:2 4 ± 2 1 ± 1 <1 - 1 ± 2 18:3 n -6 <1 - <1 ± 1 nd 6 ± 12 18:3 n-3 <1 - 3 ± 1 nd nd 20:3 n -6 nd nd nd <1 - 20:4 n -6 5 ± 3 <1 ± 1 3 ± 3 <1 -

n=3 n=4 n=4 n=4

Values are mole % for each FA detected in the phospholipid class. Values are means ± standiard deviation. Abbreviations are EtnGpl, ethanol amine glycerophospholipid; ChoGpi, choline glycerophospholipid; Ptdlns, phosphatidyl inositol; PtdSer, phosphatidyl serine. 36

Table 7 Fatty acid composition of the major phospholipid classes in ROC-1 cells

EtnGpl ChoGpi

16:0 28 ± 24 36 ± 13 16:1 5 + 8 10 ± 8 18:0 18 + 8 9 ± 2 18:1 15 + 11 36 ± 12 18:2 7 ± 9 nd 18:3 n-3 2 ± 3 <1 ± 1 20:3 n-9 4 ± 3 1 ± 0 20:3 n-3 nd <1 ± 1 20:4 n-6 21 ± 12 6 ± 5 22:6 n-3 <1 nd

n=4 n=3

Values are mole % for each FA detected in the phospholipid class. Values are means ± standard deviation. Abbreviations are, EtnGpl, ethanol amine glycerophospholipid; ChoGpi, choline glycerophospholipid. 37

Table 8 Fatty acid composition of the major phospholipid class in HUVE cells

EtnGpl ChoGpi Ptdlns PtdSer

16:0 18 ± 14 35 ± 1 21 ± 14 28 ± 14 16:1 5 ± 5 11 + 6 8 ± 8 13 + 15 18:0 24 ± 1 16 ± 1 23 ± 12 30 ± 14 18:1 30 ± 5 31 ± 1 21 ± 17 22 ± 9 18:2 3 ± 1 3 ± 1 12 ± 18 3 ± 2 18:3 n -6 2 ± 1 1 ± 0 9 ± 14 <1 -- 18:3 n-3 <1 ± 1 <1 ± 1 <1 <1 -- 20:3 n-9 nd <1 ± 1 <1 <1 -- 20:3 n-6 1 ± 0 <1 + 1 <1 <1 -- 20:4 n-6 14 ± 9 2 + 0 7 ± 7 <1 22:6 n-3 4 ± 5 <1 + 1 <1 -- <1 --

n=5 n==2 n==5 n==6

Values are mole % for each FA detected in the phospholipid class. Values are means ± standard deviation. Abbreviations are EtnGpl, ethanolamine glycerophospholipid; ChoGpi, choline glycerophospholipid; Ptdlns, phosphatidyl inositol; PtdSer, phosphati dylserine. 38

Only the ChoGpi and EtnGpl fatty acid compositions were reported

due to small sample sizes for the Ptdlns and PtdSer. These small

sample sizes caused variable results. In the EtnGpl the major FA

were 16:0, 18:0, 18:1 and 20:4 n-6 (Table 7). In the ChoGpi the

major FA were 16:0, 18:1 and 18:0 (Table 7). Very l i t t l e arachidonic

acid was found, although its proportion was variable. Very little

PUFA, excluding arachidonic acid, were found in either phospholipid

class.

The cholesterol content in ROC-1 oligodendroglia cells was 44.0 ±

5.9 nmol/mg protein.

Human Umbilical Vein Endothelial Cells

The ChoGpi was the major glycerophospholipid class in HUVE cells

(Table 3). The order of the major phospholipids in decreasing mole %

was ChoGpi > EtnGpl > CerPCho = PtdSer = lysoPtdEtn = lysoPtdCho >

Ptdlns. The lysophospholipids comprise an abnormally high mole % of

the total phospholipid. The polyphosphoinositide PtdIns4,5P2 was

reported here. This is the first report of the mole % of the

polyphosphoinositides in endothelial cells.

In HUVE cells, the major FA in decreasing order in ChoGpi were

16:0, 18:1 and 18:0. ChoGpi represents a mean of two samples. Very

little PUFA were found although arachidonic acid was found to occur

in a small but constant percentage of the total FA. In EtnGpl the

major fatty acids were 18:1, 18:0, 20:4 n-6 and 16:0. Docosa-

hexaenoic acid was found in variable amounts. All other PUFA were

found in small and variable amounts. For Ptdlns and PtdSer the FA 39 compositions were nearly identical. The PUFA were found in small but variable amounts. The major FA were 16:0, 18:0 and 18:1.

The cholesterol content of HUVE cells was 118 ± 8 nmol/mg protein.

Discussion

HUVE, ROC-1, and N1E-115 cells have a more complex phospholipid composition than the neuronal cells. The HUVE and neuronal cells, both primary cell cultures, have higher cholesterol levels and increased PUFA content. The ROC-1 and N1E-115 cells have a low PUFA content and low cholesterol levels. Interestingly, the phospholipid composition ion of the ROC-1 cell line was a hybrid of the composition of the two parent cells. The HUVE cells contain a high lysophospholipid content which may indicate an adaptation to a high pressure environment.

The phospholipid composition of primary neuronal cells and the aberrant neuroblastoma cells differ in several ways. In primary neuronal cells there were two dominating phospholipid classes, those being EtnGpl and ChoGpi. In neuronal cells, these two classes occur roughly in equal amounts. This may be due to the fetal origin of the cultures. The primary neuronal cells, while differentiated, may not have developed a more complex membrane due to th eir fetal origin or the inability to produce the complex membrane in culture. N1E-115 40 neuroblastoma cells have a more complex and varied phospholipid composition. However, the two different cell types differ in the ethanolamine plasmalogen content. The primary neuronal cells have a larger ethanolamine plasmalogen content than the N1E-115 cells. This phospholipid has been implicated in receptor transduction (Horrocks et al., 1986a,b). Thus, this difference may represent an important difference in receptor function.

The cholesterol content of the neuroblastoma cells and neuronal cells was considerably different. Cholesterol is a rigid molecule which condenses the membrane thereby decreasing its fluidity (Stubbs,

1983). The large difference between cholesterol amounts indicates the possibility of a corresponding large difference in membrane fluidity. Membrane fluidity plays an important role in receptor function (Stubbs, 1983). Changes in membrane fluidity, bulk fluidity, as well as specific domain fluidity, can cause many changes in the functional response of the cell to its environment (Spector and Yorek, 1985). Thus, this cholesterol difference could cause an inherent functional difference between neuronal and neuroblastoma cells.

The FA compositions of N1E-115 and neuronal cells were comparable except for an overall higher proportion of PUFA in neuronal cells.

Differences in FA composition, like cholesterol, can cause changes in membrane fluidity resulting in functional differences (Spector and

Yorek, 1985). However, the largest increase in acyl chain motion occurs with the first cis double bond (Stubbs et a l., 1981). This increase in acyl chain motion causes an increase in membrane fluidity over comparable chain length but saturated FA. Introduction of more double bonds does not appreciably increase acyl chain motion and fluidity. Comparing FA composition between N1E-115 and neuronal cells, the EtnGpl 18:1 levels were greater in N1E-115 cells however they had a lower 20:4 n-6 content. The FA compositions of ChoGpi,

Ptdlns and PtdSer were all similar except that Ptdlns and PtdSer contain less PUFA in the N1E-115 cells compared to neuronal cells.

The FA differences were relatively small differing only in PUFA.

Therefore these FA differences should produce only minor fluidity changes.

The ROC-1 oligodendroglia is a hybrid cell made from a calf oligodendrocyte and a C-6 glioma cell (McMorris et a l., 1981). The phospholipid composition should closely resemble one or the other or both of these cell types. The CerPCho, EtnGpl, Ptd2Gro and ChoGpi

% composition closely resembles that of the C-6 glioma (Robert et a l., 1983). These values are found in Table 8 for comparison. The

PtdgGro, CerPCho, EtnGpl and ChoGpi values are very close when comparing C-6 glioma and ROC-1 cells. The Ptdlns and PtdSer reflect the levels seen in glial cells (Table 9). These numbers are comparable to those in whole oligodendroglial cells (Deshmukh et a l.,

1988). In the whole oligodendroglia cells the Ptdlns and PtdSer were not resolved from each other and the two classes together made up

10.4% ± 0.6 of the total phospholipid. The phospholipid composition of the ROC-1 hybrid cell reflects that seen in its parent cells.

This indicates that the phospholipid composition its e lf is a hybrid. 42

Table 9 Comparison of phospholipid compositions of ROC-1, C-6 glioma, and glial cells

C-6 Glioma3 ROC-1

PtdoGro 1.6 ± 0.3 1.0 ± 0.4 CerPCho 13.1 ± 2.1 13.7 ± 2.8 EtnGpl 25.9 ± 1.8 27.6 ± 3.3 ChoGpi 45.2 ± 2.2 43.7 ± 5.1

The values represent mole percent of total phospholipid. For the C-6 glioma n=5. For the ROC-1 n=5. The C-6 glioma cell phospholipids were separated by TLC and quantitated by lipid phosphorus. The ROC-1 cell phospholipids were separated by HPLC and quantitated by phosphorus, a - from (Robert et a l., 1983). Values are means ± standard deviation.

Glialb ROC-1

Ptdlns 5.8 5.8 ± 1.4 PtdSer 6.2 5.6 ± 1.5

The values represent the mole percent of the total phospholipids for the primary rat brain glial cells n=9. For the ROC-1 cells n=5. The glial cell phospholipids were separated by two dimensional TLC and quantitated by lipid phosphorus, b - from (Witter and Debuch, 1982). Values are means ± standard deviation. 43

Compared to primary neuronal and neuroblastoma cholesterol

levels, the ROC-1 cholesterol levels are very similar to those in

N1E-115. This low cholesterol level indicates a fluidity difference

between neuronal and ROC-1 cells. However, since whole

oligodendroglia from calf brain have 60 ± 8.0 nmol/mg protein of

cholesterol (Deshmukh et al., 1988), the ROC-1 cholesterol level is

only slightly lower than that of whole oligodendroglia prepared from

calf brain. The cholesterol content of C-6 glioma cells was not

reported (Robert et a l., 1983). One could speculate that the

cholesterol content, like the phospholipid composition, may reflect

that of C-6 glioma cells or may be the average content of both cells

combined.

The FA composition of only EtnGpl and ChoGpi was reported. The

low levels of Ptdlns and PtdSer did not permit an accurate analysis

of these classes. Low levels of phospholipid, 1-10 nmol, present a difficulty due to the sensitivity limitation of the GLC. The EtnGpl

content also had some variability although the major FA were 16:0,

18:0, 18:1 and 20:4 n-6. This was comparable to those found in the other cell types although the 18:1 levels are lower. The ChoGpi were very low in PUFA and consisted primarily of 16:0, 18:1 and 18:0. The

limited amount of PUFA in aberrant cells also occurs in C-6 glioma or

NN astrocytoma cells. Differentiation of these cells causes an

increase in overall PUFA and these levels in NN cells approach the

level in primary astrocytes (Robert et a l., 1983). Thus, the R0C-1 cell may have a FA composition resembling that of its transformed parent cell. The decreases in PUFA may play a role in maintaining 44

cell transformation. As the PUFA increases in the transformed cell,

either the cell is differentiated or after the cell is differentiated

the PUFA increases as a result of the differentiation. This

indicates the importance of FA to the cell.

Several papers contain the partial phospholipid composition of

endothelial cells from various vascular sources and species

(Selivonchick and Roots, 1976; Rostogi and Nord0y, 1980; Matheson et

al., 1980a,b; Wey et al., 1986; Blank et al., 1986; Siakotos et al.,

1969). In HUVE cells a partial phospholipid composition is reported

(Rostogi and Nord^y, 1980). The values of 25.3% for EtnGpl, 36.3%

for ChoGpi and 12.8% for CerPCho are very similar to the values in

Table 3. The plasmalogen content for HUVE cells has been reported to be 42.5% PlsEtn in EtnGpl and 6.8% PlsCho in ChoGpi (Blank et a l.,

1986). These values resemble the values reported in Table 4 for HUVE cells. Rat and human brain endothelial cells have nearly the same mole % of ChoGpi and EtnGpl as HUVE cells (Siakotos et a l., 1969).

The CerPCho values in rat and human brain endothelial cells, 20.4% and 17.0% respectively, are considerably different from the values reported in Table 3. The differences may be because the brain endothelial cells were freshly isolated from tissue and were not cultured. The rat brain endothelial cells have a PlsEtn content of

46.3% of EtnGpl (Selivonchick and Roots, 1976). This value is similar to that for HUVE cells.

Polyphosphoinositide contents of endothelial cells have not been reported in the literature irrespective of species. Guinea pig brain contains 0.58% PtdIns4P and 2.58% Ptdlns(4,5)P2 (Hawthorne and Kai, 45

1970). The values for HUVE cells are within this range. Hence, the data for HUVE cell phospholipid composition compare favorably to several literature values for the partial phospholipid composition of endothelial cells from various species and vascular sources.

The possibility exists that the phospholipid composition of HUVE cells in culture does not reflect the actual composition in vivo.

Fibronectin has been shown to increase methylation of PtdEtn to

PtdCho by 10-20% thereby increasing the amount of PtdCho (Jaffe et a l., 1985). This could produce a small but possibly significant increase in PtdCho and PlsCho levels while reducing PtdEtn and PlsEtn levels.

The HUVE cells exhibited a high lysoPtdEtn and lysoPtdCho content. The position of the FA was not determined. Regardless of the FA position, lysophospholipids exhibit detergent properties on cell membranes (D'Amato et a l., 1975; Weltzien et a l., 1979a,b).

High concentrations of endogenous lysophospholipids increase the membrane fluidity of the cell giving rise to subtle changes on the cellular surface. The levels in human endothelial cells may be responsible for the apparent ease of thrombosis formation (Moncada and Vane, 1979). Increases in membrane fluidity may cause the cell membrane to be more susceptible to mechanical damage produced by shear stress (Bhagyalakshmi and Frangos, 1989; Mclntire et a l.,

1987). This increased fluidity may account for the observed morphological changes such as crater formation following mechanical trauma (Svendson et a l., 1985). Receptor activation by hormones

(Ryan and Ryan, 1985) and activity of membrane bound enzymes 46

(Weltzien, 1979a) can be affected by increased lysophospholipid

levels through an increase in membrane fluidity (D'Amato et al.,

1975; Weltzien et a l., 1979b).

The FA composition of EtnGpl, ChoGpi, Ptdlns and PtdSer was

reported. The 16:0, 18:0 and 18:1 levels were found to be the major

FA in all four phospholipid classes examined. The cells contained

l i t t l e PUFA. The majority of the PUFA were found in the EtnGpl. The

Ptdlns and ChoGpi have a limited amount of arachidonic acid. In

freshly isolated non-cultured HUVE cells, 16:0, 18:0 and 18:1 make up

a smaller percentage of the total esterified phospholipid FA than in

cells cultured in fetal bovine serum. The PUFA content of fresh

non-cultured cells is greater than that of cultured cells (Lagarde et

a l., 1984). Thus, for HUVE cells, the FA composition can be altered when grown in culture. This casts doubt on whether the FA

compositions of phospholipids in culture represent those in vivo.

The cholesterol level in HUVE cells approached that of the

primary neuronal cell cultures. The level was considerably greater than those for ROC-1 and N1E-115 cells. This higher level of cholesterol in both of the primary cell types examined indicates that differentiated cells maintain a higher level of cholesterol in order to have a more condensed, less fluid and more impermeable membrane.

Membrane fluidity may play an important role in maintaining transformed cells in the transformed state or may be involved in the actual events of transformation.

Membrane fluidity affects numerous cell functions. Diacyl- glycerols, free fatty acids, degree of unsaturation in phospholipid 47 acyl chains, cholesterol and lysophospholipids all affect membrane fluidity. Thus, the composition of lipids within the cell is an important feature in understanding the basic functioning of the cell. As information on the role of phospholipids in signal transduction increases, the overall perception of the importance of phospholipids in cellular functions will evolve from merely structural molecules to biologically active ones. CHAPTER III

ALUMINUM SILICATE

Introduction

Alzheimer's disease (AD) is a serious age-related dementia afflicting primarily elderly people. Currently no etiology for AD has been discovered, although many possibilities have been proposed.

These theories are based upon histological, biochemical and physiological data. Alzheimer's may be caused by changes in gene expression due to age with onset at different ages, a genetic inborn error, a virus, or by environmental factors such as aluminum silicates. Whatever the cause, the fact remains that this disease will grow increasingly important as the elderly population of the

United States increases.

The neuropathology of AD involves the cerebral cortex and the subcortical nuclei which innervate the cerebral cortex (German et a l., 1987). This neuropathological damage may be caused by the retrograde transport of an unidentified toxin. This toxin may induce the formation of paired helical filaments found in the tangles of AD brains. These tangles primarily affect the temporal, frontal and parietal cortex while the visual, somatosensory and motor regions are left intact. The highest density of tangles is in the hippocampus

48 49

(Wilcock, 1988). Positron emission tomagraphy (PET) has indicated a decrease in glucose metabolism in all areas affected by tangles and like the tangles does not affect the visual, somatosensory or visual areas (Milcock, 1988).

Alzheimer's disease is not only characterized by tangles, but also by neuritic placques in susceptible areas. These placques contain amyloid protein. While deposition of amyloid protein occurs during normal aging in the mammalian central nervous system, the process appears to be unregulated and the rate of deposition increased in AD (Selkoe, 1987). Structural proteins such as tubulin have impaired synthesis in AD brains due to the improper association of the tau protein and tubulin. The amyloid protein may act as a neurotoxin, thereby eliciting this and other AD pathology (Selkoe,

1987). However, aluminum silicates may also act as a neurotoxin either independently or synergistically with other agents to cause

AD. Aluminum silicates are localized in the central region of senile placques and may have a pathological action in AD (Candy et al.,

1986).

There are two types of AD. The first is characterized by an early onset of symptoms and involves greater brain atrophy, neuronal loss and histochemical and biochemical changes. Death usually occurs before 80 years of age (Kay, 1987). The principle biochemical changes are decreases in cholinergic and serotinin neurotransmitter levels and in their synthetic enzymes (Wilcock, 1988). Neuro endocrine function in AD brains appears to be normal, although

^-endorphin levels are elevated in the blood (Franceschi et al., 1988). This indicates that the large biochemical changes found in AD do not affect neuroendocrine function. Most pharmacological treatments attempt to enhance cholinergic neurotransmitter levels.

These include precursor loading with choline or lecithin, use of the choline receptor agonist arecholine or anti-cholinesterases such as tetrahydroaminoacridine or physostigmine (Wilcock, 1988). The second type of AD has a late onset and appears to follow a normal pattern of aging (Kay, 1987). Recent epidemiological studies have indicated a single gene autosomal dominant mutant may be the cause of familial

AD. The heterogeneity of AD continues to point to multi factorial causes (Kay, 1987).

The heterogeneity of AD may be explained by an environmental factor which acts as a neurotoxin. Aluminum is generally found in insoluble forms at pH 7.4, however, citrate is able to increase the solubility of aluminum in the form of the biologically active Al3+ at this pH (Martin, 1986). Aluminum silicates are also solubilized by c itrate levels found in the blood (Birchall and Chappell, 1988b).

At pH 7.4 silicates are the primary binders of Al3+, but at intracellular pH (pH <6.6) the aluminum ion preferentially binds to phosphates. Aluminum is found in food, antacid tablets, water

(Martin, 1986) as well as antiperspirants, deionizers, talc, cat litter, dental polishing compound, industrial binders and tobacco products (Roberts, 1986). The Al3+ ion intracellularly can bind to phosphates and may change the physiological functioning of the cell by binding to chromatin, adenosinetriphosphate and polyphosphoinositides (Birchall and Chappell, 1988a). While the 51 exact mechanism by which aluminum exerts its effects is unknown, its physiological effects have been well documented.

Aluminum induces encephalopathy in rats when injected intracerebroventrically (Lipman et a l., 1988). These rats have an increased startle reflex and exhibit decreased learning and memory retention. Glucose uptake is also decreased in the cortex and striatum, perhaps in the same type of neurons as those affected in AD

(Lipman et a l., 1988). This uptake may be due to aluminum-induced inhibition of brain hexokinase (Womack and Colowick, 1979). Chronic oral administration of aluminum in rats has little effect on cholinergic functioning, however, there are decreases in passive avoidance behavior, learning and memory retention (Connor et al.,

1988). To achieve a greater effect of aluminum on neurological functioning a compromised blood brain barrier may be needed.

Aluminum is thought to be the cause of dialysis-induced encephalopathy (Arieff, 1985; Alfrey et a l., 1976). This is a progressive dementia which is prevalent in patients undergoing dialysis for over two years. The initial symptoms are speech problems followed by psychosis which leads to dementia. Death generally occurs within six months after the onset of symptoms

(Arieff, 1985). Aluminum levels are increased 11-fold in gray matter following chronic hemodialysis (Alfrey et a l., 1976). This increased aluminum is the cause of the dementia.

Injection of AlCl3 intracerebroventrically into rabbits produces neurofibrillary degeneration (Beal et a l., 1989).

Choiineacetyltransferase (ChAT) activity is reduced in the entorhinal 52 cortex and hippocampus. The microtubule assembly protein tau is not found in the degenerated neurofibrils as it is in AD. However, learning and memory deficits do occur in the aluminum-injected rabbits indicating that some physiological change occurred. While not all biochemical and pathological changes found in AD occur in aluminum-induced encephalopathies, the evidence does indicate a role for aluminum as a neurotoxin eliciting AD-like symptoms. In order to achieve complete AD pathophysiology, chronic exposure to aluminum may be needed (Candy et a l., 1986).

Aluminum silicates have been hypothesized to be the actual form of aluminum to cause neuronal degeneration in AD. As previously indicated, aluminum silicates are the metal complexes which are both focally and selectively concentrated in placques. One route of entry may be through the olfactory bulb (Perl and Good, 1987; Altman, 1989;

Roberts, 1986). This has increasingly become an interesting hypothesis. In AD patients there are definite pathological changes in the nasal epithelium. These changes are in the morphology, distribution and immunoreactivity of neuronal structures associated with the epithelium (Talamo et al., 1989). Neurofilament antibodies have an increase in reactivity to these abnormal neuronal olfactory structures from AD patients. There is evidence that neurofibrillary tangles increase and the number of functioning neuronal cells in the olfactory bulb region decrease in AD patients (Esiri and Wilcock,

1984). Alzheimer's disease is known to affect the sense of smell.

This indicates a compromised olfactory region in AD patients. 53

The olfactory region is directly linked to the olfactory cortex

which is linked to the regions involved in AD (Altman, 1989).

Aluminum increases have been found in the olfactory bulb (Crapper et

a l., 1976). This has led to the hypothesis that the major route of

entry for aluminum is through the olfactory bulb (Crapper et al.,

1976; Roberts, 1986; Perl and Good, 1987). People afflicted with AD

lack effective mucosal/olfactory bulb barriers (Perl and Good,

1987). As these barriers become compromised, retrograde transport of

aluminum silicates through the olfactory tract into the brain

occurs. Once in the brain, the aluminum silicates may aid in

producing AD (Roberts, 1986). Nasal exposure of aluminum has caused

an increase in cerebral cortex aluminum levels (Perl and Good,

1987). Studies with horseradish peroxidase and other compounds

indicate a direct link between the olfactory bulb and various brain

centers affected in AD (Roberts, 1986). Thus, these studies indicate

the olfactory bulb route of aluminum silicate entry into the brain to

be very plausible.

Aluminum has numerous intracellular biochemical effects.

Aluminum binds to calmodulin with a 10 times greater affinity than

Ca2+. Aluminum induced conformational change in calmodulin which

reduces the a helical content by 30% which results in a 20-fold

increase in hydrophobic surface area (Seigle and Haug, 1983). This

displacement of Ca2+ from calmodulin will cause a disruption in all

Ca2+ calmodulin mediated events. Aluminum (Al3+) inhibits

phospholipase C activity on the polyphosphoinositides (McDonald and

Mamrack, 1988). This causes a disruption in the receptor-stimulated 54

hydrolysis of the polyphosphoinositides into diacyl glycerol and

inositol phosphates and the subsequent stimulation of protein kinase

C by diacylglycerol (Majerus et al., 1985; Nishizuka, 1984; Sekar and

Hokin, 1986; Kikkawa and Nishizuka, 1987). Aluminum also inhibits

brain hexokinase by binding to ATP to form an Al-ATP complex which

unlike the Mg-ATP complex is not capable of phosphorylating glucose

(Womack and Colowick, 1979). This effectively disrupts glucose

uptake by the brain. Aluminum is also thought to cause a structural

change in the chromatin structure leading to transcriptional and

translational disruption (Crapper McLachlan, 1986). ADP-ribo-

sylation, a mechanism important to DNA repair, is inhibited by

aluminum (Crapper McLachlan et a l., 1983).

In order to more fully understand the toxicity of aluminum

silicates to the brain and their possible role in neuronal degeneration associated with AD, the effect of aluminum silicate containing clays on cell cultures was examined. The cultures were examined for evidence of degeneration produced by the toxic clays.

To fully understand the process by which environmental minerals are toxic to cells, the effects of asbestos, silica and aluminum silicate clays must be examined.

Asbestos is a mineral with well known toxic effects (Macnab and

Harington, 1967). However, the mechanism for asbestos toxicity is not known definitively. Evidence indicates that the Mg^+ moiety of the asbestos is capable of ordering the negatively charged sialic acid on glycoproteins. This ordering produces a channel which compromises the ability of the cellular membranes to regulate ionic movement. The result is osmotic swelling and lysis of the cell

(Harington et al., 1975). Ethylenediaminetetraacetic acid (EDTA) and polyvinylpyridine-N-oxide (PVPNO) are capable of protecting the cell from osmotic lysis by binding the Mg^+ on the asbestos (Macnab and

Harington, 1967). In vivo the Mg^+ may be leached out of the asbestos fiber increasing the negative charge. As the negative charge increases so does the hemolytic activity indicating a possible role for the negative charge in cell lysis (Jayawardena and Waksman,

1977). Both of these hypotheses indicated a charge interaction with the membrane. However, the lysis due to asbestos fibers may be due to the extraction of phospholipids from the membrane thereby causing an increase in membrane permeability (Jaurand et a l., 1979). The lysis is area dependent indicating the need for a direct interaction with the membrane. Preincubation of asbestos with dipalmitoylglycerophosphocholine decreases lysis by 95 to 100%

(Jaurand et a l., 1979). This indicates the possibility of a charge interaction between the clay and the dipalmitoylglycerophosphocholine.

Asbestos causes cell lysis by several possible mechanisms. Each mechanism disrupts the ionic balance causing ion shifts to occur

(Harington et a l., 1975). Prior to these changes asbestos causes arachidonic acid release and prostaglandin E 2 formation in macrophage and fibroblast cocultures. Collagen and protein synthesis is decreased by asbestos (Goldstein et a l., 1982). This indicates that asbestos e lic its a biochemical change prior to cell lysis.

Membrane interaction may induce arachidonic acid release, however, 56 protein synthesis inhibition occurs at the level of mRNA or DNA indicating a possible intracellular role as well as an extracellular membrane interaction. The effect of asbestos on protein synthesis may be related to asbestos carcinogenicity.

Silica also possess hemolytic activity in red blood cell preparations. Silica can act as a hydrogen donor for hydrogen bonding with membrane proteins and phospholipids (Harley and

Margo!is, 1961; Nash et al., 1966). The binding to phospholipids is greater than binding to proteins (Nash et al., 1966). Silica can also denature proteins presumably through strong hydrogen binding causing a protein conformational change (Harley and Margolis, 1961).

This may be the mechanism which permits bovine serum albumin to protect the cell from lysis (Singh et a l., 1983). Particle size is important in order to achieve the best binding to phospholipids and proteins. The smaller the size the greater the reactivity (Harley and Margolis, 1961). Preincubation with phosphotidylcholine or sphingomyelin causes a decrease in ly tic activity (Singh et a l.,

1983). This is due to the absorption of the silica onto the phospholipid thereby protecting cellular phospholipids from binding with the silica. The action of silica appears to be a direct interaction of silica with phospholipids and cholesterol in the membrane to cause an increase in membrane permeability (Harington et a l., 1975).

Kaolinite is an aluminum silicate containing clay with the formula of Al 2Si205(0H)4 (Brindley and Brown, 1980). Kaolinite exhibits cytotoxicity towards macrophages, erythrocytes and epithelial cells (Oscarson et a l., 1986; Woodworth et a l., 1982;

Davies et a l., 1984). Kaolinite possess a negative charge on the particle and a positive charge on the crystal edges. Treatment of kaolinite with polyacrylic acid negates the positive charge, but does not remove the toxicity of the clay (Davies et a l., 1984). The negative charge binds with membrane phospholipids, primarily with positive charges on phosphatidylcholine and glycoproteins (Woodworth et a l., 1982). Hydrogen bonding also occurs between the clay and the membrane (Davies et a l., 1984). Polyvinyl-pyridine-N-oxide (PVPNO) reduces the hydrogen bonding of kaolinite with the membrane and reduces the negative charge, thereby protecting the membrane (Davies and Preece, 1983).

MontmorilIonite is another aluminum silicate containing clay with the formula Al2 >3gFe 3+0 68Mg0 47 (Si7 71A10 2g)O20(OH)4Xo 67

(Newman and Brown, 1987). MontmorilIonite is more cytotoxic towards epithelial cells than kaolinite. MontmorilIonite, like kaolinite, exerts its effects by hydrogen bonding reactions and negative charge interactions (Woodworth et al., 1982). The crystalline structure has the same charge arrangement as kaolinite (Woodworth et a l., 1982).

Erionite is an aluminum silic ate containing clay with the formula

(Na2K2Ca)4 >5[Al9Si27072]H2o 27 (Der et a l., 1989). Erionite has a needle, fiber-like crystalline structure. Bentonite has a formula

Al203*4 SiO2 *H20 and like montmorillonite and kaolinite, bentonite has an amorphous crystalline structure (Windholz et a l., 1983).

Montmorillonite initiated erythrocyte hemolysis is very quick occurring within 10-20 minutes. In epithelial cells the particles 58 appear to be initially phagocytized after 4 hours of incubation. By

24 hours the amount of phagocytized clays was pronounced (Woodworth et al., 1982). The negative charge on the montmorillonite plays a major role in cellular lysis. Coating the clay with aluminum decreases its toxicity due to the removal of negative charges

(Oscarson et a l., 1986). The lysis of the cells is not osmotic because no ghosts are le ft following treatment (Oscarson et a l.,

1986). As with kaolinite, there is a quick transition between viability and non-viability of the cell (Davies et a l., 1984).

Airborne mineral dusts, more than likely containing some aluminum silicate containing clays, cause an increase in platelet activating factor production by macrophages (Gercken et a l., 1988). Platelet activating factor is a potent brochoconstrictor and platelet aggregant (Braquet et a l., 1987). Kaolinite decreases amino acid incorporation into proteins and decreases non-competitive amino acid transport (Low et a l., 1980). Thus, kaolinite produces biochemical changes which precede cell death indicating the interaction with the cellular membrane produces changes within the cell other than complete lysis.

Aluminum silicate containing clays are able to catalyze several different reactions. The aluminum silicate is a very reactive species with Bronsted acid properties (Wright et al., 1985; Haag et al., 1984). Montmorillonite possesses sites which undergo cationic exchange and are capable of cleaving peptide bonds (Crapper

McLachlan, 1986). Thus, aluminum silicates are extremely reactive substances which interact with cellular membranes to cause a 59 perturbation of the membrane which eventually results in cellular lysis and death.

Aluminum silicate containing clays are potential environmental hazards. Due to the lytic activity of the clays and the evidence linking aluminum silicates to AD, a neurotoxicity study was done. In this study, various cells were used to examine the potential toxic effects of the clays on the cells. N1E-115 neuroblastoma cells were used to represent neurons. ROC-1 oligodendroglia were used to represent oligodendroglia. Due to the unavailability of cerebral microvessels, human umbilical vein endothelial cells were used to represent cerebral endothelial cells. The toxicity of the clays was assessed by fatty acid release, lactate dehydrogenose release and by morphological changes.

The hypothesis was to determine the toxicity of aluminum silicate containing clays towards ROC-1, N1E-115 and HUVE cells. These clays may cause cellular necrosis in the brain regions affected by AD and may aid in the formation of AD. Thus, the clays toxicity towards cells in culture is one step towards determining a role in the formation of AD.

Materials and Methods

CelTs

Three cell types were used in these studies. These cells were:

ROC-1 oligodendroglia, N1E-115 neuroblastoma and HUVE human umbilical 60 vein endothelial cells. Cell culture conditions and procedures have been described previously.

Addition of Silicates

Four different aluminum silicate clays indigenous to the United

States were used. These were: Montmorillonite (Wyoming), Erionite

(unknown origin), Kaolinite (South Carolina), and Bentonite

(Missouri). The clays were graciously provided by Dr. Eugene

Roberts, Chairman of the Department of Neurobiology, Beckman City of

Hope National Medical Center, Duarte, California.

The clays had been previously ground into a fine powder and purity of the clay determined. Each clay (25 mg) was transferred to a porcelain crucible to which 1 ml of PBS buffer was added. The clay was ground into a fine paste by using a glass stirring rod (ends rounded). After addition of 1 ml PBS buffer, the clay suspension was removed. The crucible was washed four times with 2 ml aliquots of

PBS buffer. This brought the final concentration of clay suspension to 2.5 mg/ml buffer. Three different final concentrations of clay suspension were added to cell culture plates, 0.1, 0.03, and 0.01 mg/ml (Woodworth et a l., 1982). The clay suspensions were mixed prior to each addition. The clay suspensions were pipetted into the culture medium and the medium gently rocked in two planes to assure proper distribution of the clay over the cells (Woodworth et al.,

1982; Oscarson et a l., 1986). The cells were then incubated for 1,

6, or 24 hours. After the proper incubation time was reached, the cells were extracted and free fatty acid levels were determined. 61

Lactate Dehydrogenase

Lactate dehydrogenase (LDH) activ ities were assayed in the cell medium as a means to determine i f the membrane had been compromised.

LDH was assayed using a kit from Sigma Diagnostic (St. Louis, MO) and is based upon the oxidation of lactate to pyruvate resulting in an increase in nicotinamideadeninedinucleotide (NADH) formation (Amador et a l., 1963). The NADH formation was determined at 340 nm in quartz cuvettes. The activity of LDH in experimental cultures was compared to that in control cultures. The following equation was used:

% LDH of control = [U LDH/mg cell protein expt ] x 100 [U LDH/mg cell protein control]

For all cell types, LDH was assayed at the clay concentration and incubation time with the greatest fatty acid release. The aluminum silicate clays had no effect on rabbit muscle LDH activity. These values are expressed as percent of control.

Cell Viability

Cell viability was assessed by using the Trypan Blue exclusion test. Intact cells are capable of excluding Trypan Blue and appear clear under phase contrast microscopy. Compromised cells with holes in th eir membrane cannot exclude the dye and appear dark under phase contrast microscopy. Using a hemacytometer, several fields of cells were counted. The number of dark cells per field were averaged as 62 were the number of clear cells per field. The percent viable was calculated using the following equation:

% viable = 100 - [# dark cells] x 100 [# clear cells] + [# dark cells]

All cell types were examined at optimal clay concentration and incubation time.

Microscopy

Cells were examined at various concentrations and incubation times with a phase contrast microscope (Nikon, Japan). Photographs were taken using a through scope camera system (Nikon, Japan).

All other methods used were previously described. These include

Bradford protein assay, TLC, GLC, FAME formation, cell harvesting and lipid extraction.

Statistics

All statistics were done using one way analysis of variance using the Statgraf® program. A multivariance statistical analysis was performed using a level of significance of 0.05 or greater.

Results

ROC-1 Oligodendroglia! Cells

The majority of the FA released after incubation with the clays were the saturated FA 16:0 and 18:0. The monounsaturated FA 18:1 was released and to a small extent 16:1. Very little polyunsaturated FA

(PUFA) were released. The order for PUFA release in decreasing 63 amounts was 20:4 n-6 > 20:3 n-6 > 18:3 n-6 > 18:3 n-3 » 22:6 n-3.

The FA were found primarily in the medium extracts. Only 5-15% of the total FA were in the cell extract and these were primarily saturated.

Bentonite (Figure 5} essentially produced no increase in FA over controls. At 1 hour, the 0.01 mg/ml and 0.03 mg/ml concentration values were significantly greater from the 0.10 mg/ml value. The

1 hour control level was abnormally high and consequently was significantly greater than all other points on the curve.

For kaolinite (Figure 6) at 1 hour, the FA released by the 0.03 mg/ml concentration were significantly greater than those released by the 0.10 mg/ml clay concentration. This trend was consistently found for each time point and appears to be time dependent. The 24 hour FA release was greater than the 6 hour release which was greater than 1 hour release.

Erionite (Figure 7) showed a dose response at 1 and 6 hours. At

6 hours, for the 0.03 mg/ml and 0.1 mg/ml clay concentrations the FA released were significantly greater than those at the 0.01 mg/ml concentration and control. By 24 hours this increase in FA levels had disappeared.

Montmorillonite (Figure 8) showed an immediate increase in FA.

At 1 hour the 0.10 mg/ml clay concentration had significantly released more FA than the control. There was no significanct difference at 6 hours but at 24 hours the FA released by the 0.01,

0.03, and 0.10 mg/ml clay concentrations were all significantly greater than the control. 64

Comparing the clay/control FA ratios, montmorillonite at all concentrations at 1 hour was significantly greater than the other clays except kaolinite at the 0.03 mg/ml concentration (Figure 9).

At 6 hours (Figure 10) the 0.01 mg/ml and 0.03 mg/ml concentrations of clay showed no significant increases in FA released. Kaolinite's greatest toxicity occurs at 0.03 mg/ml although this was not significant. The FA value for montmorillonite were significantly greater than those for bentonite and kaolinite at 0.10 mg/ml but not erionite. At 24 hours (Figure 11) the montmorillonite concentration of 0.03 mg/ml had a significantly greater FA increase than all others. At the montmorillonite concentration of 0.10 mg/ml, the FA released were significantly greater than all others. Erionite's induced FA released were significantly greater than bentonite's.

Comparing all time points in R0C-1 cells, montmorillonite consistently was significantly more toxic than the other clays.

Erionite consistently was the next toxic.

LDH was examined after 24 hours of incubation with each clay at

0.10 mg/ml. No significant increases in LDH activity were observed

(Table 10). A Trypan blue exclusion study was done. No differences in cell viability between control and experimental values were found.

Photographs were taken to examine morphological changes. As indicated in Plates I and II, very l i t t l e morphological change occurs. The cell membranes are intact even though the clays are associated with the membrane. Montmorillonite and bentonite were bound to the membrane in greater amounts than kaolinite and erionite. 65

NIE-115 Neuroblastoma Cells

The majority of the FA released after incubation with the clays were 16:0 and 18:0. Some monounsaturated FA were also detected. The monounsaturated FA were 18:1 and 16:1. Very l i t t l e PUFA were

released although the greatest amount was 20:4 n-6 followed by 20:3

n-6, 18:3 n-6 and 18:3 n-3. Minor amounts of 22:6 n-3 were found.

The FA were primarily found in the medium extracts. Only 5-15% of

the total FA released were in cell extract and these were primarily

saturated.

Bentonite (Figure 12) at a concentration of 0.10 mg/ml caused a

significant increase in total free FA over all other concentrations and controls at 1 hour. This increase fell by 6 hours tto a level which was not significantly different from controls.

Kaolinite (Figure 13) showed a unique set of values. The greatest FA values were at 1 and 6 hours with the 0.03 mg/ml clay concentration. These values were significantly greater than the FA released induced by all other concentrations and controls. There was no dose-dependent response. At 24 hours the 0.01 mg/ml clay concentration caused a FA release which was significantly greater than all other clay concentrations.

Erionite (Figure 14) showed no dose response after 1 and 6 hours of treatment. However, at 24 hours the 0.10 mg/ml clay concentration caused FA to be released at values significantly greater than the value at all other times indicating the beginning of a dose response effect. 66

Montmorillonite (Figure 15) caused an initial increase in free FA

after 1 hour as indicated by the significant difference of the 0.10

mg/ml concentration from control. This dose-dependent-like response

was not maintained at 6 and 24 hours.

Comparing clay/control FA ratios after 1 hour of incubation,

montmorilIonite's induced FA release was significantly greater than

the FA release by bentonite at 0.01 mg/ml and all other clays at 0.10

mg/ml (Figure 16). The FA release at 0.03 mg/ml of kaolinite was

significantly greater than that caused by bentonite and erionite.

This pattern for kaolinite follows that established in Figure 13. At

6 hours, montmorillonite produced a consistently lower release of

free FA than the other clays (Figure 17). At low concentrations

bentonite, kaolinite, and erionite were equally potent. Increasing

the clay concentration to 0.03 mg/ml, kaolinite becomes significantly

greater than all other clays. The order of potency was kaolinite >

bentonite > erionite » montmorillonite. At 0.10 mg/ml, bentonite,

kaolinite and erionite became equally potent again. Kaolinite

produced a significantly greater FA release than all the other clays

at 0.01 mg/ml after 24 hours of incubation (Figure 18). At 0.10

mg/ml erionite produced a significant increase in free FA whereas

bentonite, montmorillonite and kaolinite levels remained equal.

LDH activities were examined after 24 hours of incubation with

0.10 mg/ml of each clay in the cell cultures. No significant

increases in LDH activity were observed (Table 10). The same cells were used for Trypan Blue exclusion studies. No significant changes in cell viability were observed (Table 11). 67

Morphologically, the cells remained normal except that clay was associated with the membranes. This association produced l i t t l e morphological change and the cell membranes appeared normal and intact.

Human Umbilical Vein Endothelial Cells

The majority of the FA released after treatment with clays were

16:0 and 18:0. The only monounsaturated FA released were 18:1 and to a small extent 16:1. The PUFA released differed from N1E-115 and

ROC-1. The PUFA profile included 20:4 n-6, 20:3 n-6, 18:3 n-6 and

18:3 n-3, however, unlike N1E-115 and R0C-1 cells, 22:6 n-3 was found to be released in an approximate 1:1 ratio with 20:4 n-6. The FA were found primarily in the medium. Only 5-15% of the total FA released were found in the cell extracts. The FA found in the cell extracts were predominantly saturated in nature.

Bentonite (Figure 19) showed no increase in FA released at the 1 and 6 hour time points for all concentrations. At 24 hours a dose- dependent response was seen. The 0.01 mg/ml clay concentration caused a significantly greater FA release than control values. At the 0.03 mg/ml and 0.10 mg/ml clay concentrations, FA release was significantly greater than for control and 0.01 mg/ml FA levels.

Kaolinite (Figure 20) like bentonite showed no increase in FA released at the shorter time periods. By 24 hours a dose dependent effect was observed. The 0.10 mg/ml concentration of clay produced a

FA release that was significantly greater than the FA levels for 0.01 mg/ml and control cultures. 68

Erionite (Figure 21) showed no dose response effect at 1 hour

although the released FA values for the 0.03 mg/ml clay concentration

were significantly higher than the control. At 6 hours, the FA

released by all three clay concentrations were equally significant

from controls but not in a dose-response fashion. By 24 hours a

dose-response-like effect was seen starting at 0.03 mg/ml. The FA

values at the 0.10 mg/ml clay concentration was significantly greater

from the other clay concentrations and control levels.

Montmorillonite (Figure 22) showed a dose-response after 24 hours

of incubation in the HUVE cultures. The 1 and 6 hour incubations had

some sta tistic a lly significant points, however, no large changes were

seen in FA release. After 24 hours, the 0.03 mg/ml and 0.10 mg/ml

clay concentrations caused FA releases which were significantly

greater than the control and 0.01 mg/ml concentration. However, the

0.03 mg/ml and 0.10 mg/ml concentrations were not significantly

greater than each other. This indicated a dose-response to montmorillonite by the HUVE cells.

Comparing clay/control FA ratios at 1 hour, the only major

significance between clays was at the concentration of 0.03 mg/ml

(Figure 23). The order of significance was erionite > kaolinite >

bentonite > montmorillonite. At 6 hours erionite produced a

significantly greater release of FA than all other clays at each

concentration (Figure 24). However, by 24 hours the dose-response to montmorillonite became evident. At 0.01 mg/ml bentonite was

significant from all others and erionite and montmorillonite were

significant from kaolinite. At 6 hours montmorillonite induced FA 69 release was significantly greater than all others and bentonite was greater than kaolinite but not erionite. At 0.10 mg/ml the montmorillonite induced FA release was significantly greater than all others. After 24 hours of incubation, the montmorillonite induced FA release was significantly greater than all other clays at the concentrations of 0.03 and 0.10 mg/ml. Bentonite produced a significant increase in FA released at the 0.03 mg/ml concentration compared to kaolinite and erionite (Figure 25).

LDH activity was examined after 24 hours of incubating 0.10 mg/ml of each clay in the cell cultures (Table 10). Bentonite, montmorillonite and kaolinite caused significant increases in LDH in the media following clay addition. Erionite, which was effective in releasing FA, did not cause an increase in LDH release above control levels. Increases over control were 141% by bentonite, 146% by kaolinite and 154% by montmorillonite.

Morphologically the cells were drastically changed after treatment with montmorillonite, bentonite and kaolinite (Plates

III-V II). Montmorillonite caused complete lysis of the cells. Cell fragments were seen with clay adhered to both sides, however no ghost or discernible nuclei were seen. This was indicative of extensive cell damage. Bentonite and kaolinite also induced cell lysis but were not as potent as montmorillonite (Table 11). Erionite induced lysis, but few cell fragments were evident. Erionite was the least potent of all clays in causing lysis (Table 11). Erionite had a long rod-like crystalline structure, differing from the amorphous structure of the other clays. 70

Trypan blue exclusion studies were not done due to the extreme state of fragmentation of the cells. Morphologically it was evident that severe damage had occurred to the cells.

Clay Toxicity

Bentonite's greatest toxicity at 1 hour was towards HUVE and

N1E-115 cells (Figure 26). The greatest FA release in N1E-115 cells was at 0.10 mg/ml. This release was 2.2-fold over control levels.

At 6 hours bentonite's greatest toxicity was toward N1E-115 cells, however all concentrations had a 2.0 - 2.3-fold increase above controls, but exhibited no dose-response (Figure 27). The greatest

FA release at 24 hours occurred in HUVE cells. This response by the

HUVE cells to bentonite was dose-dependent (Figure 28).

K aolinite's greatest toxicity at 1 hour was towards N1E-115 cells. The greatest release (3.3-fold over controls) was at a clay concentration of 0.03 mg/ml (Figure 29). A dose-like response occurred in N1E-115 cells at 6 hours over the 0.01 mg/ml to 0.03 mg/ml concentration range. At 0.10 mg/ml this increase was not seen

(Figure 30). At 24 hours a large increase in released FA occurred at

0.01 mg/ml in N1E-115 cells. This response was not maintained at higher concentrations. The other cell types had minimal FA release (Figure 31).

Erionite's greatest toxicity at 1 hour occurred in HUVE and R0C-1 cells at an erionite concentration of 0.01 mg/ml (Figure 32). At 6 hours, the most significant increase in FA occurred in HUVE and

N1E-115 cells (Figure 33). HUVE cells had a significant increase in

FA at each concentration of clay. After treatment with 0.01 mg/ml of 71 erionite, N1E-115 cell's FA levels were significantly elevated compared to those in ROC-1 cells. The only significant FA increase at 24 hours was at the highest concentration of erionite. HUVE and

N1E-115 cells both had large significant FA increases compared to

ROC-1 cells (Figure 34).

Montmorillonite at 1 h had the largest increase in FA at the clay concentration of 0.10 mg/ml (Figure 35). N1E-115 and ROC-1 cells were significantly different from HUVE cells. After 6 hours of incubation, HUVE cells had the largest increase in FA at the clay concentration of 0.01 mg/ml. At the concentrations of 0.03 and 0.10 mg/ml, R0C-1 cells released significantly more FA than the other two cell types (Figure 36). At 24 hours, HUVE cells exhibited a dose-like response to montmorillonite. At the lowest concentration,

R0C-1 cells were significantly different from the other two cell types (Figure 37). 72

Figure 5 FA release in ROC-1 cells by bentonite. At 1 h FA

levels in control > 0.01, 0.03 > 0.10 mg/ml of clay.

At 6 h there was no significance, in FA levels. At 24 h

FA levels in controls >0.10 mg/ml of clay. The nd

indicates no data. Level of significance p < 0.05 Fatty Acid Release in ROC-1 Cells by Bentonite Fatty Acid nmol/mg protein 1 hour

125 6 hour

100 24 hour

ao 0.1 Figure 5 Bentonite mg/ml 74

Figure 6 FA release in ROC-1 cells by kaolinite. At 1 h FA

levels in 0.03 > 0.10 mg/ml of clay. At 6 h the FA

levels in control and 0.03 mg/ml >0.01 and 0.10 mg/ml

of clay. At 24 h the FA levels in 0.03 mg/ml > 1.0

mg/ml. The nd indicates no data. Level of

significance p < 0.05. Fatty Acid Release in ROC-1 Cells by Kaolinite F atty Acid nmol/mg pro tein 1 hour ------50 6 hour

40 24 hour

nd ao a oi ao3 ai Figure Kaolinite mg/ml 76

Figure 7 FA release in ROC-1 cells by erionite. At 1 h there

was no significance in FA levels. At 6 h the FA levels

in 0.03 and 0.10 mg/ml > control and 0.01 mg/ml of

clay. At 24 h FA levels in 0.01 mg/ml > 0.10 mg/ml of clay. Level of significance p < 0.05. Fatty Acid Release in ROC-1 Cells by Erionite F atty Acid nmol/mg p ro tein 1 hour 40 6 hour

24 hour 30

ao a oi a 03 a i Erionite ng/ml Figure 78

Figure 8 FA release in ROC-1 cells by montmorillonite. At 1 and

6 h there was no significance in FA levels. At 24 h FA

levels in 0.01, 0.03, and 0.10 mg/ml of clay >

controls. Level of significance p < 0.05. Fatty Acid Release in ROC-1 Cells by MontmorilIonite F atty Acid nmol/mg p ro tein 1 hour 40 6 hour

24 hour 30

OLO a 01 a i Figure 8 Montmori1Ionite «g/«l 80

Figure 9 Comparison of FA release in ROC-1 cells by bentonite,

kaolinite, montmorillonite and erionite. Incubation

time was 1 h. At 0.01 mg/ml of clay FA levels in

montmorillonite > all others and erionite > bentonite.

At 0.03 mg/ml of clay FA levels in montmorillonite and

kaolinite > bentonite and erionite. At 0.10 mg/ml of

clay FA levels in montmorillonite > erionite >

bentonite and kaolinite. Level of significance

p < 0.05. Fatty Acid Release in ROC-1 Cel1st Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite 1 hour L ]

Kaolinite 1 hour 1ZZ73

Hontaorill. 1 hour

Erionite 1 hour

i 00 a 03 Clay Concentration ag/el Figure 82

Figure 10 Comparison of FA release in ROC-1 cells by bentonite,

kaolinite, montmorilIonite and erionite. Incubation

time was 6 h. At 0.01 mg/ml of clay there was no

significance in FA levels. At 0.03 mg/ml of clay FA

levels in erionite > bentonite. At 0.10 mg/ml of clay

FA levels in erionite > all others. The nd indicates

no data. Level of significance p < 0.05. Fatty Acid Release in ROC-1 Cellsx Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite 6 hour r ] 2.0 Kaolinite 6 hour 1ZZZ3

Hontaorill. 1.5 6 hour IXXXJ

Erionite 6 hour 1.0 Y Z Z Z f t

CO00 nd a oi a 10 Clay Concentration mg/ml Figure LO 84

Figure 11 Comparison of FA release in ROC-1 cells by bentonite,

kaolinite, montmorilIonite and erionite. Incubation

time was 24 h. At 0.01 mg/ml of clay there was no

significance in FA levels. At 0.03 mg/ml of clay FA

levels in montmorillonite > all others. At 0.10 mg/ml

of clay FA levels in montmorillonite > all others and

erionite > bentonite. Level of significance p < 0.05. Fatty Acid Release in ROC-1 Cel1st Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite tzzn24 hour Koolinite 24 hour IZZZ]

Montaorill. 24 hour IKSEJ

Erionite 24 hour V 7 V Z \

a oi a 03 a 10 Figure l Clay Concentration ng/ml 86

Figure 12 FA release in N1E-115 cells by bentonite. At 1 h FA

levels in 0.10 mg/ml > all others. At 6 h FA levels in

0.01 and 0.03 mg/ml > control but not 0.10 mg/ml of

clay. At 24 h FA levels in 0.03 mg/ml > control and 0.10 mg/ml but not 0.01 mg/ml of clay. Level of

significance p < 0.05. Fatty Acid Release in N1E-115 Cells by Bentonite ^ F atty Acid nmol/mg protoin 1 hour ~

6 hour 25

24 hour 20

0 .0 aoi a 03 aio 2 Bentonite ng/ml Figure 88

Figure 13 FA release in N1E-115 cells by kaolinite. At 1 and 6 h

FA levels in 0.03 mg/ml > all others. At 24 h FA

levels in 0.01 mg/ml > all others. Level of

significance p < 0.05. Fatty Acid Release in N1E-115 Cells by Kaolinite F atty Acid nmol/mg p ro tein 1 hour

6 hour

24 hour

30 -

20 - j '

10 -

a o a oi 0.03 a 10 Kaolinite ng/nl Figure 3 90

Figure 14 FA release in N1E-115 cells by erionite. At 1 h FA

levels in 0.01 and 0.03 mg/ml > all others. At6 h FA

levels in 0.01 and 0.10 mg/ml> controls.At 24h FA

levels in 0.10 mg/ml > all others. Level of significance p < 0.05. Fatty Acid Release in N1E-115 Cells by Erionite ^ F atty Acid nmol/mg p ro tein 1 hour

8 hour 25

24 hour 20

0.0 0.01 0.03 0.10 Erionite mg/ml Figure 92

Figure 15 FA release in N1E-115 cells by montmorillonite. At 1 h

FA levels in 0.10 mg/ml > controls but not 0.01 and

0.03 mg/ml of clay. At 6 h FA levels in control > all

others. At 24 h there was no significance in FA

levels. Level of significance p < 0.05. Fatty Acid Release in N1E-115 Cells by MontnorilIonite Fatty Acid naol/ng protein 1 hour 30

6 hour 25

24 hour

0.0 aoi a 03 aio Montnoril Ionite ng/sl Figure 94

Figure 16 Comparison of FA release in N1E-115 cells by bentonite,

kaolinite, montmorillonite and erionite. The

incubation time was 1 h. At 0.01 mg/ml of clay FA

levels in montmorillonite > all others. At 0.03 mg/ml

of clay FA levels in kaolinite > bentonite and erionite

but not montmorillonite. At 0.10 mg/ml of clay FA

levels in montmorillonite > bentonite > erionite and

kaolinite. Level of significance p < 0.05. Fatty Acid Release In N1E-115 Cel 1st Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite 1 hour [ :

Kaolinite 1 hour IZZZ3

Hontaorlll. 1 hour 1ZZX3

Erionite 1 hour

I 0.01 0.03 a 10 Clay Concentration eg/el Figure 16 96

Figure 17 Comparison of FA release in N1E-115 cells by bentonite,

kaolinite, montmorillonite and erionite. The

incubation time was 6 h. At 0.01 mg/ml of clay FA

levels in bentonite, kaolinite and erionite >

montmorillonite. At 0.03 mg/ml of clay FA levels in

kaolinite > bentonite > erionite > montmorillonite. At

0.10 mg/ml of clay FA levels in erionite and bentonite

> montmorillonite but not kaolinite. Level of

significance p < 0.05. Fatty Acid Release in N1E-115 Cellsi Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite 6 hour 3.5 r z z ]

Kaolinite 6 hour 3.0 177 7 3

Montaorill. 2.5 6 hour 1X7 X3 2.0 Erionite 6 hour W 77ZK 1.5

1.0

0.01 0L03 OLIO Clay Concentration ag/nl Figure 17 98

Figure 18 Comparison of FA levels in N1E-115 cells by bentonite,

kaolinite, montmorillonite and erionite. The

incubation time was 24 h. At 0.01 mg/ml of clay FA

levels in kaolinite > bentonite and montmorillonite but

not erionite. At 0.03 mg/ml of clay there was no

significance in FA levels. At 0.10 mg/ml of clay FA

levels in erionite > all others. Level of significance p < 0.05. Fatty Acid Release in N1E-115 Cells: Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite 24 hour r r u

Kaolinite 24 hour 1ZZ Z 3

Montaorlll. 24 hour IXXX3

Erionite 24 hour Y7ZV7A

aoi 0.03 a 10 Figure 18 Cloy Concentration ag/nl 100

Figure 19 FA release in HUVE cells by bentonite. At 1 and 6 h

there was no significance in FA levels. At 24 h FA

levels in 0.10 and 0.03 mg/ml > 0.01 mg/ml > control.

Level of significance p < 0.05. Fatty Acid Release in HUVE Cells by Bentonite F atty Acid nmol/mg p ro tein 200 1 hour

175 8 hour

150 24 hour 125

100

0.0 0.01 0.1 Bentonite mg/ml Figure 9 102

Figure 20 FA release in HUVE cells by kaolinite. At 1 and 6 h

there was no significance in FA levels. At 24 h FA

levels in 0.10 mg/ml > 0.01 mg/ml of clay and control

but not 0.03 mg/ml of clay. Level of significance p < 0.05. Fatty Acid Release in HUVE Cells by Kaolinite Fatty Acid naol/sg protein 1 hour 150

6 hour 125

24 hour 100

0L0 0.1 Figure 0 Kaolinite Mg/el 104

Figure 21 FA release in HUVE cells by erionite. At 1 h FA levels

in 0.03 mg/ml > control. At 6 h FA levels in 0.01,

0.03, and 0.10 mg/ml > control. At 24 h FA levels in

0.10 mg/ml > all others. Level of significance

p < 0.05. Fatty Acid Release in HUVE Cells by Erionite Fatty Acid neol/ng protein 1 hour 125

8 hour 100

24 hour

0.0 0.01 ao3 CL1 Erionite ng/nl Figure !1 106

Figure 22 FA release in HUVE cells by montmorillonite. At 1 and

6 h there was no significance in FA levels. At 24 h FA

levels in 0.10 and 0.03 mg/ml > 0.01 mg/ml > control.

Level of significance p < 0.05. Fatty Acid Release in HUVE Cells by MontmorilIonite

Fatty Acid nmol/mg protein 1 hour 700

8 hour 600

24 hour 500

400

300

200

100 -

0.0 a oi 0.1 Figure 22 MontnorilIonite mg/ml 108

Figure 23 Comparison of FA release in HUVE cells by bentonite,

kaolinite, montmorilIonite and erionite. Incubation

time was 1 h. At 0.01 mg/ml of clay there was no

significance in FA levels. At 0.03 mg/ml of clay FA

levels in erionite > kaolinite > bentonite >

montmorilIonite. At 0.10 mg/ml FA levels in erionite >

all others. Level of significance p < 0.05. Fatty Acid Release in HUVE Cells: Comparison of Clay Type

Clay/Control Fatty Acid Levels Bentonite 1 hour r z z u

Kaolinlte 1 hour 2.5 [ 7 7 7 1

Montnorill. 1 hour 2.0 1X K E 1

Erionite 1 hour 1.5 UZZZA

1.0 ★ ★

a oi a 03 Figure 23 Clay Concentration mg/ml 110

Figure 24 Comparison of FA release in HUVE cells by bentonite,

kaolinite, montmorilIonite and erionite. Incubation

time was 6 h. At all clay concentrations FA levels in

erionite > all others. Level of significance p < 0.05. Fatty Acid Release in HUVE Cellsi Comparison of Clay Type Clay/Control Fatty Acid Levels Bentonite 6 hour { = □ 3.0 Kaolinite 6 hour ( 7 7 7 1 2.5

Hontaorill. 6 hour K X X 3 2.0

Erionite 6 hour 1.5 £ 7 7 3

1.0

a 01 OLIO Clay Concentration mg/el Figure 24 112

Figure 25 Comparison of FA release in HUVE cells by bentonite,

kaolinite, montmorilIonite and erionite. Incubation

time was 24 h. At 0.01 mg/ml of clay FA levels in

bentonite > erionite and montmorilIonite > kaolinite.

At 0.03 mg/ml of clay FA levels in montmorilIonite >

all others and bentonite > kaolinite but not erionite.

At 0.10 mg/ml of clay FA levels in montmorilIonite >

all others. Level of significance p < 0.05. Fatty Acid Release in HUVE Cells: Comparison of Clay Type

Clay/Control Fatty Acid Levels Bentonite u 24 hour Cl." .1 10

Kaolinite a 24 hour IZZZJ 8 Hontaorill. 24 hour 7 KXXJ 6 Erionite 24 hour 5 YZV77A 4

a 01 a 03 a 10 Figure !5 Clay Concentration ag/nl 114

Figure 26 Comparison of FA release by bentonite in HUVE, ROC-1

and N1E-115 cells. Incubation time was 1 h. At 0.01

and 0.03 mg/ml of clay FA levels in HUVE and N1E-115 >

ROC-1. At 0.10 mg/ml of clay FA levels in N1E-115 >

HUVE > R0C-1. Level of significance p < 0.05. Bentonite Toxicity in Cells Comparison of Cell Type Clay/Control Fatty Acid Levels HUVE 1 hour c ]

ROC-1 1 hour (7 Z Z J

N1E-115 1 hour KXXJ

LZZZ j a 01 a 03 a 10 Bentonite sg/nl Figure !6 116

Figure 27 Comparison of FA release by bentonite in HUVE, ROC-1,

and N1E-115 cells. Incubation time was 6 h. At 0.01

mg/ml of clay FA levels in N1E-115 > HUVE > ROC-1. At

0.03 and 0.10 mg/ml of clay FA levels in N1E-115 > HUVE

and ROC-1. Level of significance p < 0.05. Bentonite Toxicity in Cells Comparison of Cell Type Clay/Control Fatty Acid Level 9 HUVE 6 hour L ]

ROC-l 6 hour IZZ73

N1E-115 6 hour IXXX]

a oi a 03 a 10 Bentonite ncj/nl Figure n 118

Figure 28 Comparison of FA levels by bentonite in HUVE, ROC-1,

and N1E-115 cells. Incubation time was 24 h. At 0.01

mg/ml of clay FA levels in HUVE > N1E-115 > ROC-1. At

0.03 mg/ml of clay FA levels in HUVE > N1E-115 and

ROC-1. At 0.10 mg/ml of clay FA levels in HUVE >

N1E-115 > ROC-1. Level of significance p < 0.05. Bentonite Toxicity in Cells Connorison of Cell Type Clay/Control Fatty Acid Levels HUVE c24 hour ROC-1 24 hour IZZZ3

N1E-115 24 hour

a oi a 03 a 10 Bentonite sg/nl Figure 28 120

Figure 29 Comparison of FA levels by kaolinite in HUVE, ROC-1 and

N1E-115 cells. Incubation time was 1 h. At 0.01 mg/ml

of clay FA levels in HUVE and N1E-115 > ROC-1. At 0.03

mg/ml of clay FA levels in N1E-115 > HUVE and R0C-1.

At 0.10 mg/ml of clay there was no significance in FA

levels. Level of significance p < 0.05. KaolinitG Toxicity in Cg IIs Comparison of Cell Type Clay/Control Fatty Acid Levels HUVE 1 hour n u n

ROC-l 1 hour IZZZJ

N1E-115 1 hour [XXX]

&G3 Figure 29 Kaolinite eg/ml 122

Figure 30 Comparison of FA levels by kaolinite in HUVE, ROC-1 and

N1E-115 cells. Incubation time was 6 h . At 0.01 and

0.03 mg/ml of clay FA levels in N1E-115 > ROC-1 and

HUVE. AT 0.10 mg/ml of clay there was no significance

in FA levels. Levels of significance p < 0.05. Kaolinite Toxicity in Cells Comparison of Coll Type

Clay/Control Fatty Acid Levels HUVE 6 hour [ 1

ROC-t 6 hour (ZZZ1

H1E-115 6 hour IXXX]

% £ £ ara 2 Kaolinite ng/nl Figure 30 124

Figure 31 Comparison of FA levels by kaolinite in HUVE, ROC-1,

and N1E-115 cells. Incubation time was 24 h. At 0.01

35 of clay FA levels in N1E-115 > ROC-1 and HUVE.

At 0.03 mg/ml of clay there was no significance in FA

levels. At 0.10 mg/ml of clay FA levels in HUVE >

R0C-1 but not N1E-115. Level of significance p < 0.05. Kaolinite Toxicity in Cells Comparison of Cell Type

Clay/Cootrol Fatty Acid Levels HUVE 24 hour

ROC-1 24 hour

N1E-115 24 hour (XXX3

a oi are a 10 Kaolinite eg/el Figure 31 126

Figure 32 Comparison of FA levels by erionite in HUVE, ROC-1, and

N1E-115 cells. Incubation time was 1 h. At 0.01 and

0.03 mg/ml of clay there was no significance in FA

levels. At 0.10 mg/ml of clay FA levels in HUVE and

R0C-1 > N1E-115. Level of significance p < 0.05. Erionite Toxicity in Cells Comparison of Cell Type

Clay/Control Fatty Acid Levels HUVE 1 hour L j

ROC-1 1 hour IZZZ3

N1E-115 1 hour 1XXX3

2 0.01 a 03 a 10 Erionite ag/al Figure > 128

Figure 33 Comparison of FA levels by erionite in HUVE, ROC-1, and

N1E-115 cells. Incubation time was 6 h. At 0.01 and

0.03 mg/ml of clay FA levels in HUVE > N1E-115 and

ROC-1. At 0.10 mg/ml of clay there was no significance

in FA levels. Level of significance was p < 0.05. Erionite Toxicity in Cells Comparison of Cell Type

Clay/Control Fatty Acid Levels HUVE 6 hour n

ROC-l 6 hour 1 7 7 7 ]

N1E-115 6 hour 1X7X1

1.0 -

a 01 a 03 a 10 Erionite ag/al Figure 13 130

Figure 34 Comparison of FA levels by erionite in HUVE, ROC-1, and

N1E-115 cells. Incubation time was 24 h. At 0.01 and

0.03 mg/ml of clay there was no significance in FA

levels. At 0.10 mg/ml of clay FA levels in HUVE and

N1E-115 > R0C-1. Level of significance p < 0.05. Erionite Toxicity in Cells Comparison of Cell Type Clay/Control Fatty Acid Levels HUVE 24 hour C ]

ROC-I 24 hour IZ7 7 ]

N1E-115 24 hour 1XX X ]

% a oi a 03 a 10 Erionite ag/al Figure 4 132

Figure 35 Comparison of FA levels by montmorilIonite in HUVE,

ROC-1, and N1E-115 cells. Incubation time was 1 h. At

0.01 and 0.03 mg/ml of clay there was no significance

in FA levels. At 0.10 mg/ml of clay FA levels in

N1E-115 and ROC-1 > HUVE. Level of significance

p < 0.05. MontmorillonitQ Toxicity in Cellss Comparison of Cell Type

Clay/Control Fatty Acid Ls v q I s HUVE 1 hour L 3

ROC-1 1 hour

N1E-115 txxxi1 hour

a oi a 03 a 10 Figure 15 MontmorilIonite ag/al 134

Figure 36 Comparison of FA release by montmorilIonite in HUVE,

ROC-1, and N1E-115 cells. Incubation time was 6 h. At

0.01 mg/ml of clay FA levels in HUVE > N1E-115 and

ROC-1. At 0.03 mg/ml of clay FA levels in ROC-1 >

N1E-115 and HUVE. At 0.10 mg/ml of clay FA levels in

ROC-1 > N1E-115 but not HUVE. Level of significance

p < 0.05. MontmorilIonite Toxicity in Cg IIsi Comparison of Cell Type Clay/Control Fatty Acid Levels HUVE 6 hour c ] ROC-1 6 hour 177 7 ]

N1E-115 6 hour IXZXI

a 01 0.0 3 a 1 0 Figure 36 MontmorilIonite mg/ml 136

Figure 37 Comparison of FA release by montmorilIonite in HUVE,

ROC-1, and N1E-115 cells. Incubation time was 24 h.

At 0.01 mg/ml of clay FA levels in ROC-1 > HUVE and

N1E-115. At 0.03 and 0.10 mg/ml of clay HUVE > N1E-115

and ROC-1. Level of significance p < 0.05. MontmorilIonite Toxicity in Cell9i Comparison of Cell Type Clay/Control Fatty Acid Levels HUVE 24 hour L ] 10

ROC-l 24 hour IZZZ1 8 N1E-115 24 hour (XXX3 6

MontnorilIonite ag/al 37 138

Table 10 Clay induced LDH release in ROC-1, N1E-115, and HUVE cells

Clay (0.10 mg/ml) LDH Activity (nmol/min/mg protein)

HUVE N1E-115 R0C-1

Bentonite 1307 ± 126* 1246 + 493 222 ± 41

Kaolinite 1350 + 84* 2738 ± 623 143 ± 21

MontmorilIonite 1393 ± 41* 1790 + 524 135 + 35

Erionite 958 + 111 2001 ± 1098 163 ± 52

Control 927 + 0 1965 ± 830 190 ± 24

LDH activity was measured in media samples immediately after the experiment. Assay was done in a temperature control cuvette holder at a temperature of 30°C. The concentration of each clay was 0.10 mg/ml and the incubation time was 24 h. For each experiment n=3 except N1E-115 montmorilIonite and HUVE control n=2. The * indicates significantly greater than the control at p < 0.05. Values are means ± standard deviation. 139

Table 11 Clay induced cell death in N1E-115 and HUVE cells

N1E-1159 HUVEb % non-viable % non-viable

Control 22 0

Bentonite 23 94*

Kaolinite 28 90*

Montmorillonite 35* 99*

Erionite 16* 20* a Viability assessed by Trypan Blue exclusion. b Viability assessed by comparison of the number of cells in control vs. clay microscopic field. See Plates VI - X. Clay concentration was 0.10 mg/ml and incubation time was 24 h. Values represent percent of dead cells. The sample size was n=3. The * indicates significantly different from control at a level of p < 0.05. 140

Plate I ROC-1 cells incubated with 0.10 mg/ml of

montmorillonite for 24 h. There was no evidence of

cell damage, but cells were heavily coated with clay.

Magnification was lOOx using phase contrast microscopy. 141

Plate I

/f e , 5 ^ . ;•. > 'O:^- 142

Plate II ROC-1 cells incubated with PBS buffer for 24 h. There

was no evidence of cell damage. The cells were intact

and contained processes. Magnification was lOOx using phase contrast microscopy. Plate I I 144

Plate III HUVE cells incubated with 0.10 mg/ml of bentonite for

24 h. There was a decrease in the number of cells

compared to the control indicating cell death. Some

cells were covered with clay particles but s till appear

to be viable. Remaining cells were elongated

indicating a state of stress. Magnification was lOOx

using phase contrast microscopy.

146

Plate IV HUVE cells incubated with 0.10 mg/ml of kaolinite for

24 h. There was a decrease in the number of cells

compared to the control indicating cell death. The

lysed cells were coated with clay particles and were

light in color. Some viable cells remained but were

elongated indicating a state of stress. Magnification

was lOOx using phase contrast microscopy. Plate IV 148

Plate V HUVE cells incubated with 0.10 mg/ml of erionite for

24 h. There was a slight decrease in the number of

viable cells compared to the control. The clay was

associated with the cell membranes, however, many of

the cells appear to be viable. The cells were not

elongated indicating erionite did not cause stress on

the cells as the other clays did. Note the crystalline

structure of erionite. Magnification was 200x using phase contrast microscopy. Plate V 150

Plate VI HUVE cells incubated with 0.10 mg/ml of montmorilIonite

for 24 h. MontmorilIonite caused nearly complete cell

death. All lysed cells were covered with the

montmorilIonite. These light colored cell fragments

were covered on both sides by the clay. Several

elongated cells did appear in the field. Magnification

was lOOx using phase contrast microscopy. PI ate 152

Plate VII HUVE cells incubated with PBS buffer. These control

cells appeared to be normal. The cells were not

elongated and retained an oval shape. Magnification

was lOOx using phase contrast microscopy. Plate V II 154

Discussion

Aluminum silicate containing clays caused cellular necrosis in

HUVE cells but not ROC-1 or N1E-115 cells. The HUVE cell death occurs after 6 h, but before 24 h following addition to the cultures. ROC-1 and N1E-115 cells were perturbed by some clays as indicated by FA release, however, this perturbation did not result in cell death indicating a reversible trauma.

The aluminum silicate containing clays, bentonite, kaolinite, erionite and montmorilIonite, were tested for their toxicity towards cells in culture. The primary purpose was to determine the toxicity of the clays toward cells of neuronal origin, thereby providing a possible link to AD. An environmental causative agent such as aluminum silicate could explain the heterogeneity of AD since no clear genetic cause has been determined (German et a !., 1987). Since aluminum silicates are associated with the plaques in AD (Candy et a l., 1986), th eir toxicity towards cells needed to be examined to determine if aluminum silicates play a role in the onset of AD.

Four methods were employed to determine if cell viability changes had occurred during the various posttreatment periods. Lactate dehydrogenase activity was measured. Lactate dehydrogenase is an ubiquitous cytosolic enzyme which is commonly used as a marker for membrane integrity (Yu et a l., 1989). Trypan blue exclusion was used to further determine cell viability. A cell with a compromised membrane is unable to exclude the Trypan blue dye, while an uncompromised cell can exclude the dye for a period of time. 155

Morphological changes were monitored using phase contrast microscopy. The cells were examined for shape change and the integrity of the cellular membrane. The last method used was the quantification of FA release into the cells and medium. This method has been used to determine FA release in several central nervous system traumas, particularly cerebral ischemia (Yoshida et a l., 1980;

Ikeda et a l., 1986) and spinal cord injury (Demediuk et a l., 1985a,b;

Saunders and Horrocks, 1987). In these traumas, FA release is reversible if the release does not exceed certain levels. A good correlation has been established between the extent of trauma and the level of FA release.

Bentonite had a limited toxic effect on N1E-115 and ROC-1 cells.

In N1E-115 cells the clay caused a dose-dependent-like release of FA after 1 h of incubation (Figure 12). This release was not time- dependent. This suggests that the FA released were reacylated by the cells or otherwise metabolized. At 6 h, there were no large increases in FA levels above the control levels, indicating that no sustainable injury had occurred. These data correlated well with the morphological, LDH and Trypan blue studies which indicated no membrane damage occurred to the extent of causing the membrane to be compromised. However, in HUVE cells, a dose-dependent release of FA occurred after 24 h of incubation (Figure 19). The effect of the clay upon the cells was not immediate as indicated by the lack of morphological changes and FA release after 1 and 6 h of incubation.

By 24 h the integrity of the cells was completely lost with over 93% of the cells dead. The remaining cells were covered with bentonite 156

and appeared to be normal. The cells which were destroyed contained

no discernible intracellular organelles and the membranes were

covered on both sides with clay particles. The lytic process was so

complete that no cellular ghosts were found.

Like bentonite, kaolinite also had a limited effect on N1E-115

and ROC-1 cells. Kaolinite caused an apparent dose-dependent release

in both cell types (Figures 6 and 13). This response was greatest in

N1E-115 cells (Figures 29-31). The response in N1E-115 cells was

clearly dependent upon the concentration of the clay. The greatest

release in N1E-115 cells was at a concentration of 0.03 mg/ml at 1

and 6 h. At 24 h the maximum release was at 0.01 mg/ml (Figure 13).

As the kaolinite concentrations increase above those causing the maximal release of FA, the FA levels decrease. The same phenomena

occurred in ROC-1 cells although all FA maximums were at the clay

concentration of 0.03 mg/ml and were time dependent (Figure 6). One

explanation for these data is that at higher kaolinite concentrations

the clay irreversibly bound the released FA decreasing the amount

seen in the cells and medium. This would explain the dose-dependent

like response followed by a rapid drop in FA. However, if this were

the case one would expect to see a change in LDH levels in the medium

and changes in Trypan blue exclusion. These changes were not seen.

An extrapolation of the curves of Figures 13 and 6 would result in

approximate 6-fold and 3-fold increases in FA release respectively.

In HUVE cells, kaolinite at 0.10 mg/ml with a 24 h incubation caused

a 1.7-fold increase in FA levels which resulted in an increase in LDH

levels and 90% cell death (Figure 20, Tables 10 and 11). At 24 h the 157 curve was linear indicating that kaolinite was not binding released

FA. Although kaolinite caused only a small change in FA levels, these small changes were accompanied by a high proportion of cell mortality. Morphologically the HUVE cells were completely lysed with only membrane fragments remaining. These fragments were covered on both sides with the clay.

Erionite caused FA release in N1E-115 cells. This response occurred after 24 h of incubation and at the highest concentration of erionite. No response was seen at lower concentrations or earlier time points (Figure 14). Morphologically no major changes were seen after 24 h of incubation with 0.10 mg/ml of erionite. No significant

LDH release or cell death occurred indicating that this concentration of erionite caused only a minor perturbation of the membrane. In

ROC-1 cells, the response to erionite was immediate. FA levels increased after 1 and 6 h of incubation in a dose dependent-like manner (Figure 7). This increase was not maintained and, as in

N1E-115 cells, appeared to be only a minor perturbation of the membrane resulting in no cell death as indicated by the LDH, Trypan blue and morphological data. HUVE cells responded to erionite by releasing FA. The initial release was time-dependent, however, the lowest clay concentrations at 24 h showed very little increase in FA and were below the 1 and 6 h curves. By 24 h erionite at 0.10 mg/ml concentration increased FA release significantly. Morphologically,

20% of the cells appeared to be dead, however, the LDH levels were comparable to control levels. Thus, while erionite caused FA release in all three cell types and limited cell death in HUVE cells, these 158 changes result only in a perturbation of the membrane. This perturbation was not sufficient to cause the membrane integrity of the majority of the cells to be compromised.

Montmorillonite caused little to no FA release in N1E-115 cells

(Figure 15). Trypan blue exclusion and LDH data indicated that the cells remained viable even after 24 h of treatment with 0.10 mg/ml of montmorillonite. In ROC-1 cells, montmorillonite caused a rapid increase in released FA. This cellular response to montmorillonite was dose-dependent, but showed no time-dependence. Although the release was immediate, no LDH release or morphological changes occurred. In HUVE cells, the clay caused no release of FA at 1 and 6 h of incubation at all clay concentrations. At 24 h, there was a rapid increase in FA levels (Figure 22). This increase correlated well with LDH release and the morphological changes. Like bentonite and kaolinite, montmorillonite caused the cells to be lysed. The lysis was nearly complete with 99% of the cells dead. The only remaining visible remnants of the cells were membrane fragments covered on both sides with clay.

Although erionite caused a significant increase in FA levels over control levels at various concentrations and incubation times in HUVE cells, this release did not cause an appreciable amount of cell death. Bentonite and kaolinite were equally toxic in HUVE cells, although kaolinite's FA release was significantly less than bentonite's. Montmorillonite proved to be most toxic, and like bentonite and kaolinite, the greatest release did not occur until after 24 h of incubation. Each of these clays were associated with 159 the cell membrane after 1 h of incubation. This association continued through the 6 h of incubation time. In N1E-115 and ROC-1 cells the accumulation of clay on the cells was comparable to that seen in HUVE cells, however, the clay induced FA release in some

instances occurred quickly. This FA release indicates that the clays perturb the membrane although this perturbation did not result in cell death.

The mechanism for the cellular toxicity of the clays is not known. Several hypotheses have been put forward (Woodworth et a l.,

1982; Oscarson et a l., 1986; Davies et a l., 1984; Davies and Preece,

1983). These data indicate that FA release alone may not result in cell death. The ROC-1 and N1E-115 studies show that in itial FA release was not lethal to the cells. The released FA appeared to be reacylated, further supporting the hypothesis that clay-induced FA release does not necessarily equate with cell death. HUVE cells have a delayed release of FA, which correlated with substantial cell death. Thus, two different mechanisms may be involved in the release of the FA. The first is a rapid FA release within the first 6 h of incubation with the clays. This mechanism may be limited in its ab ility to cause enough FA release to compromise the cell membrane.

The second is a slower mechanism. This mechanism results in FA release and cell death. This mechanism is slower to develop but produces rapid cell death.

One mechanism which the clays use may be a charge interaction. A negative charge interaction of montmorillonite with a cell membrane is essential for lysis of the membrane. No proteins are needed for 160 this lysis to occur because the clay can lyse liposomes (Woodworth et a l., 1982). This association of the negatively charged clay and the membrane can be negated by the addition of Al^+ which binds to the negatively charged clay (Oscarson et a l., 1986). The process in erythrocytes appears not to involve osmotic swelling because there are no red blood cell ghosts present (Oscarson et a l., 1986). Our data with HUVE cells correlates well with this hypothesis. No cell ghosts were seen after treatment with kaolinite, bentonite and montmorillonite for 24 h at a concentration of 0.10 mg/ml. Unlike in tracheal epithelial cells, no phagocytosis of the clay particles was seen in any of the cells tested (Woodworth et a l., 1982).

Kaolinite, like montmorillonite, contains both negative and positive charges. If the positive charge on kaolinite is removed by treating with polyacrylic acid, lysis by kaolinite still occurs

(Davies et a l., 1984). Hence the positive charge is not involved in the lysis of the cell. The negative charge is more likely involved since coating montmorillonite with Al^+ limits erythrocyte lysis

(Oscarson et a l., 1986). Polyvinyl-pyridine-N-oxide (PVPNO) coated clays are not toxic. PVPNO is thought to decrease hydrogen bonding of the clay with various membrane constituents. PVPNO coated clays are phagocytized by macrophages but are not toxic (Davies et a l.,

1984). Thus, clay-induced lysis occurs after binding to the membrane via a charge interaction or through hydrogen bonding.

The time course for lysis varies depending on the complexity of the cell. Red blood cells lyse quickly with 50% of the lysis complete after 10 min. By 40 min the process is complete (Oscarson et al., 1986). Epithelial cells take longer with a maximum release of ®*Cr at 24 h (Woodworth et a l., 1982). HUVE cell lysis was complete between 6 and 24 h, putting the ly tic process on the same time scale as for the epithelial cell. Within 24 h no apparent lysis occurred in the N1E-115 and ROC-1 cells. This does not rule out lysis after longer periods of exposure. Montmorillonite in the ROC-1 and erionite in the N1E-115 cells may cause lysis with longer incubation times because each caused a dose-response-like release of

FA by the cells (Figure 8 and Figure 14). The more complex the cell, the longer the time required to develop the mechanism which induces lysis.

The clays, through a charge interaction, may interact with the membrane to cause a change in membrane dynamics. The negatively charged clays could interact with the positively charged groups in the membrane to cause a coalescing effect on the membrane.

Phosphatidylcholine and sphingomyelin make up the bulk of the outer leaflet phospholipids (Bretscher, 1973; Bergelson and Barsukov,

1977). Both have a positively charged moiety which could bind the negatively charged clays. This binding would restrict the lateral motion of these bulk phospholipids, causing a definite change in membrane fluidity. Change in membrane fluidity has been hypothesized to change receptor function (Bell et al., 1980; Raber and Bast,

1989). Many cellular mechanisms are dependent upon membrane fluidity. These include the ATPases, adenylate cyclase and transport mechanisms (Stubbs and Smith, 1984). A disruption of the normal 162

fluid state of the membrane by imposing a rigidification through clay

binding could lead to cellular dysfunction and cell death.

The mechanism for cell death from exposure to asbestos may

include binding of the Mg^+ moiety to extracellular carboxylic acid groups. This would cause a clustering of membrane proteins forming

channels permitting unregulated ion flux (Harington et a l., 1975).

EDTA has been shown to decrease the activity of asbestos, presumably

by binding the Mg^+ in the crystalline structure (Macnab and

Harington, 1967). However, leaching the Mg^+ out of the asbestos causes an increase in its negative charge and a corresponding

increase in ly tic activity (Jayawardena and Waksman, 1977). This discrepancy in charge state does not rule out a charge interaction with the membrane to cause lysis, merely i t makes the molecules different to which the asbestos binds. The outer leaflet possesses negatively charged sia lic acid groups and positively charged choline groups in abundance. Evidence does show that asbestos causes osmotic swelling in cells supporting the claim of carboxylic acid binding to form ion channels (Harington et al., 1975). However,because the negatively charged Mg^+ leached asbestos also causes cell lysis, it must do so through a interaction with positively charged groups such as phosphatidylcholine. Thus, asbestos toxicity towards cells may be due to two different mechanisms depending on the charge state of the mineral.

Silica is known to interact with membrane phospholipids and cholesterol through hydrogen bonding to the carbonyl and phosphate moieties of the phospholipid and the hydroxyl on the cholesterol 163

(Harington et al., 1975; Harley and Margolis, 1961; Nash et al.,

1966). Aluminum silicate containing clays may also interact through

hydrogen bonding to these groups. The hydrogen donor could be the

Brdnsted acid site on the clay. Thus, the mechanism by which the

aluminum silicate containing clays interact with the membrane is

unknown. Evidence indicates a charge interaction between the

negatively charged clay and positively charged outer lea fle t. The

clay may also hydrogen bond to the membrane, since the clays readily

hydrogen bond to PVPNO.

After binding to the membrane, the mechanism for cell death is

not known. We have established that aluminum silicate clays cause

the release of FA. This release will cause a fluidization of the membrane (Klausner et a l., 1980; Stubbs and Smith, 1984). As the

fluidity increases in parts of the membrane, Ca2+ entry may occur due to the permeabilization of the membrane (Yorio et a l., 1983).

The increased hydrolysis of phospholipids will also cause the membrane to become semipermeable. The method of hydrolysis could be due to the activation of phospholipases through the initial decrease

in membrane fluidity caused by the binding of the clay to the membrane. Once activated, these phospholipases will hydrolyze

phospholipids causing the membrane to become permeable. Our data

indicates the primary phospholipases activated were possibly phospholipase A2 and phospholipase Aj. These phospholipases would release the FA 16:0, 18:0 and 18:1 from the sn-1 and 20:4 n-6,

22:6 n-3, 20:3 n-6, 18:3 n-6, 18:3 n-3, and 18:2 from the sn-Z position of the phospholipid. The sn-1 position FA were released in 164 greater numbers indicating a phospholipase Aj-like mechanism. In

HUVE cells a greater release in FA from the sn-2 position indicates a phospholipase A2~like mechanism was also induced. As the membrane becomes increasingly permeable, Ca^+ entry increases resulting in eventual cell death (Piper, 1988; Laiho et al., 1983; Starke et al.,

1986; Trump and Berezesky, 1983).

Aluminum silicate clays can hydrolyze peptide bonds (Crapper

McLachlan, 1986). The clays act as Bronsted and Lewis acids (Wright et a l., 1985; Haag et a l., 1984; Crapper McLachlan, 1986). The bond enthalpy for the peptide bond is 292 kJ/mol. The bond enthalpy for the carbon-oxygen bond is 360 kJ/mol (Barrow, 1979). Thus, the clays may also be able to directly cleave the sn-1 C-0 ester bond. The choline head group could bind the clay through a charge interaction orienting the montmorillonite and other amorphous aluminum silicate clays in such a way as to cause the cleavage of the accessible sn-1

FA. The sn-1 FA in most phospholipid models is more accessible with less conformational steric hindrance than is the sn-2 position

(Stubbs and Smith, 1984). This hypothesis explains the large increase of FA that are generally found in the sn-1 position. Thus, aluminum silicate clays may produce their effect by a direct interaction with phosphatidylcholine by hydrolyzing the sn-1 FA thereby increasing membrane fluidity.

Aluminum silicate containing clays have been hypothesized to play a role in AD (Roberts, 1986; Crapper et a l., 1976; Wilcock, 1988;

Perl and Good, 1987). The main tenet of their proposals is that aluminum silicate containing clays are transported retrogradely into 165

the olfactory region within the brain. In AD patients, the olfactory

bulb is indeed compromised and shows signs of abnormal morphology

(Talamo et a l., 1989; Esiri and Wilcock, 1984; Altman, 1989). This

compromised olfactory bulb would permit the transport of the aluminum

silicates into the brain regions involved in AD pathology. Once in

the brain, aluminum silicates are thought to act as a neurotoxin to

cause or at least aid in causing the onset of AD. Aluminum silicates

are found in the senile plaques of AD patients indicating possible

involvement of the aluminum silicate in AD (Candy et a l., 1986).

Aluminum silicate containing clays were not toxic to cells of

neuronal origin. HUVE cells were very susceptible to lysis by the

clays, but are not of neural origin. The HUVE cells were primary

cells which may be one reason for their lysis by kaolinite, bentonite

and montmorillonite. Although ROC-1 cells are a hybrid cell

exhibiting many primary cell characteristics, they are still not a

true primary cell. N1E-115 cells are from a transformed cell line.

Thus, the lack of clay neurotoxicity towards N1E-115 and ROC-1 cells may be due to their origin. For unknown reasons, non-primary cells may not be susceptible to cell death induced by aluminum silicates even though the clays bind to the cell's membrane and cause a moderate amount of FA release. Therefore, i t is impossible to conclude whether aluminum silicate containing clays are toxic toward cells of neuronal origin. The clays are toxic to HUVE cells, and this may be true for cerebral endothelial cells as well.

The other aluminum hypothesis involves the solubilization and absorption of environmental aluminum, mainly through the digestive 166 system. The aluminum ion, when injected into rat brains, causes encephalopathy (Lipman et a l., 1988). Thus, i t has been hypothesized that aluminum silicates, when ingested, are solubilized by citrate

(Birchall and Chappell, 1988b; Martin, 1986). Once in the blood stream, the aluminum is carried by ferritin to the brain, crosses the blood brain barrier and enters the cells (Birchall and Chappell,

1988a,b). The Al3+ ion can bind to ATP 10^ times more strongly than the Mg2+ ion. The binding of Al3+ to ATP renders the molecule inactive. Al3+ also binds to the phosphates on the inositol ring of the phosphatidyl 4,5 b/sphosphate. The phosphayidyl

4,5 b/sphosphates are involved in signal transduction, therefore, the binding of Al3+ to the phosphates disrupts receptor function

(Birchall and Chappell, 1988a). Aluminum binds to DNA and has been shown to lim it some messenger RNA synthesis (Crapper McLachlan, 1986) and ADP-ribosylation (Seigle and Haug, 1983). It is thought that this is the primary pathophysiology of the Al3+ in the cell and may be part of the etiology of AD.

This new hypothesis is considerably different from my former hypothesis and was not tested in my system. I did not attempt to solubilize the clay using citrate. Instead the clays were used in an insoluble form and kept in a suspension, then added to the cells.

This methodology more closely follows the initial hypothesis that the clays are transported in a retrograde fashion into the brain.

More experiments are needed to fully examine the interaction of clays with cell membranes. These studies include the use of 2H-NMR to study membrane fluidity changes (Smith and Oldfield, 1984) and 167 fluorescence polarization to measure lipid order parameters (Heyn,

1979; Jahnig, 1979; Chan and Fishman, 1978). These techniques could be used to study membrane fluidity changes in cell membrane induced by clay binding. Scanning electron microscopy (SEM) should be done to see if the clays are phagocytized. SEM would also provide a detailed view of the binding of the clay to the membrane. Lipid metabolites such as prostaglandins, leukotrienes and platelet activating factor should be quantitated. These lipid mediators may be produced prior to cell death, and in vivo may cause changes in other cells. Various channel blockers should be used to determine if the clays cause ion channels to be preferentially opened or closed due to changes in membrane fluidity. Unregulated ion flux can be tested using radioactive ions. EDTA, PVPNO and coating of the clay with gold can be used to see if these agents disrupt the binding of the clay to the cell. It would be interesting to see if these agents block the action of the clay and preserve the cell. The possible phospholipase action of the clays needs to be examined in detail. If a specific phospholipase activity occurs with model membranes, this may be one mechanism through which the clays in itia te cell death.

All of these experiments would aid in understanding not only the mechanisms of clay binding, but also the mechanism by which clay induces cell death. CHAPTER IV

PRESSURE

Introduction

Spinal cord trauma is a devastating injury which often results in paralysis. Membrane damage is the first reported biochemical event during the cascade of destruction following trauma to the spinal cord. Membrane damage is characterized by an increase in free fatty acids, eicosanoid formation, phospholipid and cholesterol degradation. Maintenance of ion gradients is compromised by the decrease in ion pump activities. Further degradation of the membrane results in the inability to maintain the intracellular environment and results in cellular necrosis.

Spinal cord injury is most prevalent among young males between the ages of 20 to 34 (Bracken et a l., 1981). The annual rate of injury in the United States is 40.1 per million of population

(Bracken et a l., 1981). Worldwide rates vary between 13 to 50 cases per million people (Balentine, 1988). The primary cause of spinal cord injuries (SCI) is automobile accidents. Other sources of injury include falls, firearms, sporting activities (especially diving and football), recreational activities and workplace accidents

(Balentine, 1988). Annual direct and indirect costs have been

168 169 estimated to be $500,000 per injury. Unfortunately the cost of SCI is devastating both economically as well as socially.

Several in vivo models for inducing SCI exist. Most models use the compression of the spinal cord or impact upon the cord to produce an injury. Various animals have been used in these studies although the most prevalent are rats, dogs, and cats. A static load injury produces a compression trauma of the cord. Rats do not produce a consistent injury using a sta tic load (Black et a l., 1986), but this injury method is preferred when using cats (Anderson et al., 1980).

A mathematical model has been devised which predicts the outcome of an animal after static load injury (Kushner et a l., 1986). In this model 95% of the injury is attributed to the weight compressing the cord while 5% is due to duration of the weight on the cord. Several modifications of the weight drop impact trauma have been devised

(Ford, 1983; Black et a l., 1986). This model is capable of producing a graded injury by increasing the weight dropped or the height the weight is dropped. Each of these models have advantages and disadvantages. In today's political atmosphere perhaps the greatest disadvantage is the large number of animals used to perform experiments. This number will become larger as the search for an effective pharmacological treatment regimen continues.

Several theories have been proposed to explain the autodestructive cascade of events which lead to paralysis following

SCI. Essentially there are two schools of thought. The first maintains that the primary cause of SCI is the decrease in spinal cord blood flow resulting in an ischemic event which irreversibly 170 damages the cord. The second maintains that the primary event is the mechanical deformation of the cord which results in membrane damage.

More than likely the true etiology is a combination of these two hypotheses. However, no matter what the etiology the ultimate result is paralysis.

Following spinal cord injury the primary biochemical events are the increase in protease activity, an increase in intracellular

Ca^+ and lipid damage. Hemorrhage, metabolic changes and decreased spinal cord blood flow are secondary events which occur as a result of the primary events (Banik et a l., 1987). This theory is in agreement with the idea that the in itial sites of injury are the cellular and subcellular membranes (Saunders and Horrocks, 1987).

Thus, perhaps the focus for pharmacological treatment should be limitation of membrane breakdown and maintenance of the integrity of glial, neuronal and endothelial cells.

Lipid peroxidation is thought to play a role in SCI. The initial hemorrhage produced by trauma results in an extravasation of iron-containing blood that initiates and catalyzes lipid peroxidative reactions (Demopoulos et a l., 1984). This damage causes enzyme inhibition and microcirculatory damage leading to edema and spinal cord ischemia. Another proposal maintains that damage of the anterior spinal artery by anterior bone crush is the initial event.

Since the anterior artery accounts for 75% of the spinal cord's vascular supply, there is a concomitant increase in lactic acid and a decrease in spinal cord blood flow. As adenine nucleotide levels drop, membrane damage occurs which results in prostaglandin formation 171 and further decreases spinal cord blood flow (de la Torre, 1981). A combination of these events more than likely is the ultimate cause of spinal cord injury, although evidence indicates that the initial event is the breakdown of cellular membranes.

Free fatty acids increase during the f ir s t five minutes of recovery post-injury in the cat compression model. A large increase in arachidonic acid (20:4 n-6), palmitic acid (16:0), stearic acid

(18:0) and oleic acid (18:1) occurs following compression injury

(Demediuk et al., 1985a,b). These increases are initially found in the gray matter and spread radially into the white matter. A decrease in cholesterol (Demediuk et al., 1985a) and a rapid increase in diacylglycerols also occurs (Demediuk et a l., 1985b). A reduction in polyunsaturated fatty acids occurs after 30 minutes of recovery

(Demediuk et a l., 1985b). One source for removal of released FA is eicosanoid formation.

Thromboxane A2, a potent vasoconstrictor and platelet aggregator, is increased in SCI (Demediuk et al., 1985a,b; Demediuk and Faden, 1988; Hsu et a l., 1985b, 1986). No increases in prostacyclin, a vasodilator and platelet antiaggregant, occur following injury (Hsu et a l., 1985b, 1986; Demediuk et a l., 1985a,b;

Demediuk and Faden, 1988). Prostaglandin E 2 and prostaglandin

F2a, both vasoconstrictors, are increased 10- and 24-fold respectively (Demediuk et a l., 1985a). Peptidoleukotrienes are also increased in the cat compression model (Saunders and Horrocks, 1987) although no increase was found in the rat impact model (Demediuk and

Faden, 1988). Leukotriene B4, a chemoattractant for 172 polymorphonuclear leukocytes (PMNL), is increased following compression trauma (Means et a l., 1986). This results in PMNL in filtratio n and inflammation of the spinal cord.

Na+,K+-ATPase and Mg2+-ATPase activities decrease following

SCI (Demediuk et a l., 1985b; Faden et a l., 1987). The decrease of

Na+,K+-ATPase activity may cause edema formation. Edema formation is also due to endothelial cell damage (Hsu et al.,

1985a). There is a 20-fold increase over control of albumin extravasation at 2 hours post-injury. A slower but constant extravasation rate is seen between 2 and 18 hours post-injury.

Ca2+ levels are also increased during SCI (Hsu et al., 1985a;

Balentine, 1983). These excessive Ca2+ levels are lethal to glia and neurons and may have an impact on cell membrane structure and function (Balentine, 1983).

Ca2+ channel blockers have been used to alleviate the effect of increased tissue Ca2+. Nimodipine, a Ca2+ channel blocker, increases spinal cord blood flow by 18-25% only if mean systemic blood flow is maintained at normal levels (Guha et a l., 1987).

Nicardipine, another Ca2+ channel blocker, did not have an effect on spinal cord blood flow or on morphological changes (Black et a l.,

1988). Diltiazem and nifedipine, calcium channel blockers, cause an increase in spinal cord blood flow following SCI (Hall and Wolf, 1986).

Naloxone, an opiate receptor antagonist, has been reported to increase spinal cord blood flow (Faden et a l., 1981), but this result has not been reproducible (Robertson et a l., 1986; Wallace and Tator, 1986; Haghighi and Chehrazi, 1987). Methyl prednisolone sodium succinate, a synthetic glucocorticoid, has been used successfully to tre a t SCI in experimental animals (Anderson and Means, 1985; Means et al., 1981; Anderson et al., 1985). Post-injury treatment with methyl prednisolone sodium succinate has increased clinical recovery in cats (Means et a l., 1981). One possible mechanism for this action is the preservation of membrane cholesterol and the limitation of fatty acid degradation through lipid peroxidation. Prevention of necrosis and paralysis is the result of this decrease in lipid peroxidation (Anderson and Means, 1985). Maiondialdehyde, an indicator of lipid peroxidation, is also reduced by methyl- prednisolone (Anderson and Means, 1985).

Numerous pharmacological treatment regimens have been tried following SCI in an effort to lim it membrane breakdown and maintain physiological parameters such as spinal cord blood flow. Many of these treatments have had limited success. Membrane degradation is known to occur, but the contribution of each cell type in the spinal cord to the overall injury is unknown. Without this information, it is d iffic u lt to form an effective treatment regimen. Thus, an in vitro model for SCI needs to be developed in order to define the contribution of different cell types to the injury and to screen pharmacological agents for effectiveness in treating SCI. I have initiated the development of such a model which utilizes an increase in atmospheric pressure over the cells to mimic a mechanical compression. To further understand this model one must examine the 174 effects of pressure on model membranes, biological membranes, protein activities, ion transport and finally biological organisms.

The effects of pressure were first examined on frog spinal cord preparations. At 50 atm the effects of pressure begin. At 150-250 atm convulsions begin and at 300 atm paralysis occurs (Ebbecke,

1936a,b). This indicates pressure has an adverse effect on the normal functioning of the central nervous system. Pressure undoubtedly affects membrane fluidity and consequently normal membrane functions.

Using alkanes as a molecular model, pressure was found to increase the randomization of the molecules. The molecules did not straighten out to form straight chains as hypothesized (Schoen et al., 1977). This is not the case in more complex molecular mixtures. Dipalmitoylglycerophosphocholine (Pam2GroPCho) undergoes an increase in transition temperature under pressure. This increase in the transition temperature from the gel to liquid crystalline state indicates an ordering of the membrane (Liu and Kay, 1977).

This ordering may be produced through a lateral compression of the membrane. X-ray diffraction studies of Pam2GroPCho indicate that as pressure increases there is a compression of the membrane not perpendicular to the plane of the membrane, but instead within the plane of the membrane (Stamatoff et a l., 1978). This lateral compressibility causes acyl chain straightening and acyl chain ordering, thereby decreasing movement within the plane of the membrane. Polarized phase fluorometric studies have indicated a minimum volume is obtained under pressure by the side to side 175 arrangement of a ll -trans fatty acids. There is also an elevation in the transition temperature in Pan^GroPCho membranes under pressure

(Lakowicz and Thompson, 1983).

A more physiologic study was performed using nitroxide labeled egg yolk phosphatidylcholine. A rapid and linear increase in membrane order occurs under pressure. Fatty acid molecular motion decreases possibly modifying the functioning of membrane bound proteins (Chin et a l., 1976). Low levels of cholesterol have l i t t l e buffering effect on these molecular order changes induced by pressure. As the cholesterol mole percent approaches 50%, membrane compressibility decreases with the increasing order that cholesterol imparts to the membrane. As the cholesterol/phospholipid ratio increases so does membrane viscosity, however, pressure continues to increase membrane rigidity over that of a non-pressurized system

(Chong and Cossins, 1984). Microviscosity of Pan^GroPCho membranes increases with increases in pressure from 5 to 24 atm. Small pressures of 10 atm are capable of increasing the transition temperature creating a change in acyl chain packing and order (Barkai et al., 1983). The ultimate result is a decrease in membrane fluidity. Thus, with increases in pressure, phospholipid membranes become more ordered and this is reflected in acyl chain straightening and decrease in motion of the acyl chains resulting in a decrease in fluidity and a lateral compression of the membrane.

Hydrogen bonding is an important part of phospholipid- phospholipid, phospholipid-cholesterol and phospholipid-protein interactions. In an Fourier transform infrared (FTIR) spectroscopic 176 study using various phosphatidylcholines, pressure strengthened hydrogen bonding. Cholesterol's hydrogen bond with the sn-2 carbonyl group is strengthened, but there is no effect upon the hydrogen bond of the phosphate group. The phosphate group hydrogen bonds primarily with water (Wong et al., 1989). Pressure strengthens the hydrogen bonds of 1,2-dipalmitoyl-sn-glycerol liposomes. The intermolecular space between the molecules decreases. FTIR indicates that the sn-1 carbonyl hydrogen bonds with the sn-1 hydrogen of a neighboring dipalmitoyl-sn-glycerol (Mushayakarara et a l., 1986b). The sn-1 carbonyl of dipalmi toy!glycerophosphoethanolamine (Pam2GroPEtn) binds with the hydrogen from the sn-1 position of a neighboring

Pam2GroPEtn. Pressure increases the strength of these bonds (Wong and Mantsch, 1988). Pressure also increases the strength of the hydrogen bonds between triacetylglycerol and the neighboring triacetylglycerol molecule. Pressure strengthens the mechanical attractive forces, that is pressure restricts the vibrational frequency of the carbonyl carbon and restric ts the motion of the carbonyl oxygen enhancing the strength of the hydrogen bond

(Mushayakarara et a l., 1986b).

Increased pressure causes changes in ion transport. In red blood cells pressure causes an increase in passive K+ leak and reduces

Na+,K+-cotransport. Na+,K+-ATPase activity is also reduced

(Hall et a l., 1982; Hall and Ellory, 1986). The effects on

Na+,K+-ATPase are seen below 100 atm of pressure. Stretch- activated and inactivated channels may be modified by changes in atmospheric pressure. These channels open or close by a physically 177

stretching of the membrane, hence they are called mechanoreceptive

channels (Morris and Sigurdson, 1989). Stretch-activated channels

increase Ca2+ entry into the neuron. These channels may exist in

neuroblastoma-glioma hybrid cells (Yang and Sachs, 1989). These

channels may be activated during increased atmospheric pressure

further changing normal cation transport in cells under pressure.

Many proteins are affected by pressure changes. The primary

pressure effect is on ligand-enzyme binding through a protein

conformational change which changes the ligand binding site (Low and

Somero, 1975). Packing of protein can change causing different amino

acid residues to be exposed promoting or interfering with ligand

binding. Chymotrypsin is inactivated by pressure. Pressure causes a

decrease in protein volume (Heremans, 1982). Ligand binding is also

changed. There are two types of binding: soft binding and hard

binding. Soft binding sites rotate around a backbone bond which

permits the reduction of binding site under pressure due to

rotational changes. Hard binding sites do not decrease in size with

pressure due to their rigidity. Pressure disrupts binding to these

sites by an unknown mechanism (Heremans, 1982). Sarcoplasmic

reticulum ATPase undergoes a quarternary structural change which

causes the dissociation of the oligomeric complex during ultra

centrifugation or pressure. If Ca2+ is in its binding site this

conformational change is prevented. A change in the lipid

environment may be a contributing factor (Champeil et a l., 1981).

Activity of Ca2+-ATPase decreases with pressure. Evidence

indicates a change in the lipid environment may induce this decrease in Ca2+-ATPase activity. This fluidity change causes a conformational change in the protein decreasing its activity

(Heremans and Wuytack, 1980). Na+,K+-ATPase activity decreases as pressure increases. The Arrhenius plot has a break which corresponds to the temperature transition indicative of dependence on lipid environment. p-Nitrophenylphosphatase activity also decreases under increased pressure (DeSmedt et a l., 1979).

Acetylcholinesterase activity decreases with increased pressure due to the decrease in hydrophobic binding between substrate and protein

(Hochachka, 1974).

The activities of most terrestrial polymeric enzymes decrease as pressure increases. This is due to oligomeric dissociation.

However, enzymes from abyssal organisms show the opposite effect. At low pressure citrate synthetase polymerizes into a 270 kDa protein.

At high pressure the enzyme depolymerizes into a two subunit enzyme with much greater activity (Hochachka, 1975). Deep sea organisms have adapted to extreme pressures. Shallow water organisms do not adapt as indicated by the induction of tremors at 50 atm and convulsions at 75-100 atm of pressure (MacDonald and Gilchrist,

1982). During actin formation, deep sea organisms have l i t t l e changes in a H and a S indicating less reliance on hydrophobic interaction for polymerization (Swezey and Somero, 1982). The homeoviscous theory maintains that these deep water organisms must change their basal membrane fluidity to compensate for the increased ordering effect of pressure (Cossins and MacDonald, 1984). To compensate for high pressure, deep sea organisms should have 179 decreased cholesterol levels and increased cis fatty acids to cause an increase in overall membrane fluidity. Cis double bonds put a kink in the linearity of the acyl chains creating a significant amount of free volume. One would also expect a decrease in the content of large membrane-bound proteins which increase fluidity (MacDonald,

1986).

While abyssal organisms maintain a more fluid membrane to compensate for increased pressure, terrestrial organisms have no need for such an adaptation. Thus, when terrestrial membranes are subjected to an increase in pressure their biophysical reaction is to decrease the membrane's fluidity. Such a decrease interrupts and generally limits protein-1igand interactions causing a large decrease in activity. A theory has been proposed which indicates the membrane must maintain a fluid state in order to have normal membrane-bound protein activities. The membrane is continually undulating and these waves cause a cooperative conformational change in the hydrophobic region of the protein (Haines, 1982). Thus, as membrane fluidity decreases cooperative waves would decrease resulting in decreased protein activity.

The hypothesis was to determine the effect of pressure on ROC-1,

N1E-115, and HUVE cells in an effort to establish an in vitro model of SCI. Increased pressure was used to cause a mechanical deformation of the cell membrane. Pressure has been shown to laterally condense lipid bilayers causing changes in membrane dynamics and membrane protein activities. 180

Materials and Methods

Cells

Four cell types were used in the pressure studies. Preliminary work was done with primary murine neuronal-enriched cultures. These cultures proved to be d ifficu lt to produce in sufficient quantity to meet the demands of the project. Therefore, N1E-115 neuroblastoma cells were used in place of the primary murine neuronal enriched cultures. For the same reason ROC-1 oligodendroglia were chosen over primary oligodendroglia. Primary human umbilical vein endothelial cells were used because they were readily available and had exhibited good growth characteristics to meet the demands of the project.

Astrocytes were not tested due to their unavailability.

These cell types represent the majority of the cell types making up the spinal cord tissue. The methodology for culturing these cells has been described previously.

Pressure Device

Pressure was induced by increasing atmospheric pressure within a stainless steel chamber. Hi-purity compressed air was used as a gas source. Two different pressure chambers were used. The first achieved pressures up to 20 atmospheres and could hold a single cell culture 100 mm plate. This chamber was provided by Dr. Douglas K.

Anderson, Veterans Administration Hospital, Cincinnati, Ohio.

Pressure was measured using a general service pressure gauge (Omega

Engineering, Stanford, CT). The second chamber could hold three cell culture plates. The chamber was precision machined by Adaptive 181

Machine Technologies, Inc., Columbus, Ohio. Pressure was measured

using a silicone filled pressure gauge (Pressure Devices, Inc., West

Union, SC).

Pressure was increased at a steady rate over 6 seconds to obtain

the desired pressure. Pressure durations of 1, 3, 7, and 10 minutes

were used over a pressure range of 1, 5, 10, 15, and 20 atmospheres

to determine the optimal pressure duration for largest increase in

fatty acids. Recovery times of 1 and 10 minutes were used in these

experiments. Recovery times used with the optimal duration were 30,

60, 180, 360, 720, and 1440 minutes. Control cultures were placed

into the chamber for the desired length of time, removed and

permitted to recover for the time corresponding to the experimental

cell culture plates. Depressurization of the chamber occurred at a

constant rate over 10 seconds.

Lipid Extraction

Lipid extraction was done as previously described, except medium was saved to determine lactate dehydrogenase levels and lipids were

extracted using a two phase system (Bligh and Dyer, 1959). To each media sample 3.75 vol of chloroform: methanol (1:2 v/v) was added in

a separatory funnel. The contents were thoroughly mixed and 1.25 vol of water added to form two phases. The lower chloroform phase contains the lipids.

Lactate Dehydrogenase

LDH assays were performed as previously described. The levels of

LDH were assayed 2-3 hours after completion of the experiment to ensure enzyme stability. 182

S tatistics

All statistics were done using one way analysis of variance using the Statgraf® program. A multi variance statistical analysis was performed using a level of significance of 0.05 or greater.

Other Procedures

All other procedures were performed as previously described.

These procedures include TLC, GLC, FAME formation and Bradford assay of protein.

Results

Initial pressure experiments were done using primary neuronal cells. These cells could not be supplied in sufficient quantities to meet the high number of cell cultures needed for the project. In order to meet th is demand, N1E-115 neuroblastoma cells were used to substitute for the neuronal cells. ROC-1 oligodendroglia cells were used to represent primary oligodendroglia. Primary human umbilical vein endothelial cells were used to represent the vascular component.

The effect of pressure duration was examined to determine if increased duration correlated with increased LDH and FA release.

Pressure studies were done using pressure of 5, 10, 15, and 20 atm and pressure durations of 1, 3, 5, and 10 min. Recovery times were 1 and 10 min for these initial experiments. Using the optimal pressure and pressure duration, longer recovery times were examined to see if the injury produced was sustainable. These recovery times were 1, 10, 30, 60, 360, 720 and 1440 min. 183

ROC-1 Oligodendroglia1 Cells

A 1 min pressure duration caused a release of FA. After 1 min of recovery, the highest FA release measured was at 5 and 15 atm (Figure

36). A recovery time of 10 min produced a greater release of FA.

The 10 atm release of FA was significantly greater than the releases at all other pressures (Figure 38). The FA release at 10 atm 10 min recovery was significantly greater than the 1 min recovery FA release. The release at 10 min was a 7.5-fold increase over the 1 min release. This indicates the injury was sustained.

A 3 min pressure duration produced FA release after 10 min recovery but not after 1 min. The greatest increase in FA (3-fold) was at 10 atm of pressure (Figure 39). Significant increases in FA occurred at 5 and 15 min of pressure. The release of FA at 10 atm following 10 min of recovery was significantly greater than those levels released after 10 min of recovery at other pressures. A sustainable injury was produced as indicated by the time-dependence of the FA release at 10 atm.

A 5 min pressure duration produced no FA release after 1 min of recovery. However, after 10 min of recovery, significant increases in FA occurred at 5 and 15 atm. These releases resulted in a

2.5-fold increase in FA levels (Figure 40).

After 10 min of pressure duration, no significant increases in FA levels occurred following 1 min of recovery. A recovery time of 10 min at 15 atm produced a significant increase in FA compared to 5 atm

10 min recovery (Figure 41). This increase was not significant when compared to the release of FA following 1 min of recovery. 184

Thus, in ROC-1 cells, the greatest release of FA following

pressure trauma occurred following 1 min of pressure and 10 min of

recovery (Figure 42). The FA release pattern seen after 3 min of

pressure resembled that of the 1 min pressure duration except the FA

release was attenuated. Following a 1 min recovery, the greatest FA

release was seen after 1 min of pressure. The largest releases were

at 5 and 15 atm (Figure 43).

The effects of longer recovery times were examined. A pressure

of 10 atm with a duration of 3 min was used. The initial release of

FA occurred 10 min post-injury. This release decreased by 30 min to

control levels. At 1440 min an increase in FA was measured (Figure

44). The only two significant increases were at 10 and 1440 min

following trauma.

The majority of the FA released were saturated. Of the PUFA

released, the majority was 20:4 n-6. When large quantities of 20:4

n-6 were released, small amounts of 22:6 n-3 were detected. The release of PUFA was variable, however, the largest releases correlated with the pressures, durations and recovery times which

produced the greatest FA release.

The only significant release of LDH was after 5 min of 20 atm of pressure and a 10 min recovery period. No other significant increase occurred (Appendix B).

N1E-U5 Neuroblastoma Cells

A 1 min pressure duration caused a release of FA at 15 atm

(Figure 45). The maximum release was measured at 10 atm following 10 min of recovery. This was a significant increase over the 1 min 185

recovery FA level at the corresponding pressure. A significant

increase in FA occurred at 15 atm following 1 min of recovery. The

FA acid levels at this pressure did not increase with an increase in

recovery time.

At a 3 min pressure duration, there was no increase in FA levels

at any pressure following 1 min of recovery (Figure 46). A 6-fold

increase in FA occurred at 15 atm of pressure following a 10 min of

recovery. This increase was significant compared to the FA released

by all the other pressures following 10 min of recovery as well as

the corresponding pressure at 1 min of recovery.

Only moderate increases in FA levels occurred following the 5 min pressure duration (Figure 47). After 1 min of recovery there were no

releases in FA over control levels, although several pressures had

significant differences. The largest FA increase was at 20 atm of pressure. The FA releases at 10 and 20 atm of pressure and 10 min recovery were significantly greater than those at 5 atm and 10 min of

recovery.

After 10 min of pressure, there were only moderate increases in

FA levels (Figure 49). The only significant increase was at 5 atm,

10 min of recovery. This level was significantly greater than the FA level released at 20 atm.

The greatest FA increase in N1E-115 cells occurred following 15

atm of pressure for a duration of 3 min (Figure 49). Only small to moderate increases occurred at other pressures and pressure durations following 10 min of recovery. A small significant increase in FA 186 levels was detected at 15 atm following 1 min of pressure (Figure

50). No other significant increases were observed.

The effects of longer recovery times were examined. The initial increase in FA levels was at 10 min (Figure 51). This level was significantly greater than the FA levels at all other recovery times. At 720 min a significant increase was detected. This level was significantly greater than at all other recovery times excluding

10 min.

The majority of the FA released were saturated or monounsaturated. Of the PUFA released, the majority was 20:4 n-6. When large quantities (0.5-1.0 nmol/mg protein) of 20:4 n-6 were released, small amounts of 22:6 n-3 were detected. The release of PUFA was variable, however, the largest releases correlated with the pressure durations and recovery times which produced the greatest FA release.

Small but significant increases in LDH were found at several pressures and pressure durations (Appendix B). These increases in

LDH did not correlate with increases in FA levels. An increase in

LDH occurred after 1 min of 15 atm of pressure following 10 min of recovery. Increases were also detected at 20 atm following 5 min of pressure at both 1 and 10 min recovery times. LDH was increased following a pressure 15 atm for 3 min and a 1440 min recovery time.

HUVE - Human Umbilical Vein Endothelial Cells

A 1 min pressure duration produced a large significant increase in FA at 5 atm and 1 min recovery (Figure 52). This increase was substantially decreased after 10 min of recovery. The FA levels at 5 187 atm and 10 min of recovery were greater than those measured at 15 atm and 10 min recovery.

A pressure duration of 3 min produced only small increases in FA levels after the cells were subjected to 15 atm of pressure (Figure

53). The FA release after 1 min of recovery at 15 atm was significantly greater than the levels at 5 and 10 atm following 1 min of recovery. The FA release following 10 min recovery at 15 atm was significantly greater than the level at 5 atm at the same recovery time. However, the FA levels after 10 min recovery were less than those after 1 min of recovery.

Only small increases in FA were measured after a pressure duration of 5 min (Figure 54). The FA levels at 10 atm were significantly greater than those at other pressures. The 1 min recovery time FA levels at 10 atm were significantly greater than those at 5 and 20 atm. The 10 min recovery time FA levels at 10 atm were significantly greater than those at 5 and 15 atm. These levels at 10 min were less than those after 1 min of recovery time indicating reacyl ation.

In HUVE cells, the largest FA increases occurred after a 10 min pressure duration (Figure 55). The largest increases were at 10 and

15 atm. The FA release at 1 min of recovery following 10 atm was significantly greater than those following 5 and 20 atm FA levels.

Following 10 min of recovery, the 15 atm FA values were significantly greater than those at 20 and 5 atm. The FA levels at 10 atm were significantly greater than the 5 atm levels following 10 min of recovery. 188

The largest increase in FA occurred at 5 atm following a 1 min

pressure duration and 1 min recovery time (Figure 56). This increase

was not sustainable as indicated by the decreased FA levels following

10 min of recovery (Figure 57). The largest sustainable increase in

FA occurred at 15 atm and a 10 min pressure duration. At this

pressure, the largest FA increase was seen after a 1 min recovery

time. The 10 min recovery time levels were slightly less than the 1

min recovery time levels.

The effects of longer recovery times were examined using a

pressure of 15 atm and a duration of 10 min (Figure 58). The FA

levels at 1, 10, and 1440 min were significantly greater than all

other recovery times. The initial FA release was decreased by 30

min. Between 720 and 1440 min there was an increase in FA released.

The levels at 1440 min was greater than the levels seen at 1 and 10

min recovery.

The majority of the FA released were saturated or monounsaturated

FA. Unlike R0C-1 and N1E-115 cells, large amounts of PUFA were

released. These released PUFA in decreasing order were 22:6 n-3 >

20:4 n-6 > 18:2 > 18:3 n-6 = 18:3 n-3 = 20:3 n-6. The release of

PUFA correlated with the pressure durations and recovery times which produced the greatest FA release.

The LDH levels were assayed in control and experimental medium.

Following 15 atm of pressure for a duration of 10 min, a significant

increase in LDH activity was detected at 1440 min (Appendix B). This

significance correlates with the FA release seen. Another

significant increase in LDH was seen at 20 atm with a pressure duration of 5 min following 10 min recovery. 189

Figure 38 Pressure induced FA release in ROC-1 c ells. Pressure

duration was 1 min followed by 1 and 10 min recovery

times. After 1 min of recovery the FA release at 5 and

15 atm > all others. After 10 min of recovery FA

release at 10 atm > 5 atm and 15 atm > 20 atm. Level

of significance p < 0.05. Pressure Induced FA Release In ROC-1 Cells* 1 rain Duration Pressure/Control FA Ratio I ain 8 ^Recovery^ 7 10 Bin Recovery (Z Z Z 3 6

4 ★ ★ **

10 1 5 20 Proesuna atm 191

Figure 39 Pressure induced FA release in ROC-1 cells. Pressure

duration was 3 min followed by 1 and 10 min recovery

times. After 1 min of recovery there was no

significance in FA levels. After 10 min of recovery FA

release at 10 atm > 5 atm and 15 atm > 20 atm. Level

of significance p < 0.05. Pressure Induced FA Release In ROC-1 Cel 1st 3 min Duration Pressure/Control FA Ratio 1 ain Recovery 4.0

3.5 10 ain Recovery 17 Z Z 1 3.0

2.5

★★ ★ ★ 1.5

1.0

5 -

10 15 20 Pressure ate Figure 39 193

Figure 40 Pressure induced FA release in ROC-1 cells. Pressure

duration was 5 min followed by 1 and 10 min recovery

times. After 1 min of recovery there was no

significance in FA levels. After 10 min of recovery FA release at 5 and 15 atm > 20 atm. Level of

significance p < 0.05. Pressure Induced FA Release In ROC-1 Cell9: 5 min Duration Pressure/Control FA Ratio 1 Bin ^Recovary^ 3.0

10 Bln Recovery 2.5 1 Z 7 Z 3

2.0

1.5

1.0 -

5 10 15 Pressure atm Figure 10 195

Figure 41 Pressure induced FA release in ROC-1 cells. Pressure

duration was 10 min followed by 1 and 10 min recovery

times. After 1 min of recovery there was no

significance in FA levels. After 10 min of recovery FA

release at 15 atm > 5 atm. Level of significance

p < 0.05. Pressure Induced FA Release In ROC-1 Cells: 10 rain Duration Preoaure/Contro1 FA Ratio 1 sin ^Recovery^

2.0 10 Bin Recovery I7 7 Z ]

1.5

1.0

CD 10 Figure 41 Preeoura at® 197

Figure 42 Comparison of pressure induced FA release in ROC-1

cells at various pressure durations. The recovery time

was 10 min. After 1 and 3 min durations FA release at

10 atm > all others. After 5 min duration FA release

at 5 atm > 20 atm. After 10 min duration FA release at

15 atm > 5 atm. Level of significance p < 0.05. Pressure Induced FA Release In ROC-1 Cells* 10 min Recovery Pressure/Control FA Ratio 5 ata 8 PrMsure (ZZZ3 7 10 ata Proesura 1ZZZI 6

15 ata PrtwsurQ [XXXI

20 ata 4 P ressu re V Z Z 7 A 3

1 3 5 10 Pressure Duration nin Figure 42 199

Figure 43 Comparison of pressure induced FA release in ROC-1

cells at various pressure durations. The recovery time

was 1 min. After 1 min duration FA release at 5 and 15

atm > 10 and 20 atm. After 3, 5, and 10 min durations

there was no significance in FA levels. Level of

significance p < 0.05. Pressure InducGd FA Release In ROC-1 Cellsi 1 min Recovery Pressure/Control FA Ratio 5 a ta Pressure ]

10 ata Pressure r/ 7 7 j

15 ata Pressure IXXXJ

20 ata Pressure

3 5 Figure 3 Pressure Duration Min 201

Figure 44 Comparison of pressure induced FA release in ROC-1

cells at various recovery times. The pressure was 10

atm and the duration was 3 min. The FA release at 10

and 1440 min > all other recovery times. Level of

significance p < 0.05. Pressure Induced FA Release In ROC-1 Cells: 3 rain Duration Pressure/Control FA Ratio

3.0 ★

★ 2.5

2.0

1.5

1.0

1 10 30 60 360 720 1440 Figure 44 Recovery Tine nin 203

Figure 45 Pressure induced FA release in N1E-115 cells. Pressure

duration was 1 min followed by 1 and 10 min recovery

times. After 1 min of recovery FA release at 15 atm >

all others. After 10 min of recovery FA release at

10 atm > all others. Level of significance p < 0.05. Pressure Induced FA Release In N1E—115 Cellsx 1 min Duration Pressure/Control FA Ratio 1 aln Recovery i ^ 3.0 10 ain Recovery tZ Z Z 3 2.5

2.0

1.5

1.0

510 15 20 Pressure atm Figure *5 205

Figure 46 Pressure induced FA release in N1E-115 cells. Pressure

duration was 3 min followed by 1 and 10 min recovery

times. After 1 min of recovery there was no

significance in FA levels. After 10 min of recovery FA

release at 15 atm > all others. Level of significance

p < 0.05. Pressure Induced FA Release In N1E-115 Cells: 3 min Duration Preesuro/Control FA Ratio 1 Bin ^Recovery^ 6

10 ain Recovary 5 X 7 7 7 \

5 10 15 Pressure ata 46 207

Figure 47 Pressure induced FA release in N1E-115 cells. Pressure

duration was 5 min followed by 1 and 10 min recovery

times. After 1 min of recovery FA release at 20 atm >

5 and 10 atm and 15 and 10 atm > 5 atm. After 10 min

of recovery FA release at 10 and 20 atm > 5 atm. Level

of significance p < 0.05. Pressure Induced FA Release In N1E-115 Cellsi 5 min Duration Pressure/Control FA Ratio I Bin Recovery i— ° i 2.0

10 Bin Recovery 1ZZ Z 3 1.5 -

1.0 ★ ★

5 10 15 20 Pressure atm

Figure 47 209

Figure 48 Pressure induced FA release in N1E-115 cells. Pressure

duration was 10 min followed by 1 and 10 min recovery

times. After 1 min of recovery there was no

significance in FA levels . After 10 min of recovery

FA release at 5 atm > 20 atm. Level of significance p < 0.05. Pressure Induced FA Release In N1E-115 Cells: 10 min Duration Pressure/Control FA Ratio 1 ain ^Recovery^

10 Bin Recovery 1.5 177Z3

1.0

ro 5 10 15 20 t —4 o Pressure atm

Figure 48 211

Figure 49 Comparison of pressure induced FA release in N1E-115

cells at various pressure durations. The recovery time

was 10 min. After 1, 5, and 10 min durations there was

no significance in FA levels. After 3 min duration FA

release at 15 atm > all others. Level of significance p < 0.05. Pressure Induced FA Release In N1E-115 Cells> 10 min Recovery

Preosuro/Control FA Ratio 5 ota PT&89UTQ ! = □

10 ata Pressure 1Z Z Z ]

15 ata PTG83UTO 1X K S 1

20 ata IVe8SUTQ Y 7 Z Z Z A

no I—I ro 3 5 Pressure Duration sin Figure 19 213

Figure 50 Comparison of pressure induced FA release in N1E-115

cells at various pressure durations. The recovery time

was 1 min. After 1 min duration FA release at 15 atm >

all others. After 3, 5, and 10 min durations there was

no significance in FA levels. Level of significance

p < 0.05. Pressure Induced FA Release In N1E-115 Cells: 1 rain Recovery

Pressure/Control FA Ratio 5 ata Pressure r j

10 ata Pressure (ZZZ3

15 ata Prossura IXXX] '4.

20 a ta ProssirQ E Z Z 2 i I i'4 4 i i '4. 4

Pressure Duration nin Figure 50 215

Figure 51 Comparison of pressure induced FA release in N1E-115

cells at various recovery times. The pressure was 15

atm and the duration was 3 min. The FA release at 10

min > 720 min > all others. Level of significance

p < 0.05. Pressure Induced FA Release In N1E-115 Cells* 3 min Duration

Pressure/Control FA Ratio 7 ★

★ ★ 3

1 10 30 60 360 720 1440 £ cn Recovery Tine nln

Figure 51 217

Figure 52 Pressure induced FA release in HUVE cells. Pressure

duration was 1 min followed by 1 and 10 min recovery

times. After 1 min of recovery FA release at 15 atm >

5 and 10 atm. After 10 min of recovery FA release at

5 atm > 15 atm. Level of significance p < 0.05. Pressure Induced FA Release In HUVE Cellsi 1 sin Duration

Preesuro/Control FA Ratio 1 Bin Recovery r ~ T 3 4 10 sin Recovery 177 2 3 3

10 15 20 Proesura ata Figure 52 219

Figure 53 Pressure induced FA release in HUVE cells. Pressure

duration was 3 min followed by 1 and 10 min recovery

times. After 1 min of recovery FA release at 15 atm >

5 and 10 atm. After 10 min of recovery FA release at

15 atm > 5 atm. Level of significance p < 0.05. Pressure Induced FA Release In HUVE Cellsi 3 min Duration PreesurQ/Contro1 FA Ratio 1 Bin ^Recovery^ 1 .5

10 a in Recovery tZZ ZD

1.0

ro ro o 5 10 15 Pressure atm Figure >3 221

Figure 54 Pressure induced FA release in HUVE cells. Pressure

duration was 5 min followed by 1 and 10 min recovery

times. After 1 min of recovery FA release at 10 atm >

5 and 20 atm. After 10 min of recovery FA release at

10 atm > 5 and 15 atm. Level of significance p < 0.05. Pressure Induced FA Release In HUVE Cellsi 5 min Duration Prwesura/Contro1 FA Ratio 1 sin r~Recovery

10 Bin Recovery IZZZJ 1.0

s 10 15 Prossuro atn Figure 54 223

Figure 55 Pressure induced FA release in HUVE cells. Pressure

duration was 10 min followed by 1 and 10 min recovery

times. After 1 min of recovery FA release at 10 atm >

5 and 20 atm. After 10 min of recovery FA release at

15 atm > 5 and 20 atm and 10 atm > 5 atm. Level of

significance p < 0.05. Pressure Induced FA Release In HUVE Cells* 10 min Duration Pressure/Control FA Ratio 1 sin ^Recovery^ 3.0

10 Bin Recovery 1ZZ Z ] 2.5

1.5

1.0

5 -

ro ro4* 5 10 15 20 Pressure atm Figure >5 225

Figure 56 Comparison of pressure induced FA release in HUVE cells

at various pressure durations. Recovery time was

1 min. After 1 min durationFA release at 5 atm > all

others and 15 atm > 20 atm. After 3 min duration FA

release at 15 atm > 10 atm. After 5 and 10 min

duration FA release at 10 atm > 5 and 20 atm. Level of

significance p < 0.05. Pressure Induced FA Release In HUVE Cellsx 1 min Recovery Pressure/Control FA Ratio 5 ata Pressure c J 10 ata Pressure 1ZZZ1

15 ata Pressure K X X 3

20 ata Pressure

3 5 Pressure Duration Bin Figure 56 227

Figure 57 Comparison of pressure induced FA release in HUVE cells

at various pressure durations. Recovery time was 10

min. After 1 min duration FA release at 5 atm > 15

atm. After 3 min duration FA release at 15 atm > 5

atm. After 5 min duration FA release at 10 atm > 5 and

15 atm. After 10 min duration FA release at 15 atm > 5

and 20 atm and 10 atm > 5 atm. Level of significance p < 0.05. Pressure Induced FA Release In HUVE Cellsi 10 min Recovery

Prossure/Control FA Ratio 5 ata Pressure c J 10 ata Pressure IZZZ3

15 ata Pressure (XXX3

20 ata 1 Pressure V///M 1 i i i i I i 3 5 Pressure Duration min Figure 17 229

Figure 58 Comparison of pressure induced FA release in HUVE cells

at various recovery times. The pressure was 15 atm and

the duration was 10 min. The FA release at 1, 10, and

1440 min > all others. Level of significance p < 0.05. Pressure Induced FA Release In HUVE Cells: 10 min Duration Preoaurw/Control FA Ratio

3L0

2.5

2.0

1.5

1.0

60 360 Recovery Tine nin Figure 58 231

Discussion

Pressure caused a release in FA in all cell types. This release

was dependent upon the level of pressure and the pressure duration.

ROC-1 and N1E-115 cells underwent a reversible trauma at short

pressure durations. HUVE cells underwent a reversible trauma at

longer pressure durations, although with increased recovery time this

trauma begins to be irreversible. The level of pressure needed to

e lic it the greatest FA release was 10 atm for ROC-1, 15 atm for

N1E-115 and 15 atm for HUVE cells. The pressure duration was 1, 3,

and 10 min respectively for each optimal pressure.

No studies have been published studying the effect of pressure on

cell cultures, although the physical effects of pressure on various

model membrane systems have been extensively studied. The only

physiological studies have been done using deep sea organisms. It

was my desire to examine the effects of increased pressure on cell

cultures. The ultimate goal is to use pressure to induce a

mechanical trauma in cell cultures which is comparable to that seen

during SCI. A cell culture model would offer several advantages to

trauma produced in vivo. These advantages include the ease of manipulating the extracellular medium, the ability to focus on a

single cell type's reaction to trauma and the decrease in animal use.

Initial experiments were done on primary neuronal cells using a

prototype pressure chamber. From these initial experiments, the

integral role of pressure duration in producing trauma was discovered. Short pressure durations of 30 sec or less produced no 232 sustainable trauma. Pressure durations exceeding 1 min produced a sustainable injury. A sustainable injury was defined as an injury which caused the cells to continue to release FA as injury time increased. However, while these initial experiments were useful, the primary neuronal culture supply could not meet the demands of the project.

The pressure chamber used in the neuronal studies had several flaws. First the chamber was difficult to open quickly. This difficulty limited recovery times to those greater than 1 min. There was no baffling system to reduce the possible shear forces formed by the pressure wave. These shear forces could induce an undesirable artifac t in the system.

There also were several safety flaws. F irst, the pressure gauge was located in a position which forced the operator to position himself over the instrument to read the pressure gauge. There were no safety measures designed to compensate for system failure. If the chamber failed, a potentially hazardous situation would occur.

To eliminate these problems, a new chamber was fabricated. This new design met the needs of the project. The vessel could hold up to

3 cell cultures to increase experimental efficiency. A baffling system was in place to reduce shear forces. A breech design was used to permit the quick removal of the cells after pressurization.

The effects of pressure on ROC-1 cells were studied. The greatest FA release was found after a 1 min pressure duration at 10 atm. This pressure produced a sustainable injury. A 3 min pressure duration affected the cells in the same way except the FA release was 233

attenuated. As pressure duration increased there was no

corresponding increase in FA released as seen at 1 and 3 min

durations. The 5 min duration had a moderate release of FA at

several pressures. However, the pattern established in figure 38 and

39 was not reproduced at the longer pressure durations. The initial

increases at a 10 min pressure duration were not sustainable.

Between 1 and 10 min recovery, the FA were either metabolized or

reacylated. Thus, the largest sustainable increases in FA occurred

at the duration times of 1 and 3 min.

The pressure of 10 atm and a duration of 3 min was used to

examine recovery times. The increase in FA from 1 to 10 min of

recovery indicates a possibility of a sustainable injury. However,

the injury produced was only a minor perturbation of the membrane.

After 30 min of recovery, FA levels returned to control levels and

remained at those levels until 720 min. Between 720 and 1440 min an

increase in FA occurred. This biphasic release was not expected.

Unfortunately time points beyond 24 h were not studied, which may

indicate the sustainability of the increase in FA.

One possible explanation for this biphasic response is that after

the cells are traumatized, there is an initial release of FA. This

release could be through a phospholipase Aj and phospholipase A2

mechanism which is activated by a compression of the membrane by the

pressure. The FA released include saturated, monounsaturated and

polyunsaturated FA which are released primarily by phospholipases.

The activation of phospholipases was transitory because by 30 min FA

values were at control levels. As the recovery time increases, the 234 cell undergoes repair to remove the accumulated FA through the metabolism or reacyl ation of the released FA. However, another increase in FA occurred at 24 h. This increase indicates phospholipases were reactivated by an unknown mechanism which causes increased FA release.

The release of FA did not correlate with the release of LDH. A

3.8-fold increase in FA did not result in the loss of membrane integrity because LDH levels did not increase. Thus, while the membrane was perturbed by an increase in atmospheric pressure, this injury did not result in cell death within the time points investigated.

As in ROC-1 cells, the largest increases in FA levels in traumatized N1E-115 cells occurred within the first 3 min of pressure duration. Pressure durations of 5 and 10 min produced minimal increases in FA (Figure 47 and 48). The largest increase occurred at

15 atm following a 3 min pressure duration (Figure 45). This large increase in FA consisted primarily of saturated FA, monounsaturated

FA, and arachidonic acid. The large amount of arachidonic acid indicated a possible activation of the phospholipase by the mechanical deformation of the cell membrane.

Longer duration times showed the injury was not sustainable. By

30 min, the FA were reacylated or metabolized. However, like ROC-1 cells, the release of FA was biphasic. Unlike ROC-1 cells, the increased release at 720 min was also transitory. By 24 h the FA released at 720 min had reached control levels. This quick recovery 235

reinforces the idea that the membrane was merely perturbed since FA

levels at 24 h reached control levels.

LDH activity did not correlate well with the FA data. The lack

of increase in LDH activity indicates the cell membrane remains

intact although fluidity changes may have occurred.

HUVE cells need an extended duration of pressure in order to

sustain FA release. Pressure durations of 1, 3, and 5 min caused

only minimal releases of FA (Figures 52, 53, and 54). A pressure

duration of 10 min causes an increase in fatty acids (Figure 55).

Levels were increased at 5, 10, and 15 atm following 1 min of

recovery. The increase at 5 atm was unsustainable and the fatty

acids were reacylated. The higher pressures also showed signs of

reacylation because the 10 min recovery values were slightly less

than the initial FA release.

Values for FA at longer recovery times confirm that the injury

was unsustainable. However, like the ROC-1 cells, the release of FA was biphasic with an increase in FA after 1440 min of recovery. This

increase in FA after 1440 min of recovery was greater than any of the

FA releases in the R0C-1 and N1E-115 cells. This increase was not

statistically significant from the FA releases at 1 and 10 min

recovery times in HUVE cells. However, the trend for increasing the

FA release appears to be established.

In HUVE cells there was a large increase in PUFA as well as

saturated and monounsaturated FA. There were higher levels of PUFA

following pressure trauma in HUVE cells than in R0C-1 and N1E-115 cells. In order to get these larger increases, phospholipase 236 must be activated or a mechanism which couples several lipases and phospholipases or several phospholipiases such as phospholipase A} followed by lysophospholipase. Lysophospholipase may be activated which removes the FA from the lysophospholipid resulting from increased phospholipase Aj activity during pressure trauma.

Phospholipase Aj activity was evident due to the large increases in saturated and monounsaturated FA. However, this pathway may not be plausable due to the high endogenous basal levels of lysophospholipids. These levels indicate either low amounts of lysophospho lipase or low enzyme activity.

The LDH activity measured at the 1440 min recovery time indicates the traumatized cell was becoming compromised. The LDH levels of the medium had increased above control levels. As the cell membrane loses its integrity, the cytosolic enzyme LDH leaks out into the extracellular space confirming an injury had occurred greater than a mere perturbation of the cell membrane.

One explanation for the increased pressure duration needed to produce an injury in HUVE cells may be found in its phospholipid composition. HUVE cells contain a high lysophospholipid content.

Lysophospholipids act as detergents and fluidize the membrane

(D'Amato et a l., 1975; Weltzien, 1979a; Weltzien et a l., 1979b).

This is done by their ability to increase the intermolecular space in the membrane (Lee, 1983). The increase in fluidity would cause a greater pressure or a longer pressure duration to be needed to compress the membrane. Since HUVE cells are normally found in a high pressure environment, their adaptation to this environment would not 237 be unexpected. Deep sea organisms adapt to their environment by increasing membrane fluidity (Cossins and MacDonald, 1984). A high lysophospholipid content may be one adaptation the endothelial cell has developed to protect its e lf from high pressure trauma.

ROC-1 and N1E-115 cells do not contain a large amount of lysophospholipids. Furthermore, their FA composition is relatively saturated. As the number of cis double bonds increase, there is a concomitant increase in membrane fluidity (Stubbs, 1983; Stubbs et a l., 1981). Phospholipids of N1E-115 and ROC-1 cells have a small unsaturated FA content. This coupled with the lack of lysophospho lipids may explain why these cells have the greatest pressure effect at shorter pressure durations. Conversely, as one increases pressure and pressure duration, a less fluid membrane will undergo physical changes much quicker than a more fluid membrane. As the pressure duration increase, there may be an over-ordering of the membrane and a subsequent inactivation of the phospholipases. This could explain lack of an increase in FA release at longer pressure durations in N1E-115 and ROC-1 cells.

Membrane fluidity has an important role in maintaining proper cell functioning (Spector and Yorek, 1985). As FA levels increase, they intercalate into the membrane causing an increase in membrane flu idity (Michel1 et a l., 1975; Klausner et a l., 1980; Stubbs and

Smith, 1984). This causes the cell to become more permeable to ions

(Yorio et a l., 1983). These fluidity changes can also disrupt proper receptor function (Bell et a l., 1980; Raber and Bast, 1989). Thus, 238 the maintenance of membrane fluidity is essential for proper cellular functioning.

When model membranes are subjected to a high pressure environment there is a lateral compression of the membrane (Stamatoff et a l.,

1978). As pressure increases there is a rapid and linear increase in membrane order (Chin et a l., 1976). This increase in order is due to the decrease in acyl chain motion (Chin et a l., 1976). Acyl chains condense and form an all trans configuration to obtain a minimum volume (Lakowicz and Thompson, 1983). This condensation causes an increase in the microviscosity of the membrane (Barkai et a l.,

1983). As pressure increases, the membrane must compromise its normal fluid state in response to the pressure. This response is an ordering of the membrane into the most compact orientation possible.

As the membrane becomes more ordered, the membrane fluidity decreases causing possible changes in receptor functioning (Bell et a l., 1980;

Raber and Bast, 1989) and normal cellular functioning (Spector and Yorek, 1985).

Stretch-activated channels could be opened during a pressure insult due to changes in membrane structure (Morris and Sigurdson,

1989; Yang and Sachs, 1989). Opening of these channels would cause an unregulated influx of Ca^+ into the cell. Increased pressure is known to increase ion transport in red blood cells (Hall et a l.,

1982; Hall and Ellory, 1986). These increases in ion transport and non-regulated ion flux will disrupt the normal ionic balance of the cell. As Ca^+ enters the cell, the result will be a Ca^+ mediated death of the cell (Balentine, 1983; Siesjo, 1988). The 239 ultimate result of a pressure injury could very well be cell death by an unregulated Ca2+ influx. The membrane changes induced by pressure could cause an increase in Ca2+ influx by physically opening a closed Ca2+ channel, opening a stretch-activated channel or causing an increased permeability of the cell by membrane degradation.

Pressure also effects hydrogen bonding. As pressure increases so does the hydrogen bonding in the membrane (Mushayakarara et a l.,

1986a,b; Wong and Mantsch, 1988; Wong et a l., 1989). The increase in hydrogen bonding aids the ordering and condensing of the membrane.

Ordering of the membrane affects membrane bound enzymes. As pressure increases the Na+,K+-ATPase activities in red blood cell membranes drops (Hall and Ellory, 1986). Ca2+-ATPase activity also decreases with increases in pressure (Heremans and Wuytack, 1980).

Acetylcholinesterase activity also decreases with increases in pressure (Hochachka, 1974). Many of these changes are due to the change in protein volume (Heremans, 1982). These changes cause a disruption of the binding of the ligand to the enzyme through a conformational change in the protein (Low and Somero, 1975).

Thus, pressure causes physical changes to occur within the membrane which results in changes in ion flux and membrane bound enzyme activity. Hydrogen bonding increases to aid in membrane condensation. Acyl chains straighten to form a closely packed condensed membrane using minimal space. This acyl chain packing is due to the overall response of the cell to pressure. As the pressure increases, there is lateral compressibility of the membrane. This 240 compressibility causes many events to occur which will undoubtedly lead to the demise of the cell.

During pressure-induced trauma in cell cultures, FA were released indicating an activation of phospholipases. This increase in activity of phospholipases may be due to the actual pressure or pressure effects on the membrane. Evidence indicates that activation of phospholipases alone may not result in cell death. Initially released FA were quickly reacylated or metabolized. The release was biphasic, and the greatest trauma occured in the HUVE cells. The more fluid endothelial cells respond to the pressure trauma only after a long duration of pressure. The ROC-1 and N1E-115 respond to the short pressure duration more than longer pressure durations.

These differences may be due to the origin of the cells and to their membrane fluidity.

Future studies must be done to determine the complete effects of pressure on cells. Currently, cell death does not result from the pressure-induced insult. The question arises, can pressure sufficiently traumatize cells to result in cell death? I believe the answer is yes. Longer recovery times should be examined. HUVE cells appear to begin their demise by 24 h post-trauma. Longer pressure durations may also result in the quicker death of the HUVE cells.

ROC-1 and N1E-115 cells must undergo pressure trauma using a shorter pressure duration. Membrane fluidity changes should be assessed to determine the actual physical state of the membrane. Ca^+ influx studies should be performed to assess the role of Ca^* in pressure induced trauma. CHAPTER V

CONCLUSION

Cell cultures were used to examine the effects of pressure and

aluminum silicate containing clays on cell viability. The ultimate

goal of the pressure studies is to develop an in vitro model of SCI.

This model would aid in understanding the type of cells injured in

SCI and to what extent pharmacological treatment regimens could limit

cellular necrosis. Aluminum silicate containing clays were examined

for their toxicity towards cells of neuronal origin. The toxicity of

the clays to these cells was an important step in determining a role

for the clays in AD.

The phospholipid composition of N1E-115, ROC-1, HUVE, and

neuronal cells was reported. HUVE cells have a high mole % of

lysophospholipids and contained a large percentage of PUFA. These

two factors could fluidize the membrane. The phospholipid

composition of the ROC-1 cells was a hybrid of the two parent cells.

The N1E-115 and neuronal cells had greatly different phospholipid

composition. This difference could be due to developmental stage, origin or primary versus transformed state of the cells. All of the

cells from neuronal origin had low levels of PUFA and this was

reflected in the FA released during the pressure studies.

241 242

Aluminum silicate containing clays were not toxic to ROC-1 or

N1E-115 cells at the concentrations and times studied. Primary cells may react differently to the clays and may explain the apparent

resistance of the N1E-115 and ROC-1 cells to lysis by the clays.

This may be the case since HUVE cells, a primary cell, were very

susceptible to clay induced cell lysis. The clays with an amorphic

structure were most toxic to the cells. The order of toxicity was montmorilIonite > kaolinite = bentonite » erionite.

The clays caused total cell lysis in HUVE cells leaving only cell fragments with no discernible nuclei. The mode of action of the clays is not known. However, the clays possess an overall negative charge and possess the ability to cleave bonds. Thus, the clays may elicit their effect by ordering the membrane to activate lipases and/or phospholipases, open ion channels or by directly acting as phospholipases cleaving the FA from the sn-1 position of phospholipids. Hydrolysis of these FA would cause an increase in fluidity and result in the eventual death of the cell.

Increasing the atmospheric pressure over cell cultures induces a reversible trauma. A short pressure duration induces the greatest FA release in ROC-1 and N1E-115 cells. These pressure durations only perturbed the membrane since no LDH was released. Shorter pressure durations and longer recovery times may result in cell death. In

HUVE cells, a long pressure duration was needed to elicit the pressure effects. After 24 h of recovery, LDH was beginning to be released indicating membrane breakdown was beginning. Longer pressure durations and recovery times may produce greater cell death. 243

Pressure induced PUFA release in HUVE cells but not in ROC-1 and

N1E-115 cells. Saturated and monounsaturated FA were released in all three cell types in large quantities. Thus in HUVE, ROC-1 and

N1E-115 cells a phospholipase A^ like mechanism was activated. In

HUVE cells a phospholipase A2 and lysophospholipase like mechanism apparently was activated by an increase in pressure.

As pressure increases there is a lateral compression of the membrane possibly inducing a conformational change and altering their phospholipase and lipase activities. Pressure may also cause an unregulated ion flux resulting in the activation of lipases and phospholipases. Regardless, either mechanism would produce cellular necrosis. Currently, the model does not cause cell death which would support the conformational change theory. The question of how long of a duration and what pressure is needed to achieve cell death must be answered. These questions must be answered to fully examine the pressure induced model of cell trauma.

Thus, cell culture models are useful tools to examine complex biological questions such as the response of a cell to trauma.

However, these models do not always accurately represent the in vivo situation. They are not a replacement for in vivo models but merely permit increased understanding of biological processes. APPENDIX A

ALUMINUM SILICATE

244 Table 12

Clay Induced Fatty Acid Release in ROC-1 Cells

Time (h) Concentration of Bentonite mg/ml

None 0.01 0.03 0.10

1 h 133.00 ± 10.10 24.00 ± 14.00 28.00 ± 6.00 8.00 ± 0.60

6 h 11.00 ± 1.00 * NA 22.00 ± 13.00 16.00 ± 8.00

24 h 30.00 ±10.00 * 35.00 ± 3.00 38.00 ± 2.00 * 14.00 ± 2.00

Time (h) Concentration of Kaolinite mg/ml

None 0.01 0.03 0.10

1 h 19.00 ± 10.00 NA 18.00 ± 2.00 * 11.00 ± 1.00 *

6 h 23.00 ± 5.00 14.00 ± 10.00 33.00 ± 2.00 * 18.00 ± 6.00

24 h 35.00 ± 3.00 36.00 ± 14.00 52.00 ± 29.00 23.00 ± 3.00

Fatty acid values represent total fatty acid release in medium and cells. Values are expressed as nmol/mg protein. All clays were incubated with the cells for 1, 6, or 24 hours at the concentrations of 0.01, 0.03, or 0.10 mg/ml as previously described. The * indicates n = 2. Values are means ± standard deviation. NA is not available. Table 12 (continued)

Clay Induced Fatty Acid Release in ROC-1 Cells

Time (h) Concentration of Erionite mg/ml

None 0.01 0.03 0.10

l h 32.31 ± 0.75* 16.90 ± 9.20 30.80 ± 29.89 34.35 ± 1.78

6 h 21.71 ± 0.78 20.96 ± 1.85 41.72 ± 17.79 55.77 ± 3.45 *

24 h 14.10 ± 4.20 25.19 ± 11.03 16.68 ± 6.32 9.16 ± 5.78

Time (h) Concentration of Montmorillonite mg/ml

None 0.01 0.03 0.10

1 h 15.54 ± 3.06 20.88 ± 2.47 31.14 ± 9.48 48.48 ± 12.56 *

6 h 25.95 ± 14.17 21.46 ± 4.16 36.81 ± 11.12 35.51 ± 14.96

24 h 10.25 ± 4.72 23.00 ± 7.99 32.55 ± 10.70 25.41 ± 4.89

Clay Induced Fatty Acid Release in N1E-115 Cells

Time (h) Concentration of Bentonite mg/ml

None 0.01 0.03 0.10

1 h 10.15 ± 3.14 6.40 ± 3.35 7.65 ± 2.24 21.70 ± 5.53

6 h 3.99 ± 1.28 8.86 ± 2.75 9.26 ± 3.55 6.73 ± 1.52

24 h 6.81 ± 2.96 7.60 ± 0.47 10.57 ± 3.79 5.49 ± 1.15

Time (h) Concentration of Kaolinite mg/ml

None 0.01 0.03 0.10

1 h 21.73 ± 10.76 20.47 ± 0.31 68.50 ± 31.34 14.08 ± 2.42

6 h 14.83 ± 8.13 30.23 ± 14.68 49.35 ± 13.66 20.84 ± 10.40

24 h 12.69 ± 1.77 61.02 ± 6.82 * 18.59 ± 4.42 16.25 ± 6.34

ro Fatty acid values represent total fatty acid release in medium and cells. Values are expressed as nmol/mg protein. All clays were incubated with the cells for 1, 6, or 24 hours at the concentrations of 0.01, 0.03, or 0.10 mg/ml as previously described. The * indicates n = 2. Values are means ± standard deviation. Table 13 (continued)

Clay Induced Fatty Acid Release in N1E-115 Cells

Time (h) Concentration of Erionite mg/ml

None 0.01 0.03 0.10

1 h 7.33 ± 1.11 9.11 ± 1.70 9.16 ± 4.84 5.55 ± 0.57

6 h 5.27 ± 2.03 9.13 ± 1.17 7.40 ± 2.09 8.03 ± 2.30

24 h 5.53 ± 1.74 11.19 ± 5.68 5.13 ± 0.40 * 25.53 ± 5.61 *

Time (h) Concentration of Montmorillonite mg/ml

None 0.01 0.03 0.10

1 h 5.68 ± 2.14 10.29 ± 8.53 4.15 ± 0.36 * 16.00 ± 2.12

6 h 16.13 ± 10.92 7.63 ± 4.10 5.36 ± 0.23 8.05 ± 1.36

24 h 4.83 ± 2.02 5.81 ± 1.83 3.59 ± 0.57 4.53 ± 0.95

Clay Induced Fatty Acid Release in HUVE Cells

Time (h) Concentration of Bentonite mg/ml

None 0.01 0.03 0.10

1 h 98.09 ± 73.89 87.57 ± 16.20 62.82 ± 14.98 60.67 ± 15.49

6 h 66.86 ± 11.25 61.70 ± 28.89 67.19 ± 11.70 * 57.10 ± 5.98

24 h 41.85 ± 4.88 117.98 ± 14.41 171.86 ± 69.84 193.81 ± 6.23 *

Time (h) Concentration of Kaolinite mg/ml

None 0.01 0.03 0.10

1 h 78.86 ± 31.85 59.25 ± 17.97 71.85 ± 10.82 53.40 ± 16.00 **

6 h 112.59 ± 14.15 * 61.16 ± 17.73 63.73 ± 28.77 70.76 ± 25.31

24 h 115.33 ± 15.57 * 58.26 ± 16.87 104.27 ± 37.23 115.04 ± 23.24 *

Fatty acid values represent total fatty acid release in medium and cells. Values are expressed as nmol/mg protein. All clays were incubated with the cells for 1, 6, or 24 £ hours at the concentrations of 0.01, 0.03, or 0.10 mg/ml as previously described. The *° * indicates n = 2. The ** indicates that only the medium was used for the assay. Values are means ± standard deviation. Table 14 (continued)

Clay Induced Fatty Acid Release in HUVE Cells

Time (h) Concentration of Erionite mg/ml

None 0.01 0.03 0.10 1 h 59.98 ± 18.94 54.17 ± 15.67 * 97.15 ± 9.67 76.59 ± 13.28

6 h 52.32 ± 18.55 115.06 ± 36.12 118.82 ±24.37 * 148.13 ± 20.34 *

24 h 38.72 ± 8.99 67.76 ± 12.09 72.54 ± 13.48 124.13 ± 47.07

Time (h) Concentration of Montmori 11 onite mg/ml

None 0.01 0.03 0.10

1 h 38.13 ±14.71 * 41.39 ± 2.49 * 25.13 ± 2.08 37.94 ± 10.81 6 h 65.38 ± 14.97 72.58 ± 35.42 41.95 ± 13.70 62.13 ± 13.48

24 h 72.98 ± 12.32 103.78 ± 15.34 435.11 ± 59.66 * 512.59 ± 99.98 *

PRESSURE

251 Table 15

Pressure Induced Fatty Acid Release in ROC-1 Cells

Pressure (atm) Duration of Pressure (min)

1 3 5 10

1 3.60 ± 1.07 9.85 ± 3.84 3.70 ± 1.46 1.98 ± 0.46

5 6.34 ± 1.49 4.90 ± 0.81 2.25 ± 0.82 4.29 ± 4.98

10 3.67 ± 1.41 5.53 ± 1.03 1.56 ± 0.96 1.90 ± 1.04

15 6.21 ± 1.27 4.35 ± 0.77 1.84 ± 0.16 1.36 ± 0.44

20 2.15 ± 0.80 3.72 ± 0.49 1.12 ± 0.05 1.64 ± 0.23

Fatty acid values represent total fatty acid release in cells and medium. Values are expressed as nmol/mg protein. Recovery time was 1 min. Values are means ± standard deviation. All numbers are n=3.

ro in ro Table 16

Pressure Induced Fatty Acid Release in ROC-1 Cells

Pressure (atm) Duration of Pressure (min)

1 3 5 10 1 1.81 ± 1.92 6.00 * 1.61 ± 0.24 2.18 ± 1.81

5 6.82 ± 4.39 8.92 ± 3.49 4.28 ± 3.28 1.15 ± 0.17

10 13.54 ± 6.48 17.50 ± 5.35 2.08 ± 1.40 1.44 ± 0.10

15 6.55 ± 0.91 9.60 ± 3.29 3.49 ± 2.29 3.13 ± 3.36

20 1.41 ± 0.34 2.97 ± 2.90 1.19 ± 0.32 1.73 ± 0.99

Fatty acid values represent total fatty acid release in cells and medium. Values are expressed as nmol/mg protein. Recovery times were 10 min. Values are means ± standard deviation. All numbers are n=3. The * indicates n=2.

no u i co Table 17

Pressure Induced LDH Release in ROC-1 Cells

Duration (min) Pressure (atm)

1 5 10 15 20

1 119 ± 43 132 ± 10 167 ± 44 185 ± 48 161 ± 42

3 132 ± 33 194 ± 46 265 ± 66 197 ± 103 245 ± 131

5 176 ± 73 143 ± 39 180 ± 43 203 ± 84 226 ± 49

10 215 ± 21 267 ± 26 161 ± 24 212 ± 23 187 ± 36

The values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The values are n=3. Recovery time was 1 min. Values are means ± standard deviation.

ro -P*u i Table 18

Pressure Induced LDH Release in ROC-1 Cells

Duration (min) Pressure (atm)

1 5 10 15 20

1 138 ± 55 198 ± 7 * 200 1 3 199 ± 40 170 ± 58

3 264 ± 100 264 ± 68 290 ± 113 297 ± 77 212 ± 52

5 183 ± 54 225 ± 84 166 ± 38 165 ± 32 258 ± 17 **

10 274 ± 56 274 ± 36 211 ± 62 173 ± 38 193 ± 22

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The values are n=3. Recovery time was 10 min. The * indicates n=2. The ** indicates the value is significant from control at a level of p < 0.05. Values are means ± standard deviation.

ro (71 (71 256

Table 19

Pressure Induced FA Release in ROC-1 Cells As A Function of Recovery Time

Recovery Time (min) Pressure (atm)

10 1

1 5.53 ± 1.03 9.85 ± 3.84

10 17.53 ± 5.35 6.15 ± 0.51

30 1.28 ± 0.40 1.93 ± 1.27

60 1.35 ± 0.28 1.43 ± 0.21

360 1.48 ± 0.08 2.53 ± 0.83

720 2.34 ± 0.93 3.67 ± 1.82

1440 4.39 ± 3.87 1.68 ± 1.20

Fatty acid values represent fatty acid release in medium and cells. Values are expressed as nmol/mg protein. The pressure duration was 3 min. Values are means ± standard deviation. All numbers are n=3. 257

Table 20

Pressure Induced LDH Release in ROC-1 Cells As A Function of Recovery Time

Recovery Time (min) Pressure (atm)

10 1

1 265 ± 66 163 ± 33

10 290 ± 113 204 ± 100

30 127 ± 14 107 ± 5

60 146 ± 43 119 ± 20

360 121 ± 8 140 ± 22

720 148 ± 30 129 ± 7

1440 209 ± 51 180 ± 5

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The pressure duration was 3 min. Values are means ± standard deviation. All numbers are n=3. Table 21

Pressure Induced Fatty Acid Release in N1E-115 Cells

Pressure (atm) Duration of Pressure (min)

1 3 5 10

1 10.26 ± 5.66 10.26 ± 5.66 9.64 ± 1.93 4.96 ± 0.54

5 5.74 ± 1.76 7.88 ± 1.99 3.82 ± 1.38 5.46 ± 0.51

10 6.31 ± 1.30 14.63 ± 2.55 6.46 ± 2.32 5.70 ± 0.15

15 16.40 ± 1.38 11.02 ± 1.01 8.66 ± 1.94 7.01 ± 2.70

20 4.93 ± 0.67 10.73 ± 8.04 10.10 ± 3.35 7.77 ± 3.12

Fatty acid values represent total fatty acid release in medium and cells. Values are expressed as nmol/mg protein. Recovery time was 1 min. Values are means ± standard deviation. All numbers are n=3.

ro CJl 00 Table 22

Pressure Induced Fatty Acid Release in N1E-115 Cells

Pressure (atm) Duration of Pressure (min)

1 3 5 10

1 7.51 ± 1.47 7.51 ± 1.47 6.64 ± 3.96 7.26 ± 1.89

5 6.75 ± 2.16 7.13 ± 1.25 5.98 ± 1.70 11.25 ± 7.12

10 23.50 ± 1.50 10.27 ± 8.70 10.25 ± 1.54 9.22 ± 1.47

15 10.12 ± 6.33 43.80 ± 16.47 9.16 ± 3.24 8.18 ± 3.19

20 6.37 ± 2.93 4.23 ± 1.98 11.11 ± 3.58 6.55 ± 1.06

Fatty acid values represent total fatty acid release in medium and cells. Values are expressed as nmol/mg protein. Recovery times were 10 min. Values are means ± standard deviation. All numbers are n=3.

ro (71 to Table 23

Pressure Induced LDH Release in N1E-115 Cells

Duration (min) Pressure (atm)

1 5 10 15 20

1 244 ± 129 174 + 44 278 ± 144 294 ± 86 221 ± 146

3 219 ± 46 162 ± 45 207 ± 17 164 ± 65 198 ± 27

5 103 ± 10 80 ± 4 117 ± 21 113 ± 6 * 172 ± 49 **

10 139 ± 39 116 ± 27 109 ± 7 135 ± 7 181 ± 84

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The values are n=3. Recovery time was 1 min. The * indicates n=2. The ** indicates the value is significant from control at a level of p < 0.05. Values are means ± standard deviation.

ro cr> o Table 24

Pressure Induced LDH Release in N1E-115 Cells

Duration (min) Pressure (atm)

1 5 10 15 20

1 157 ± 44 157 ± 8 197 ± 112 224 ± 20 ** 233 ± 64

3 206 ± 80 298 + 230 206 ± 140 251 ± 71 148 ± 13

5 70 ± 36 91 + 11 119 ± 13 140 ± 28 191 ± 47 **

10 137 ± 14 143 ± 13 115 ± 13 162 ± 37 124 ± 62

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The values are n=3. Recovery time was 10 min. The ** indicates the value was significant from control at a level of p < 0.05. Values are means ± standard deviation. 262

Table 25

Pressure Induced FA Release in N1E-115 Cells As A Function of Recovery Time

Recovery Time (min) Pressure (atm)

15 1

1 11.02 ± 1.01 9.64 ± 1.93

10 43.80 ± 16.47 6.64 ± 3.96

30 3.90 ± 0.35 5.48 ± 2.06

60 9.32 ± 2.49 7.14 ± 2.70

360 7.91 ± 2.51 7.69 ± 1.62

720 37.64 ± 19.29 11.85 ± 1.97

1440 7.69 ± 0.97 4.44 ± 0.76

Fatty acid values represent fatty acid release in medium and cells. Values are expressed as n/mol/mg protein. The pressure duration was 3 min. Values are means ± standard deviation. All numbers are n=3. 263

Table 26

Pressure Induced LDH Release in N1E-115 Cells As A Function of Recovery Time

Recovery Time (min) Pressure (atm)

15 1

1 164 + 65 219 ± 46

10 251 ± 71 206 ± 80

30 131 ± 14 CM CM o ± 83

60 170 ± 83 151 + 63

360 255 ± 21 315 ± 56

720 282 + 38 304 + 46

1440 638 ± 184 ** 279 + 66 *

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The pressure duration was 3 min. Values are means ± standard deviation. All numbers are n=3. The * indicates n=2. The ** indicates the value was significant from its matched control at p < 0.05. Table 27

Pressure Induced Fatty Acid Release in HUVE Cells

Pressure (atm) Duration of Pressure (min)

1 3 5 10

1 16.08 ± 3.82 23.02 ± 9.21 41.50 ± 24.90 23.42 ± 7.58 5 65.35 ± 10.07 19.05 ± 5.37 27.53 ± 5.63 55.71 ± 7.74

10 16.47 ± 4.36 14.26 ± 7.90 48.09 ± 19.42 66.06 ± 5.82

15 21.48 ± 8.90 28.36 ± 1.85 35.12 ± 2.32 67.19 ± 15.64

20 8.97 ± 4.39 21.66 ± 3.29 32.72 ± 0.66 35.87 ± 8.17

ro cn Table 28

Pressure Induced Fatty Acid Release in HUVE Cells

Pressure (atm) Duration of Pressure (min)

1 3 5 10

1 13.13 ± 2.34 27.28 ± 4.95 35.15 ± 1.47 33.77 ± 6.94

5 22.58 ± 7.66 17.42 ± 5.98 30.18 ± 4.14 15.30 ± 8.47

10 17.10 ± 4.15 17.14 ± 1.06 36.21 ± 0.16 78.59 ± 4.99

15 14.17 ± 3.80 31.22 ± 4.40 26.31 ± 3.73 83.09 ± 15.04

20 18.77 ± 6.53 26.09 ± 2.68 36.97 ± 8.36 45.46 ± 18.89

ro (71a i Table 29

Pressure Induced LDH Release in HUVE Cells

Duration (min) Pressure (atm)

1 5 10 15 20

1 685 ± 60 758 ± 82 737 ± 85 630 ± 66 642 ± 43

3 968 ± 214 956 ± 52 979 ± 89 967 ± 98 990 ± 85

5 1307 ± 439 834 ± 75 1047 ± 133 984 ± 85 934 ± 185

10 836 ± 65 863 ± 174 887 ± 110 688 ± 101 790 ± 76

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The values are n=3. Recovery time was 1 min. Values are means ± standard deviation.

CTlro o i Table 30

Pressure Induced LDH Release in HUVE Cells

Duration (min) Pressure (atm)

1 5 10 15 20

1 657 ± 74 763 ± 15 633 ± 44 623 ± 116 741 ± 89

3 749 ± 66 794 ± 184 837 ± 131 970 ± 149 930 ± 178

5 841 ± 276 1121 ± 353 1313 ± 476 ** 1076 ± 62 1010 ± 85

10 975 ± 274 736 ± 89 1072 ± 507 1029 ± 271 1288 ± 471

O)ro 268

Table 31

Pressure Induced FA Release in HUVE Cells As A Function of Recovery Time

Recovery Time (min) Pressure (atm)

15 1

1 67.19 ± 15.64 23.42 ± 7.58

10 83.09 ± 15.04 33.77 ± 6.94

30 33.80 ± 1.86 48.74 ± 21.20

60 39.49 ± 13.68 49.89 ± 30.44

360 67.38 + 33.74 65.65 ± 16.33

720 133.54 ± 62.61 139.90 ± 12.94

1440 72.33 ± 39.45 22.68 ± 7.30

Fatty acid values represent fatty acid release in medium and cells. Values are expressed as n/mol/mg protein. The pressure duration was 10 min. Values are means ± standard deviation. A ll numbers are n=3. 269

Table 32

Pressure Induced LDH Release in HUVE Cells As A Function of Recovery Time

Recovery Time (min) Pressure (atm)

15 1

1 688 ± 101 836 ± 65

10 1029 ± 271 975 ± CM *3*

30 510 ± 265 598 ± 250

60 633 ± 515 794 ± 117

360 1037 ± 233 1063 ± 130

720 1934 ± 233 1936 ± 112

1440 1394 ± 317 ** 985 ± 133

Values represent U/mg protein. A unit is defined as 1 nmol/min/mg protein. The pressure duration was 10 min. Values are means ± standard deviation. All numbers are n=3. The ** indicates the value was significant from matched control at a level of p < 0.05. BIBLIOGRAPHY

270 271

BIBLIOGRAPHY

Akesson B., Elovsson J. and Arvidsson G. (1970) Initial incorporation into rat liver glycerolipids of intraportally injected [3H]glycerol. Biochimica et Biophysica Acta 210, 15-27.

Alfrey A.C., LeGendre G.R. and Kaehney W.D. (1976) The dialysis encephalopathy syndrome: Possible aluminum intoxication. N. Engl. J. Med. 294, 184-188.

Altman J. (1989) A nose for Alzheimer's disease? Nature 337, 688- 688.

Amador E., Oorfman L.E. and Uacker E.W.C. (1963) Serum lactic dehydrogenase: An analytical assessment of current assays. C lin . Chem. 9, 391.

Amar S., Kitabgi P. and Vincent J.P. (1987) Stimulation of inositol phosphate production by neurotensin in neuroblastoma N1E-115 cells: Implication of GTP-binding proteins and relationship with the cyclic GMP response. Journal of Neurochemistry 49, 999-1006.

Anderson D.K. and Means E.D. (1985) Iron-induced lipid peroxidation in spinal cord: Protection with mannitol and methyl prednisolone. J. Free Rad. B io l. Med. 1, 59-64.

Anderson D.K., Means E.D. and Water T.R. (1980) Spinal cord energy metabolism in normal and post laminectomy cats. J. Neurosurg. 52, 387-391.

Anderson D.K., Saunders R.D., Demediuk P., Dugan L., Braughler J.M., Hall E.D., Means E.D. and Horrocks L.A. (1985) Lipid hydrolysis and peroxidation in injured spinal cord: Partial protection with methyl prednisolone or vitamin E and Selenium. Central Nervous System Trauma 2, 257-267.

Arieff A.I. (1985) Aluminum and pathogenesis of dialysis encephalopathy. Am. J. Kidney 6, 317-321.

Balentine J.D. (1983) Calcium toxicity as a factor in spinal cord injury. Survey and Synthesis of Pathology Research 2, 184-193. 272

Balentine J.D., Greene W.B. and Bornstein M. (1988) In vitro spinal cord trauma. Lab. In vest. 58, 93-99.

Balentine J.D. (1988) Impact injuries of the spine and spinal cord, in Forensic Neuropathology (Leestma J.E. and Kirkpatrick J.B., eds.), pp. 254-275. Raven Press, New York.

Banik N.L., Happel R.D., Sostek M.B., Chiu F.C. and Hogan E.L. (1987) Ca2+-mediated degradation of central nervous system (CNS) proteins: Topographic and species variation. Metabolic Brain Disease 2, 117-126.

Banik N.L., Hogan E.L. and Hsu C.Y. (1987) The multimolecular cascade of spinal cord injury: Studies on prostanoids, calcium, and proteinases. Meurochem. Pathol. 7, 57-77.

Barkai G., Goldman B., Mashiach S. and Shinitzky M. (1983) The effect of physiologically relevant pressures on the phase transition of liposomes containing dipalmitoyl phosphatidylcholine. J. C olloid Interface Sci. 94, 343-347.

Barrow G.M. (1979) Physical Chemistry. McGraw-Hill Book Company, New York, NY.

Beal M.F., Mazurek M.F., Ellison D.W., Kowall N.W., Solomon P.R. and Pendlebury W.W. (1989) Neurochemical characteristics of aluminum-induced neurofibrillary degeneration in rabbits. Neuroscience 29, 339-346.

Bell R.L., Baenziger N.L. and Majerus P.W. (1980) Bradykinin-stimulated release of arachidonate from phosphatidyl inositol in mouse fibrosarcoma cells. Prostaglandins 20, 269-274.

Bergelson L.D. and Barsukov L.I. (1977) Topological asymmetry of phospholipids in membranes. Science 197, 224-230.

Bhagyalakshmi A. and Frangos J.A. (1989) Mechanism of shear-induced prostacyclin production in endothelial cells. Biochem. Biophys. Res. Commun. 158, 31-37.

Birchall J.D. and Chappell J.S. (1988) The chemistry of aluminum and silicon in relation to Alzheimer's disease. C lin . Chem. 34, 265-267.

Birchall J.D. and Chappell J.S. (1988) Aluminum, chemical physiology, and Alzheimer's disease. Lancet i i , 1008-1010. 273

Black P., Markowitz R.S., Cooper V., Mechanic A., Kushner H., Damjanov I ., Finkelstein S.D. and Wachs K.C. (1986) Models of spinal cord injury: Part 1. Static load technique. Neurosurgery 19, 752-762.

Black P., Markowitz R.S., Finkelstein S.D., McMonagle-Strucko K. and Gillespie J.A. (1988) Experimental spinal cord injury: Effect of a calcium channel antangonist (nicardipine). Neurosurgery 22, 61-66.

Blank M.L., Spector A.A., Kaduce T.L. and Snyder F. (1986) Composition and incorporation of [3H]arachidonic acid into molecular species of phospholipid classes by cultured human endothelial cells. Biochimica et Biophysica Acta 877, 211-215.

Bligh E.G. and Dyer W.J. (1959) A rapid method of total lipid extraction and purification. Canadian Journal o f Biochemistry and Physiology 37, 911-917.

Boonstra J ., Nelemans S.A., Feijen A., Bierman A., Van Zoelen E .J.J., Van Der Saag P.T. and de Laat S.W. (1982) Effect of fatty acids on plasma membrane lipid dynamics and cation permeability in neuroblastoma cells. Biochim. Biophys. Acta 692, 321-329.

Bowman R.E. and Wolf R.C. (1962) A rapid and specific ultramicro method for total serum cholesterol. C lin . Chem. 8, 302-309.

Bracken M.B., Freeman D.H. and Hellenbrand K. (1981) Incidence of acute traumatic hospitalized spinal cord injury in the United States, 1970-1977. J. Epidemiol. 113, 615-622.

Bradford M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254.

Braquet P., Touqui L., Shen T.Y. and Vargaftig B.B. (1987) Perspectives in platelet-activating factor research. Pharmacol. Rev. 39, 97-145.

Bretscher M.S. (1973) Membrane structure: Some general principles. Science 181, 622-629.

Brindley G.W., and Brown G. (1980) Crystal structures of clay minerals and their x-ray diffraction. Ballantyne Ltd., London and Colchester.

Brockerhoff H. (1975) Determination of the positional distribution of fatty acids in glycerolipids. Heth. Enzymol. 35, 315-325. 274

Buchanan M.R., Butt R.W., Magas Z., Van Ryn J ., Hirsh J. and Nazir D.J. (1985) Endothelial cells produce a lipoxygenase derived chemo-repellent which influences platelet/endothelial cells interactions - effect of aspirin and salicylate. Thrombosis and Haemostasis 53, 306-311.

Candy J.M., Klinowski J ., Perry R.H., Perry E.K., Fairbairn A., Oakley A.E., Carpenter T.A., Atack J.R ., Blessed G. and Edwardson J.A. (1986) Aluminosilicates and senile plaque formation in Alzheimer's disease. Lancet i, 354-357.

Champeil P., Buschlen S. and Guillain F. (1981) Pressure-induced inactivation of sarcoplasmic reticulum adenosine triphosphate during high-speed centrifugation. Biochemistry 20, 1520-1524.

Chan P.H., Chen S.F. and Yu A.C.H. (1988) Induction of intracellular superoxide radical formation by arachidonic acid and by polyunsaturated fatty acids in primary astrocytic cultures. Journal of Neurochemistry 50, 1185-1193.

Chan P.H. and Fishman R.A. (1978) Brain edema: Induction in cortical slices by polyunsaturated fatty acids. Science 201, 358-360.

Chan P.H., Kerlan R. and Fishman R.A. (1983) Reductions of Y-aminobutyric acid and glutamate uptake and (Na+,K+)-ATPase activity in brain slices and synaptosomes by arachidonic acid. Journal of Neurochemistry 40, 309-316.

Chin J.H., Trudell J.R. and Cohen E.N. (1976) The compression-ordering and solubility-disordering effects of high pressure gases on phospholipid bilayers. L ife S ci. 18, 489-498.

Chong P.L.G. and Cossins A.R. (1984) Interacting effects of temperature, pressure and cholesterol content upon the molecular order of dioleoylphosphatidylcholine vesicles. Biochim. Biophys. Acta 772, 197-201.

Clark M.A., Littlejohn D., Mong S. and Crooke S.T. (1986) Effect of leukotrienes, bradykinin and calcium ionophore A 23187 on bovine endothelial cells: Release of prostacyclin. Prostaglandins 31, 157-166.

Clendenon N.R., Palayoor S.T. and Gordon W.A. (1985) Influence of CDPcholine on ATPase activity in acute experimental spinal cord trauma, in Novel Biochemical, Pharmacological and Clinical Aspects of Cytidinediphosphocholine (Zappia V., Kennedy E.P., Nilsson B.I. and Galletii P., eds.), pp. 275-284. Elsevier, Amsterdam.

Connor D.J., Jope R.S. and Harrell L.E. (1988) Chronic, oral aluminum administration to rats: Cognition and cholinergic parameters. Pharmacol. Biochem. Behav. 31, 467-474. 275

Cossins A.R. and MacDonald A.G. (1984) Homeoviscous theory under pressure. II. The molecular order of membranes from deep-sea fish. Biochimica et Biophysica Acta 776, 144-150.

Crapper McLaehlan D.R. (1986) Aluminum and Alzheimer's disease. Neurobiol. Aging 7, 525-532.

Crapper McLaehlan D.R., Dam T.V., Farnell B.J. and Lewis P.N. (1983) Aluminum inhibition of ADP-ribosylation in vivo and in vitro. Neurobehavioral Toxicology Teratology 5, 645-647.

Crapper D.R., Krishnan S.S. and Quittkat S. (1976) Aluminum, neurofibrillary degeneration and Alzheimer's disease. Brain 99, 67-80.

Crutchley D.J., Ryan J.W., Ryan U.S. and Fisher G.H. (1983) Bradykinin induced release of prostacyclin and thromboxanes from bovine pulmonary artery endothelial cells. Biochimica et Biophysica Acta 751, 99-107.

D'Amato R.A., Horrocks L.A. and Richardson K.E. (1975) Kinetic properties of plasmalogenase from bovine brain. J. Neurochem. 24, 1251-1255.

Davies R., G riffiths D.M., Johnson N.F., Preece A.W. and Livingston D.C. (1984) The cytotoxicity of kaolin towards macrophages in vitro. Br. J . Exp. Pathol. 65, 453-466.

Davies R. and Preece A.W. (1983) The electrophoretic mobilities of minerals determined by laser Doppler velocimetry and their relationship with the biological effect of dusts towards macrophages. Clinical Physics of Physiological Measures 4, 129-140. de la Torre J.C. (1981) Spinal cord injury: Review of basic and applied research. Spine 6, 315-335.

Demediuk P., Anderson D.K., Horrocks L.A. and Means E.D. (1985) Mechanical damage to murine neuronal-enriched cultures during harvesting: Effects on free fatty acids, diglycerides, Na+,K+-ATPase, and lipid peroxidation. In Vitro Cellular & Developmental Biology 21, 569-574.

Demediuk P., Saunders R.D., Anderson D.K., Means E.D. and Horrocks L.A. (1985) Membrane lipid changes in laminectomized and traumatized cat spinal cord. Proc. N a tl. Acad. S ci. USA 82, 7071-7075.

Demediuk P., Saunders R.D., Clendenon N.R., Means E.D., Anderson D.K. and Horrocks L.A. (1985) Changes in lipid metabolism in traumatized spinal cord. Prog. Brain Res. 63, 211-226. 276

Demediuk P. and Faden A.I. (1988) Traumatic spinal cord injury in rats causes increases in tissue thromboxane but not peptidoleukotrienes. J. Neurosci. Res. 20, 115-121.

Demopoulos H.6., Flamm E.S., Seligman M.L., Ransohoff J. and Naftchi N.E. (1984) Molecular pathogenesis of spinal cord degeneration after traumatic injury, in (pp. 45-64. , .

Der W.A., Howie R.A., and Zussman J. (1989) Framework silic a te s . Longmans Green & Co., London.

Derian C.K. and Moskowitz M.A. (1985) Bradykinin (BK) stimulates polyphosphoinositide metabolism in cultured endothelial cells. Fed. Proc. 44, 4862.

Derian C.K. and Moskowitz M.A. (1986) Polyphosphoinositide hydrolysis in endothelial cells and carotid artery segments. J. B io l. Chem. 261, 3831-3837.

Deshmukh D.S., Vorbrodt A.W., Lee P.K., Bear W.D. and Kuizon S. (1988) Studies on the submicrosomal fractions of bovine oligodendroglia: Lipid composition and glycolipid biosynthesis. Neurochem. Res. 13, 571-582.

DeSmedt H., Borghgraef R., Ceuterick F. and Heremans K. (1979) Pressure effects on lipid-protein interactions in (Na+,K+)-ATPase. Biochimica et Biophysica Acta 556, 479-489.

Drejer J ., Benveniste H., Diemer N.H. and Schousboe A. (1985) Cellular origin of ischemia-induced glutamate release from brain tissue in vivo and in vitro. J. Neurochem. 45, 145-151.

Dugan L.L., Bazan N.G. and Horrocks L.A. (1986) Extraction and HPLC separation of polyphosphoinositides by neutral solvents. Trans. Am. Soc. Neurochem. 17, 138-138.

Dugan L.L., Demediuk P., Pendley II C.E. and Horrocks L.A. (1986) Separation of phospholipids by high pressure liquid chromatography: All major classes, including ethanol amine and choline plasmalogens, and most minor classes, including lysophosphatidylethanolamine. J. Chromatogr. 378, 317-327.

Ebbecke (J. (1936) Einwirkung hoher Drucke auf glattmusklige Organe (Froschmagenpraparat). Archives fu r de gesamte Physiologie von P flu g er 237, 771-784.

Ebbecke U. (1936) Uber das Verhalten des Zentralnervensystems (Ruckenmarksfrosch) unter der Einwirkung hoher Drucke. Archives fur de gesamte Physiologie von P fluger March 10, 785-789. 277

El-Fakahany E. and Richelson E. (1980) Regulation of muscarinic receptor-mediated cyclic GMP synthesis by cultured mouse neuroblastoma cells. Journal of Neurochemistry 35, 941-948.

Esiri M.M. and Wilcock G.K. (1984) The olfactory bulb in Alzheimer's disease. J. Neurol. Neurosurg. Psychiatry 47, 56-60.

Faden A .I., Chan P.H. and Longar S. (1987) Alterations in lipid metabolism, Na+,K+-ATPase activity, and tissue water content of spinal cord following experimental traumatic injury. Journal of Neurochemistry 48, 1809-1816.

Faden A .I., Jacobs T.D. and Holaday J.U. (1981) Opiate antagonists improves neurologic recovery recovery after spinal injury. Science 211, 493-494.

Ford R.U.J. (1983) A reproducible spinal cord injury model in the cat. J. Neurosurg. 59, 268-275.

Franceschi M., Perego L., Ferini-Strambi L., Smirne S. and Canal N. (1988) Neuroendocrinological function in Alzheimer's disease. Neuroendocrinology 48, 367-370.

Gercken G., Kahlert H. and Gulyas H. (1988) Mineral dust-induced release of platelet activating factor (PAF) by alveolar macrophages. J. Aerosol Sci. 19, 1141-1144.

German D.C., White C.L. and Sparkman D.R. (1987) Alzheimer's disease: Neurofibrillary tangles in nuclei that project to the cerebral cortex. Neuroscience 21, 305-312.

Gimbrone M.A., Cotran R.S. and Folkman J. (1974) Human vascular endothelial cells in culture. J. Cell Biol. 60, 673-684.

Goldberg W.J., Dorman R.V., Dabrowiecki Z. and Horrocks L.A. (1985) The effects of ischemia and CDPamines on Na+,K+-ATPase and acetylcholinesterase activities in rat brain. Neurochem. Pathol. 3, 237-248.

Goldstein R.H., Miller K., Glassroth J., Linscott R., Snider G.L., Franzblau C. and Polgar P. (1982) Influence of asbestos fibers on collagen and prostaglandin production in fibroblast and macrophage co-cultures. J . Lab. C lin . Med. 100, 778-785.

Gonatas N.K., Hirayama M., Stieber A. and Silberberg D.H. (1982) The ultrastructure of isolated rat oligodendroglial cell cultures. J. Neurocytol. 11, 997-1008.

Gorman R.R., Oglesby T.D., Bundy G.L. and Hopkins N.K. (1985) Evidence for 15-HETE synthesis by human endothelial cells. P la te le ts and Vascular Occlusion 72, 708-712. 278

Guha A., Tator C.H. and Piper I. (1987) Effect of a calcium channel blocker on posttraumatic spinal cord blood flow. J. Neurosurg. 66, 423-430.

Haag W.O., Lago R.H. and Weisz P.B. (1984) The active site of acidic aluminosilicate catalysts. Nature 309, 589-591.

Haghighi S.S. and Chehrazi B. (1987) Effect of naloxone in experimental acute spinal cord injury. Neurosurgery 20, 385-388.

Haines T.H. (1982) A model for transition state dynamics in bilayers. Biophys. J. 37, 147-148.

Hall A.C. and Ellory J.C. (1986) Hydrostatic pressure effects on membrane transport in red cells, in Topics in Lipid Research: From Structural Elucidation to Biological Function (Klein R.A. and Schmitz B., eds.), pp. 331-336. Royal Society of Chemistry, .

Hall A.C., Ellory J.C. and Klein R.A. (1982) Pressure and temperature effects on human red cell cation transport. J. Hembr. B io l. 68, 47-56.

Hall E.D. and Wolf D.L. (1986) A pharmacological analysis of the pathophysiological mechanisms of posttraumatic spinal cord ischemia. J. Neurosurg. 64, 951-961.

Halldorsson H., Kjeld M. and Thorgeirsson G. (1988) Role of phosphoinositides in the regulation of endothelial prostacyclin production. Arteriosclerosis 8, 147-154.

Hammerson F. and Hammerson E. (1985) Some structural and functional aspects of endothelial cells. Basic Research in Cardiology 80, 491-501.

Hara A. and Radin N.S. (1978) Lipid extraction of tissues with a low-toxicity solvent. Anal. Biochem. 90, 420-426.

Harington J.S ., Allison A.C. and Badami D.V. (1975) Mineral fibers: Chemical, physiochemical, and biological properties. Adv. Pharmacol. Chemother. 12, 291-402.

Harley J.D. and Margolis J. (1961) Haemolytic activity of colloidal silic a . Nature 189, 1010-1011.

Hawthorne J.N. and Kai M. (1970) Metabolism of phosphoinositides, in Handbook of Neurochemistry (Lajtha A., ed.), pp. 491-508. Plenum Press, New York. 279

Heliums J.D., Peterson D.M., Stathopoulos N.A., Moake J.L. and Giorgio T.D. (1987) Studies on the mechanisms of shear-induced platelet activation, in Cerebral Ischemia and Hemorheology (Hartmann A. and Kuschinsky W., eds.), pp. 80-89. Springer-Verlag Berlin, Heidelberg.

Heremans K. (1982) High pressure effects on proteins and other biomolecules. Annu. Rev. Biophys. Bioeng. 11, 1-21.

Heremans K. and Wuytack F. (1980) Pressure effect on the Arrhenius discontinuity in Ca+-ATPase from sarcoplasmic reticulum. FEBS L e tt. 117, 161-163.

Heyn M.P. (1979) Determination of lipid order parameters and rotational correlation times from fluorescence depolarization experiments. FEBS L e tt. 108, 359-364.

Hillered L. and Chan P.H. (1988) Role of arachidonic acid and other free fatty acids in mitochondrial dysfunction in brain ischemia. J. Neurosci. Res. 20, 451-456.

Hochachka P.W. (1974) Temperature and pressure adaptation of the binding site of acetylcholinesterase. Biochem. J. 143, 535-539.

Hochachka P.W. (1975) How abyssal organisms maintain enzymes of the "right" size. Comp. Biochem. Physiol. B 52, 39-41.

Hong S.L. and Deykin D. (1982) Activation of phospholipase A 2 and C in pig aortic endothelial cells synthesizing prostacyclin. J. B io l. Chem. 257, 7151-7154.

Horrocks L.A., Dorman R.V. and Porcellati G. (1984) Fatty acids and phospholipids in brain during ischemia, in Cerebral Ischemia (Bes A., Braquet P., Paoletti R. and Siesjo B.K., eds.), pp. 211-222. Elsevier, Amsterdam.

Horrocks L.A., Harder H.W., Mozzi R., Goracci G., Francescangeli E., Porcellati S. and Nenci G.G. (1986) Receptor mediated degradation of choline plasmalogen and glycerophospholipid methylation: A new hypothesis, in Enzymes of Lipid Metabolism, Vol. 2 (Freysz L., Dreyfus H., Massarelli R. and Gatt S., eds.), pp. 707-711. Plenum Press, New York.

Horrocks L.A., Yeo Y.K., Harder H.W., Mozzi R. and Goracci G. (1986b) Choline plasmalogens, glycerophospholipid methylation, and receptor-mediated activation of adenylate cyclase, in Advances in Cyclic Nucleotide and Protein Phosphorylation Research (Greengard P. and Robinson G.A., eds.), pp. 263-292. Raven Press, New York. 280

Hsu C.Y., Halushka P.V., Hogan E.L., Banik N.L., Lee W.A. and Perot P.L. (1985) Alteration of thromboxane and prostacyclin levels in experimental spinal cord injury. Neurology 35, 1003-1009.

Hsu C.Y., Halushka P.V., Hogan E.L. and Cox R.D. (1986) Increased thromboxane level in experimental spinal cord injury. J. Neurol. Sci. 74, 289-296.

Hsu C.Y., Hogan E.L., Gadsden Sr. R.H., Spicer K.M., Shi M.P. and Cox R.D. (1985) Vascular permeability in experimental spinal cord injury. J. Neurol. Sci. 70, 275-282.

Huang H.M. and Gibson G.E. (1989) Phosphatidyl inositol metabolism during in vitro hypoxia. Journal of Neurochemistry 52, 830-835.

Ikeda M., Yoshida S., Busto R., Santiso M. and Ginsberg M.D. (1986) Polyphosphoinositides as a probable source of brain free fatty acids accumulated at the onset of ischemia. J. Neurochem. 47, 123-132.

Ingerman-Wojenski C., Silver M.J., Smith J.B. and Macarck E. (1981) Bovine endothelial cells in culture produce thromboxane as well as prostacyclin. J. Clin. Invest. 67, 1292-1296.

Jaffe E.A., Hoyer L.W. and Wachmar R.L (1973) Synthesis of antihemophilic factor antigen by cultured human endothelial cells. J. Clin. Invest. 52, 2757-2764.

Jaffe S., Novak P., Sorgente N. and Kalra V.K. (1985) Phospholipid methylation in isolated luminal membranes of bovine aortic endothelial cells: Effect of prostacyclin. Fed. Proc. 44, 388.

Jahnig F. (1979) Structural order of lipids and proteins in membranes: Evaluation of fluorescence anisotropy data. Proc. Natl. Acad. S ci. USA 76, 6361-6365.

Jakubovicz D.E., Grinstein S. and Klip A. (1987) Cell swelling following recovery from acidification in C-6 glioma cells: an in vitro model of postischemic brain edema. Brain Res. 435, 138-146.

Jaurand M.C., Magne L. and Bignon J. (1979) Inhibition by phospholipids of haemolytic action of asbestos. Br. J. Indust. Ned. 36, 113-116.

Jayawardena A.N. and Waksman B.H. (1977) Surface charge and asbestos toxicity. Nature 265, 537-541.

Johns R.A., Izzo N.J., Milner P .J., Loeb A.L. and Peach M.J. (1988) Use of cultured cells to study the relationship between arachidonic acid and endothelium-derived relaxing factor. Am. J . Ned. S ci. 31, 287-292. 281

Jones M., Keenan R.W. and Horowitz P. (1982) Use of 6-p-toluidino-2-naphthalenesulfonic acid to quantitate lipids after thin-layer chromatography. J. Chromatogr. 237, 522-524.

Kay D.W.K. (1987) Heterogeneity in Alzheimer's disease: Epidemiological and family studies. Trends Neurosci. 10, 194-195.

Kikkawa U. and Nishizuka Y. (1987) Inositol phospholipid turnover and protein kinase C in stimulus-response coupling. Biochem. Soc. Trans. 15, 124.

Koh J.Y. and Choi D.W. (1987) Quantitative determination of glutamate mediated cortical neuronal injury in cell culture by lactate dehydrogenase efflux assay. J. Neurosci. Meth. 20, 83-90.

Klausner R.D., Kleinfeld A.M., Hoover R.L. and Karnovsky M.J. (1980) Lipid domains in membranes; Evidence derived from structural perturbations induced by free fatty acids and lifetim e heterogeneity analysis. J. B io l. Chem. 255, 1286-1295.

Kushner H., Markowitz R.S., Mechanic A. and Black P. (1986) Models of spinal cord injury: Part 2. A mathematical model. Neurosurgery 19, 763-766.

Lagarde M., Sicard B., Guichardant M., Felisi 0. and Dechavanne M. (1984) Fatty acid composition in native and cultured human endothelial cells. In V itro 20, 33-37.

Laiho K.U., Berezesky I.K. and Trump B.F. (1983) The role of calcium in cell injury. Survey and Synthesis of Pathology Research 2, 170-183.

Lakowicz J.R. and Thompson R.B. (1983) Differential polarized phase fluorometric studies of phospholipid bilayers under high hydrostatic pressure. Biochimica et Biophysica Acta 732, 359-371.

Lee A.G. (1983) Lipid phase transitions and mixtures, in Membrane Fluidity in Biology, Vol. 2 (Aloia R.C., ed.), pp. 43-88. Academic Press, New York.

Lipman J .J ., Colowick S.P., Lawrence P.L. and Abumrad N.N. (1988) Aluminum induced encephalopathy in the rat. L ife S ci. 42, 863-875.

Liu N.I. and Kay R.L. (1977) Redetermination of the pressure dependence of the lipid bilayer phase transition. Biochemistry 16, 3484-3486.

Lo W.W.Y. and Fan T.P.D. (1987) Histamine stimulates inositol phosphate accumulation via the Hi-receptor in cultured human endothelial cells. Biochem. Biophys. Res. Commun. 148, 47-53. 282

Low M.G., Ferguson Futerman A.H. and Silman I. (1986) Covalently attached phosphatidyl inositol as a hydrophobic anchor for membrane proteins. Trends Biochem. S c i. 11, 212-215.

Low P.S. and Somero G.N. (1975) Pressure effects on enzyme structure and function in vitro and under simulated in vivo conditions. Comp. Biochem. Physiol. B 52, 67-74.

Low R.B., Leffingwell C.M. and Bulman C.A. (1980) Effects of kaolinite on amino acid transport and incorporation into protein by rabbit pulmonary alveolar macrophages. Arch. Environ. Health 35, 217-223.

Loffelholz K. (1989) Receptor regulation of choline phospholipid hydrolysis: A novel source of diacylglycerol and phosphatidic acid. Biochem. Pharmacol. 38, 1543-1549.

MacDonald A.G. (1986) The role of membrane lipids in the adaptation of deep sea organisms to high pressure, in Topics in Lipid Research: From Structural Elucidation to Biological Function (Klein R. and Schmitz B., eds.), pp. 319-330. Royal Society of Chemistry.

MacDonald A.G. and G ilchrist I. (1982) The pressure tolerance of deep sea amphipods collected at th eir ambient high pressure. Comp. Biochem. Physiol. A 71, 349-352.

Maciag T., Cerundolo J ., Ilsley S., Kelly P.R. and Forand R. (1979) An endothelial cell growth factor from bovine hypothalmus: Identification and partial characterization. Proc. N a tl. Acad. S ci. USA 76, 5674-5678.

MacNab G. and Harington J.S. (1967) Haemolytic activity of asbestos and other mineral dusts. Nature 214, 522-523.

Majerus P.W., Wilson D.B., Connolly T.M., Bross T.E. and Neufeld E.J. (1985) Phosphoinositide turnover provides a link in stimulus-response coupling. Trends Biochem. S ci. 10, 168-171.

Majewska M.D., Strosznajder J. and Lazarewicz J. (1978) Effect of ischemic anoxia and barbiturate anesthesia on free radical oxidation of mitochondrial phospholipids. Brain Res. 158, 423-434.

Marcus A .J., Weksler B.B. and Jaffe E.A. (1978) Enzymatic conversion of prostaglandin endoperoxide H 2 and arachidonic acid to prostacyclin by cultured human endothelial cells. J. Biol. Chem. 253, 7138-7141.

Martin R.B. (1986) The chemistry of aluminum as related to biology and medicine. C lin . Chem. 32, 1797-1806. 283

Martin T.W. and Wysolmerski R.B. (1987) Ca2+-dependent and Ca2+-independent pathways for release of arachidonic acid from phosphatidyl inositol in endothelial cells. J. B io l. Chem. 262, 13086-13092.

Matheson D.F., Oei R. and Roots B.I. (1980) Influence of diet on the acyl composition of phospholipids in endothelial cells and mitochondria of rat brain. Neurochem. Res. 5, 43-59.

Matheson D.F., Oei R. and Roots B.I. (1980) Changes in the content and acyl group composition of glycerophospholipids of brain endothelial cells of the developing rat. Neurochem. Res. 5, 683-695.

McCarthy K.D. and deVellis J. (1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. J. C ell B io l. 85, 890-902.

McDonald L.J. and Mamrack M.D. (1988) Aluminum affects phosphoinositide hydrolysis by phosphoinositidase C. Biochem. Biophys. Res. Commun. 155, 203-208.

Mclntire L.V., Frangos J.A., Rhee B.G., Eskin S.G. and Hall E.R. (1987) The effect of fluid mechanical stress on cellular arachidonic acid metabolism. Ann. N. Y. Acad. S ci. 516, 513-524.

McMorris F.A., Miller S.L., Pleasure D. and Abramsky 0. (1981) Expression of biochemical properties of oligodendrocytes in oligodendrocyte x glioma cell hybrids proliferating in vitro. Exp. C ell Res. 133, 395-404.

Means E.D., Anderson D.K., Demediuk P., Saunders R.D. and Horrocks L.A. (1986) Inflammation in experimental spinal cord injury. Exp. Brain Res. 13, 74-87.

Means E.D., Anderson D.K., Waters T.R. and Kalaf L. (1981) Effect of methyl prednisolone in compression trauma to the feline spinal cord. J. Neurosurg. 55, 200-208.

Meier P., Sachse J.H., Brophy P .J., Marsh D. and Kothe G. (1987) Integral membrane proteins significantly decrease the molecular motion in lipid bilayers: A deuteron NMR relaxation study of membranes containing myelin proteolipid apoprotein. Proc. N a tl. Acad. S ci. USA 84, 3704-3708.

Mersel M., Lelong I ., Hindelang C., Sariieve L.L. and Vincendon G. (1987) Isolation of plasma membranes from neurons grown in primary culture. Anal. Biochem. 166, 246-252. 284

Michel1 R.H., Allan D. and Finean J.B. (1975) Significance of minor glycerolipids in membrane structure and function, in Advances in Experimental Medicine and Biology, Vol. 72 (Porcellati G., Amaducci L. and Gal1i C., eds.), PP* 3-13. Plenum Press, New York.

Moncada S. and Vane J.R. (1979) Arachidonic acid metabolites and the interactions between platelets and blood-vessel walls. N. Engl. J. Med. 300, 1142-1147.

Morris C.E. and Sigurdson W.J. (1989) Stretch-inactivated ion channels coexist with stretch-activated ion channels. Science 243, 807-809.

Murphy E.J., Anderson D.K., Means E.D. and Horrocks L.A. (1987a) Pressure induced trauma in neuronal cell cultures: A novel model for spinal cord injury. J. Neurochem. 48, 550-A.

Murphy E.J., Stephens R., Jurkowitz M.S. and Horrocks L.A. (1987b) Plasmalogen quantitation in cell culture lipids by acid hydrolysis followed by high performance liquid chromatography. J. Neurochem. 48, 560-B.

Murphy E .J., Anderson D.K., Means E.D. and Horrocks L.A. (1988) Pressure induced trauma in neuronal cell cultures. Trans. Am. Soc. Neurochem. 19, 175.

Murphy E.J., Anderson D.K., Means E.D. and Horrocks L.A. (1989) Pressure induced trauma in ROC-1 oligodendroglial cell cultures. Trans. Am. Soc. Neurochem. 20, 141.

Murphy E.J. and Horrocks L.A. (1989) Mechanisms of action of CDPcholine and CDPethanolamine on fatty acid release during ischemia of brain. In Lipid Mediators in Ischemic Brain Damage and Experimental Epilepsy, Bazan, N.G. (Ed.), Karger Press, Basel, Switzerland (in press).

Mushayakarara E.C., Wong P.T.T. and Mantsch H.H. (1986a) Detection by high pressure infrared spectrometry of hydrogen-bonding between water and triacetyl glycerol. Biochem. Biophys. Res. Commun. 134, 140-145.

Mushayakarara E.C., Wong P.T.T. and Mantsch H.H. (1986b) The effect of pressure on the hydrogen bonding between carbonyl and hydroxyl moieties in 1,2-dipalmitoylglycerol: A Fourier-transform infrared spectroscopy study. Biochimica et Biophysica Acta 857, 259-264.

Nash T., Allison A.C. and Harington J.S. (1966) Physio-chemical properties of silica in relation to its toxicity. Nature 210, 259-261. 285

Nawroth P.O., Stern D.M. and Dietrich M. (1985) Thromboxane production by perturbed bovine aortic endothelial cells in culture. B lut 50, 7-11.

Newman A.C.D., and Brown G. (1987) The chemical composition of clays. Longman Scientific and Technical, Essex, England.

Nishizuka Y. (1984) Turnover of inositol phospholipids and signal transduction. Science 225, 1365-1370.

Oscarson D.W., Van Scoyoc G.E. and Ahlrichs J.L. (1986) Lysis of erythrocytes by silicate minerals. Clays Clay Miner. 34, 74-80.

Pelech S.L. and Vance D.E. (1989) Signal transduction via phosphatidylcholine cycles. Trends Biochem. S ci. 14, 28-30.

Perl D.P. and Good P.F. (1987) Uptake of aluminum into central nervous system along nasal-olfactory pathways. Lancet 1, 1028.

Phelps B.M., Primakoff P., Koppel D.E., Low M.G. and Myles D.G. (1988) Restricted lateral diffusion of PH-20, a Pi-anchored sperm membrane protein. Science 240, 1780-1782.

Piper H.M. (1988) Metabolic processes leading to myocardial cell death. Methods Achiev. Exp. Pathol. 13, 144-180.

Pleasure D., Abramsky 0., Silberberg D., Quinn B., Parris J. and Saida T. (1977) Lipid synthesis by an oligodendroglial fraction in suspension culture. Brain Res. 134, 377-382.

Polak P.E. and Szuchet S. (1988) Plasma membrane of cultured oligodendrocytes: I. Isolation, purification, and initial characterization. G lia 1, 39-53.

Raber J. and Bast A. (1989) Changes in receptor response by the effect of disease on membrane fluidity. Med. Hypoth. 28, 169-171.

Ransom B.R., Neale E., Henkart M., Bullock P.N. and Nelson P.G. (1977) Mouse spinal cord in cell culture: I. Morphology and intrinsic neuronal electrophysiologic properties. J. Neurophysiol. 40, 1132-1150.

Ranson B.R., Neale E., Herkart M., Bullock P.N. and Nelson P.G. (1977) Mouse spinal cord in cell culture: I. Morphology and intrinsic neuronal electrophysiological properties. J. Neurophysiol. 40, 1132-1150.

Resink T .J., Grigorian G.Y., Moldabaeva A.K., Danilov S.M. and Buhler F.R. (1987) Histamine-induced phosphoinositide metabolism in cultured human umbilical vein endothelial cells. Association with thromboxane and prostacyclin release. Biochem. Biophys. Res. Commun. 144, 438-446. 286

Rhoads D.E., Kaplan M.A., Peterson N.A. and Raghupathy E. (1982) Effects of free fatty acids on synaptosomal amino acid uptake systems. Journal of Neurochemistry 38, 1255-1260.

Rhoads D.E., Osburn L.D., Peterson N.A. and Raghupathy E. (1983) Release of neurotransmitter amino acids from synaptosomes: Enhancement of calcium-independent efflux by oleic and arachidonic acid. Journal of Neurochemistry 41, 531-537.

Richelson E. (1979) The use of cultured cells in neurobiological studies, in International Review of Biochemistry, Physiological and Pharmacological Biochemistry (Tipton K.F., ed.), pp. 81-120. University Park Press, Baltimore, MD.

Robert J ., Montaudon D. and Hugues P. (1983) Incorporation and metabolism of exogenous fatty acids by cultured normal and tumoral glial cells. Biochimica et Biophysica Acta 752, 383-395.

Roberts E. (1986) Alzheimer's disease may begin in the nose and may be caused by aluminosilicates. Neurobiol. Aging 7, 561-567.

Robertson C.S., Foltz R., Grossman R.G. and Goodman J.C. (1986) Protection against experimental ischemic spinal cord injury. J. Neurosurg. 64, 633-642.

Rostogi B.K. and Nord0y A. (1980) Lipid composition of cultured human endothelial cells. Thromb. Res. 18, 629-641.

Rouser G., Siakotos A. and Fleischer S. (1969) Quantitative analysis of phospholipids by thin layer chromatography and phosphorus analysis of spots. Lipids 1, 85-86.

Rudge J.S ., Smith G.M. and Silver J. (1989) An in vitro model of wound healing in the CNS: Analysis of cell reaction and interaction at different ages. Exp. Neurol. 103, 1-16.

Ryan U.S. and Ryan J.W. (1985) Relevance of endothelial surface structure to the activity of vasoactive substances. Chest 88, 2035-2075.

Sanderman H. (1978) Regulation of membrane enzymes by lipids. Biochimica et Biophysica Acta 515, 209-237.

Saunders R. and Horrocks L.A. (1987) Eicosanoids, plasma membranes, and molecular mechanisms of spinal cord injury. Neurochem. Pathol. 7, 1-22.

Schoen P.E., Priest R.G., Sheridan J.P. and Schnur J.M. (1977) Pressure-induced changes in molecular conformation in liquid alkanes. Nature 270, 412-414. 287

Schror K. (1985) Prostaglandins, other eicosanoids and endothelial c ells. Basic Research in Cardiology 80, 502-514.

Seigle N. and Haug A. (1983) Aluminum interactions with calmodulin: Evidence for altered structure and function from optical and enzymatic studies. Biochimica et Biophysica Acta 744, 36-45.

Sekar M.C. and Hokin L.E. (1986) The role of phosphoinositides in signal transduction. J. Hembr. B io l. 89, 193-210.

Selivonchick D.P. and Roots B.I. (1976) Lipid and fatty acyl composition of rat brain capillary endothelia isolated by a new technique. Lipids 12, 165-169.

Selkoe D.J. (1987) Deciphering Alzheimer's disease: The pace quickens. Trends Neurosci. 10, 181-184.

Shier U.T. and DuBourdieu D.J. (1982) Role of phospholipid hydrolysis in the mechanism of toxic cell death by calcium and ionophore A23187. Biochem. Biophys. Res. Commun. 109, 106-112.

Siakotos A.N., Rouser G. and Fleischer S. (1969) Isolation of highly purified human and bovine brain endothelial cells and nuclei and their phospholipid composition. Lipids 4, 234-239.

Siesjo B.K. (1988) Historical overview: Calcium, ischemia, and death of brain cells. Part VIII. Neurological Disease. Ann. N. Y. Acad. S ci. 522, 638-661.

Singh S.V., Viswanathan P.N. and Rahman Q. (1983) Interaction between erythrocyte plasma membrane and silic a dusts. Environmental Health Perspectives 51, 55-60.

Smith R.L. and Oldfield E. (1984) Dynamic structure of membranes by deuterium NMR. Science 225, 280-288.

Snider R.M., McKinney M., Forray C. and Richelson E. (1984) Neurotransmitter receptors mediate cyclic GMP formation by involvement of arachidonic acid and lipoxygenase. Proc. N a tl. Acad. S ci. USA 81, 3905-3909.

Snider R.M. and Richelson E. (1983) Thrombin stimulation of guanosine 3 ', 5 '-monophosphate formation in murine neuroblastoma cells (clone N1E-115). Science 221, 566-568.

Snider R.M. and Richelson E. (1984) Bradykinin receptor-mediated cyclic GMP formation in a nerve ceil population (murine neuroblastoma clone N1E-115). Journal of Neurochemistry 43, 1749-1754.

Spector A.A. and Yorek M.A. (1985) Membrane lipid composition and cellular function. J. Lip id Res. 26, 1015-1035. 288

Stamatoff J ., Guillon D., Powers L. and Cladis P. (1978) X-ray diffraction measurements of dipalmitoylphosphatidyl choline as a function of pressure. Biochem. Biophys. Res. Commun. 85, 724-728.

Starke P.E., Hoek J.B. and Farber J.L. (1986) Calcium-dependent and calcium-independent mechanisms of irreversible cell injury in cultured hepatocytes. J. B io l. Chem. 261, 3006-3012.

Stubbs C.D. (1983) Membrane fluidity: Structure and dynamics of membrane lipids. Essays Biochem. 19, 1-39.

Stubbs C.D., Kouyama T., Kinosita Jr. K. and Ikegami A. (1981) Effect of double bonds on the dynamic properties of the hydrocarbon region of lecithin bilayers. Biochemistry 20, 4257-4262.

Stubbs C.D. and Smith A.D. (1984) The modification of mammalian membrane polyunsaturated fatty acid composition in relation to membrane fluidity and function. Biochimica et Biophysica Acta 779, 89-137.

Svendson W., Dalen H., Mol and J. and Engodal H. (1985) Crater formation and endocytosis in human endothelial cells in vitro. Arteriosclerosis 58, 27-37.

Swezey R.R. and Somero G.N. (1982) Polymerization thermodynamics and structural stabilities of skeletal muscle actins from vertebrates adapted to different temperatures and hydrostatic pressures. Biochemistry 21, 4496-4503.

Talamo B.R., Rudel R.A., Kosik K.S., Lee V.M.Y., Neff S., Adelman L. and Kauer J.S. (1989) Pathological changes in olfactory neurons in patients with Alzheimer's disease. Nature 337, 736-739.

Thomas J.M.F., Hullin F., Chap H. and Douste-Blazy L. (1984) Phosphatidylcholine is the major phospholipid providing arachidonic acid for prostacyclin synthesis in thrombin-stimulated human endothelial cells. Thromb. Res. 34, 117-123.

Thornton S.C., Mueller S.N. and Levine E.M. (1983) Human endothelial cells: Use of heparin in cloning and long-term serial cultivation. Science 222, 623-624.

Trump B.F. and Berezesky I.K. (1983) The role of calcium deregulation in cell injury and cell death (introduction). Survey and Synthesis of Pathology Research 2, 165-169.

Wallace M.C. and Tator C.H. (1986) Failure of naloxone to improve spinal cord blood flow and cardiac output after spinal cord injury. Neurosurgery 18, 428-432. 289

Weksler B.B., Marcus A.J. and Jaffe E.A. (1977) Synthesis of prostaglandin I 2 (prostacyclin) by cultured human and bovine endothelial cells. Proc. N a tl. Acad. S ci. USA 74, 3922-3926.

Weltzien H.U. (1979) Cytolytic and membrane-perturbing properties of lysophosphatidylcholine. Biochimica et Biophysica Acta 559, 259-287.

Weltzien H.U., Richter G. and Ferber E. (1979) Detergent properties of water-soluble choline phosphatides. J. B io l. Chem. 254, 3652-3657.

Wey H.E., Jakubowski J.A. and Deykin D. (1986) Incorporation and redistribution of arachidonic acid in diacyl and ether phospholipids of bovine aortic cells. Biochimica et Biophysica Acta 878, 380-386.

Whatley R.E., Nelson P., Zimmerman G.A., Stevens D.L., Parker C.J., McIntyre T.M. and Prescott S.M. (1989) The regulation of platelet-activating factor production in endothelial cells: The role of protein kinase C. J. B io l. Chem. 264, 6325-6333.

Wilcock G.K. (1988) Alzheimer's disease - current issues. Q. J. Med. 66, 117-124.

Windholz M., Budavari S., Blumetti R.F., and Otterbein E.S. (1983) The Merck Index: An encyclopedia of chemicals, drugs, and biological s. Merck & Co., Inc., Rahway, NJ.

Witter B. and Debuch H. (1982) On the phospholipid metabolism of glial cell primary cultures: Cell characterization and their utilization of 1-alkyl-glycerophosphoethanolamine. Journal o f Neurochemistry 38, 1029-1037.

Womack F.C. and Colowick S.P. (1979) Proton-dependent inhibition of yeast and brain hexokinases by aluminum in ATP preparations. Proc. N a tl. Acad. S ci. USA 76, 5080-5084.

Wong P.T.T., Capes S.E. and Mantsch H.H. (1989) Hydrogen bonding between anhydrous cholesterol and phosphatidylcholines: An infrared spectroscopic study. Biochimica et Biophysica Acta 980, 37-41.

Wong P.T.T. and Mantsch H.H. (1988) High-pressure infrared spectroscopic evidence of water binding sites in 1,2-diacyl phospholipids. Chemistry and Physics of Lipids 46, 213-224.

Woodworth C.D., Mossman B.T. and Craighead J.E. (1982) Comparative effects of fibrous and nonfibrous minerals on cells and liposomes. Environ. Res. 27, 190-205. 290

Wright P.A., Thomas J.M., Cheetham A.K. and Nowak A.K. (1985) Localizing active sites in zeolitic catalysts: Neutron powder profile analysis and computer simulation of deuteropyridine bound to gallozeolite-L. Nature 318, 611-614.

Yamaja Setty B.N., Stuart M.J. and Walenga R.W. (1985) Formation of 11-hydroxyeicosatetraenoic acid and 15-hydroxyeicosatetraenoic acid in human umbilical arteries is catalyzed by cyclooxygenase. Biochimica et Biophysica Acta 833, 484-494.

Yang X.C. and Sachs F. (1989) Block of stretch-activated ion channels in xenopus oocytes by gadolinium and calcium ions. Science 243, 1068-1071.

Yorio T., Torres S. and Tarapoom N. (1983) Alteration in membrane permeability by diacylglycerol and phosphatidylcholine containing arachidonic acid. Lipids 18, 96-99.

Yoshida S., Ikdeda M., Busto R., Santiso M., Martinez E. and Ginsberg M.D. (1986) Cerebral phosphoinositide, triacylglycerol, and energy metabolism in reversible ischemia: Origin and fate of free fatty acids. J. Neurochem. 47, 744-757.

Yoshida S., Inoh S., Asano T., Sano K., Kubota M., Shimazaki H. and Ueta N. (1980) Effect of transient ischemia on free fatty acids and phospholipids in the gerbil brain. J. Neurosurg. 53, 323-331.

Yu A.C.H., Gregory G.A. and Chan P.H. (1989) Hypoxia-induced dysfunctions and injury of astrocytes in primary cell cultures. J. Cereb. Blood Flow Hetab. 9, 20-28.

INFORMATION to USERS the Most Advanced Technology Has Been Used to Photo­ Graph and Reproduce This Manuscript from the Microfilm Master

INFORMATION to USERS the Most Advanced Technology Has Been Used to Photo Graph and Reproduce This Manuscript from the Microfilm Master