Distribution and detection of Clostridium perfringens type A enterotoxin after intraperitoneal and intragastic administr...

MISSING IMAGE

Material Information

Title:
Distribution and detection of Clostridium perfringens type A enterotoxin after intraperitoneal and intragastic administration using the murine model
Physical Description:
Book
Language:
English
Creator:
Keller, Andreas Markus
Publication Date:

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
aleph - 27624669
oclc - 37236147
System ID:
AA00026794:00001


This item is only available as the following downloads:


Full Text









DISTRIBUTION AND DETECTION OF CLOSTRIDIUM PERFRINGENS
TYPE A ENTEROTOXIN AFTER INTRAPERITONEAL AND
INTRAGASTRIC ADMINISTRATION USING THE MURINE MODEL













By


ANDREAS MARKUS KELLER












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA

1997























This dissertation is dedicated to my parents, Karl-Heinz and Charlotte; to my wife
Cecilia; to my daughter Dominik; to my Onkel Peter; to my Omi Lotti; and with
remembrance to my late sister Sylvia; and Opi Paul and Omi Elisabeth, for their
never ending love, patience, enthusiasm, dedication, understanding and support.
Thank you, God, for such a wonderful family.













ACKNOWLEDGMENTS

I would like to thank Dr. James A. Lindsay, my major advisor (I prefer the

German word "Dr. Vater," since advisor is a too simplistic a term to describe a

mentor, role model and benefactor), from the bottom of my heart, for his never-

ending support, patience, persistence, guidance, motivation, commitment,

encouragement, enthusiasm, and enlightenment in every step of my doctoral research,

graduate program, and personal life. I am extremely grateful to the other members

of my doctoral committee, Dr. Douglas L. Archer, Dr. Sean F. O'Keefe, Dr. Mark

L. Tamplin, and Dr. Ramon D. Littell, for their interest, advice, suggestions, review

of manuscript, and supportive role in my research. I am deeply appreciative of the

Food Science and Human Nutrition Department, USDA, and NIH for providing me

with graduate assistantships and all other funding. I would like to thank F. Morgan

Wallace for his advice, support, and friendship. I also give my sincerest thanks to

Annette S. Mach for conducting the tissue culture studies, and above all her

suggestions, assistance and guidance in the lab, making it a fun, efficient, sound, and

safe environment to work in. Finally, my sincerest thanks go to my friend Dr.

Antonio A. Figueiredo.





iii













TABLE OF CONTENTS


ACKNOWLEDGMENTS .................................... iii

LIST OF TABLES ............... ........................ vi

LIST OF FIGURES .................................... vii

ABSTRACT ............................ ............. viii

CHAPTERS

1 INTRODUCTION ...... .......... .............. 1

2 REVIEW OF LITERATURE ....................... 5
Clostridium perfringens ....................... 5
Classification ............................. 5
C. perfringens Foodborne Illness ................. 9
H istory ............................. 9
Type A Foodborne Illness:
General Characteristics ................. 11
Type A FBI Outbreaks .................. 12
Identification of C. perfringens FBI Outbreaks . 13
C. perfringens and Pathogenicity ................ 15
Virulence Factors Contributing
to C. perfringens FBI ..... ........... 15
Molecular Biology of CPE ................ 16
Sporulation and CPE ................... 17
Biochemistry of CPE ................... 19
Structure-Function and Vaccines ............ 21
Silent CPE Genes ..................... 23
CPE Intragastric Mechanisms of Action ....... 24
Characteristics of the CPE Binding, Complex
Formation and Insertion ................ 25
CPE and Medicine .................... 28
Activation of CPE ..................... 29

iv









Enhancement of CPE Activity and Human
Non-Foodborne Disease ................ 30
Role of CPE in Non-Foodborne Diseases (SIDS) . 32


3 DETECTION OF CLOSTRIDIUM PERFRINGENS TYPE A
ENTEROTOXIN AFTER IN VITRO BINDING TO MURINE
TISSUES .... .. ............................. 34
Introduction ............................ 34
Materials and Methods ...................... 35
Results and Discussion ...................... 41


4 DETECTION AND DISTRIBUTION OF CLOSTRIDIUM
PERFRINGENS TYPE A ENTEROTOXIN AFTER IN VIVO
INTRAPERITONEAL ADMINISTRATION INTO SWISS
W EBSTER MICE ............................... 51
Introduction ............................. 51
Materials and Methods ..................... 52
Results and Discussion ...................... 56

5 DETECTION AND DISTRIBUTION OF CLOSTRIDIUM
PERFRINGENS TYPE A ENTEROTOXIN AFTER IN VIVO
INTRAGASTRIC ADMINISTRATION INTO SWISS
W EBSTER MICE .................... .......... 68
Introduction ............................. 68
Materials and Methods ...................... 71
Results and Discussion ...................... 76

6 SUMMARY AND CONCLUSIONS .................. 91

REFERENCES ............................... 98

BIOGRAPHICAL SKETCH ..................... .111









v













LIST OF TABLES

Table page

3.1 Detection of unbound CPE by ELISA in murine organ
tissues after in vitro interaction .......................... 44

4.1. ELISA detection of unbound CPE in murine organ tissues
after in vivo IP administration ................ ........... 58

4.2 Detection of unbound CPE by ELISA in murine organ tissues
after in vivo IP administration: time study ................ 61

4.3 ELISA and Western immunoblot detection of CPE (free toxin)
CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine
organ tissues after in vivo IP administration ................. 63

5.1 Detection of unbound CPE by ELISA in murine organ tissues after
in vivo IG administration .............................. 78

5.2 ELISA and Western immunoblot of CPE (free toxin) CPE:R1
(small complex) and CPE:R1:R2 (large complex) in murine
organ tissues after IG administration ....................... 79

5.3 ELISA detection of unbound CPE in murine organ tissues after
in vivo IG administration: time study ...................... 82

5.4 ELISA and Western immunoblot of CPE (free toxin) CPE:R1
(small complex) and CPE:R1:R2 (large complex) in murine
organ tissues after IG administration ................... 83

5.5 ELISA detection of unbound CPE in murine organ tissues after
in vivo IG administration: time study ................. . 85

5.6 ELISA, Western immunoblot and Vero cell detection of: CPE (free toxin)
CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine
organ tissues after in vivo IG administration .............. . 88

vi













LIST OF FIGURES

Figure page

3.1 Western immunoblot of lung tissue ....................... 48

5.1 SDS-PAGE of tissue supernatants ........................ .86

5.2 Western immunoblot of tissue supernatants ................... 87

5.3 Vero cell assay of murine tissues ......................... 89

6.1 Murine model for CPE distribution after IG administration ......... 93
(a) Nonlethal murine model ...................... 93
(b) Sublethal murine model ...................... 93
(c) Sudden death murine model ................... 95
(d) Nonabrupt death murine model ................. 95























vii













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


DISTRIBUTION AND DETECTION OF CLOSTRIDIUM PERFRINGENS
TYPE A ENTEROTOXIN AFTER INTRAPERITONEAL AND
INTRAGASTRIC ADMINISTRATION USING THE MURINE MODEL

By

ANDREAS MARKUS KELLER

May, 1997


Chairman: Dr. James A. Lindsay
Major Department: Food Science and Human Nutrition


Clostridium perfringens has been described as the most important anaerobic

pathogen of man, and is considered the most common cause of enteric diseases in

animals. Virulency of the bacterium is related to the production of at least 15

different protein-toxins, many of which are lethal. Diseases associated with

C. perfringens infections and production of these protein-toxins include myonecrosis

(gas gangrene), necrotic enteritis, antibiotic associated diarrhea, sudden infant death

syndrome and food poisoning in man and animals.




viii







To understand these changes, the whole body distribution of CPE after either

intraperitoneal and intragastric administration was determined, using the murine

model. Results showed that CPE appeared to have three different modes of

distribution and activity which were time and concentration dependent. Nonlethal

levels induced enterotoxigenic symptoms, while sublethal levels induced symptoms

described as parasympathomimetic. Administration of lethal levels induced two

patterns of death, first, a sudden death induced within minutes of CPE administration,

and second, a nonabrupt death that required several hours for manifestation. Animals

expressed symptoms of respiratory distress, shock and multiple organ failure, similar

to the action of a superantigen.

From the murine model studies, the following parallels may be suggested for

CPE toxicosis in humans. Nonlethal levels of CPE causes a toxicosis similar to a

self-limiting foodborne illness. Systemic absorption of sublethal levels of CPE induce

a neurotoxicosis, from which healthy individuals would likely recover. However,

death could occur in immunocompromised persons or the elderly. Systemic

absorption of lethal levels of CPE will cause death in both healthy and

immunocompromised individuals. The finding of a "sudden" pattern of death after

lethal ingestion is pivotal, since this mimics the suggested response of some at-risk

infants to CPE toxicosis, and supports the role of CPE as a trigger in some cases of

the sudden infant death syndrome.




ix













CHAPTER 1
INTRODUCTION


Clostridium perfringens foodborne illness (FBI) is associated with

enterotoxin(s) from type A strains and is the third most common cause of bacterial

FBI in the USA following Salmonella spp. and Staphylococcus aureus. Annual costs

of FBI associated with C. perfringens in the USA and Canada are estimated to be

higher than $200 million. Mortality rates from C. perfringens FBI are dependent on

age and immune status, and the debilitated, immunocompromised, young and elderly

are at high risk. In the USA mortality rates may be as high as 4% (Janssen et al.,

1996). Besides causing foodborne illness C. perfringens has been associated with

some unusual disease states, for example wound infections and sudden infant death

syndrome (SIDS) (Lindsay et al., 1993, 1994; CDC, 1994; Lindsay, 1996). Studies

indicate that C. perfringens infections and type A enterotoxin appear to be associated

with 50-80% of the approximately 7,000 SIDS deaths/year in the USA (Lindsay,

1996.

Consequently, there has been a dramatic increase of interest in the bacterium's

pathogenicity, virulence determinants, and the signals controlling expression of these

determinants. Studies show that C. perfringens type A enterotoxin (CPE), which is

produced in the highest amounts during bacterial sporulation, may have the ability to


1







2

modulate the host defense system, by acting as a superantigen and exerting

immunomodulatory effects on various lymphoid cell populations, thus playing an

important role in the overall pathogenesis of the organism (Lindsay, 1996).

Superantigens may activate and stimulate up to 1 in 5 T-cells as compared to a

classical antigen which normally stimulates 1 in 10,000 T-cells. Stimulation of

T-cells may lead to the induction of cytokines, such as interferon-y, tumor necrosis

factor-B, interleukins, and others, usually in a cascade. Cytokines may cause a

decrease in blood pressure, shock, respiratory distress, multiple organ failure and

death (Lindsay, 1996). Most T-cells activated by superantigens are useless in fighting

infections, and even worse they could unleash an autoimmune attack, driving the

immune system into a self destructive frenzy, hurting the individual instead of

protecting. Superantigens also have the ability to trigger the cell death of cells they

excite, thus weakening the body's defense system (Johnson et al., 1992).

Studies on superantigens have predominantly focussed on the effects of

staphylococcal enterotoxins in animals as a model of human toxicosis (Cerami, 1992;

Tracey and Cerami, 1993; Fleischer, 1994). Still numerous important questions such

as: the functional role of bacterial superantigens, and how superantigens with different

structures can interact with major histocompatibility complex (MHC) and T-cell

receptors remain unanswered.

Recent data indicates that CPE has superantigenic properties; however, it has

not been determined whether the toxin's enterotoxigenic, cytotoxigenic and







3

parasympathomimetic properties are linked to superantigenicity (Bowness et al., 1992;

Lindsay, 1996). Currently there are several areas under study relative to CPE: i. the

mechanics of pathogenicity; ii. the structure of the CPE genes and the signals

controlling regulation and expression; iii. the mechanics of CPE action at the cellular

level and interactions with host molecules; iv. and the mechanisms/role of

superantigenicity. Various laboratories are currently working on: i. cloning the

enterotoxin genes to determine structure and function, and mechanisms for regulation;

ii. CPE receptor binding and gross mechanisms of action; iii. and CPE's mechanisms

of superantigenicity.

The overall aim of this dissertation was to investigate the pathophysiological

responses after intraperitoneal and intragastric administration of CPE using the murine

model. This evaluation may explain how CPE becomes distributed during toxicosis,

which organs were specifically affected, and describe the numerous sequela and

pathophysiological changes that may lead to death. Within the overall aim was the

development of a murine model that would explain the mode of CPE distribution after

intragastric administration, thus being able to draw a comparison with the SIDS model

proposed by Lindsay et al. (1994), and Lindsay (1996), and possibly determining the

events involved and leading to infant death.







4

The specific objectives to this study were as follows:

1. To identify the murine organs and tissues that bind CPE after in vivo

administration of CPE.

2. To determine and compare the number of CPE cell receptors in each

organ.

3. After intraperitoneal CPE administration, to investigate the whole-body

distribution of CPE.

4. Describe the animals' symptoms during toxicosis after intraperitoneal

and intragastric CPE administration, physical and pathophysiological

changes, and findings at necropsy.

5. After intragastric CPE administration investigate the whole-body

distribution of CPE .

6. Propose a murine model for the whole-body distribution of CPE after

intragastric administration.



The results will hopefully elucidate the CPE distribution mode, delineating the

affected organs and pathophysiological changes possibly responsible for the illness and

death. Ultimately the murine model will provide a platform for analogy in the

development of human FBI and SIDS.












CHAPTER 2
REVIEW OF THE LITERATURE

Clostridium perfringens

Clostridium perfringens (Clostridium welchii) has been described as the most

important anaerobic pathogen of man (Lindsay, 1996) and is considered the most

common cause of enteric diseases in animals (Hobbs et al., 1953; Bartoszcze et

al.,1990; Songer, 1996). Virulence of the bacterium is related to the production of

at least 15 different protein toxins, many of which are lethal. Diseases associated

with C. perfringens infections and production of these protein toxins include

myonecrosis (gas gangrene), necrotic enteritis, antibiotic associated diarrhea, sudden

infant death syndrome and food poisoning in man (MacLennan, 1962; Smith, 1979;

Fekety et al., 1980; McDonel, 1980; Rood and Cole, 1991; Lindsay, 1996) and lamb

dysentery, ovine, bovine and equine enterotoxemia, and pulpy kidney disease of sheep

and other animals (McDonel, 1980; Niilo, 1980; Sterne, 1981; Songer, 1996).



Classification

C. perfringens strains are initially classified into a series of different types (A-

E) based upon their production of one or more of the major toxins, alpha, beta,

epsilon and iota. Toxin production is verified by neutralization with type specific

antisera using mice; however, the process is very tedious, expensive and relies upon

5







6
the unnecessary use of animals. PCR-typing techniques are becoming more available

and have proven to be reliable under some circumstances. However, considerable

discussion has ensued as to whether using a classification method based on four toxins

is correct since it appears that not all strains contain the alpha-toxin (phospholipase

C) a current defining characteristic of C. perfringens (Lindsay, 1996).

C. perfringens is the most widely distributed pathogenic bacterium. The

organism is a Gram-positive, rod-shaped, variably sized (0.6-2.4 x 1.3-1.9 /m)

encapsulated, nonmotile, spore former, occurring singly or in pairs. Vegetative cells

are mostly square-ended rods but some strains have rounded ends. The bacterium

usually grows very quickly and can have a generation time of 7 minutes in an optimal

meat-containing environment (Labbe, 1989). Although C. perfringens can be isolated

relatively easily, colony appearance on solid medium varies with organism type.

Isolation requires the differential use of various antibiotics, the presence or absence

of iron and sulfite, and incubation temperature. C. perfringens are resistant to many

antibiotics which inhibit other anaerobes or facultative anaerobes. Sulfites are

reduced to sulfides which in turn react with iron, forming a precipitate that renders

C. perfringens colonies black. Selective media commonly used for isolation and

enumeration are sulfite polymyxin sulfadiazine (SPS) agar; tryptone sulfite neomycin

(TSN) agar; Shahidi Ferguson perfringens (SFP) agar; D-cycloserine blood agar;

oleandomycin polymyxin sulfadiazine perfringens (OPSP) agar; tryptose sulfite

cycloserine (TSC) agar; and egg yolk free tryptose sulfite cycloserine (EY-free TSC)







7

agar. Although colony growth is good at 37C, incubation at 46C especially on TSN

is highly selective. If the organism is present in foods or feces in the spore state,

samples are usually heat shocked at 75C for 15 minutes before plating. Lindsay

(1996) suggested that since spore heat resistance is correlated with CPE synthesis,

that spores should be heat shocked at both low (750C) and high (100C) temperatures

to ensure complete activation of all sub-populations within a sample.

C. perfringens is not a strict anaerobe since growth occurs between +125 and

-125 millivolts (Eh). Thus the organism is described as aerotolerant. Vegetative cells

are sensitive to high Eh during lag- and early log-phase, but oxygen extends the lag

phase and growth can be stimulated by lowering the Eh. Oxygen peroxides, however,

reduce colony growth, and cell/spore counts. The optimum growth temperature

varies from 43-47C based on organism type, and the To is usually 200C and the T.,

50C. Thus the vegetative cells are heat tolerant. Some strains are known to grow

slowly at 150C but these are the exception. Refrigeration at < 5C and freezing can

decrease the number of vegetative cells and spores. Vegetative cells are very

sensitive to acid environments during log phase, but during stationary phase cells are

resistant. The optimum pH range for growth and toxin production differ. Optimum

growth occurs at pH 6.0-7.5, while pH under 5 or above 8.3 are extremely

inhibitory. The optimum pH for toxin production is 7.0 for alpha toxin, 7.5 for beta,

and 7.2 for epsilon and theta (Hobbs, 1979; Labbe, 1989). Optimal toxin production

occurs between 30-46C. The water activity (a) range for growth is 0.93-0.97







8

depending on the solute controlling the a, of the substrate. During sporulation, the

a, is a more significant growth limiting factor than for vegetative growth (Hobbs,

1979; Labbe, 1989). Many strains tolerate curing agents and smoking when a

suitable growth temperature and pH are maintained. Complete growth inhibition

occurs at 8% NaCI, Ig/kg NaNo3, 400mg/kg NaNO2, and with a combination of

5.3% NaCI with 25mg/kg nitrite (Hobbs, 1979; Labbe, 1989).

In nature, C. perfringens is usually found as a spore which becomes

metabolically active only when it encounters a suitable substrate. Thus the bacterium

is regarded as a r-strategist (Lindsay, 1996). Spores produced in sporulation media

are subterminal and oval in shape. Sporulation is strain, temperature and medium

dependent. A range of 32-40C is appropriate for most strains (Lindsay, 1996) and

maximum spore production is reached in 6-8 hrs. Many C. perfringens strains have

different nutrient requirements, thus the choice of sporulation medium is critical.

Duncan and Strong (DS) medium (Duncan and Strong, 1968) with some minor

adjustments is used by most laboratories. The addition of either starch, raffinose,

amylopectin, amylose, glucose, maltose and methylxanthines to DS is known to

increase spore production, however, complete sporulation is never observed (Labbe,

1989; Lindsay, 1996). Five to ten percent sporulation is considered usual, and 50%

exceptional. Some strains are known to have almost no spore formation (< 0.001 %)

even under the optimal conditions. C. perfringens spores are relatively heat resistant

D1oo 15 min, however, strain variation is known. Non-hemolytic strains have a







9
decimal reduction time (DioooC) of 6-17 min, whereas hemolytic strains have a D,0oC

of 0.1-0.5 min. Heat activation is also strain dependent. Some strains only require

60C for 5 min, although 10-20 min at 75-800C is usual (Labbe, 1989). Some type

A strains are highly heterogeneous with a sub-population activated at 750C, and

another at 100C (Lindsay, 1996). Spore radiation resistance varies from 1.2 to 3.4

kGy, and radiation resistance parallels heat resistance (Labbe, 1989). C. perfringens

spores are also highly resistant to curing agent at concentrations of 21.5 % NaCI, 1.8

g/1 NaNO3 and 1.2 g/1 NaO, (Labbe, 1989). Phenolic antioxidants are inhibitory or

bacteriocidal depending on the compound and concentration used (Labbe, 1989).



C. perfringens Foodborne Illness

History

Although an outbreak of C. perfringens foodborne illness (FBI) was first

described by Klein in 1895, it was not until 1943 when Knox and McDonald in the

United Kingdom and McClung (1945) in the United States made the association

between the organism and the FBI. Persons afflicted with the illness expressed

symptoms of severe abdominal pain, mild chronic to explosive diarrhea accompanied

by nausea. From 1947-9 there were many reported outbreaks of FBI in Germany

with a large number of fatalities. Unlike the illnesses described in the UK and USA

a few years earlier, the German patients suffered from severe gross hemorrhagic

enteritis (enteritis necroticans: Darmbrand) where the bowel was completely







10

desquamated. Thus it was obvious that C. perfringens was responsible for two

completely different types of FBI. The dilemma was finally resolved by Hobbs et al.

in 1953 who unequivocally showed that type A strains were responsible for the mild

form of FBI, and type C strains for the more dramatic-lethal form. Subsequent

studies by Duncan and Strong (1969), Hobbs (1979), Niilo (1975), and Tsai and

Riemann (1975a; 1975b) showed that the type A FBI was caused by an enterotoxin,

now termed CPE, produced during cell sporulation. In 1967 a milder form of

necrotic enteritis was observed in New Guinea. This type of outbreak was

subsequently found to be common, and coincided with traditional pig feasting.

A heat sensitive strain of C. perfringens type C was found to be the causative agent

of the disease, now referred to as pigbel (Murrell and Walker, 1991). Studies during

the last 20 years have shown that Darmbrand and pigbel result from the consumption

of type C vegetative cells which proliferate in the gut producing beta-toxin. The diet

of persons afflicted with the diseases is usually lacking in proteases needed for

enzymatic digestion of the beta-toxin, or as in New Guinea, protease inhibitors are

present in the bowel due to the consumption of sweet potatoes, the staple diet of the

natives. Darmbrand and pigbel are now relatively rare due to immunization of the

susceptible populations with beta-toxoid. However, in some inhospitable regions of

New Guinea where cannibalism still occurs the disease is common and death usually

occurs after infection (Murrell, 1989).







11

Type A Foodborne Illness: General Characteristics

While type C necrotic enteritis is rare, type A FBI is very common and these

strains are associated with about 10% (third most common in the USA) of all bacterial

FBI (Bean and Griffin, 1990). An infection may be caused by the ingestion of foods

contaminated with > 106-107 vegetative cells/gram (Hobbs, 1979). This

contamination usually results from temperature abuse of a prepared food. Vegetative

cells that survive the stomach acids, enter the small intestine, multiply and sporulate.

During sporulation some cells synthesize CPE, and upon cell lysis the enterotoxin is

released in the intestinal lumen where it attaches to villous enterocytes to act

cytotoxically and histopathologically, or is also absorbed systemically where it may

additionally be parasympathomimetic, cardiotrophic or superantigenic (Lindsay,

1996). Illness usually develops 8-24 hours after ingestion, and is resolved 12-24

hours after onset (McDonel, 1980; Labbe, 1989). The classical symptoms are severe

abdominal cramps with mild to explosive diarrhea, accompanied by nausea. Vomiting

is rare, although fever may occur. Death rates are usually low in immunocompetent

individuals, however, in the young, elderly or immunocompromised persons the

mortality rate is 3-4% (Janssen et al., 1996).









Type A FBI Outbreaks

From 1973-87 the Center for Disease Control (CDC) reported 1994 outbreaks

of C. perfringens type A FBI in the USA. These outbreaks involved 12,234 cases

and 12 deaths. The actual number of FBI cases/year is estimated to be around

650,000 with 7-10 fatalities/year (Bean and Griffin, 1990). C. perfringens outbreaks

are often large, with a mean of 25 cases/outbreak (Todd, 1989; Bean and Griffin,

1990; CDC, 1994). There have been several serious outbreaks during the last two

years traced to St. Patrick's Day meals, the first, in Cleveland, Ohio, involving 156

persons. Corned beef boiled for three hours was cooled slowly at room temperature

before refrigeration. Four days later the meat was reheated at 48.8"C and consumed

in sandwiches several hours later after preparation. The second occurred in Virginia

where 86 people became ill. Commercially prepared, frozen-brined corned beef was

cooked in large pieces, refrigerated and reheated with a heat lamp for 90 minutes

before consumption (CDC, 1994). The third outbreak in a British hospital, involved

17 patients. Cooked vacuum-packed pork was cooled slowly at a commercial meat

preparation facility, transported to the hospital, slowly reheated then consumed.

Temperature abuse during cooking (inadequate cooking) and improper cooling

or holding temperatures account for 97% of type A FBI outbreaks. Since the

bacterium is heat tolerant and its spores have an high resistance, incomplete cooking

of contaminated foods may not kill all vegetative cells and will likely promote

germination of spores and rapid outgrowth. Other factors often associated with type







13

A FBI include contaminated equipment, and improper personal hygiene. Prevention

and control of type A FBI can be accomplished by thoroughly cooking foods and

ensuring that high internal temperatures necessary to destroy the bacterium's spores

are achieved. Cooked foods should be cooled quickly and either stored at

refrigeration temperatures or consumed immediately.



Identification of C. perfringens FBI Outbreaks

To successfully identify C. perfringens type A FBI outbreaks several

bacteriological criteria have to be fulfilled. Public health agencies may identify an

outbreak by 1) finding the contaminated food and determining that it contains

> 105 C. perfringens vegetative cells/gram, 2) finding that patients have > 103

C. perfringens spores/gram feces, 3) finding that patients express the same serotype

of C. perfringens or, the same serotype should be found in both contaminated food

and in patient feces (Stringer, 1985; McClane, 1992). A problem with these criteria

is that many elderly individuals naturally have high C. perfringens spore counts in

their stool. Further, many isolates from the U.S. and Japan are unusual and cannot

be serotyped (Saito, 1990; McClane, 1992). Thus, it is now recommended that the

diagnostic criteria for C. perfringens type A food poisoning outbreaks should include

1) the presence of CPE in feces FBI patients, and 2) isolation of C. perfringens type

A strains carrying the cpe gene from feces or foods associated with the FBI.

Detection of CPE in feces is a partial indicator of C. perfringens type A FBI and







14

several serologic assays (ELISA/reverse-passive latex agglutination:RPLA) are

commercially available. A problem with these assays is that fecal samples have to

be collected and examined immediately (Bartholomew et al., 1985; Stringer, 1985;

Birkhead et al., 1988; Labbe, 1989; McClane, 1992). There is also the problem that

a contaminating molecule produced by all C. perfringens strains (CPE positive and

negative strains) cross reacts with antisera to CPE. Thus CPE detection should be

corroborated with cpe gene probe assays, which are relatively easy to perform

(Labbe, 1989; Kokai-Kun and McClane, 1996; Lindsay, 1996). However, there is

an additional concern. In a recent survey of C. perfringens associated FBI cases

where all the microbiological and serological criteria were met, cpe gene probe

analysis revealed only 59% of associated strains were cpe-positive. Lindsay (1996)

recently argued that results using even the latest molecular biology identification

methods for C. perfringens should be treated with caution.

The reasons for this caution are as follows. All enterotoxigenic C. perfringens

type A carry the cpe gene, however, CPE positive isolates account for < 5% of all

C. perfringens isolates found globally. Thus strains capable of causing FBI present

a very small sub-group within the ubiquitous and regularly encountered C. perfringens

(Van Damme-Jongsten et al., 1989; Saito, 1990; Kokai-Kun et al., 1994). CPE-A

positive strains have been isolated from feces collected from 6% of healthy food

handlers, suggesting that humans may serve as reservoirs for C. perfringens type A

strains (Hobbs, 1979; Saito, 1990). The issue of whether animals are indeed potential







15
reservoirs of enterotoxigenic C. perfringens remains questionable. In human FBI

causing strains cpe is chromosomally encoded, whereas in veterinary isolates cpe is

plasmid-bome (Cornillot et al., 1995). Only one plasmid encoded strain has been

isolated from human C. perfringens FBI cases (Katayama et al., 1996).



C. perfringens and Pathogenicity

Virulence Factors Contributing to C. perfringens FBI

The association between cpe and C. perfringens FBI, although not formally

proven by Koch's molecular hypothesis, is accepted by researchers. This acceptance

is based on strong evidence from a number of sources, such as epidemiological

studies where patients have detectable enterotoxin levels in their stools (80-100% of

patients fecal samples tested CPE positive) (Bartholomew et al., 1985; Birkhead et

al., 1988). CPE ingestion is known to cause illness with the same symptoms in

experimental animals (Hobbs et al., 1953; McDonel and Duncan, 1975; Bartholomew

et al., 1985; Birkhead et al., 1988). When either purified CPE or C. perfringens type

A strains were fed to human volunteers they developed FBI symptoms, but patients

fed CPE negative strains showed no illness (Skjelkvale and Uemura, 1977). It has

also been shown that the effects of CPE in animal models can be neutralized with

CPE specific antiserum (Hauschild et al., 1971).







16
Molecular Biology of CPE

The complete cpe gene has been cloned and sequenced allowing studies on the

regulation and expression of the gene with regards to sporulation, and for the

construction of various nucleic acid probes (Van Damme-Jongsten et al., 1989;

Czeczulin et al., 1993; Corillot et al., 1995; Lindsay, 1996). In type-A food

poisoning isolates the cpe gene is present as a single chromosomal copy in the

hypervariable C region of the chromosome. This hypervariable region is thought to

contain mobile genetic elements (transposon or lysogenized phage) that allow the

transfer of cpe to other C. perfringens strains (Canard et al., 1992). The plasmid

encoded veterinary isolates are thought to represent an example of this type of

mobility. The cpe gene can also be transferred and expressed in non-enterotoxigenic

type A, B and C strains. A comparison of open reading frame (ORF) sequence data

strongly suggests that CPE produced from different type A strains, or different

variants of the same strain are similar, but not identical (Czeczulin et al., 1993,

1996). Although microsequence variation occurs, the cpe-ORF generally appears to

be highly conserved. Enterotoxigenic veterinary and food poisoning strains have

slightly different regulatory factors preceding the cpe ORF (Brynestad et al., 1994).

A 45-base pair insertion about 265 nucleotides upstream from the start of the cpe gene

has been detected in three strains, implying at least two types of cpe-promoter regions

in foodborne isolates (Melville et al., 1994). The insertion does not appear to alter

the starting point of cpe transcription in C. perfringens, although transcription was







17

significantly altered when the cpe gene was cloned into B. subtilis. The reason for

this promoter diversity remains undetermined, although geographical segregation

appears not to be a factor. The cpe gene may be associated with at least two

repetitive sequences, one of which is a known insertion sequence (IS1151) (Daube et

al., 1993; Brynestad et al., 1994; Cornillot et al., 1995). In human FBI strains, when

cpe was chromosomally located, several factors were observed: first, the gene was

preceded by a repeated sequence, the HindIEI repeat and ORF3, which is homologous

to a gene present on an IS element, and second, the gene was always present on a 5

kb NruI fragment. In contrast, strains isolated from domestic livestock where cpe

was plasmid encoded, the gene was close to IS1151, not linked to the HindIII repeat,

generally preceded by ORF3, but never encoded on a Nru-fragment (Comillot et al.,

1995).



Sporulation and CPE

The molecular basis for regulation of CPE expression is not completely

understood. Several key factors are now accepted: a) CPE expression is sporulation

associated, b) CPE is not a structural component of the spore coat and, c) CPE is not

a post-translationally processed product of a 52 kDa precursor molecule. The

association between sporulation and CPE synthesis was first made by Duncan et al.

(1972) who showed that CPE synthesis could be blocked by early stage sporulation

mutants spoO but not but not by later stage spoV mutants. Subsequent work







18

confirmed these observations, indicating that although some CPE was synthesized

during vegetative growth due to leaky gene regulation, CPE synthesis was indeed

sporulation controlled, possibly by a global regulator such as a sigma factor (Duncan

et al., 1972; McDonel, 1986; Czeczulin et al., 1993; Kokai-Kun et al., 1994;

Lindsay, 1996). CPE expression is an exclusive trait of sporulating cells since

although the cpe gene can be transferred to E. coli, transformed cells do not express

the toxin. Enterotoxin production starts soon after sporulation is committed.

Synthesis peaks 6-8 hrs into sporulation (Smith and McDonel, 1980; McDonel, 1986)

and CPE comprises up to 10-20% of the total cellular protein content (Czeczulin et

al., 1993).

The cpe mRNA may be transcribed as a monocistronic message, and its

transcription starts approximately 200 base pairs (bp) upstream of the CPE translation

start site. This promoter region does not show any significant sequence homology

with other known bacterial promoters (Melville et al., 1994; Czeczulin et al., 1996).

The cpe mRNA has an exceptionally long half-life of 58 minutes in sporulating cells,

and this could in part be the reason for high CPE expression (Labbe and Duncan,

1977). Increased mRNA stability may result from a stem-loop structure localized 36

bp downstream of the 3' end of the cpe ORF. This region may also function as a

rho-independent transcriptional terminator to regulate expression (Czeczulin et al.,

1993). A similar situation occurs in the seb gene, where a palindromic sequence

occurs 40 nucleotides downstream from the TGA stop codon. The hairpin-structure







19

is followed by an thymine-rich region that may function as a rho-independent

transcriptional terminator (Gaskill and Khan, 1986; Jones and Khan, 1986).

CPE does not fit the classical definition of an exotoxin because the molecule

does not contain a leader sequence, nor is it transported through the cell membrane.

CPE production takes place in the mother cell cytoplasm, where it accumulates.

When CPE synthesis is excessive, paracrystalline inclusion bodies are often formed

or toxin is trapped between spore coat layers. Toxin is only released as a

consequence of mother cell lysis at stage VII of sporulation (McDonel, 1980; Labbe,

1989; Czeczulin et al., 1993; Lindsay, 1996).



Biochemistry of CPE

CPE is a single polypeptide composed of 319 amino acids. The molecule has

an isoelectric point of 4.3 and a M, of 35,317 Daltons (McClane, 1992). The protein

is heat labile and is inactivated by heating for 5 min at 600C (McDonel, 1986). The

toxin is denatured by pH extremes, but some proteolytic enzymes do not affect its

stability (McDonel, 1986). Indeed trypsinization activates CPE threefold in vitro by

cleavage of the first 23 amino acid residues respectively from the N-terminus.

Chymotryptic removal of the first 36 amino acid residues from the N-terminus

increases cytotoxicity twofold. These data strongly support the hypothesis that

proteolytic digestion of the toxin in the intestinal lumen activates CPE during FBI

(Granum and Richardson, 1991; McClane, 1992).







20
There are some data indicating that at least two different CPE molecules are

synthesized, the well-characterized CPE also known as CPE-A and the less common

CPE-86. The CPE-86 toxin is derived from a C. perfringens coatless spore mutant.

Some data indicated a high degree of nucleic acid sequence and N-terminal protein

homology between CPE-A and CPE-86 (Wojciechowski, 1995). However, other data

from amino acid analysis and matrix-assisted laser desorption (MALD) ionization

mass spectroscopy comparisons suggested distinct structural differences (Lindsay et

al., 1985; Wojciechowski, 1995; Lindsay, 1996). These conflicting results led

Wojciechowski (1995) to suggest that although the nucleic acid sequences were the

same, any differences in protein structure either resulted from post-translational

modification, or conformational differences as a result of post-translational

modification. This hypothesis remains to be proven (Lindsay, 1996). CPE-A and

CPE-86 appear to have similar mechanisms of action with regard to in vitro cell

proliferation (McClane and McDonel, 1979; Lindsay, 1988) and cytokine modulation

in several different human and animal cell lines (Mach and Lindsay, 1994, 1997).

Although CPE-86 is biologically more active in vivo and in vitro (McClane and

McDonel, 1979; Lindsay et al., 1985; Lindsay, 1988), both toxins are equally

mitogenic in the J774M4 cell line (Mach and Lindsay, 1997).

The structure-function relationship for CPE is complicated in part by the

inconsistencies in structural data. Examination of CPE-A secondary structure by UV

circular dichroism predicted totally different values to those derived from the amino







21

acid composition (Granum and Harbitz, 1985; Lindsay et al., 1985). Additionally,

attempts to predict CPE-A secondary structure using 9 different structural models

showed conflicting putative conformations that were consistent in only two areas

(Granum and Stewart, 1992). Possibly the high percentage of hydrophobic amino

acids in both protein-toxins causes equivocal results by different methods. The 3D

structure of CPE-A examined by differential spectroscopy, and by titration of

accessible amino groups suggests a 2-domain structure, where the N-terminal portion

of the protein is cytotoxic and the C-terminal portion is the binding portion (Granum

and Whitaker, 1980; Whitaker and Granum, 1980; Granum and Stewart, 1992). This

concept is strongly supported by MAb and PAGE studies (Lindsay et al., 1985;

Hanna et al., 1992). X-ray analysis of either protein-toxin could possibly provide

more definitive information. Unfortunately, although crystals of CPE-A have been

produced, they are too small in both quantity and size for analysis (Granum and

Stewart, 1992).



Structure-Function and Vaccines

Through the use of recombinant peptide fragments, and the generation of

monoclonal antibodies, the regions of activity within the CPE protein have been partly

mapped (Horiguchi et al., 1987; Hanna et al., 1989, 1991, 1992; Granum and

Richardson, 1991; Hanna and McClane, 1991). Although CPE is a single unit

polypeptide, similar to most bacterial toxins, CPE presents two distinct regions, a







22

hydrophobic toxic fragment and a hydrophilic binding fragment. As previously noted

proteolytic digestion of the N-terminus with either trypsin or chymotrypsin activates

CPE, with a concomitant increase in CPE cytotoxicity (Granum et al., 1980). Amino

acids 37 through 171 in the first half of the protein contain sequences for insertion

and cytotoxicity (Granum et al., 1981; Granum and Richardson, 1991). Amino acids

290-319 of the C-terminus contains the receptor binding region, and the receptor

moiety per se is neither capable of insertion nor is it cytotoxic (Horiguchi et al.,

1987; Hanna et al., 1989, 1992; Hanna and McClane, 1991). Other regions of the

protein remain undefined and may be necessary for large complex formation. There

has been considerable discussion as to whether a vaccine for CPE can be produced

(Hanna et al., 1989, 1992; Mietzer et al., 1992). Purified CPE may present as many

as five epitopes scattered throughout different regions of the enterotoxin. The

monoclonal antibody 3C9 neutralizes CPE cytotoxicity by blocking receptor binding

(Wnek et al., 1985). Since CPE C-terminal fragments are not cytotoxic and have

neutralizing epitopes, they have potential for CPE vaccine construction. Immunity

to C. perfringens type A FBI may also require a vaccine that would stimulate the

production of high titers of secretory IgA anti-CPE in the intestinal lumen.

For a number of reasons, however, the production of a CPE vaccine may

simply not be possible. In both human and animal studies, administration of CPE

only induces transient immunity. Circulating antibodies to CPE are only present for

a few weeks and there is no long term immunity (Bouvier-Colle et al., 1989; Hoffian







23
et al., 1987). Indeed there is some suggestion that the first toxicosis induces

sensitivity to subsequent intoxication. This is similar in manner to the toxicosis

presented after administration of Staphylococcus aureus enterotoxins (SEs). In one

sense a lack of long-term immunity is not surprising since both groups of toxins are

known to be superantigens. That CPE possess superantigenic activity is perhaps the

most important finding about the nature of the molecule, since it may explain some

of the unusual aspects of the toxin's activities, and the relationship CPE has with

other non-foodborne diseases such as SIDS. It is very difficult, albeit impossible to

generate long term immunity to a superantigen. Normal antigens activate 1:10,000

T cells while superantigens may generate 1:5. While this over stimulation of the

immune system generates the release of high concentrations of cytokines, it does not

allow for the generation of long term immunity since the immune response is so

rapid. CPE-A over-stimulates a selected group of T cell receptors, namely VB 6.9

and VB 22 (Bowness et al., 1992), and out-competes SEA for binding sites.



Silent CPE Genes

Although microsequence variation of various cpe genes has been observed, the

cpe ORF is generally highly conserved. The CPE sequence has little homology with

other proteins, except for slight homology with a C. botulinum complexing protein,

and a 5 amino acid sequence from the Vibrio cholera B subunit (Czeczulin et al.,

1993, 1996; Hauser et al., 1994; Lindsay, 1996). Recent studies by Lindsay (1996)







24

indicates that a silent CPE gene is encoded within the intervening sequence of the Iota

toxin gene from C. perfringens type E. There is 90.4% homology between the cpe

gene sequence and that found associated with the iota toxin (UIa) (Lindsay, 1996).

The potentially derived amino acid sequence from UIa shows 81.7% identity with

CPE. Minor modifications (only single base changes) could increase the sequence

identities to 96.8%. Lindsay (1996) calculated that the evolutionary relationship

between the two sequences was 10 PAM (accepted point mutations per 100 residues)

for the nucleic acid sequence (9.6 changes/100 nucleotides) and a PAM of 21 for the

amino acid sequence (18.3 changes/100 amino acids). This argued that the

evolutionary relationship was too close to simply be fortuitous. Comparison of the

sequences upstream of the cpe and pr-cpe (IUa) genes showed similar regulatory Hpr

sequences. This strengthened the evolutionary relationship between pr-cpe and cpe,

and suggested that cpe is on a mobile genetic element which extends at least 500

nucleotides upstream of the cpe ORF (Lindsay, 1996).



CPE Intragastric Mechanisms of Action

The CPE mode of action in vitro and in vivo has few similarities with other

bacterial toxins, thus it presents an unique toxicosis (McClane, 1994, 1997; Lindsay,

1996). Since CPE causes fluid and electrolyte losses in the GI tract of humans and

animals, it has been classified as an enterotoxin (McDonel, 1986). Based on in vivo

animal model studies, CPE appears to alter the fluid and electrolyte balance







25
compromising the villus integrity, thus breaking down the normal GI absorption and

secretion mechanism, which is pathologically manifested as diarrhea (Sherman et al.,

1994; Lindsay, 1996). Animal studies also show that CPE targets the small intestine

with high affinity causing gross desquamation to the intestinal villi. Histopathological

damage can occur very rapidly, thus CPE is also considered cytotoxic (McDonel and

Duncan, 1975; McDonel, 1986; Sherman et al., 1994).



Characteristics of CPE Binding. Complex Formation and Insertion

The current model based on the extensive studies from McClane's lab is that

CPEs action is a multi-step process with four events. Briefly, CPE first binds to a

50 kDa membrane protein creating a 90 kDa small complex. Second, the structural

change of CPE contained in the small complex, which results from either CPE or

small complex insertion into the cellular membrane, or from small complex

conformational changes. Third, the formation of a 160 kDa large complex resulting

from the small complex binding to a 70 kDa membrane protein (Wnek and McClane,

1989; McClane, 1997). The fourth and final step is the loss of plasma membrane

permeability properties, caused by either the large complex directly serving as a pore

or simply by directly affecting membrane permeability (Wieckowski et al., 1994,

McClane, 1994, 1997; Czeczulin et al., 1996).

The CPE receptor(s) is found in several different cell types of mammalian

species and due to it's broad distribution the receptor is thought to have an important







26
physiological role, but not be essential for cell viability (McDonel and McClane,

1979; McDonel, 1980; Wnek and McClane, 1986; McClane et al., 1988b; Sugii and

Horiguchi, 1988; McClane, 1994). Receptors is in plural since there is some

disagreement as to whether more than one receptor exists. Kinetic studies from

different labs found contradictory results (McDonel, 1986; McClane, 1994). Affinity

chromatography studies of brush border membranes (BBM) and Vero cells strongly

suggest two proteinaceous CPE receptors of 50 kDa and 70 kDa (Wnek and McClane,

1983; Wnek et al., 1985; Sugii and Horiguchi, 1988). Additional support for the 50

kDa protein came from immunoprecipitation analysis of a 90 kDa CPE-mammalian

protein complex (small complex) that consisted stoichiometrically of one 50 kDa

membrane protein and one CPE molecule (Wieckowski et al., 1994). The 160 kDa

protein was determined from molecular weight analysis of the large protein complex

indicating the complex was stoichiometrically composed of one 50 KDa protein, one

CPE molecule and a 70 kDa protein. CPE binding is receptor specific and saturable

(106 CPE receptors per cell). Binding occurs rapidly (15-30 min) and in one tissue

culture system binding appeared to be temperature sensitive (CPE binding at 40C is

lower than at 370C) (McDonel and Duncan, 1975; McClane et al., 1988b; McClane,

1994; Wieckowski et al., 1994).

The action mode of many membrane-active toxins, including CPE, often

involves plasma membrane-permeability imbalances through toxin insertion into the

lipid bilayer (McDonel, 1980; McClane, 1994). The McClane model suggests that







27

after CPE binds to receptors it remains inserted in the lipid bilayer, toxin is neither

dissociated nor internalized (McClane, 1994). Binding to receptor(s) is a two-step

process during which CPE acquires amphiphilic capabilities (Wieckowski et al.,

1994). As previously noted, binding is irreversible, protease resistant and possibly

temperature dependent (McClane et al., 1988b; McClane, 1994). CPE is inserted

into the membrane after binding and is entrapped within the lipid bilayer. Although

the CPE-small complex undergoes conformational changes, it is unknown whether the

entire CPE-small complex or CPE alone remains entrapped (Czeczulin et al., 1996).

Unlike small complex formation, CPE-large complex formation is temperature

dependent. Large complex is formed above 24C but not at 4C, thus at 4C CPE

cytotoxicity is blocked. Cytotoxicity can be unblocked by transfer to higher

temperatures. Although the 70 kDa protein is part of the large complex, it is

unknown whether the molecule is a receptor. Possibly it is simply a functional

membrane protein brought into close contact with the CPE-small complex through

stearic attraction, CPE insertion into lipid bilayer, or conformational charge in the

small complex (Kokai-Kun and McClane, 1996).

Large complex formation is thought to either directly affect plasma membrane

permeability (influx and efflux) by functioning as a membrane "pore" (ion-permeable

channel) or by interfering with membrane pump regulation via continuous pump

activation (Sugimoto et al., 1988; Czeczulin et al., 1996; Kokai-Kun and McClane,

1996). The permeability changes may develop within 5 minutes, and restrict passage







28

to ions and small amino acids (less than 200 Daltons) that transit through membrane

lesions of about 0.5 nm2 in size (McDonel and McClane, 1979; McDonel, 1986;

McClane, 1994; Czeczulin et al., 1996; Kokai-Kun and McClane, 1996). Since

eukaryotic cells have a lower intracellular ion concentration than the external medium

in order to maintain a normal colloid-osmotic equilibrium (McClane, 1994),

permeability changes allow for a rapid ion influx into the cell cytoplasm. This initial

influx is calcium ion dependent, requiring elevated intracellular levels of Ca2 ions.

High levels of calcium ions may lead to collapse of the cytoskeleton (McClane,

1994). A rapid influx of small molecules causes the plasma membrane to "stretch"

facilitating the additional influx of macromolecules > 5 kDa. This leads to gross

osmotic changes and cell destabilization, often seen as membrane blebs. Loss of

essential amino acids causes secondary effects of inhibition of DNA, RNA and protein

synthesis, and thus the cells become nonviable (McClane and McDonel, 1979;

McDonel, 1986; McClane et al., 1988a; Hulkower et al., 1989).



CPE and Medicine

Although the above described CPE-induced changes were elucidated from in

vitro cell studies, it is believed that the same effects occur in vivo during CPE induced

FBI. CPE affects intestinal epithelial cells in a similar fashion. Villous epithelial

cells present small molecule permeability changes, disruption of normal villus

integrity and function, morphologic damage, cell lysis and net secretion of fluids and







29

electrolytes (McClane, 1994). The ability of CPEs to kill a wide variety of cells both

in vivo and in vitro has also led to the evaluation of CPE-86 as anti-neoplastic agent

(Lindsay, 1996). Preliminary studies showed that CPE-86 was destructive to Lewis

lung carcinoma cells in vivo and neoplastic cell lines (P388, B16-F1, and L110). The

mechanisms of neoplastic cell death were the same as in Vero cells with membrane

permeability changes, bleb formation, inhibition of nucleic acid and protein synthesis,

and subsequent cell death. Bacterial toxins have a great potential as anti-cancer drugs

(Pastan et al., 1995; Lindsay, 1996), and several plant and bacterial toxins (ricin,

Pseudomonas exotoxin A) have previously been suggested as therapeutic agents for

cancer treatment via linkage to cell specific antibodies (magic bullets). Although the

data with CPE-86 is preliminary it suggests that the C. perfringens CPEs have a

potential use in medicine.



Activation of CPE

CPE causes dose dependent death in mice where mice die quietly with

symptoms of respiratory interference and shock, while in vitro, CPE induces dose

dependent morphological damage, inhibition of nucleic acid synthesis, modulation of

membrane transport, lysis and cell death in Vero cells (McClane and McDonel, 1979;

Mach and Lindsay, 1994; Lindsay, 1996). Human fetal ileal cells (FI) (ATCC

CCL241) are resistant to the action of CPE, even at toxin levels of 1 gg/well. The

mechanism of this resistance is unknown, and it can be considered very strange that







30
cells from the actual tissue affected in vivo by CPE should not be affected in vitro.

A possible explanation is that since the FI cells are derived from fetal tissue they

are likely undifferentiated. In this state they are not susceptible, and require

differentiation to become susceptible.

Studies by McClane et al., (1987) and Mach and Lindsay (1994) indicated that

the activity of CPE can be dramatically exacerbated both in vivo and in vitro by the

presence of interferon-gamma (IFN-y). In vivo, the presence of IFN-- can decrease

the LD50 1,000 fold and reduce the time to death 360 fold. In vitro both Vero and

FI cells either pretreated with IFN--y then CPE, or IFN--y combined with CPE showed

dramatic sensitization with a several log fold increase in CPE activity.



Enhancement of CPE Activity and Human Non-Foodbome Disease

Both groups proposed that enhancement of activity likely resulted from IFN-y

sensitizing cells to the action of CPE. In vitro, IFN-y possibly acts by sensitizing

cells to the action of CPE resulting in death, possibly by the same mechanisms

(inhibition of protein synthesis and destruction of the cell membrane) but at lower

toxin concentrations. Mach and Lindsay (1994) and Lindsay et al., (1994) suggested

that these observations assisted in the formulation of a toxico-hypothesis for the

sudden infant death syndrome (SIDS). The observation that a majority of victims

have infections in the two weeks prior to death is known (Morris, 1987; Murrell et

al., 1987; Murrell et al., 1992). Viruses, bacteria and bacterial toxins are all known







31

inducers of IFNs (Collier and Kaplan, 1985; Pestka et al., 1987; Ijzermans and

Marquet, 1989; Baron et al., 1991; Chonmaitree and Baron, 1991). Therefore, there

may be a link between prior infection increasing IFN levels, sensitization to bacterial

toxins and SIDS. Although studies conducted over 15 years ago in relation to

viraemia and SIDS showed no evidence of a systemic viral infection or elevated

interferon at death (Ray and Hebestreit, 1971; Seto and Carver, 1978). It is possible

that increased circulating IFN levels might not have been observed, since the immune

response to the initial antigen had abated. Alternately, the tissue-cell-methodology

for detecting IFNs may not have been very sensitive. Howatson (1992) recently

examined the relationship between viral infection and the production of IFN-a. He

concluded that the abnormal presence of IFN-ae in neurones of the medulla of the

brain stem suggested that it was premature to discount a viral hypothesis for some

proportion of SIDS cases. Jakeman et al., (1991) recently showed that the toxicity

of several bacterial toxins could be significantly exacerbated in infant ferrets by a

previous infection with influenza virus. These authors speculated a role in SIDS by

some mechanism(s), including an enhancement of cell permeability which may allow

increased or more rapid uptake of a toxin, resulting in death by the same mechanism

but at lower concentrations. Mach and Lindsay (1994) argued that by extension, it

is likely that the same situation occurs in human infants. They proposed that a

window of vulnerability occurs in the life of some infants due to immunological

immaturity, which predisposes them to infection. In the weeks prior to death, these







32
infants suffer from an infection which induces the synthesis of IFNs, sensitizing the

infant to a later albeit more virulent infection which may act as a trigger for sudden

death. See the following section for further discussion of the CPE toxico/SIDS

hypothesis.



Role of CPE in Non-Foodborne Diseases (SIDS)

CPE has recently been implicated in a very unusual disease state, the sudden

infant death syndrome (SIDS). SIDS is defined as the sudden death of an infant

from one month to one year of age, which remains unexplained after a complete post-

mortem examination, including an investigation of the death scene and a review of the

case history" (Hoffman et al., 1988; Zylke, 1989). In the United States SIDS claims

the lives of 6-7,000 infants (1/1,000 live births) (Willinger, 1989) and remains the

number one cause of post-neonatal infant mortality (Lindsay, 1996). While a large

number of theories have been proposed for SIDS, the reason for this large number

of infant deaths remains unresolved (Staton, 1980; Valdes-Dapena, 1980; Thach et

al., 1988; Verrier and Kirby, 1988; Spika et al.,1989; Wilkinson, 1992). There are

many epidemiological indices found in SIDS victims, however, the most striking are

the age at death, and that > 85 % of SIDS victims were ill in the two weeks prior to

death. It has been suggested that some infants present a "window of vulnerability"

where physiological discrepancies or abnormalities make them susceptible (Stephens,

1990; Lindsay et al., 1992, 1993; Murrell et al., 1993; Murrell et al., 1994; Lindsay,







33

1996). The nature of these abnormalities are as yet unresolved, however, they may

possibly explain why SIDS fatalities occur within such a narrow time frame

(Willinger, 1989). Although premature infants are more susceptible to SIDS,

prematurity is not a determinant, since gestational age is not related to age at death

(Buck et al., 1988; Grether and Schulman, 1989). This might suggest that a common

factor or factors are responsible for, or triggers the biochemical changes that lead to

death.

Common pathological indices observed at autopsy include thymic petechiae

(pinpoint hemorrhages) and patchy pulmonary edema, muscle fiber necrosis of the

diaphragm with a histopathology suggesting hypoxia, astrogliosis of the brainstem,

leukomalacia (Jones and Weston, 1976; Valdes-Dapena, 1983; Krous, 1984; Guilian

et al., 1987; Beckwith, 1988; Hollander, 1988; Gillan et al., 1989; Guntheroth, 1989;

Kariks, 1989; Bruce and Becker, 1992). It has been suggested that these indices

might result from or be caused by infections. It should be stressed that the presence

of a particular pathogen within an infant is not predictive of a disease state, or SIDS.

Most infants have infections during their first year of life, yet only a small percentage

die from SIDS (Aron, 1983; Bettelheim et al., 1990; Blackwell et al., 1994, 1995).

Therefore it has been suggested by several authors that a small sub-population of

infants are predisposed to SIDS and when the correct "conditions" occur death is a

likely sequelae.












CHAPTER 3
DETECTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN
AFTER IN VITRO BINDING TO MURINE TISSUES



Introduction

In humans, a Clostridium perfringens foodborne illness (FBI) occurs 6-12 hours

after ingestion of vegetative cells. Upon entering the host C. perfringens responds

to changes in environmental stress by initiating the sporulation cycle with concomitant

production of enterotoxin (CPE), and subsequent induction of FBI. In the small

intestine, CPE utilizes a unique mechanism of action to directly affect the plasma

membrane of cells leading to inhibition of macromolecular synthesis, morphological

damage and cell lysis (cytotoxicity), and fluid loss (enterotoxigenicity).

Desquamation of villous cells allows CPE to be absorbed and systemically distributed

throughout the body, causing various pathophysiological (neurotoxigenic) and immune

responses (superantigenic), which may lead to severe illness or death.

In order to study the pathophysiological and immune responses generated by

CPE, the whole body distribution of toxin after absorption must be determined, which

was the overall goal of this dissertation. To fully elucidate toxin distribution,

administration had to be performed both intraperitoneally, and through the natural

port of entry, the digestive tract (intragastrically). The aim of the work described in


34







35

this chapter was to determine, using in vitro techniques and the murine model, which

tissues had the ability to bind intact CPE molecules.



Materials and Methods

Murine Model

During the last 25 years there has been a large number of animal species used

as model systems to study the action of CPE. Dr. Lindsay's lab has placed a

particular emphasis on mice for a number of reasons. First, the species use in a wide

range of published toxicological studies, and the relative inexpensive cost of

purchasing and maintaining large numbers of animals. Second, large numbers of

animals that are genetically similar increases the validity of results. Third, mice have

been extensively used by other researchers to examine the effects of bacterial toxins.

Fourth, the mode of action, symptoms and pathophysiological changes caused by CPE

in the murine model strongly mimic those observed in human cases of CPE toxicosis.

Taking all these factors into consideration, we believe that the mouse is an excellent

model, and its use can be justified.



C. perfringens Enterotoxin (CPE)

Purified, freeze dried CPE was gratefully obtained from Dr. Bruce McClane,

University of Pittsburgh. For administration, toxin was resuspended in phosphate

buffered saline-Tween (PBS-TW) (0.15 M NaCI, 0.01 M Na2HPO4, 0.01 M







36

NaH2PO4, and 0.2% Tween 20, pH 7.2) and standardized to a final concentration of

1 pg/il of protein. Protein determinations were made by the method of Lowry et al.

(1951) with bovine serum albumin as the standard.



Preparation of Antisera to CPE

Three milligrams of purified CPE were supplied as antigens. Individual 4-5

kg New Zealand White rabbits were used to generate antisera. One milliliter of

Freud's Complete adjuvant was emulsified with 1 mg (300 pl) of the antigen and 5

injections (3 intradermal, 1 subcutaneous, 1 intramuscular) were administered to the

rabbits. The injections were repeated using 1 mg of antigen and 1 ml of Freud's

Incomplete adjuvant 30 days later and again 14 days after that. To monitor the titre,

test bleeds from the ear vein were done every 7 days after an injection. After the

third series of injections, the rabbit was anesthetized with Ketamine and Rompun and

bled by cardiac puncture. Serum was separated by centrifugation and stored at -200C.



Biotinvlation of Antisera

Antisera to CPE was biotinylated using NHS-LC-Biotin (Pierce, Rockford, IL).

One hundred microliters of 0.05 M bicarbonate buffer, pH 8.5, was added to 2 mg

of antisera. NHS-LC-biotin (0.04 mg) was added to the antisera and incubated on ice

with mixing every 20 minutes for 2 hours. Unreacted biotin was removed by the

addition of PBS to a total volume of 1 ml followed by dialysis in 6-8,000 wt cut off







37

dialysis tubing (Spectrum Medical Industries, Inc., Los Angeles, CA) against two 500

ml volumes of PBS for 16 hours. The biotin labeled antisera was aliquoted and

stored at -200C.



Animals

Three week old (12-13g) male Swiss Webster (SW) mice used in this study

were obtained from Harlan-Sprague Dawley, through the University's Department of

Animal Resources. The Department of Animal Resources is an IACUC Veterinary

controlled facility. Animal Resources (AR) order, install, raise and maintain the

animals within the Food Science and Human Nutrition (FSHN) Departmental Animal

Facility as prescribed by the University of Florida IACUC. Animals are kept on a

12/12 light-dark cycle at 250C and are examined on a daily basis by AR who also

change bedding and provide food and water. All animal studies were performed

within the FSHN Animal Facility.



Animal Tissue Preparation

Mice were allowed to reach body weights of between 15-18 g before

experimental use. This was usually 4-6 days after delivery, which allowed time for

the animals to become accustomed to their surroundings and being handled, thus

reducing any contributing stress factors. Animals were euthanized by CO2

asphyxiation followed by cervical dislocation. Mice were dissected and all organs







38

(brain, thymus, heart, lung, liver, kidney and small bowel) were isolated, washed in

PBS-TW, weighed and transferred to sterile 15 ml polypropylene tubes. Three

milliliters of PBS-TW were immediately added and each organ was homogenized and

disrupted on ice using a Polytron (Brinkmann Instruments) three times for 15 seconds.

All individual organs from each mouse were kept separately in polypropylene tubes,

to which CPE resuspended in PBS-TW was added at a standardized concentration,

vortexed and allowed to bind.



CPE:Tissue Binding

To determine the level of CPE binding to tissue from each organ, five separate

experiments (E1-E5) were performed, each requiring 8 mice. In four experiments the

toxin concentration varied from 40 ng (El) to 80 ng (E2), 1 tig (E3) and 2.5 yg (E4)

of CPE/ml of tissue homogenate, with incubation for 1 hour at 27C (lab controlled

to this temperature). In the fifth experiment (E5) the CPE concentration was 2.5

Atg/ml tissue homogenate, with incubation at 40C for 4 hours. During the incubation

period, samples were gently vortexed every 10 minutes to ensure complete

distribution of toxin and tissue. Each experiment contained an equivalent number of

negative controls, that is, tissues treated with only PBS-TW and no (zero) CPE.

After incubation, a 1 ml sample from each individual organ was removed, and the

remaining homogenate was stored at -700C. The 1 ml isolated supernatant sample







39
was centrifuged at 12,000 x g for 30 minutes at 4C, and the supernatant collected

and stored at -700C.



Enzyme Linked Immunosorbent Assay (ELISA)

The ELISAs were conducted using the methodology described by Crowther

(1995). Before the binding experiments were performed, extensive preliminary

studies (concentration vs concentration analysis) were undertaken to determine the

optimum levels and incubation times for each of the ELISA components and steps.

A 200 gl supernatant sample (At: defined as soluble protein-antigens) from each of

the tissues was thoroughly mixed with 200 /l of coating buffer (CB: 1.5 mM Na2CO3,

3.3 mM NaHCO3, 3 mM NaN3). Triplicate 100 pl aliquots were added to individual

wells of a polystyrene 96 well ImmunolonTM 2 flat bottom plate (Dynatech

Laboratories, Chantilly, VA), and allowed to passively adsorb for 16 hours at 40C.

Any unbound sample was then aspirated, and each well was washed five times (X 5)

with 100 pl PBS-TW. Primary polyclonal antibody (lAb) to CPE previously diluted

to 1/10-3 in PBS-TW was added to each well (100 pl/well) and allowed to adsorb for

16 hours at 4C. Any unbound loAb was then aspirated, and each well was washed

X 5 with 100 pl PBS-TW. Secondary (2Ab) goat-anti rabbit IgG antibody labeled

with the enzyme alkaline phosphatase (AP) (Sigma Chemical Co., St. Louis, MO)

diluted to 1/10-3 in PBS-TW was added to each well (100 pl/well) and allowed to

adsorb for 2 hours at 27C. Wells were then aspirated and washed as described







40

previously. The At: 1Ab:2Ab-AP complex was detected by the addition of 100 pl

of alkaline phosphatase buffer (50 mM Na2CO3, 50 mM NaCO3, 0.5 mM MgCl2)

containing 100 Ag of p-nitrophenyl phosphate (Sigma Chemical Co.). The enzyme

reaction was stopped after 1 hour by the addition of 100 pl of 1.0 N NaOH.

Presumptive CPE positive samples are indicated by a sample color change from clear

to yellow, which was quantitated by spectrophotometric analysis at 405 nm (Bio-Rad

ELISA reader, model 2550).



Western Immunoblot Analysis

Tissue proteins and any CPE contained in the supernatant were separated by

polyacrylamide gel electrophoresis (PAGE) with sodium dodecyl sulfate (SDS) using

Bio-Rad Mini-Protean II slab gels and the buffered system of Laemmli (1970).

Stacking and separating gels were 4.0% and 10% polyacrylamide, respectively.

Prestained protein molecular weight markers (27-180 kDa) were obtained from Sigma

Chemical Co. Gels were stained with 0.15% Coomassie Brilliant Blue R-250

overnight after being used for Western immunoblots. After soaking in transfer buffer

(0.192 M glycine, 0.025 M Tris, 20% methanol) for 30 minutes, gels were

immunoblotted using a Bio-Rad mini Trans-blot electrophoretic transfer cell and

nitrocellulose membranes (Stratagene, La Jolla, CA) according to manufacturers

directions. After transfer, the membrane was blocked with PBS containing 0.1%

Tween 20 and 7.0% casein for 1 hour. The membrane was drained and soaked for







41

16 hours in a 103 dilution of biotinylated antisera in blocking buffer. The membrane

was then washed and soaked in a 1 mg/ml solution of strepavidin (Sigma Chemical

Co.) for 2 hours. After a second washing step with PBS-TW the membrane was

soaked and rinsed 3 times for 5 minutes with substrate buffer (100 mM Tris, 100 mM

NaCI, 5 mM MgC12, pH 9.5). The protein bands were made visible by soaking the

membrane in 10 ml of substrate buffer containing 0.1 mg/ml of nitroblue tetrazolium

and 0.05 mg/ml of 5-bromo-4-chloro-3-indolyl phosphate for 30 minutes at 40C.



Statistical Analysis and Interpretation

Statistical analysis of ELISA readings was performed using an one-way

analysis of variance (ANOVA) with the STATISTICA for Windows software

program, release 4.5 (Copyrightc StatSoft, Inc. 1993). When a significant difference

was found (p < 0.05), Post hoc comparisons were done by using Tukey's honest

significant difference (HSD) test, choosing an alpha level for critical ranges

set at (a = 0.05).



Results and Discussion

ELISA Data Interpretation and Difficulties

The aim of these experiments was to determine, using in vitro techniques, to

which tissues (organs) within the mouse the CPE molecule had the highest affinity.

Previous studies (Wnek and McClane, 1989; McClane and Wnek, 1990) have







42

indicated that one CPE molecule binds to a single 50 kDa receptor (R1) on the cell

surface to form the CPE-R1 "small complex". Formation of small complex occurs

at both temperatures 4C and 27C. Large complex formation occurs when the

hydrophobic portion (N-terminal:toxic) of CPE is internalized and interacts with a 70

kDa protein (receptor R2) to form CPE-R1-R2. This interaction (large complex

formation) is irreversible and does not occur at 4C.

Several definitions were made for the ELISA results. First, an ELISA was

considered statistically positive after performing an ANOVA (p < 0.05) and a Post-

hoc comparison (HSD test, a = 0.05). The positive ELISA (a = 0.05) indicated

that after the tissue-CPE interaction was completed, the isolated supernatant still

contained free CPE molecules. Second, an ELISA was considered statistically

negative after performing an ANOVA and a Post-hoc comparison (a = 0.05) where

no significant difference was detected between control and CPE treated tissues. Thus

it appears that a given tissue supernatant contained a greater number of R1 receptors

than toxin which bound all available CPE molecules. Third, that if the experiments

were conducted with fresh organ tissues at 27C, then CPE binding to receptor R1

would form the small complex (CPE-R1). This small complex would, within the 1

hour incubation period conformationally change and interact with receptor R2 to form

the large complex (CPE-R1-R2). Since we were using polyclonal antibodies to CPE,

a number of epitopes on the CPE molecule were available for interaction (binding).

We simply did not know whether the interaction of CPE with R1 on the membrane







43

left any regions of the toxin molecule available for interaction with the lAb. This

is not an unlikely situation since the N-terminal hydrophobic portion of CPE does not

appear to bind to R1, and thus could be available. However, since the binding of

CPE to R1 is temperature dependent and rapid, we believed that if CPE did bind to

R1 to form the small complex, then large complex formation would be relatively

rapid, and thus no portion of the CPE molecule would be available to bind to the

lAb. Additionally, if the amount of tissue from a specific organ was kept constant

and the amount of toxin varied, it might be possible to approximate CPE saturation

levels for a specific tissue. That is, since the molecular weight of CPE is known, it

was possible to calculate the number of CPE molecules/mg protein, and thus the

number of R1 receptors/organ.

Initially there were some difficulties in conducting the ELISAs, particularly for

liver and kidney tissues which had a higher degree of non-specific binding. This was

possibly a function of some tissues having both a higher amount and more varied

number of tissue proteins, which may have caused some stearic hindrance.

Alternately some proteins may have had a higher non-specific affinity for the lAb,

caused by conserved protein regions common to both CPE and tissue proteins,

resulting in stearic interference. Fortunately, most of the non-specific binding

problems were resolved by using purified lAb, and additional blocking steps which

significantly reduced background interference. Despite these problems, we believe

the ELISA results are valid for all tissues.







44


Table 3.1 Detection of unbound CPE by ELISA in murine organ
tissues after in vitro interaction.


CPE concentration and binding conditions



Organ 40 ng 80 ng 1.0 ug 2.5 ug/27C 2.5ug/4C

Brain -- -- -- ++ ++

Thymus -- ++ + +

Heart -- ++ ++ ++

Lung -- ++ ++ ++

Liver

Kidney

Bowel





Binding conditions: 40 ng/ml CPE at 27C for 1 hour
80 ng/ml CPE at 270C for 1 hour
1.0 Ig/ml CPE at 270C for 1 hour
2.5 /g/ml CPE at 27C for 1 hour
2.5 1/g/ml CPE at 40C for 4 hours

(--) = no free CPE detected [a = 0.05]
(+ +) = free CPE detected [a = 0.05]







45

CPE Binding

A summary of the five experiments described in the materials and methods is

shown in Table 3.1. In experiments El (40 ng/ml) and E2 (80 ng/ml) no free CPE

was observed in the tissue supernatant from any mouse organ. This suggested that

neither of these toxin levels saturated the R1 sites in any organ tissue. The molecular

weight (MW) of a compound is the sum of the atomic weights of the atoms in the

molecule. It is the number of grams containing Avagadro's number of molecules

(6.022 x 1023) (a mole is Avagadro's number of molecules). Thus, since the MW of

the CPE molecule is 35,317 (Czeczulin et al 1993) then the minimum number of R1

receptors in each of the organ tissues examined in this study was > 8.5 x 109 /mg.

Experiment E3 examined a 12.5 fold increase in CPE to 1,000 ng/ml. Results

indicated that at this level the heart and lung showed free CPE at a significance level

of a = 0.05. This suggested that the receptor-saturation threshold for CPE in the

these organs was between 80-1,000 ng/ml, and the number of R1 receptors per mg

of tissue was between > 8.5 x 109 /mg and < 1.5 x 10" /mg. No free CPE was

detected with the brain, thymus, liver, kidney and bowel. Experiment 4 examined

a 2.5 fold increase in CPE from E3, and a 32 fold increase over E2. The results

obtained were consistent with E3 indicating that 2,500 ng/ml of CPE saturated R1

receptors in thymus, heart, lung and brain tissue. Receptor-saturation threshold for

CPE in the brain and thymus was between 1,000-2,500 ng/ml, and the number of R1

receptors per mg of tissue was between > 4.3 x 1010 /mg and < 1.5 x 10" /mg for







46
the brain and between > 2.0 x 10" /mg and < 5.0 x 10" /mg for the thymus. No

free CPE was found with liver, kidney and bowel tissues indicating that the number

of CPE Rl-receptors in these tissues was > 1.5 x 10" /mg tissue. These results

confirm previous conclusions by McDonel (1980): first, that the receptor saturation

threshold for liver, kidney and bowel tissues was > 1.5 x 10"/mg of tissue protein,

and there appeared to be no difference between the number of CPE-receptors between

these organ tissues. Second, he also considered that competitive or inhibitory

factor(s) released from the brain homogenate into the reaction buffer prevented

significant CPE-receptor binding to this specific tissue. In this present study

significant binding of CPE to brain tissue was found, which contradicts McDonel

(1980). However, it should be noted that in this current study the methods used were

far more sensitive and the binding technique was different. McDonel manipulated the

tissue through a greater number of steps which may have caused the release of the

"inhibitory" CPE-binding factor he suggested.

One of the initial assumptions was that even if CPE bound to R1, the small

complex formed might still allow additional binding of lAb to any available regions

of the CPE molecule. We argued however, that under our experimental conditions

any small complex formed would conformationally change and become large complex

very rapidly. Thus lAb binding to small complex was not an issue. Small complex

formation is temperature independent, however, large complex formation is

temperature dependent. In E5 we examined whether saturating levels of CPE (2,500







47

ng/ml) gave the same results at 4C when compared to using 2,500 ng/ml CPE at

27C (E4). Table 3.1 shows that indeed the results from E4 and E5 are directly

comparable, where in both experiments free toxin is only found in the supernatants

of heart, lung, thymus and brain tissues, but not in liver, kidney and bowel.

There are several possible alternatives. If McClane and Wnek (1990) are

correct and small complex formation is temperature independent and large complex

formation temperature dependent, it simply makes no difference, since no free toxin

is found in some tissues. That is, in liver, kidney and bowel tissues small complex

formation does not appear to allow any residual lAb binding to CPE. The small

complex must simply be in an unavailable conformational state. Although it cannot

be equivocally stated that this situation also occurs with heart, lung, thymus and brain

tissues, it would seem reasonable to assume that CPE binding in these tissues is no

different than in liver, kidney and bowel.

Western Immunoblots were performed to determine if the technique could

distinguish between control and CPE treated tissue. Only lung tissue was examined

since this organ bound a significant amount of toxin, yet had a low level of interfering

soluble proteins compared to liver and kidney. Preliminary studies indicated that with

the Western immunoblot technique, the lower limit of CPE detection either as pure

toxin or unbound CPE was 50 ng/ml. Results shown in Figure 3.1 lane C indicated

that control (untreated) tissue showed no positive bands, indicating that contaminating

tissue proteins from the lung did not bind non-specifically to the lAb. When CPE







48































Figure 3.1 Western immunoblot of lung tissue: CPE, lung tissue without CPE, and
pure CPE detected with a 10-3 dilution of antisera to CPE. The amount
of protein (CPE) in each sample is noted in parenthesis. Lane A:
purified CPE (28 /ig); lane C: lung tissue supernatant without
CPE (0 ag); lane L: lung tissue supernatant with CPE (100 ng).







49
was in excess to the receptors within the lung (see Table 3.1 treatment 1.0 Ag, and

Figure 3.1 lane L) free toxin was clearly detected as a Mr 35,000 band (lane L)

exactly at the same position as pure CPE (lane A). Strong antibody positive bands

were also detected at approximately M, 87,000 (small complex) and 160 kDa (large

complex) (lane L).

These in vitro binding studies suggested that the liver, kidney and bowel had

the highest number of CPE receptors within their tissue (> 4.5 x 10'3 receptors/

organ), followed by brain and thymus (_ 4.5 x 1013 receptors/organ), and heart and

lung (5 1.7 x 10'" receptors/organ). These results are not unexpected since

introduction of CPE to the body occurs via the gastrointestinal tract, and metabolism

and detoxification of CPE likely occurs via the kidneys and liver. Indeed Skjelkvale

et al. (1980) in his studies with '25I-labeled CPE suggested that the kidney was a

target organ for CPE binding and that this organ contained the highest level of CPE

receptors. The studies conducted herein, confirm that the kidney is an organ that

contains a large number of CPE receptors, however, there is no indication that this

organ is the main target. Skjelkvale et al. (1980) suggested that CPE is metabolized

and expelled from the body within urine, which would make the kidney an organ of

focus. However, further in vivo studies are required to confirm this. The liver and

small bowel would also be organs of focus since cells from these tissues are highly

susceptible to the action of CPE (Lindsay and Dennison, 1986a, 1986b; Kokai-Kuhn

and McClane, 1996; Mach and Lindsay 1997). Although our data strongly suggest







50
that liver, kidney and bowel tissues contain the highest number of receptors, and

require very high levels of CPE to reach receptor saturation, the results do not

necessarily translate to the in vivo model. That is, these organs may be the main

focus of CPE distribution but not necessarily CPE activity in vivo. It is possible that

low levels of CPE within a specific tissue have a more dramatic effect on the host,

than high levels of CPE in another. For example, high levels of CPE in the bowel

may be enterotoxigenic and cytotoxic causing diarrhea, fluid loss and tissue

desquamation. However, low level CPE binding and activity in brain and lung tissue

may affect neurologic and respiratory status. Additionally, CPE binding to heart

tissue may have cardiotrophic effects. Thus, enterotoxigenicity may be transient, but

alteration in neurologic, respiratory or cardiac status may be lethal. Chapters 4 and

5 detail our studies to determine the in vivo distribution of CPE after intraperitoneal

and intragastric administration and the consequential effects.













CHAPTER 4
DETECTION AND DISTRIBUTION OF CLOSTRIDIUM PERFRINGENS
TYPE A ENTEROTOXIN AFTER IN VIVO INTRAPERITONEAL
ADMINISTRATION INTO SWISS WEBSTER MICE


Introduction

McDonel (1980) indicated that in vitro, CPE binds with different specificities

to various organ tissues. Results suggested that liver and kidney tissues contained the

largest number of CPE receptors, and that both organs were specific sites of CPE

attack that required > 0.5 jAg of CPE/mg of tissue to reach receptor saturation.

There is however, an anomaly in McDonel's supposition. The data suggested that to

reach complete saturation of liver and kidney tissues in vivo required lethal levels of

CPE to be administered and absorbed, an unlikely circumstance in most CPE induced

foodborne illnesses. Skjelkvale et al. (1980) using the murine model attempted to

determine the distribution and levels of CPE binding after intravenous (IV)

administration of radioiodinated CPE. Results suggested that the liver and kidneys

were specific organs of CPE accumulation and attack, and that a major fraction of

CPE was rapidly metabolized and excreted in urine. This apparently confirmed

McDonel's findings, however, alternative explanations are possible. First, decay of

radiolabeled toxin could have spread radioactivity to organs where CPE was not

apparent, leading to false positives. Second, CPE was not administered through its


51







52

natural port of entry, that is intragastrically (IG). It is not known whether CPE

administered both IP and IG give similar results. It could be argued however, that

this second point is moot, since Tsai and Riemann (1975a) showed that CPE orally

administered to mice was present in blood within minutes. Thus, IG and IP

administration may be similar, however, CPE is known to be structurally altered by

proteolytic enzymes in the small bowel before systemic absorption (Granum and

Richardson, 1991). Third, CPE distribution in mice was only monitored during the

two hours after IV administration, and only at a single CPE dose. It is possible that

there are various CPE distribution patterns which are toxin dose and time dependent.

This chapter describes the approaches to determine the organ/tissue distribution of

varying doses of non-labelled CPE toxin after IP administration into Swiss Webster

mice.



Materials and Methods

CPE-Toxin and Antisera

Freeze dried CPE obtained from Dr. Bruce McClane was prepared for IP

administration as described in Chapter 3. The biological activity (specific activity)

of preparations was examined before use by Vero cell analysis (Mach and Lindsay,

1994) and was standardized to 4,000 EU/Al toxin. The methods to produce antisera

to CPE are described in Chapter 3.







53

Animal preparation and IP enterotoxin administration

Animals were purchased and maintained as described in Chapter 3. As all

animals were bred from the same line and obtained from the same source it was

assumed that they were genetically similar. Gross examination indicated no apparent

phenotypic differences. When animals weighed 15-18 g (25-30 days old), they were

randomized and grouped six/cage. To reduce any administration differences, one

person (Keller) held the mice in the correct alignment, while another (Lindsay)

measured the correct volume and performed the administration. Animals were

administered CPE IP in the left side of the peritoneal cavity, using a 1 ml tuberculin

syringe and a 27 gauge needle. After injection animals were returned to their

respective cages and monitored every 15 minutes.



Study la: CPE Concentration

This study was performed to determine the following: i. the CPE concentration

at a time which caused death within 72 hours; ii. the CPE concentration at which the

pathophysiological changes observed during the toxicosis could be strongly predicted

without interference from cases of random death within the treated mouse population.

These data would then allow animal necropsy and tissue sampling with reliability.

CPE was administered IP at various levels ranging from 0.1 to 5 mouse lethal

doses (MLD): nonlethal 0.1 MLD (0.5 /g/250 tl) to 0.25 MLD (1.25 J/g/250 1l);

sublethal 0.5 MLD (2.5 #tg/250 il) to 1.0 MLD (5.0 /ig/250 D:); lethal 2.0 MLD







54
(10.0 Ag/250 A1) to 5.0 MLD (50.0 Ag/250 il); and control: 250 il of PBS. Twelve

mice were used at each treatment level, and 3 control mice were administered PBS.

Any pathophysiological changes were monitored every 15 minutes from To

(immediately after administration) to T72 hours, or time to death, which ever came

first. Mice were necropsied and all organs (brain, thymus, lung, heart, liver, kidney,

and small bowel) were isolated, washed in PBS-TW, weighed and transferred to

sterile 15 ml polypropylene tubes and stored at -70C for later analysis. Blood and

urine (where possible) were also collected. Before storage, blood samples were

gently centrifuged at 200 x g to collect serum. Both serum and urine were stored

at -700C for analysis.



Study lb: CPE Distribution vs Time

Based on data obtained in Study la (see Results and Discussion for detailed

data) a CPE concentration of 10 ig CPE/250 /l PBS was chosen for IP administration

in the distribution versus time study. Twenty four animals were prepared, and

administered CPE IP in a single dose as described above. Six similar weighted

animals were administered PBS as controls. At six time intervals 0.25, 1, 2, 3, 4,

and 5 hours, four animals were randomly chosen from the toxin administered group

and one from the control group and euthanized. Mice were necropsied and all organs

(brain, thymus, lung, heart, liver, kidney, and small bowel) were collected as

described above. Blood and urine (where possible) were also collected. Before







55

storage, blood samples were gently centrifuged at 200 x g to remove red blood cells,

and the serum collected and stored at -70C, as was any urine sample.



Immunological Methods

To prepare organ tissue samples for ELISA, 3 ml of PBS-TW was added and

each organ was homogenized and disrupted using a Polytron as described previously.

Serum and urine were examined without PBS-TW dilution. The materials and

methods to conduct the ELISA and Western immunoblot are described previously in

Chapter 3.



Interpretation of ELISA Assays and Statistical Analysis

Several assumptions made with regards to the interpretation of the ELISA

assays, based on data presented in Chapter 3, and on recently published studies by

Kokai-Kuhn and McClane (1996). First, that the ELISA assay could detect both free

CPE and CPE-R1 receptor bound small complex noting that the small complex moiety

had to be in the soluble fraction. Second, that CPE polyclonal antibodies (CPE-pAb)

used in the ELISA assay had a higher affinity for free CPE, and thus would

preferentially bind to this moiety. This is also stoichiometrically logical since CPE

binding to R1 would reduce the number of epitopes available for CPE-CPEpAb

binding. Third, low (sublethal) levels of CPE would not saturate R1 receptors in

most organs, while high (lethal) levels of CPE would saturate R1 receptors in most







56

organs. Thus, when administering CPE IP at high levels, unbound CPE would

preferentially bind to the CPE-pAb resulting in a positive ELISA (a = 0.05). See

Chapter 3 for discussion. If the ELISA was positive at a = 0.2, CPE may be

detected as a mixture of unbound, small complexed and large complexed CPE.

Alternately, at sublethal CPE levels (without receptor saturation) CPE detection might

be predominantly as a mixture of small complex and large complex forms. Statistical

analysis was performed as described in Chapter 3.



Results and Discussion

Symptoms During Toxicosis

Immediately after CPE-A administration the animals presented an accelerated

heart rate and a spasmatic breathing pattern (hyperpnea). Animals exhibited arched

backs, ruffled-opaque fur, opaque eyes, disorientation, loss of appetite for both food

and water, and a requirement for group association in the corners of the cage

(gathering). The CPE toxin appeared to express apparent neurotoxicity where animals

exhibited flaccid paralysis, suggesting that the vagus nerve had possibly been

compromised. At no time did animals present any apparent signs of pain, and until

recovery or death the animals would cuddle and rest. When animals recovered from

the toxicosis they showed no signs of any long term pathophysiological damage, and

ate and drank as before CPE administration. Similar symptoms in animals due to

CPE toxicosis were observed by Tsai and Riemann (1975a, 1975b) and Skjelkvale et







57

al. (1980) although not in as much detail. The toxicosis presented differences in onset

time and recovery dependent on the route of administration and the purity of CPE

preparations.



Pathological Findings at Necropsy

At necropsy the liver, thymus and kidneys showed petechiae. Some animals

had blood in the bladder, while others had an enlarged liver and/or kidneys. At high

toxin levels the animals had a mushy bowel with distinct signs of proteolysis of the

intestinal walls. The integrity of the blood vessels seemed to be compromised as

blood was found in the peritoneal and thoracic cavity. It was extremely difficult to

collect blood, and the average blood volume collected was about 1 ml. This was

likely due to two factors: first, the animals were very small, and only had a total

blood volume of 1.5-3.0 ml. Second, blood coagulation seemed to be accelerated in

CPE treated animals compared to control animals. This observation has been

confirmed in studies by Wallace et al. (1997). The liver in CPE treated animals was

heavier than in controls, suggesting that the toxicosis causes a higher metabolic rate

(detoxifying mechanism) which is an attempt to clear CPE from the body. Thus more

blood and/or fluids were present in the organ(s), and either upon organ failure or

death the fluids were trapped.







58


Table 4.1 ELISA detection of unbound CPE in murine organ
tissues after in vivo IP administration.


CPE dose (MLD*) administered




Organ 0.1 0.25 0.5 1.0 2.0 5.0

Brain -- -- -- -- ++ ++

Thymus -- ++ ++ ++ ++

Heart

Lung -- + + + ++ + +

Liver

Kidney -

Bowel -- -- -- ++

Blood -- -- ++ + + +

Urine -- ++ ++ ++ ++


CPE dose administered: 0.1 MLD (0.5 gg/250 /l)
0.25 MLD (1.25 /g/250 j1)
0.5 MLD (2.5 j/g/250 pl)
1.0 MLD (5.0 /g/250 il)
2.0 MLD (10.0 /g/250 pl)
5.0 MLD (50.0 gg/250 pl)
control: 250 /1 of PBS

Mouse Lethal Dose.

(--) = no free CPE detected [c = 0.05]
(+ +) = free CPE detected [a = 0.05]







59

CPE Concentration Study

When CPE was administered at sublethal doses (0.1 and 0.25 MLD) none of

the ELISA assays for any of the seven organs examined or blood (as serum) and urine

was statistically significant at a = 0.05 level (Table 4.1). This suggested two

alternatives: first, CPE at this low level was being rapidly metabolized by the liver

and kidney, and excreted in the urine before the animals were necropsied and

screened by ELISA. Second, the low CPE levels did not saturate cellular receptors

as suggested in the section Interpretation of ELISA Assays. Thus neither unbound

CPE nor small complex was available for detection. These results are consistent with

the pathological findings as none of the animals died within the 72 hour time period

after CPE administration. Indeed, while animals presented clear symptoms of CPE

induced toxicosis they were able to recover within 8 to 12 hours.

CPE administered at sublethal-lethal levels (0.5 to 1.0 MLD) was detected as

free toxin in thymus and lung tissue, and urine (a = 0.05). All other organ tissues

(brain, heart, liver, kidney, and bowel) were not statistically significant (a = 0.05).

Serum was found to contain free toxin (a = 0.05) at 1.0 MLD but not at 0.5 MLD.

Animal sensitivity and resistance to CPE between 0.5-1.0 MLD showed varied

responses. Two patterns of sensitivity and resistance were apparent. First, 30-50 %

of animals appeared resistant since they effectively recovered from the toxicosis

within 8-12 hours of CPE administration. Animals sensitive to CPE died

predominantly within the first 24 hours. This bimodal distribution of effects was







60

consistent through several studies, and was in part not unexpected, since the toxin

concentration was near the LD50. Normally within a homogeneous population of mice

(same species, similar age and weight) two subgroups are apparent that differ in their

ability to metabolize a toxin. The nature or reason for this variability remains

unknown, although it is likely that detoxifying mechanism(s) of an individual animal

may be less effective and compromised.

All mice administered CPE at 2.0 MLD died within 12 hours, with about 80%

of mice dying between 8 and 9 hours. At 5 MLD all mice died within 15 minutes.

Free CPE was detected (a = 0.05) in brain, thymus, lung and bowel tissues, and

blood and urine. No free toxin was apparent in heart, liver and kidney tissues, which

was in part unexpected. Several alternatives were suggested: first, none of these

organs were saturated with toxin at these levels, and any toxin present in these tissues

was present in a form (possibly large complex ?) that was undetectable. Second, due

to multiple organ failure, any damaged tissue containing CPE was removed from the

organ via the circulation. Third, the animals died from the CPE toxicosis before

detectable levels of free toxin were reached. This latter alternative is applicable for

CPE at 5 MLD where the toxin likely acted as a neurotoxin.







61


Table 4.2 Detection of unbound CPE by ELISA in murine organ
tissues after in vivo IP administration: time study.


CPE binding time in hours




Organ 0.25 1.0 2.0 3.0 4.0 5.0

Brain -- -- -- ++ ++ ++

Thymus -- -

Heart -- -- ++ ++ ++

Lung -- -- +

Liver -- ++ ++ ++

Kidney

Bowel -

Blood ++ + +/-



Treatments: administration of 10 /g/250 pl (2 MLD*) CPE, animals necropsied
at 0.25, 1.0, 2.0, 3.0, 4.0, and 5.0 hours; control: 250 1l of PBS.
* Mouse Lethal Dose.

(--) = no free CPE detected [a = 0.2]
(+/-) = free CPE detected [ac = 0.2]
(+) = free CPE detected [a = 0.1]
(+ +) = free CPE detected [a = 0.05]







62

Time Study

Based on data obtained from the concentration study a toxin level of 2 MLD

(50 pg/250 1) was chosen for the time study. This level was chosen for two reasons:

first, CPE could be detected with statistical significance and reliability by ELISA, and

second, after toxin administration, most mice died within a narrow time-frame 8-9

hours and there was no random death during first 5 hours.

Results from the time study at the single CPE level are shown in Table 4.2.

Free toxin was detected in serum within 15 minutes of CPE administration, however,

toxin was apparently cleared within three hours. Toxin was detected (a = 0.05) in

brain, heart, and liver tissues 3 hours after administration, and in the lungs (a = 0.1)

after 5 hours. CPE was not detected in thymus, kidney or bowel tissue (a = 0.2).

Since CPE is systemically absorbed and distributed, its presence in brain, heart, liver

and lung tissue was not unexpected. However, apparent lack of presence in thymus

and kidney was unexpected. Lack of presence in bowel was not unexpected since,

this tissue contains a high number of receptors. More importantly however, for toxin

to be present in ileal tissue would require translocation of toxin from the peritoneal

cavity to the lumen, against the normal translocation gradient. Alternately, lack of

free toxin in thymus and kidney may also suggest that in these tissues, that if CPE is

present, it is in the large complex form and therefore inaccessible to CPE-pAb.

Another consideration can be made. It is possible that toxin distribution and binding

does not follow a set distribution pattern. That is distribution and binding may







63


Table 4.3 ELISA and Western immunoblot detection of CPE (free toxin)
CPE:R1 (small complex) and CPE:R1:R2 (large complex) in
murine organ tissues after in vivo IP administration.


CPE concentration and binding conditions



Western immunoblot

Organ ELISA 35 85 160

Brain -- -

Thymus --

Heart -

Lung --

Liver

Kidney 0 U

Bowel --


CPE dose administered: 0.1 MLD (0.5 gg/250 /l)
control: 250 Il of PBS

ELISA
(--) = no free CPE detected [a = 0.2]

Western immunoblot
35: Mr 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1);
160: M, 160 kDa (CPE:R1:R2).
(--) = no band detected
(U) = band detected







64

depend on how CPE was systemically absorbed, the time needed for toxin to be

absorbed, exact location of IP administration, or other factors inherent to the animal

itself, for example the metabolism of the individual animal.

It is interesting to compare the results obtained for the concentration and time

studies at the same CPE levels. It appears that there are differences in toxin

distribution patterns regarding thymus, heart, and liver tissues. CPE was detected in

the free form in the time study, but not in the concentration study in heart tissue,

while the liver and thymus were negative (a = 0.1) One possible explanation for

these differences was that the sample size of the two studies was different. Also, the

time allowed for CPE distribution and subsequent binding to receptors was different.

In the time study the samples were collected within the first 5 hours, while in the

concentration studies samples were collected between 8-12 hours. Since it was not

known how time affected CPE distribution and binding, we chose not to directly

compare the ELISA data from the time and concentration studies, but to further

analyze each separately.



Western Immunoblot

From the in vitro studies it was presupposed that CPE bound to various organ

tissues within the mouse, albeit with different affinities. It was also assumed that

CPE bound initially to receptor R1 forming small complex, and then to R2 forming

large complex. The ELISA data predicts the presence of CPE on various organs.







65

However, in order to confirm CPE presence and determine the conformation and

configuration state (small or large complex) of the toxin it was necessary to conduct

analysis by Western immunoblot. It should be noted that sample preparation for each

method was significantly different. ELISA samples required no special preparation

steps, and simply screened the supernatant for soluble protein (antigen). Western's

required tissue samples to be boiled in the presence of B-mercaptoethanol and SDS,

allowing the release of tissue associated proteins. In order to initially determine the

sensitivity of the Western immunoblot assay, all tissue samples from the 0.1 MLD

concentration study were screened for the presence of CPE. These samples were

chosen since they represented the lowest level of CPE administration, and the ELISA

assay was not able to detect free CPE in any samples. Previous work had indicated

that the sensitivity of the ELISA assay in this study was > 100 pg CPE/ml, whereas

the sensitivity for the Western assay 1 pg CPE/ml. It was assumed that if CPE

was present in tissue it had to be found either as the small (85 kDa) or more likely

the large (160 kDa) complex. A summary of the analyses is shown in Table 4.3. As

previously noted, at 0.1 MLD the ELISA did not detect any unbound toxin as the Mr

35 kDa protein in any tissue. However, Western analysis of CPE treated animals

showed a 35 kDa protein band in the heart and lung which was not present in control

(untreated animals). Whether toxin was present in these organs as free (unbound)

toxin not detectable by ELISA, or as loosely bound small complex which dissociated

upon sample preparation is unknown. A 85 kDa band, presumably small complex







66

was detected in the brain, thymus, heart, kidney and bowel tissues. This band was

not present in control (untreated samples). A 160 kDa band presumably large

complex was present in brain and kidney tissues, and again this band was not present

in control tissues.

The Western analyses suggest that either the ELISA assay was less sensitive

to CPE detection when toxin was administered at nonlethal levels or that at nonlethal

levels CPE was bound as small and large complex, and thus not available for

detection. The latter is more likely since the number of R1 receptors exceeds the

number of CPE molecules, and thus receptor saturation was not reached. The

Western assay was very informative since it allowed determination of the state and

type of CPE association to specific receptors (either small or large complex). Both

the ELISA and Western assays compliment each other since the ELISA determines

receptor saturation threshold, while the Western determines the type and distribution

of CPE-receptor interaction. The type of CPE-receptor complex formed appeared to

be specific within individual organs, that is, different organs showed different CPE-

receptor complex patterns. These patterns may suggest that toxin distribution within

an individual tissue is both CPE-concentration dependent and cell type specific.

The intraperitoneal studies showed that CPE is systemically distributed

throughout the body within minutes of administration. The pathophysiological effects

of this rapid distribution were both concentration and time dependent. At nonlethal

levels CPE was not detected in blood, however, it is likely that at these concentrations







67

the toxin was rapidly bound to receptors within the liver and kidneys, and metabolized

and excreted via urine. The basis for this assumption is first, that both the liver and

kidney contain the highest number of CPE-receptors. Second, the receptors within

these organs appear to have the highest affinity for CPE (McDonel, 1980). Third,

these two organs are normally associated with the removal of toxic products from the

body. At sublethal levels CPE appears to have several modes of action. First, as a

cytotoxic-enterotoxin causing diarrhea accompanied by gross tissue damage to the

bowel. Second, as a neurotoxin since animals presented symptoms usually termed

parasympathomimetic (cholinergic). This neurotoxic effect, confirms previous

descriptions associated with CPE toxicosis in animals and humans (Murrell et al.

1987; Lindsay, 1996). At lethal levels CPE appears to act as both a neurotoxin and

a superantigen. Death is rapid, and symptoms of bradycardia, hyperpenia and flaccid

paralysis were apparent, leading to multiple organ failure, shock and death. These

symptoms suggest a complete collapse of the sympathetic nervous system

accompanied by the massive induction of cytokines at the cellular level. While these

findings apply to intraperitoneal administration of CPE it remains to be determined

whether similar whole body distribution patterns and responses occur after intragastric

administration. This is critical since CPE is normally synthesized, activated and

systemically absorbed in the small bowel. This is the focus of Chapter 5.











CHAPTER 5
DETECTION AND DISTRIBUTION OF CLOSTRIDIUM PERFRINGENS
TYPE A ENTEROTOXIN AFTER IN VIVO INTRAGASTRIC
ADMINISTRATION INTO SWISS WEBSTER MICE


Introduction

The effects of in vivo oral administration of C. perfringens type A vegetative

cells, spores suspensions, sterile culture filtrates, and crude cell preparations

containing type A enterotoxin (CPE) into various animal models has been studied

previously by several groups (Canada and Strong, 1965; Weiss et al., 1965;

Hauschild et al., 1971; Uemura et al., 1975; Tsai and Riemann, 1975a, 1975b).

Unfortunately, either the basis for the studies were questionable or the results were

inconclusive. Canada and Strong (1965) force-fed germ-free mice with live

C. perfringens type A suspensions, and measured the heat resistance of the spores

recovered after passage through the animals GI tract. As would be expected the

recovered spores had the same heat resistance as those initially fed the animals. What

this study showed or proved is moot. In other studies it appeared that either the

administration or absorption levels were too low, thus enterotoxin could not be

detected in serum, and the only signs of toxicosis were emesis and diarrhea (Weiss

et al., 1965; Hauschild et al., 1971; Uemura et al., 1975).

Tsai and Riemann (1975a, 1975b) examined the responses of mice to oral

challenge with C. perfringens type A vegetative cells, spore suspensions, and a crude


68







69
C. perfringens cell supernatant containing unquantified levels of CPE. Challenge with

vegetative cells and spore suspensions showed the bacterium to be present in the

gastrointestinal (GI) tract, heart, lung, liver, kidney, spleen, and blood. The presence

in various organs is interesting, however, there was no indication that the bacterium

was actually bound within tissue. Presence could merely have been a function of

systemic distribution via blood, and lack of washing prior to analysis. Unfortunately

there was no determination of the level of CPE in any of these tissues or within

blood. Administration of crude supernatants containing CPE, and for that matter any

other C. perfringens toxins, showed CPE present in serum. Mice also exhibited

diarrhea, and mucosal enteritis with edema and hyperemia of the small intestine.

Whether these pathophysiological responses were due to CPE or any of the

other C. perfringens toxins was not determined.

The work of Tsai and Riemann (1975a) is perhaps the best of the previous

studies, since it did show that CPE can be found in blood after oral administration.

However, several criticisms of their work can be made. First, it is impossible to

determine the level of toxin absorption let alone the amount of CPE produced in the

small intestine simply based on feeding live cultures. There is significant variation

in sporulation between C. perfringens strains in vitro, and the level of sporulation and

CPE production in vivo is simply unknown, except for the 8239 lab strain. Even for

this strain there are many variables that interplay in vivo. Second, were there any

other C. perfringens toxins produced and absorbed, since this bacterium may produce







70

up to 15 different toxins, many of which appear to act synergistically? Third, the use

of crude toxin containing supematants is fraught with problems, since as noted, many

toxins act synergistically. In a crude extract how can any one toxin be shown to

cause a specific pathological or histopathological response. Thus pure preparations

of a specific toxin have to be used to determine cause and effect. Fourth, the time

frame used to study any pathological effects needs to be extended. Distribution

studies should at least mimic FBI symptoms which usually take 1-24 hours, or

possibly longer to elicit. Further, a range of toxin doses should be administered to

determine if there are dose and time dependent responses. Fifth, toxin detection or

effect should not be investigated by a single type of assay, since a single assay may

provide either or both, false positives and/or false negatives. Thus, several methods

of analyses must be used, one of which must be biological. Finally, any method

used, has to have the ability to be statistically validated. This chapter reports studies

to determine the whole body distribution of CPE toxin after intragastric administration

through the natural port of entry, that is intragastrically, using the murine model.

The criticisms and requirements discussed above have been incorporated into the

study, in order to specifically define the response.







71

Materials and Methods



CPE-Toxin and Antisera

Freeze dried CPE obtained from Dr. Bruce McClane was prepared for IG

administration as described in Chapter 3. The specific activity of preparations was

examined before use by Vero cell analysis (Mach and Lindsay, 1994) and was

standardized to 4,000 EU/pl toxin. The methods to produce antisera to CPE are

described in Chapter 3.



Animal Preparation and IG Enterotoxin Administration

Animals were purchased and maintained as described in Chapter 3. All mice

were bred from the same line and obtained from the same source. It was assumed

that they were genetically similar, and after gross examination no phenotypic

differences were apparent. When animals weighed 15-18 g (25-30 days old), they

were randomized and grouped six/cage. To reduce any administration differences,

one person (Keller) performed all toxin preparation and oral administration (gavage).

Animals were IG administered CPE via a 250 pl volume into the oesphagous and

stomach, using a 1 ml tuberculin syringe and a 27 gauge, 2 mm ball-ended Popper

needle. Preliminary studies indicated that it was not necessary to neutralize stomach

acid using 50 mM bicarbonate buffer prior to CPE administration. Prior

administration was observed to upset the mice and vary the biological response after







72

CPE administration. Additionally, not using bicarbonate buffer more closely

mimicked the true toxicosis, since it allowed protease activation of the toxin in the

small intestine. CPE retained biological activity in the presence of PBS, and since

the volume administered was quite small, delivery was relatively straight-forward.

After CPE administration mice were returned to their respective cages and monitored

every 15 minutes.



Study la: CPE Concentration

This study was performed to determine several factors. First, the CPE

concentration that caused death within 72 hours, which would allow a direct

comparison between the IP MLD obtained in Chapter 4, and the IG MLD. Second,

the CPE concentration at which any pathophysiological changes observed during

toxicosis could be strongly predicted without interference from cases of random death

within the treated mouse population. This would then allow an IG time concentration

study to be conducted similar to that performed in Chapter 4. Thus CPE

administration, necropsies and tissue sampling could be done within a specific time

frame with reliability.

CPE was administered IG at various levels ranging from 25-75 jig

CPE (0.1-0.25 MLD) in 250 xl PBS. Twelve mice were used at each treatment

level, and 3 control mice were administered PBS alone. Mice were monitored for

pathophysiological changes every 15 minutes from To (immediately after







73

administration) to T72 hours, or time to death, which ever came first. Mice were

necropsied and all organs (brain, thymus, lung, heart, liver, kidney and small bowel)

were isolated, washed in PBS-TW, weighed and transferred to sterile 15 ml

polypropylene tubes and stored at -700C for analysis.



Study Ib: CPE Distribution vs Time

In retrospect the data obtained in Study la (see Results and Discussion for

details) proved to be of limited value, since toxin concentrations within this range

were nonlethal. Rather than conduct additional studies using large numbers of

animals to determine the IG MLD, two mice were each administered CPE IG at

concentrations of 100, 125, 150, 250 and 500 /pg/250 l and the time to death

determined. This two mouse method has been used previously to determine the MLD

of C. botulinum neurotoxin with good statistical reliability (Sugiyama, 1986). Based

on these additional data two CPE concentrations, 150 pg CPE/250 l1 PBS (0.5 MLD)

and 500 Ag CPE/250 pl PBS (2.0 MLD) were chosen for the distribution vs time

studies. For each time study, ten mice were administered CPE IG in a single dose,

nine similar weighted control mice were administered PBS alone.

Each toxin level was initially designed as a distribution vs time study,

however, for reasons discussed in the Results and Discussion section each study

presented its own unique problem. Unexpectedly at each of the CPE concentrations,

two distinct time to death patterns were apparent. That is, despite the genotypic and







74

phenotypic similarity of the mouse populations used for each study, half the

population died within one time period, the other half within a different time period.

Therefore the data were analyzed taking into account the time to death variable. Mice

administered CPE were necropsied immediately upon death. At the same time one

matched control mouse was euthanized and necropsied. All organs (brain, thymus,

lung, heart, liver, kidney, spleen, stomach, and small bowel) were collected and

stored as described previously.



Immunological Methods and Statistical Analysis

To prepare organ tissue samples for ELISA, 3 ml of PBS-TW was added, and

each organ was homogenized and disrupted using a Polytron as described previously.

Serum and urine were not examined. The materials and methods required to conduct

the ELISA and Western immunoblot are described previously in Chapter 3. The

same assumptions for the interpretation of the ELISA assay and statistical analysis as

described in Chapter 4 were used in the IG studies.



Vero Cell Assay

Tissue suspensions of all organs were tested for biologically active CPE using

the Vero cell assay (African Green Monkey kidney cells:ATCC CCL81) (Mach and

Lindsay, 1994). Cells were cultured in Sarstedt (Newton, NC) 75 cm2 flasks and

maintained in modified Eagles medium (MEM) containing 10% fetal bovine serum







75

(FBS), 2 mM L-Glutamine, and antibiotic/antimycotic (Sigma Chemical Co.). Cells

were incubated as an adherent monolayer in a humidified incubator with 5%

CO,/95% air at 37C. When confluent, cells were removed with trypsin-EDTA

(Sigma Chemical Co.), diluted to 10 ml in MEM and plated 0.2 ml/well (10' cells/ml)

in Costar 48 well cluster dishes. Cluster dishes were incubated as described above

for 24 hours, the media removed and each well washed twice with sterile 15 mM

PBS. Earl's balanced salts solution (free of FBS) containing 0.28 mM phenol red

(BSS-PR) was added to the wells (final well volume of 0.2 ml). To test the biological

activity (cytotoxicity) of any CPE in the high-speed centrifuged tissue supernatant,

100 Il of BSS-PR was replaced with 100 jl of tissue supernatant and incubated for

2 hours. Cells were examined for morphological damage and viability with a Nikon

phase contrast microscope. A sample was considered biologically active if it caused

> 50% cell lysis in duplicate wells. To determine whether cell death was caused by

CPE, all suspected CPE positive tissues as determined by a positive ELISA, positive

Western and positive Vero cell assay, were re-assayed in the Vero cell assay using

CPE neutralizing antibody. Tissue supernatants were treated with a 1.100 dilution

of CPE primary antibody, thus neutralizing any available toxin. Any sample which

was ELISA, Western and Vero cell assay positive, and negative in the Vero cell assay

after neutralization with anti-CPE antibody, was considered a confirmed positive.







76

Results and Discussion

Symptoms During Toxicosis

To induce a toxicosis by IG administration required a 10-15 fold higher CPE

level. Upon IG CPE administration, mice presented similar symptoms and

pathophysiological changes as observed for IP CPE administration, see Chapter 4 for

detailed description. Mice administered nonlethal CPE levels fully recovered from

the toxicosis within 6-8 hours. Animals administered sublethal levels of CPE,

presented two response modes. Approximately 25 % of the mice died within 4 hours,

while the remaining 75% recovered within 6-8 hours. Although the animals used

were genetically and phenotypically similar, this response was not unexpected since

the amount of toxin administered was near the 0.5 MLD. Administration of lethal

levels of CPE caused the death of all animals, however, two modes of death were

apparent. First, a "sudden" death induced within minutes of toxin administration, and

second, a nonabrupt death that required several hours for manifestation.



Pathological Findings at Necropsy

At nonlethal CPE levels there was no apparent gross histopathological tissue

damage to organs, other than the bowel. At sublethal and lethal CPE levels the liver,

thymus, and kidneys showed petechiae, and the liver and kidneys were enlarged.

Enlarged organs suggest hyperactivity which leads to organ failure. Additionally, the

GI tract exhibited signs of extensive proteolysis (mushy and spongy) to the







77

submucosa, and without care upon dissection the tissue would disintegrate. It was

apparent that the damage to the digestive tract would significantly compromise

metabolism to the point where animals had very little chance of recovery. This type

of bowel tissue damage has been observed for other C. perfringens infections, for

example pigbel, where recovery from the disease requires surgery. Collection of

blood was not possible since the blood vessels in the peritoneal cavity showed signs

of hemorrhage. Coagulated blood was found in the peritoneal cavity, lungs, liver,

kidneys and digestive tract. This may have contributed to a decrease in blood volume

within the animal. Wallace et al. (1997) observed similar responses in their work,

and preliminary data indicated that the coagulation factors appeared to be affected

during CPE-toxicosis. No urine could be collected during necropsy likely due to

kidney organ failure, or possibly loss of involuntary control of the urinary system.



CPE Concentration Study

The ELISA assay did not detect free CPE administered at 0.1 MLD (nonlethal)

in any organ-tissue sample (Table 5.1). This was expected since nonlethal levels

would not have saturated tissue receptors, and thus toxin was not available for

detection. At 0.25 MLD the brain and kidney had free CPE at a significance level

of ca = 0.2, but all the other organs showed no free CPE (Table 5.1). This may

suggest that at 0.25 MLD the CPE threshold for receptor saturation in brain and

kidney tissues was reached. However, since the kidney is the main organ of CPE








78




Table 5.1 Detection of unbound CPE by ELISA in murine
organ tissues after in vivo IG administration.


CPE dose administered


Organ 0.1 MLD 0.25 MLD

Brain +/-

Thymus

Heart

Lung

Liver

Kidney +/-

Bowel



CPE dose administered: 0.1 MLD (25 Ig/250 1l)
0.25 MLD (75 tg/250 (1)
controls 250 pl of PBS

Statistics
(-) = no free CPE detected [a = 0.2]
(+/-) = free CPE detected [a = 0.2]








79

Table 5.2 ELISA and Western immunoblot of CPE (free toxin)
CPE:R1 (small complex) and CPE:R1:R2 (large complex)
in murine organ tissues after IG administration.


CPE concentration and binding condition

0.25 MLD


Organs ELISA Western
35 85 160
Brain +/- --

Thymus -

Heart -- -

Lung -

Liver -

Kidney +/- -

Bowel -


CPE binding conditions: administration of 0.25 MLD (75 gg/250 Il)
controls were administered 250 1l of PBS

Statistics: (--) no free CPE detected [a = 0.2]
(+/-) free CPE detected [a = 0.2]

Western: 35: Mr 35 kDa (free CPE); 85: Mr 85 kDa (CPE:R1)
160: Mr 160 kDa (CPE:R1:R2); (0) band; (--) no band







80

detoxification and excretion, free CPE might be expected. Whether the CPE-

receptors in brain tissue were saturated is unknown, however, based on the in vitro

studies this would appear likely.

Western immunoblots of tissues from animals administered 0.25 MLD showed

CPE not as free CPE, but rather as a combination of small (CPE-R1) and large (CPE-

R1-R2) complex (Table 5.2). These results confirm Kokai-Kun and McClane's

(1996) studies showing that the ELISA assay could detect CPE when it is exposed but

sequestered in small or large complex. The statistical significance of the ELISA

results are lower than perhaps expected (a = 0.2) however, the binding coefficient

of free CPE to the primary antibody is higher than bound toxin. Thus, as noted

previously in Chapters 3 and 4, the sensitivity of the ELISA is lower when it is

detecting CPE as a mixture of complexes and free toxin. An interesting and as yet

unexplained finding was that lung and liver tissues contained CPE predominantly as

large complex.

Sensitivity differences between the ELISA assay and the Western immunoblot

have been discussed previously in Chapter 4. However, it is important to stress that

while the assays are different, that is, the immunoblot specifically discerns between

free CPE and CPE-complexed forms, the ELISA predominantly detects free CPE in

tissue supernatants (due to a higher binding coefficient). Thus the methods become

complimentary.







81

Time Study

Since there was negligible tissue damage upon administration of nonlethal

amounts of CPE no time studies were conducted at this level. ELISA results of

administration of sublethal CPE levels (0.5 MLD) showed free toxin (significant at

a = 0.05 level) in thymus, heart, and bowel tissues, and (significant at a = 0.1

level) in kidneys (Table 5.3). Brain, lung, liver, stomach and spleen as assayed by

ELISA contained no free CPE. As discussed previously two patterns of responses

were found at sublethal levels, recovery or death. Comparison of CPE time-

distribution patterns between these two groups showed remarkable similarities except

for the kidneys. Animals that died within 4 hours after toxin administration had free

CPE, while kidneys of animals that recovered from the toxicosis had no significant

free CPE. This is not unexpected since animals that recovered from the toxicosis

obviously metabolized and excreted toxin via urine.

Western immunoblots confirmed the presence of free CPE in all organs except

in bowel and spleen tissues (Table 5.4). Brain, thymus, heart, lung, liver, and kidney

tissues also had CPE in the small and large complex form. The stomach tissue

contained only free CPE strongly suggesting that this organ has no CPE receptors.

Free CPE in stomach tissue confirmed that intragastric administration was successful,

and that the presence of CPE in lung tissue was the result of systemic distribution not

faulty administration. The bowel had no CPE in any form (free or complexes) as

detected by Western immunoblot. Possibly, the gross destruction of the small bowel








82



Table 5.3 ELISA detection of unbound CPE in murine organ tissues
after in vivo IG administration: time study.


CPE binding time in hours


Organ 4 hr. 72 hr.

Brain

Thymus + + + +

Heart + + + +

Lung

Liver

Kidney +

Bowel + + + +

Stomach

Spleen




CPE binding conditions: administration of 0.5 MLD (150 #ig/250 dl)
animals necropsied either at 4 or 72 hours
controls were administered 250 p1 of PBS

Statistics: (-) = no free CPE detected [a = 0.1]
(+) = free CPE detected [a = 0.1]
(++) = free CPE detected [a = 0.05]







83

Table 5.4 ELISA and Western immunoblot of CPE (free toxin)
CPE:R1 (small complex) and CPE:R1:R2 (large complex)
in murine organ tissues after IG administration.


CPE concentration and binding condition

0.5 MLD


Organs ELISA Western
35 85 160
Brain -- 0 M

Thymus ++ + N E

Heart ++ +

Lung --

Liver -- U U

Kidney +

Bowel ++ -

Stomach -

Spleen


CPE binding conditions: administration of 0.5 MLD (150 jig/250 Il)
animals necropsied at either 4 and 72 hours
controls were administered 250 J1 of PBS

Statistics: (--) no free CPE detected [a = 0.2]
(+) free CPE detected [a = 0.1]
(++) free CPE detected [a = 0.05]

Western: 35: M, 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1);
160: Mr 160 kDa (CPE:R1:R2); (m) band; (--) no band







84

resulted in the complete elimination of any toxin from this tissue. Desquamated tissue

was simply either excreted or proteolytically digested before necropsy.

Lethal CPE administration of 2 MLD also showed a bimodal time response

where approximately 50 % of the animals died within the 15 minutes, referred to as

"sudden" death, while the remaining 50% died a nonabrupt death requiring several

hours for lethal outcome. ELISA results showed free CPE in thymus, liver, kidney,

stomach, and bowel tissue (a = 0.05), and at a = 0.1 heart tissue (Table 5.5).

Brain, lung, and spleen tissue had no detectable levels of free CPE. Higher free CPE

levels were expected in liver and kidney tissues since these are the organs of

detoxification and excretion. Another hypothesis could be proposed that, immediately

after absorption in the small intestine CPE reaches the liver and kidney via portal vein

compromising these organs. This in turn causes multiple organ failure, shock, and

death. Liver and kidney failure could reduce blood pressure and blood flow to other

organs. As blood is entrapped in failing organs less CPE reaches other tissues, thus

not saturating their receptors. Detection of higher levels of free CPE in more organs

than previously observed for lower toxin levels could simply be a function of toxin

concentration. Simply stated, as more CPE becomes available a greater number of

organ-tissues would reach their CPE receptor-saturation threshold. Western

immunoblot results showed CPE both as free toxin, small and large complex in all

organs except the stomach and spleen (Figures 5.1, 5.2 and Table 5.6).







85



Table 5.5 ELISA detection of unbound CPE in murine organ
tissues after in vivo IG administration: time study.



Organ Sudden Nonabrupt

Brain

Thymus + + + +

Heart +/-

Lung

Liver + + + +

Kidney + + + +

Bowel -- ++

Stomach + + + +

Spleen



CPE binding conditions: administration of 2.0 MLD (500 gg/250 l)
controls were administered 250 pl of PBS
Sudden death: death within 15 minutes
Nonabrupt death: several hours for lethal outcome

Statistics: (--) = no free CPE detected [a = 0.2]
(+/-) = free CPE detected [a = 0.2]
(++) = free CPE detected [a = 0.05]







86



































Figure 5.1 SDS-PAGE of tissue supernatants: tissue protein analysis of mice
administered 2 MLD IG of CPE. Tissue supernatants from: (C)
control tissue homogenate (no CPE), (U) lung, (H) heart, (B) brain,
(M) high molecular weight standard, (T) thymus, (V) liver, (K) kidney,
(S) spleen, (A) CPE.







87



































Figure 5.2 Western immunoblot of tissue supernatants: mice administered 2 MLD
IG of CPE. Toxin detected with polyclonal antisera to CPE. Tissue
supernatants from: (C) control tissue homogenate (no CPE), (U) lung,
(H) heart, (B) brain, (M) high molecular weight standard, (T) thymus,
(V) liver, (K) kidney, (S) spleen, (A) purified CPE.







88

Table 5.6 ELISA, Western immunoblot and Vero cell detection of: CPE (free
toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in
murine organ tissues after in vivo IG administration.


CPE concentration and binding conditions for 2MLD

Organs ELISA Western Cell death
35 85 160
Brain -

Thymus ++

Heart +/-

Lung --

Liver + + 0 0 0 0

Kidney ++ N aU 0

Bowel ++ M a U

Stomach + + -- -

Spleen -- -- 0


CPE binding conditions: administration of 2.0 MLD (500 /ig/250 /l)
controls were administered 250 pl of PBS

Statistics: (--) no free CPE detected [a = 0.2]
(+/-) free CPE detected [a = 0.2]
(+ +) free CPE detected [a = 0.05]

Western: 35: M, 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1);
160: M, 160 kDa (CPE:R1:R2); (U) band; (--) no band

Cell death: (m) in Vero cell assay, sample supernatant caused > 50% cell lysis
and cell lysis was completely inhibited by neutralization with
primary antibody to CPE. (0) no cell lysis







89








































Figure 5.3 Vero cell assay of murine tissues: (A) liver, (B) brain, and (C) bowel
tissue supernatants. Animals were administered 2 MLD IG of CPE.







90
The stomach had only free CPE confirming the ELISA results, and the spleen had no

detectable CPE in any form (free or complexed).

The Vero cell analyses concurred with both the ELISA and Western

immunoblot results. Free and biologically active CPE was detected in thymus, heart,

bowel, and stomach tissue (Table 5.6 and Figure 5.3). Brain, and lung tissue had

free biologically active CPE, which confirmed the Western results. The Vero cell

assay did not detect free biologically active CPE in liver, kidney or spleen tissues,

and these negative results were consistent and reproducible for all samples tested. It

would appear that unlike other organ tissues except the stomach, the spleen has no

receptors for CPE.

In summary, these findings confirmed the need to use all three assays (ELISA,

Western and Vero cell) to elucidate the state of the CPE molecule within any tissue.

Although free CPE (35 kDa) appeared to be present within a specific tissue based on

the ELISA and Western assays, it did not necessarily mean that the toxin was

biologically active. Indeed, free toxin was detected in liver and kidney tissues by

ELISA and Western but neither of these tissues contained biologically active CPE.











CHAPTER 6
CONCLUSIONS AND SUMMARY

Clostridium perfringens type A enterotoxin is a known causal agent of

foodborne illness in man, and has also recently been implicated as a trigger in the

sudden infant death syndrome (SIDS). The pathophysiological changes associated

with either disease state appear very complex. Diarrhea and histopathological tissue

damage to many organs within the body are a function of the cytotoxigenic and

enterotoxigenic nature of the CPE molecule. However, other symptoms observed

during the toxicosis suggest both neurotoxigenic (cholinergic) and superantigenic

activities. The goal of this research was to better understand the nature of these CPE

induced pathophysiological disturbances, by determining where CPE was specifically

distributed during toxicosis. The murine model was chosen for this study since as a

model it offered many unique genotypic and phenotypic advantages.

In vitro tissue studies showed CPE binding to cellular membrane protein

receptors in brain, thymus, heart, lung, liver, kidney, and bowel. Liver, kidney and

bowel tissues contained the highest number of CPE-receptors.

Intraperitoneal CPE administration revealed that toxin was systemically and

rapidly absorbed, being distributed throughout the body within minutes of

administration. CPE induced pathophysiological effects were both concentration and

time dependent. All organ tissues within the body except the spleen and stomach


91




Full Text

PAGE 1

DISTRIBUTION AND DETECTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN AFTER INTRAPERITONEAL AND INTRAGASTRIC ADMINISTRATION USING THE MURINE MODEL By ANDREAS MARKUS KELLER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1997 i

PAGE 2

This dissertation is dedicated to my parents, Karl-Heinz and Charlotte; to my wife Ceciha; to my daughter Dominik; to my Onkel Peter; to my Omi Lotti; and witii remembrance to my late sister Sylvia; and Opi Paul and Omi Elisabeth, for their never ending love, patience, enthusiasm, dedication, understanding and support. Thank you, God, for such a wonderful family.

PAGE 3

ACKNOWLEDGMENTS I would like to thank Dr. James A. Lindsay, my major advisor (I prefer the German word "Dr. Vater," since advisor is a too simplistic a term to describe a mentor, role model and benefactor), from the bottom of my heart, for his neverending support, patience, persistence, guidance, motivation, commitment, encouragement, enthusiasm, and enlightenment in every step of my doctoral research, graduate program, and personal life. I am extremely grateful to the other members of my doctoral committee. Dr. Douglas L. Archer, Dr. Sean F. O'Keefe, Dr. Mark L. Tamplin, and Dr. Ramon D. Littell, for their interest, advice, suggestions, review of manuscript, and supportive role in my research. I am deeply appreciative of the Food Science and Human Nutrition Department, USDA, and NIH for providing me witii graduate assistantships and all other funding. I would like to thank F. Morgan Wallace for his advice, support, and friendship. I also give my sincerest thanks to Armette S. Mach for conducting the tissue culture smdies, and above all her suggestions, assistance and guidance in the lab, making it a fun, efficient, sound, and safe environment to work in. Finally, my sincerest thanks go to my friend Dr. Antonio A. Figueiredo. iii

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGMENTS iii LIST OF TABLES vi LIST OF FIGURES vii ABSTRACT viii CHAPTERS 1 INTRODUCTION 1 2 REVIEW OF LITERATURE 5 Clostridium perfringens 5 Classification 5 C. perfringens Foodborne Illness 9 History 9 Type A Foodbome Illness: General Characteristics 11 Type A FBI Outbreaks 12 Identification of C. perfringens FBI Outbreaks ... 13 C. perfringens and Pathogenicity 15 Virulence Factors Conttibuting to C. perfringens FBI 15 Molecular Biology of CPE 16 Sporulation and CPE 17 Biochemistry of CPE 19 Structure-Function and Vaccines 21 Silent CPE Genes 23 CPE Intragastric Mechanisms of Action 24 Characteristics of the CPE Binding, Complex Formation and Insertion 25 CPE and Medicine 28 Activation of CPE 29 iv

PAGE 5

Enhancement of CPE Activity and Human Non-Foodborne Disease 30 Role of CPE in Non-Foodborne Diseases (SIDS) 32 3 DETECTION OF CLOSTRIDIUM PERFRINGENS TY?E A ENTEROTOXIN AFTER IN VITRO BINDING TO MURINE TISSUES 34 Introduction 34 Materials and Methods 35 Results and Discussion 41 4 DETECTION AND DISTRIBUTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN AFTER IN VIVO INTRAPERITONEAL ADMINISTRATION INTO SWISS WEBSTER MICE 51 Introduction 51 Materials and Methods 52 Results and Discussion 56 5 DETECTION AND DISTRIBUTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN AFTER IN VIVO INTRAGASTRIC ADMINISTRATION INTO SWISS WEBSTER MICE 68 Introduction 68 Materials and Methods 71 Results and Discussion 76 6 SUMMARY AND CONCLUSIONS 91 REFERENCES 98 BIOGRAPHICAL SKETCH Ill V

PAGE 6

LIST OF TABLES Table page 3.1 Detection of unbound CPE by ELISA in murine organ tissues after in vitro interaction 44 4.1. ELISA detection of unbound CPE in murine organ tissues after in vivo IP administration 58 4.2 Detection of unbound CPE by ELISA in murine organ tissues after in vivo IP administration: time smdy 61 4.3 ELISA and Western immunoblot detection of CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after in vivo IP administration 63 5.1 Detection of unbound CPE by ELISA in murine organ tissues after in vivo IG administration 78 5.2 ELISA and Western immunoblot of CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after IG administration 79 5.3 ELISA detection of unbound CPE in murine organ tissues after in vivo IG administration: time study 82 5.4 ELISA and Western immunoblot of CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after IG administration 83 5.5 ELISA detection of unbound CPE in murine organ tissues after in vivo IG administration: time study 85 5.6 ELISA, Western immunoblot and Vero cell detection of: CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after in vivo IG administration 88 vi

PAGE 7

LIST OF FIGURES Figure page 3.1 Western immunoblot of lung tissue 48 5.1 SDS-P AGE of tissue supernatants 86 5.2 Western immunoblot of tissue supernatants 87 5.3 Vero cell assay of murine tissues 89 6.1 Murine model for CPE distribution after IG administration 93 (a) Nonlethal murine model 93 (b) Sublethal murine model 93 (c) Sudden death murine model 95 (d) Nonabrupt death murine model 95 vii

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DISTRffiUTION AND DETECTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN AFTER INTRAPERITONEAL AND INTRAGASTRIC ADMINISTRATION USING THE MURINE MODEL By ANDREAS MARKUS KELLER May, 1997 Chairman: Dr. James A. Lindsay Major Department: Food Science and Human Nutrition Clostridium perfringens has been described as the most important anaerobic pathogen of man, and is considered the most common cause of enteric diseases in animals. Virulency of the bacterium is related to the production of at least 15 different protein-toxins, many of which are lethal. Diseases associated with C. perfringens infections and production of these protein-toxins include myonecrosis (gas gangrene), necrotic enteritis, antibiotic associated diarrhea, sudden infant death syndrome and food poisoning in man and animals. viii

PAGE 9

To understand these changes, the whole body distribution of CPE after either intraperitoneal and intragastric administration was determined, using the murine model. Results showed that CPE appeared to have three different modes of distribution and activity which were time and concentration dependent. Nonlethal levels induced enterotoxigenic symptoms, while sublethal levels induced symptoms described as parasympathomimetic. Administration of lethal levels induced two patterns of death, first, a sudden death induced within minutes of CPE administration, and second, a nonabrupt death that required several hours for manifestation. Animals expressed symptoms of respiratory distress, shock and multiple organ failure, similar to the action of a superantigen. From the murine model studies, the following parallels may be suggested for CPE toxicosis in humans. Nonlethal levels of CPE causes a toxicosis similar to a self-limiting foodbome illness. Systemic absorption of sublethal levels of CPE induce a neurotoxicosis, from which healthy individuals would likely recover. However, death could occur in immunocompromised persons or the elderly. Systemic absorption of lethal levels of CPE will cause death in both healthy and immunocompromised individuals. The finding of a "sudden" pattern of death after lethal ingestion is pivotal, since this mimics the suggested response of some at-risk infants to CPE toxicosis, and supports the role of CPE as a trigger in some cases of the sudden infant death syndrome. ix

PAGE 10

CHAPTER 1 INTRODUCTION Clostridium perfringens foodborne illness (FBI) is associated with enterotoxin(s) from type A strains and is the third most common cause of bacterial FBI in the USA following Salmonella spp. and Staphylococcus aureus. Annual costs of FBI associated with C. perfringens in the USA and Canada are estimated to be higher than $200 million. Mortality rates from C. perfringens FBI are dependent on age and immune stams, and the debilitated, immunocompromised, young and elderly are at high risk. In the USA mortality rates may be as high as 4% (Janssen et al., 1996). Besides causing foodborne illness C. perfringens has been associated with some unusual disease states, for example wound infections and sudden infant death syndrome (SIDS) (Lindsay et al., 1993, 1994; CDC, 1994; Lindsay, 1996). Studies indicate that C. perfringens infections and type A enterotoxin appear to be associated with 50-80% of the approximately 7,000 SIDS deaths/year in the USA (Lindsay, 1996. Consequently, there has been a dramatic increase of interest in the bacterium's pathogenicity, virulence determinants, and the signals conttolling expression of these determinants. Smdies show that C. perfringens type A enterotoxin (CPE), which is produced in the highest amounts during bacterial sporulation, may have the ability to 1

PAGE 11

2 modulate the host defense system, by acting as a superantigen and exerting immunomodulatory effects on various lymphoid cell populations, thus playing an important role in the overall pathogenesis of the organism (Lindsay, 1996). Superantigens may activate and stimulate up to 1 in 5 T-cells as compared to a classical antigen which normally stimulates 1 in 10,000 T-cells. Stimulation of T-cells may lead to the induction of cytokines, such as interferon-7, nimor necrosis factor-6, interleukins, and others, usually in a cascade. Cytokines may cause a decrease in blood pressure, shock, respiratory distress, multiple organ failure and death (Lindsay, 1996). Most T-cells activated by superantigens are useless in fighting infections, and even worse they could unleash an autoimmune attack, driving the immune system into a self destructive frenzy, hurting the individual instead of protecting. Superantigens also have the ability to trigger the cell death of cells they excite, thus weakening the body's defense system (Johnson et al., 1992). Studies on superantigens have predominantly focussed on the effects of staphylococcal enterotoxins in animals as a model of human toxicosis (Cerami, 1992; Tracey and Cerami, 1993; Fleischer, 1994). Still numerous important questions such as: the functional role of bacterial superantigens, and how superantigens with different structures can interact with major histocompatibility complex (MHC) and T-cell receptors remain unanswered. Recent data indicates that CPE has superantigenic properties; however, it has not been determined whether the toxin's enterotoxigenic, cytotoxigenic and

PAGE 12

3 parasympathomimetic properties are linked to superantigenicity (Bowness et al., 1992; Lindsay, 1996). Currently there are several areas under smdy relative to CPE: i. the mechanics of pathogenicity; ii. the structure of the CPE genes and the signals controlling regulation and expression; iii. the mechanics of CPE action at the cellular level and interactions with host molecules; iv. and the mechanisms/role of superantigenicity. Various laboratories are currently working on: i. cloning the enterotoxin genes to determine structure and function, and mechanisms for regulation; ii. CPE receptor binding and gross mechanisms of action; iii. and CPE's mechanisms of superantigenicity. The overall aim of this dissertation was to investigate the pathophysiological responses after intraperitoneal and intragastric administration of CPE using the murine model. This evaluation may explain how CPE becomes distributed during toxicosis, which organs were specifically affected, and describe the numerous sequela and pathophysiological changes that may lead to death. Within the overall aim was the development of a murine model that would explain the mode of CPE distribution after intragastric administration, thus being able to draw a comparison with the SIDS model proposed by Lindsay et al. (1994), and Lindsay (1996), and possibly determining the events involved and leading to infant death.

PAGE 13

4 The specific objectives to this study were as follows: 1. To identify the murine organs and tissues that bind CPE after in vivo administration of CPE. 2. To determine and compare the number of CPE cell receptors in each organ. 3. After intraperitoneal CPE adminisu-ation, to investigate the whole-body distribution of CPE. 4. Describe the animals' symptoms during toxicosis after intraperitoneal and intragastric CPE administration, physical and pathophysiological changes, and findings at necropsy. 5. After intragastric CPE administration investigate the whole-body distribution of CPE 6. Propose a murine model for the whole-body distribution of CPE after intragastric administration. The results will hopefully elucidate the CPE distribution mode, delineating the affected organs and pathophysiological changes possibly responsible for the ilhiess and death. Ultimately the murine model will provide a platform for analogy in the development of human FBI and SIDS.

PAGE 14

CHAPTER 2 REVIEW OF THE LITERATURE Clostridium perfringens Clostridium petfringens {Clostridium welchii) has been described as the most important anaerobic pathogen of man (Lindsay, 1996) and is considered the most common cause of enteric diseases m animals (Hobbs et al., 1953; Bartoszcze et al.,1990; Songer, 1996). Virulence of the bacterium is related to the production of at least 15 different protein toxins, many of which are lethal. Diseases associated with C. perfringens infections and production of these protein toxins include myonecrosis (gas gangrene), necrotic enteritis, antibiotic associated diarrhea, sudden infant death syndrome and food poisoning in man (MacLennan, 1962; Smith, 1979; Fekety et al., 1980; McDonel, 1980; Rood and Cole, 1991; Lindsay, 1996) and lamb dysentery, ovine, bovine and equine enterotoxemia, and pulpy kidney disease of sheep and other animals (McDonel, 1980; Niilo, 1980; Sterne, 1981; Songer, 1996). Classification C. perfringens strains are initially classified into a series of different types (AE) based upon thenproduction of one or more of the major toxins, alpha, beta, epsilon and iota. Toxin production is verified by neutralization with type specific antisera using mice; however, the process is very tedious, expensive and reUes upon 5

PAGE 15

6 the unnecessary use of animals. PCR-typing techniques are becoming more available and have proven to be reliable under some circumstances. However, considerable discussion has ensued as to whether using a classification method based on four toxins is correct since it appears that not all strains contain the alpha-toxin (phospholipase C) a current defining characteristic of C. perfringens (Lindsay, 1996). C. perfringens is the most widely distributed pathogenic bacteriimi. The organism is a Gram-positive, rod-shaped, variably sized (0.6-2.4 x 1.3-1.9 /xm) encapsulated, nonmotile, spore former, occurring singly or in pairs. Vegetative cells are mostly square-ended rods but some strains have roimded ends. The bacteriimi usually grows very quickly and can have a generation time of 7 minutes in an optimal meat-containing environment (Labbe, 1989). Although C. perfringens can be isolated relatively easily, colony appearance on soUd mediimi varies with organism type. Isolation requires the differential use of various antibiotics, the presence or absence of iron and sulfite, and incubation temperature. C. perfringens are resistant to many antibiotics which inhibit other anaerobes or facultative anaerobes. Sulfites are reduced to sulfides which in turn react with iron, forming a precipitate that renders C. perfringens colonies black. Selective media commonly used for isolation and enumeration are sulfite polymyxin sulfadiazine (SPS) agar; tryptone sulfite neomycin (TSN) agar; Shahidi Ferguson perfringens (SFP) agar; D-cycloserine blood agar; oleandomycin polymyxin sulfadiazine perfringens (OPS?) agar; tryptose sulfite cycloserine (TSC) agar; and egg yolk free tryptose sulfite cycloserine (EY-free TSC)

PAGE 16

7 agar. Although colony growth is good at 37C, incubation at 46C especially on TSN is highly selective. If the organism is present in foods or feces in the spore state, samples are usually heat shocked at 75C for 15 minutes before plating. Lindsay (1996) suggested that since spore heat resistance is correlated with CPE synthesis, that spores should be heat shocked at both low (75C) and high (lOOT) temperamres to ensure complete activation of all sub-populations within a sample. C. perfringens is not a strict anaerobe since growth occurs between + 125 and -125 millivolts (Eh). Thus the organism is described as aerotolerant. Vegetative cells are sensitive to high Eh during lagand early log-phase, but oxygen extends the lag phase and growth can be stimulated by lowering the Eh. Oxygen peroxides, however, reduce colony growth, and cell/spore counts. The optimimi growth temperature varies from 43-47C based on organism type, and the T^„ is usually 20C and the T^ 50C. Thus the vegetative cells are heat tolerant. Some strains are known to grow slowly at 15C but these are the exception. Refrigeration at < 5C and freezing can decrease the number of vegetative cells and spores. Vegetative cells are very sensitive to acid environments during log phase, but during stationary phase cells are resistant. The optimimi pH range for growth and toxin production differ. Optimum growth occurs at pH 6.0-7.5, while pH under 5 or above 8.3 are extremely inhibitory. The optimum pH for toxin production is 7.0 for alpha toxin, 7.5 for beta, and 7.2 for epsilon and theta (Hobbs, 1979; Labbe, 1989). Optimal toxin production occurs between 30-46C. The water activity (a^) range for growth is 0.93-0.97

PAGE 17

8 depending on the solute controlling the of die substrate. During sporulation, the a^ is a more significant growth limiting factor than for vegetative growth (Hobbs, 1979; Labbe, 1989). Many strains tolerate curing agents and smoking when a suitable growth temperature and pH are maintained. Complete growth inhibition occurs at 8% NaCl, Ig/kg NaNo3, 400mg/kg NaNO^, and with a combination of 5.3% NaCl widi 25mg/kg nitrite (Hobbs, 1979; Labbe, 1989). In namre, C. perfringens is usually found as a spore which becomes metabolically active only when it encounters a suitable substrate. Thus the bacterium is regarded as a r-strategist (Lindsay, 1996). Spores produced in sporulation media are subterminal and oval in shape. Sporulation is strain, temperature and medium dependent. A range of 32-40C is appropriate for most strains (Lindsay, 1996) and maximum spore production is reached in 6-8 hrs. Many C. perfringens strains have different nutrient requirements, thus the choice of sporulation mediimi is critical. Duncan and Strong (DS) medium (Duncan and Strong, 1968) with some minor adjustments is used by most laboratories. The addition of either starch, raffinose, amylopectin, amylose. glucose, maltose and methylxanthines to DS is known to increase spore production, however, complete sporulation is never observed (Labbe, 1989; Lindsay, 1996). Five to ten percent sporulation is considered usual, and 50% exceptional. Some strains are known to have almost no spore formation ( < 0.001 %) even under die optimal conditions. C. perfringens spores are relatively heat resistant Dioo 15 min, however, strain variation is known. Non-hemolytic strains have a

PAGE 18

9 decimal reduction time (Dioo''C) of 6-17 min, whereas hemolytic strains have a Dioo^C of 0.1-0.5 min. Heat activation is also strain dependent. Some strains only require 60C for 5 min, although 10-20 min at 75-80C is usual (Labbe. 1989). Some type A strains are highly heterogeneous with a sub-population activated at 75C, and another at 100C (Lindsay, 1996). Spore radiation resistance varies from 1.2 to 3.4 kGy, and radiation resistance parallels heat resistance (Labbe, 1989). C. perfringens spores are also highly resistant to curing agent at concentrations of 21.5% NaCl, 1.8 g/1 NaNOj and 1.2 g/1 Na02 (Labbe, 1989). PhenoHc antioxidants are inhibitory or bacteriocidal depending on the compound and concentration used (Labbe, 1989). C. perfringens Foodbome Illness History Although an outbreak of C. perfringens foodbome illness (FBI) was first described by Klem in 1895, it was not until 1943 when Knox and McDonald in the United Kingdom and McClung (1945) in the United States made the association between the organism and the FBI. Persons afflicted with the iUness expressed symptoms of severe abdominal pain, mild chronic to explosive diarrhea accompanied by nausea. From 1947-9 there were many reported outbreaks of FBI in Germany with a large number of fataUties. Unlike the iUnesses described in the UK and USA a few years earher, the German patients suffered from severe gross hemorrhagic enteritis (enteritis necroticans: Darmbrand) where the bowel was completely

PAGE 19

10 desquamated. Thus it was obvious that C. perfringens was responsible for two completely differem types of FBI. The dilemma was finally resolved by Hobbs et al. in 1953 who unequivocally showed that type A strains were responsible for the mild form of FBI, and type C strains for the more dramatic-lethal form. Subsequent smdies by Duncan and Strong (1969), Hobbs (1979), Niilo (1975), and Tsai and Riemann (1975a; 1975b) showed that the type A FBI was caused by an enterotoxin, now termed CPE, produced during cell sporulation. In 1967 a milder form of necrotic enteritis was observed in New Gumea. This type of outbreak was subsequently found to be common, and coincided with traditional pig feasting. A heat sensitive strain of C. perfringens type C was found to be the causative agent of the disease, now referred to as pigbel (Murrell and Walker, 1991). Studies during the last 20 years have shown that Darmbrand and pigbel result from the consumption of type C vegetative cells which proliferate in the gut producmg beta-toxin. The diet of persons afflicted with the diseases is usually lacking in proteases needed for enzymatic digestion of the beta-toxin, or as in New Guinea, protease inhibitors are present in the bowel due to the consumption of sweet potatoes, the staple diet of the natives. Darmbrand and pigbel are now relatively rare due to immunization of the susceptible populations with beta-toxoid. However, in some inhospitable regions of New Guinea where cannibalism still occurs the disease is common and death usually occurs after infection (Murrell, 1989).

PAGE 20

11 Type A Foodbome Illness: General Characteristics While type C necrotic enteritis is rare, type A FBI is very common and these strains are associated with about 10% (third most common in die USA) of all bacterial FBI (Bean and Griffin, 1990). An infection may be caused by the ingestion of foods contaminated with > 10*10'' vegetative cells/gram (Hobbs, 1979). This contamination usually results from temperature abuse of a prepared food. Vegetative cells that survive the stomach acids, enter the small intestine, multiply and sponilate. During sporulation some cells synthesize CPE, and upon cell lysis die enterotoxin is released in die intestinal liunen where it attaches to villous enterocytes to act cytotoxically and histopathologically, or is also absorbed systemically where it may additionally be parasympatiiomimetic, cardiotropic or superantigenic (Lindsay, 1996). lUness usually develops 8-24 hours after ingestion, and is resolved 12-24 hours after onset (McDonel, 1980; Labbe, 1989). The classical symptoms are severe abdominal cramps widi mild to explosive diarrhea, accompanied by nausea. Vomiting is rare, aldiough fever may occur. Deadi rates are usually low in immunocompetent individuals, however, m the young, elderly or immunocompromised persons the mortahty rate is 3-4% (Janssen et al., 1996).

PAGE 21

12 Type A FBI Outbreaks From 1973-87 the Center for Disease Control (CDC) reported 1994 outbreaks of C. perfringens type A FBI in the USA. These outbreaks involved 12,234 cases and 12 deaths. The actual number of FBI cases/year is estimated to be around 650,000 with 7-10 fatalities /year (Bean and Griffin, 1990). C. perfringens outbreaks are often large, with a mean of 25 cases/outbreak (Todd, 1989; Bean and Griffin, 1990; CDC, 1994). There have been several serious outbreaks during the last two years traced to St. Patrick's Day meals, the first, in Cleveland, Ohio, involving 156 persons. Corned beef boiled for three hours was cooled slowly at room temperature before reftigeration. Four days later the meat was reheated at 48.8C and consumed in sandwiches several hours later after preparation. The second occurred in Virginia where 86 people became ill. Commercially prepared, frozen-brined corned beef was cooked in large pieces, refrigerated and reheated with a heat lamp for 90 minutes before consumption (CDC, 1994). The third outbreak in a British hospital, involved 17 patients. Cooked vacuum-packed pork was cooled slowly at a commercial meat preparation facility, transported to the hospital, slowly reheated then consumed. Temperamre abuse during cooking (inadequate cooking) and improper cooling or holding temperamres account for 97% of type A FBI outbreaks. Since the bacterium is heat tolerant and its spores have an high resistance, mcomplete cooking of contaminated foods may not kill all vegetative cells and will Ukely promote germination of spores and rapid outgrowth. Other factors often associated with type

PAGE 22

13 A FBI include contaminated equipment, and improper personal hygiene. Prevention and control of type A FBI can be accomplished by thoroughly cooking foods and ensuring that high internal temperamres necessary to destroy the bacterium's spores are achieved. Cooked foods should be cooled quickly and either stored at refrigeration temperatures or consumed immediately. Identification of C. perfrinsens FBI Outbreaks To successftiUy identify C. perfringens type A FBI outbreaks several bacteriological criteria have to be fulfilled. Public healtii agencies may identify an outbreak by 1) finding the contaminated food and determining that it contains > 10^ C. perfringens vegetative cells/gram, 2) finding that patients have > 10^ C. perfringens spores/gram feces, 3) finding that patients express the same serotype of C. perfringens or, the same serotype should be found in both contaminated food and in patient feces (Stringer, 1985; McClane, 1992). A problem with these criteria is that many elderly individuals naturally have high C. perfringens spore counts in their stool. Further, many isolates from the U.S. and Japan are unusual and cannot be serotyped (Saito, 1990; McClane, 1992). Thus, it is now recommended that the diagnostic criteria for C. perfringens type A food poisoning outbreaks should include 1) the presence of CPE in feces FBI patients, and 2) isolation of C. perfringens type A strains carrying the cpe gene from feces or foods associated witii the FBI. Detection of CPE in feces is a partial indicator of C. perfringens type A FBI and

PAGE 23

14 several serologic assays (ELIS A/reverse-passive latex agglutination: RPLA) are commercially available. A problem with these assays is that fecal samples have to be collected and examined immediately (Bartholomew et al., 1985; Stringer, 1985; Birkhead et al., 1988; Labbe, 1989; McClane, 1992). There is also the problem that a contaminating molecule produced by all C. perfringens strains (CPE positive and negative strains) cross reacts widi antisera to CPE. Thus CPE detection should be corroborated with cpe gene probe assays, which are relatively easy to perform (Labbe, 1989; Kokai-Kun and McClane, 1996; Lindsay, 1996). However, there is an additional concern. In a recent survey of C. perfringens associated FBI cases where all the microbiological and serological criteria were met, cpe gene probe analysis revealed only 59% of associated strains were cpe-positive. Lindsay (1996) recently argued that results using even the latest molecular biology identification methods for C. perfringens should be treated with caution. The reasons for this caution are as follows. AU enterotoxigenic C. perfringens type A carry the cpe gene, however, CPE positive isolates account for < 5% of all C. perfringens isolates found globally. Thus strains capable of causing FBI present a very small sub-group within the ubiquitous and regularly encountered C. perfringens (Van Damme-Jongsten et al., 1989; Saito, 1990; Kokai-Kun et al., 1994). CPE-A positive strains have been isolated from feces collected from 6% of healthy food handlers, suggesting that humans may serve as reservoirs for C. perfringens type A strains (Hobbs, 1979; Saito, 1990). The issue of whether animals are indeed potential

PAGE 24

15 reservoirs of enterotoxigenic C. perfringens remains questionable. In human FBI causing strains cpe is chromosomally encoded, whereas in veterinary isolates cpe is plasmid-bome (Comillot et al., 1995). Only one plasmid encoded strain has been isolated from himian C. perfringens FBI cases (Katayama et al., 1996). C. perfringens and Pathogenicity Virulence Factors Contributing to C. perfringens FBI The association between cpe and C. perfringens FBI, although not formally proven by Koch's molecular hypothesis, is accepted by researchers. This acceptance is based on strong evidence from a number of sources, such as epidemiological stodies where patients have detectable enterotoxm levels in their stools (80-100% of patients fecal samples tested CPE positive) (Bartholomew et al., 1985; Birkhead et al., 1988). CPE ingestion is known to cause illness with the same symptoms in experimental animals (Hobbs et al., 1953; McDonel and Duncan, 1975; Bartholomew etal., 1985; Birkhead etal., 1988). When either purified CPE or C. perfringens type A strains were fed to human volunteers they developed FBI symptoms, but patients fed CPE negative strains showed no iUness (Skjelkvale and Uemura, 1977). It has also been shown that the effects of CPE in animal models can be neutrahzed with CPE specific antiserum (Hauschild et al., 1971).

PAGE 25

16 Molecular Biology of CPE The complete cpe gene has been cloned and sequenced allowing studies on the regulation and expression of the gene with regards to sporulation, and for the construction of various nucleic acid probes (Van Danune-Jongsten et al., 1989; CzeczuUn et al., 1993; Comillot et al., 1995; Lindsay, 1996). hi type-A food poisoning isolates the cpe gene is present as a single chromosomal copy in the hypervariable C region of the chromosome. This hypervariable region is thought to contain mobile genetic elements (transposon or lysogenized phage) that allow the transfer of cpe to other C. perfringens strains (Canard et al., 1992). The plasmid encoded veterinary isolates are thought to represent an example of this type of mobility. The cpe gene can also be transferred and expressed in non-enterotoxigenic type A, B and C strains. A comparison of open reading frame (ORF) sequence data strongly suggests that CPE produced from different type A strains, or different variants of the same strain are sumlar, but not identical (CzeczuUn et al., 1993, 1996). Aldiough microsequence variation occurs, the cpe-ORF generally appears to be highly conserved. Enterotoxigenic veterinary and food poisoning strains have sUghtly different regulatory factors preceding the cpe ORF (Brynestad et al., 1994). A 45-base pair insertion about 265 nucleotides upstream from the start of the cpe gene has been detected in three strains, implying at least two types of cpe-promoter regions in foodbome isolates (Melville et al., 1994). The insertion does not appear to alter the starting point of cpe transcription in C. perfringens, although transcription was

PAGE 26

17 significantly altered when the cpe gene was cloned into B. subtilis. The reason for this promoter diversity remains undetermined, although geographical segregation appears not to be a factor. The cpe gene may be associated with at least two repetitive sequences, one of which is a known insertion sequence (ISI151) (Daube et al., 1993; Brynestad et al., 1994; ComiUot et al., 1995). In human FBI strains, when cpe was chromosomally located, several factors were observed: first, the gene was preceded by a repeated sequence, the HindUl repeat and 0RF3, which is homologous to a gene present on an IS element, and second, the gene was always present on a 5 kb Nrul fragment. In contrast, strains isolated from domestic hvestock where cpe was plasmid encoded, the gene was close to 157757, not hnked to the Hindm repeat, generally preceded by 0RF3, but never encoded on a A^rw-fragment (ComiUot et al., 1995). Sporulation and CPE The molecular basis for regulation of CPE expression is not completely understood. Several key factors are now accepted: a) CPE expression is sporulation associated, b) CPE is not a structural component of the spore coat and, c) CPE is not a post-translationally processed product of a 52 kDa precursor molecule. The association between sporulation and CPE synthesis was first made by Duncan et al. (1972) who showed that CPE synthesis could be blocked by early stage sporulation mutants spoO but not but not by later stage spoV mutants. Subsequent work

PAGE 27

18 confirmed these observations, indicating that although some CPE was synthesized during vegetative growth due to leaky gene regulation, CPE synthesis was indeed sporulation controlled, possibly by a global regulator such as a sigma factor (Duncan et al., 1972; McDonel, 1986; Czeczuhn et al., 1993; Kokai-Kun et al., 1994; Lindsay, 1996). CPE expression is an exclusive trait of sporulating ceUs since although the cpe gene can be transferred to E. coli, transformed cells do not express the toxin. Enterotoxin production starts soon after sporulation is committed. Synthesis peaks 6-8 hrs into sporulation (Smith and McDonel, 1980; McDonel, 1986) and CPE comprises up to 10-20% of the total cellular protein content (Czeczulin et al., 1993). The cpe mRNA may be transcribed as a monocistronic message, and its transcription starts approximately 200 base pairs (bp) upstream of the CPE translation start site. This promoter region does not show any significant sequence homology with other known bacterial promoters (Melville et al., 1994; Czeczuhn et al., 1996). The cpe mRNA has an exceptionally long half-Ufe of 58 minutes in sporulating cells, and this could in part be the reason for high CPE expression (Labbe and Duncan, 1977). Increased mJlNA stability may result from a stem-loop strucmre localized 36 bp downstream of the 3' end of the cpe ORE. This region may also fimction as a rho-independent transcriptional terminator to regulate expression (Czeczuhn et al., 1993). A similar situation occurs in the seb gene, where a palindromic sequence occurs 40 nucleotides downstream from the TGA stop codon. The hairpin-structure

PAGE 28

19 is followed by an thymine-rich region that may function as a rho-independent transcriptional terminator (Gaskill and Khan, 1986; Jones and Khan, 1986). CPE does not fit the classical definition of an exotoxin because the molecule does not contain a leader sequence, nor is it transported through the cell membrane. CPE production takes place in the mother cell cytoplasm, where it accimiulates. When CPE synthesis is excessive, paracrystalline inclusion bodies are often formed or toxin is trapped between spore coat layers. Toxin is only released as a consequence of mother cell lysis at stage VII of sporulation (McDonel, 1980; Labbe, 1989; Czeczuhn et al., 1993; Lindsay, 1996). Biochemistry of CPE CPE is a single polypeptide composed of 319 amino acids. The molecule has an isoelectric point of 4.3 and a of 35,3 17 Daltons (McClane, 1992). The protein is heat labile and is inactivated by heating for 5 min at 60C (McDonel, 1986). The toxin is denatured by pH extremes, but some proteolytic enzymes do not affect its stability (McDonel. 1986). Indeed trypsinization activates CPE threefold w v/fr
PAGE 29

20 There are some data indicating that at least two different CPE molecules are synthesized, the well-characterized CPE also known as CPE-A and the less common CPE-86. The CPE-86 toxin is derived from a C. perfringens coatless spore mutant. Some data indicated a high degree of nucleic acid sequence and N-terminal protein homology between CPE-A and CPE-86 (Wojciechowski, 1995). However, other data from amino acid analysis and matrix-assisted laser desorption (MALD) ionization mass spectroscopy comparisons suggested distinct strucmral differences (Lindsay et al., 1985; Wojciechowski, 1995; Lindsay, 1996). These conflicting results led Wojciechowski (1995) to suggest that although the nucleic acid sequences were the same, any differences in protein strucmre either resulted from post-translational modification, or conformational differences as a result of post-translational modification. This hypothesis remains to be proven (Lindsay, 1996). CPE-A and CPE-86 appear to have similar mechanisms of action with regard to in vitro cell proliferation (McClane and McDonel, 1979; Lindsay, 1988) and cytokine modulation in several different human and animal cell lines (Mach and Lindsay, 1994, 1997). Although CPE-86 is biologically more active in vivo and in vitro (McClane and McDonel, 1979; Lindsay et al., 1985; Lindsay, 1988), both toxins are equally mitogenic in the J774M<^ cell Une (Mach and Lindsay, 1997). The stnicmre-function relationship for CPE is complicated m part by the inconsistencies in strucmral data. Examination of CPE-A secondary strucmre by UV circular dichroism predicted totaUy different values to those derived from the amino

PAGE 30

21 acid composition (Granum and Harbitz, 1985; Lindsay et al., 1985). Additionally, attempts to predict CPE-A secondary strucmre using 9 different strucmrai models showed conflicting putative conformations that were consistent in only two areas (Granimi and Stewart, 1992). Possibly the high percentage of hydrophobic amino acids in both protein-toxins causes equivocal results by different methods. The 3D strucmre of CPE-A examined by differential spectroscopy, and by titration of accessible amino groups suggests a 2-domaui strucmre, where the N-terminal portion of the protein is cytotoxic and the C-terminal portion is the binding portion (Granum and Whitaker, 1980; Whitaker and Granum, 1980; Granum and Stewart, 1992). This concept is strongly supported by MAb and PAGE smdies (Lindsay et al., 1985; Hanna et al., 1992). X-ray analysis of either protein-toxin could possibly provide more definitive information. Unfortunately, although crystals of CPE-A have been produced, they are too small in both quantity and size for analysis (Granum and Stewart, 1992). Strucmre-Fimction and Vaccines Through the use of recombinant peptide fragments, and the generation of monoclonal antibodies, the regions of activity within the CPE protein have been partly mapped (Horiguchi et al., 1987; Hanna et al., 1989, 1991, 1992; Granum and Richardson, 1991; Hanna and McClane, 1991). Although CPE is a single unit polypeptide, similar to most bacterial toxins, CPE presents two distinct regions, a

PAGE 31

22 hydrophobic toxic fragment and a hydrophiUc binding fragment. As previously noted proteolytic digestion of the N-terminus with either trypsin or chymotrypsin activates CPE, with a concomitant increase in CPE cytotoxicity (Granum et al. 1980). Amino acids 37 through 171 in the first half of die protein contain sequences for insertion and cytotoxicity (Granum et al., 1981; Granimi and Richardson, 1991). Amino acids 290-319 of the C -terminus contains the receptor binding region, and the receptor moiety per se is neither capable of insertion nor is it cytotoxic (Horiguchi et al., 1987; Hanna et al., 1989, 1992; Hanna and McClane, 1991). Other regions of the protein remain imdefined and may be necessary for large complex formation. There has been considerable discussion as to whether a vaccine for CPE can be produced (Hanna et al., 1989, 1992; Mietzer et al., 1992). Purified CPE may present as many as five epitopes scattered throughout different regions of the enterotoxin. The monoclonal antibody 3C9 neutralizes CPE cytotoxicity by blocking receptor binding (Wnek et al., 1985). Since CPE C-terminal fragments are not cytotoxic and have neutralizing epitopes, they have potential for CPE vaccine construction. Immunity to C. perfringens type A FBI may also require a vaccine that would stimulate the production of high titers of secretory IgA anti-CPE in the intestinal Imnen. For a nimiber of reasons, however, the production of a CPE vaccine may simply not be possible. In both human and animal smdies, administration of CPE only induces transient immunity. Circulating antibodies to CPE are only present for a few weeks and there is no long term immunity (Bouvier-Colle et al., 1989; Hoffman

PAGE 32

23 et al., 1987). Indeed there is some suggestion that die first toxicosis induces sensitivity to subsequent intoxication. This is similar in manner to the toxicosis presented after administration of Staphylococcus aureus enterotoxins (SEs). In one sense a lack of long-term immunity is not surprising since both groups of toxins are known to be superantigens. That CPE possess superantigenic activity is perhaps the most important finding about the nature of the molecule, since it may explain some of the unusual aspects of the toxin's activities, and the relationship CPE has widi other non-foodbome diseases such as SIDS. It is very difficult, albeit impossible to generate long term immunity to a superantigen. Normal antigens activate 1:10,000 T cells while superantigens may generate 1:5. While this over stimulation of the immune system generates the release of high concentrations of cytokines, it does not allow for the generation of long term immunity since the immune response is so rapid. CPE-A over-stimulates a selected group of T cell receptors, namely Vfi 6.9 and VB 22 (Bowness et al., 1992), and out-competes SEA for binding sites. Silent CPE Genes Although microsequence variation of various cpe genes has been observed, the cpe ORE is generally highly conserved. The CPE sequence has Uttle homology with other proteins, except for slight homology with a C. botulinum complexing protein, and a 5 amino acid sequence from the Vibrio cholera B subunit (Czeczuhn et al., 1993, 1996; Hauser et al., 1994; Lindsay, 1996). Recent smdies by Lindsay (1996)

PAGE 33

24 indicates that a silent CPE gene is encoded within the intervening sequence of the Iota toxin gene from C. perfringens type E. There is 90.4% homology between the cpe gene sequence and that found associated with the iota toxin (Ula) (Lindsay, 1996). The potentially derived amino acid sequence from Ula shows 81.7% identity with CPE. Minor modifications (only single base changes) could increase the sequence identities to 96.8%. Lindsay (1996) calculated that die evolutionary relationship between the two sequences was 10 PAM (accepted point mutations per 100 residues) for the nucleic acid sequence (9.6 changes/ 100 nucleotides) and a PAM of 21 for the amino acid sequence (18.3 changes/100 amino acids). This argued fliat the evolutionary relationship was too close to sunply be formitous. Comparison of the sequences upstream of the cpe and pr-cpe (lUa) genes showed similar regulatory Hpr sequences. This strengthened the evolutionary relationship between pr-cpe and cpe, and suggested that cpe is on a mobile genetic element which extends at least 500 nucleotides upstream of the cpe ORE (Lindsay, 1996). CPE hitragastric Mechanisms of Action The CPE mode of action in vitro and in vivo has few sunilarities with other bacterial toxins, thus it presents an unique toxicosis (McClane, 1994, 1997; Lindsay, 1996). Smce CPE causes fluid and electrolyte losses in the GI tract of humans and animals, it has been classified as an enterotoxin (McDonel, 1986). Based on in vivo animal model smdies, CPE appears to alter the fluid and electrolyte balance

PAGE 34

25 compromising the villus integrity, thus breaking down the normal GI absorption and secretion mechanism, which is pathologically manifested as diarrhea (Sherman et al., 1994; Lindsay, 1996). Animal smdies also show that CPE targets the small mtestine widi high affinity causing gross desquamation to the intestinal villi. Histopatiiological damage can occur very rapidly, thus CPE is also considered cytotoxic (McDonel and Duncan, 1975; McDonel, 1986; Sherman etal., 1994). Characteristics of CPE Binding. Complex Formation and Insertion The current model based on the extensive smdies from McClane's lab is that CPEs action is a multi-step process with four events. Briefly, CPE first binds to a 50 kDa membrane protein creating a 90 kDa small complex. Second, the stnictoral change of CPE contained in the small complex, which results from either CPE or small complex insertion into the cellular membrane, or from small complex conformational changes. Third, the formation of a 160 kDa large complex resulting from the small complex binding to a 70 kDa membrane proteiu (Wnek and McClane, 1989; McClane, 1997). The fourth and final step is the loss of plasma membrane permeability properties, caused by either the large complex directly serving as a pore or simply by directly affecting membrane permeabiHty (Wieckowski et al., 1994, McClane, 1994, 1997; Czeczulin et al. 1996). The CPE receptor(s) is found in several different cell types of mammaliaTi species and due to it's broad distribution the receptor is thought to have an important

PAGE 35

26 physiological role, but not be essential for cell viability (McDonel and McClane, 1979; McDonel, 1980; Wnek and McClane, 1986; McClane et al., 1988b; Sugii and Horiguchi. 1988; McClane, 1994). Receptors is in plural since there is some disagreement as to whether more than one receptor exists. Kinetic studies from different labs found contradictory results (McDonel, 1986; McClane, 1994). Affinity chromatography smdies of brush border membranes (BBM) and Vero cells strongly suggest two proteinaceous CPE receptors of 50 kDa and 70 kDa (Wnek and McClane, 1983; Wnek et al., 1985; Sugii and Horiguchi, 1988). Additional support for the 50 kDa protein came from inmiunoprecipitation analysis of a 90 kDa CPE-mammalian protein complex (small complex) that consisted stoichiometrically of one 50 kDa membrane protein and one CPE molecule (Wieckowski et al., 1994). The 160 kDa protein was determined from molecular weight analysis of the large protein complex indicating the complex was stoichiometrically composed of one 50 KDa protein, one CPE molecule and a 70 kDa protein. CPE bmding is receptor specific and saturable (10* CPE receptors per cell). Binding occurs rapidly (15-30 min) and in one tissue culmre system binding appeared to be temperamre sensitive (CPE binding at 4C is lower dian at 37C) (McDonel and Duncan, 1975; McClane et al., 1988b; McClane, 1994; Wieckowski etal., 1994). The action mode of many membrane-active toxins, including CPE, often involves plasma membrane-permeability imbalances through toxin insertion into the Upid bilayer (McDonel, 1980; McClane, 1994). The McClane model suggests that

PAGE 36

27 after CPE binds to receptors it remains inserted in the lipid bilayer, toxin is neither dissociated nor intemaUzed (McClane, 1994). Binding to receptor(s) is a two-step process during which CPE acquires amphiphilic capabilities (Wieckowski et al., 1994). As previously noted, binding is irreversible, protease resistant and possibly temperature dependent (McClane et al., 1988b; McClane, 1994). CPE is inserted into the membrane after binding and is entrapped witiiin the Upid bilayer. Although the CPE-small complex imdergoes cortformational changes, it is unknown whether the entire CPE-small complex or CPE alone remains entrapped (Czeczulin et al., 1996). UnUke small complex formation, CPE-large complex formation is temperature dependent. Large complex is formed above 24C but not at 4C, thus at 4C CPE cytotoxicity is blocked. Cytotoxicity can be unblocked by transfer to higher temperamres. Although the 70 kDa protein is part of the large complex, it is imknown whether the molecule is a receptor. Possibly it is simply a fimctional membrane protein brought into close contact with the CPE-small complex through stearic attraction, CPE insertion into Upid bilayer, or conformational charge in the small complex (Kokai-Kun and McClane, 1996). Large complex formation is thought to either directly affect plasma membrane permeability (influx and efflux) by ftmctioning as a membrane "pore" (ion-permeable channel) or by interfering with membrane pump regulation via continuous pump activation (Sugimoto et al., 1988; Czeczuhn et al., 1996; Kokai-Kun and McClane, 1996). The permeability changes may develop within 5 minutes, and restrict passage

PAGE 37

28 to ions and small amino acids (less than 200 Daltons) that transit through membrane lesions of about 0.5 nm' in size (McDonel and McClane, 1979; McDonel, 1986; McClane, 1994; Czeczulin et al, 1996; Kokai-Kun and McClane, 1996). Since eukaryotic cells have a lower intracellular ion concentration than the external mediimi in order to maintain a normal colloid-osmotic equilibrium (McClane, 1994), permeability changes allow for a rapid ion influx into the cell cytoplasm. This initial influx is calcium ion dependent, requirmg elevated intracellular levels of Ca^^ ions. High levels of calciimi ions may lead to collapse of the cytoskeleton (McClane, 1994). A rapid influx of small molecules causes the plasma membrane to "stretch" facilitating the additional influx of macromolecules > 5 kDa. This leads to gross osmotic changes and cell destabilization, often seen as membrane blebs. Loss of essential amino acids causes secondary effects of inhibition of DNA, RNA and protein synthesis, and thus the cells become nonviable (McClane and McDonel, 1979; McDonel, 1986; McClane etal., 1988a; Hulkower etal., 1989). CPE and Medicine Although the above described CPE-induced changes were elucidated from in vitro cell studies, it is beheved that the same effects occur in vivo during CPE induced FBI. CPE affects intestinal epithelial cells in a sunilar fashion. Villous epithelial cells present small molecule permeability changes, disruption of normal villus integrity and ftmction, morphologic damage, ceU lysis and net secretion of fluids and

PAGE 38

29 electrolytes (McClane, 1994). The ability of CPEs to kill a wide variety of cells both in vivo and in vitro has also led to the evaluation of CPE-86 as anti-neoplastic agent (Lindsay, 1996). PreUminary studies showed that CPE-86 was destructive to Lewis lung carcinoma cells in vivo and neoplastic cell hues (P388, B16-F1 and LI 10). The mechanisms of neoplastic cell death were the same as in Vero cells with membrane permeabihty changes, bleb formation, inhibition of nucleic acid and protein synthesis, and subsequent cell death. Bacterial toxins have a great potential as anti-cancer drugs (Pastan et al., 1995; Lindsay, 1996), and several plant and bacterial toxins (ricin, Pseudomonas exotoxin A) have previously been suggested as therapeutic agents for cancer treatment via Unkage to cell specific antibodies (magic bullets). Although the data with CPE-86 is preUminary it suggests that the C. perfringens CPEs have a potential use in medicine. Activation of CPE CPE causes dose dependent death in mice where mice die quietly with symptoms of respiratory interference and shock, while in vitro, CPE induces dose dependent morphological damage, inhibition of nucleic acid synthesis, modulation of membrane transport, lysis and cell death in Vero cells (McClane and McDonel, 1979; Mach and Lindsay, 1994; Lmdsay, 1996). Human fetal ileal ceUs (FT) (ATCC CCL241) are resistant to the action of CPE, even at toxin levels of 1 /xg/well. The mechanism of this resistance is unknown, and it can be considered very strange that

PAGE 39

30 ceils from the actual tissue affected in vivo by CPE sliould not be affected in vitro. A possible explanation is that since the FT cells are derived from fetal tissue they are hkely undifferentiated. In this state they are not susceptible, and require differentiation to become susceptible. Studies by McClane et al., (1987) and Mach and Lindsay (1994) mdicated that the activity of CPE can be dramatically exacerbated both in vivo and in vitro by the presence of interferon-gamma (IFN-7). In vivo, the presence of IFN-7 can decrease the LD50 1 ,000 fold and reduce die time to death 360 fold. In vitro both Vero and FI cells either pretreated widi IFN-7 then CPE, or IFN-7 combined with CPE showed dramatic sensitization with a several log fold increase in CPE activity. Enhance ment of CPE Activity and Human Non-Foodbome Disease Both groups proposed that enhancement of activity likely resulted from IFN-7 sensitizing cells to the action of CPE. In vitro, IFN-y possibly acts by sensitizing cells to the action of CPE resulting in death, possibly by the same mechanisms (inhibition of protein synthesis and destruction of the cell membrane) but at lower toxin concentrations. Mach and Lindsay (1994) and Lindsay et al., (1994) suggested that these observations assisted in the formulation of a toxico-hypothesis for the sudden infant death syndrome (SIDS). The observation that a majority of victims have infections in the two weeks prior to death is known (Morris, 1987; Murrell et al., 1987; Murrell et al., 1992). Viruses, bacteria and bacterial toxins are all known

PAGE 40

31 inducers of IFNs (Collier and Kaplan, 1985: Pestica et al., 1987; Ijzermans and Marquet, 1989; Baron etal., 1991; Chonmaitree and Baron, 1991). Therefore, there may be a link between prior infection increasing IFN levels, sensitization to bacterial toxins and SIDS. Although studies conducted over 15 years ago in relation to viraemia and SIDS showed no evidence of a systemic viral infection or elevated interferon at death (Ray and Hebestreit, 1971 ; Seto and Carver, 1978). It is possible that increased circulating IFN levels might not have been observed, since die immune response to the initial antigen had abated. Alternately, the tissue-cell-methodology for detecting BFNs may not have been very sensitive. Howatson (1992) recently examined the relationship between viral infection and the production of IFN-a. He concluded that the abnormal presence of IFN-a in neurones of the medulla of the brain stem suggested that it was premamre to discount a viral hypothesis for some proportion of SEDS cases. Jakeman et al., (1991) recently showed that the toxicity of several bacterial toxins could be significantly exacerbated in infant ferrets by a previous infection with influenza virus. These authors speculated a role in SIDS by some mechanism(s), including an enhancement of cell permeability which may allow increased or more rapid uptake of a toxin, resulting in death by the same mechanism but at lower concentrations. Mach and Lindsay (1994) argued that by extension, it is hkely that the same situation occurs in human infants. They proposed that a window of vuhierability occurs in the life of some infants due to immunological immamrity, which predisposes them to infection. In the weeks prior to death, these

PAGE 41

32 infants suffer from an infection which induces the synthesis of IFNs, sensitizing the infant to a later albeit more virulent infection which may act as a trigger for sudden death. See the following section for further discussion of the CPE toxico/SIDS hypothesis. Role of CPE in Non-Foodbome Diseases (SIPS) CPE has recently been implicated in a very unusual disease state, the sudden mfant death syndrome (SIDS). SIDS is defined as the sudden death of an infant from one month to one year of age, which remains imexplained after a complete postmortem examination, including an investigation of the death scene and a review of the case history" (Hoffman et al., 1988; Zylke, 1989). In the United States SIDS claims the Uves of 6-7,000 infants (1/1,000 Uve births) (Wilhnger, 1989) and remains the nimiber one cause of post-neonatal infant mortality (Lindsay, 1996). While a large number of theories have been proposed for SIDS, die reason for this large number of infant deaths remains unresolved (Staton, 1980; Valdes-Dapena, 1980; Thach et al., 1988; Verrier and Kirby, 1988; Spika et al.,1989; Wilkinson, 1992). There are many epidemiological indices found m SIDS victims, however, the most striking are the age at death, and that > 85 % of SIDS victims were ill in the two weeks prior to death. It has been suggested that some infants present a "window of vuhierability" where physiological discrepancies or abnormalities make them susceptible (Stephens, 1990; Lindsay etal., 1992, 1993; Murrell etal., 1993; MurreU etal., 1994; Lindsay,

PAGE 42

33 1996). The nature of these abnormaUties are as yet unresolved, however, they may possibly explain why SIDS fatalities occur within such a narrow time frame (WiUinger, 1989). Although premamre infants are more susceptible to SIDS, prematurity is not a determinant, since gestational age is not related to age at death (Buck et al., 1988; Grether and Schulman, 1989). This might suggest that a common factor or factors are responsible for, or triggers the biochemical changes that lead to death. Common pathological indices observed at autopsy include thymic petechiae (pinpoint hemorrhages) and patchy pulmonary edema, muscle fiber necrosis of the diaphragm with a histopathology suggesting hypoxia, astrogUosis of the brainstem, leukomalacia (Jones and Weston, 1976; Valdes-Dapena, 1983; Krous, 1984; Guilian etal., 1987; Beckwith, 1988; Hollander, 1988; Gillanetal., 1989; Guntheroth, 1989; Kariks, 1989; Bruce and Becker, 1992). It has been suggested that these indices might result from or be caused by infections. It should be stressed that the presence of a particular pathogen within an infant is not predictive of a disease state, or SIDS. Most infants have infections during their first year of hfe, yet only a small percentage die from SIDS (Amon, 1983; Bettelheim et al., 1990; BlackweU et al., 1994, 1995). Therefore it has been suggested by several authors that a small sub-population of infants are predisposed to SIDS and when the correct "conditions" occur death is a likely sequelae.

PAGE 43

CHAPTER 3 DETECTION OF CLOSTRIDIUM PERFRINGENS TY?E A ENTEROTOXIN AFTER IN VITRO BINDING TO MURINE TISSUES Introduction In humans, a Clostridium perfringens foodbome illness (FBI) occurs 6-12 hours after ingestion of vegetative cells. Upon entering the host C. perfringens responds to changes in environmental stress by initiating the sporulation cycle with concomitant production of enterotoxin (CPE), and subsequent induction of FBI. In the small intestine, CPE utilizes a unique mechanism of action to directly affect the plasma membrane of cells leading to inhibition of macromolecular synthesis, morphological damage and cell lysis (cytotoxicity), and fluid loss (enterotoxigenicity). Desquamation of villous cells allows CPE to be absorbed and systemically distributed throughout the body, causing various pathophysiological (neurotoxigenic) and immune responses (superantigenic), which may lead to severe illness or death. In order to suidy the pathophysiological and immune responses generated by CPE, the whole body distribution of toxin after absorption must be determined, which was the overall goal of this dissertation. To fully elucidate toxin distribution, administration had to be performed both intraperitoneally, and tfirough the natural port of entry, the digestive tract (intragastrically). The aim of the work described in 34

PAGE 44

35 this chapter was to determine, using in vitro techniques and the murine model, which tissues had the abihty to bind intact CPE molecules. Materials and Methods Murine Model During the last 25 years there has been a large number of animal species used as model systems to study the action of CPE. Dr. Lindsay's lab has placed a particular emphasis on mice for a number of reasons. First, the species use in a wide range of published toxicological smdies, and the relative inexpensive cost of purchasing and maintaining large numbers of animals. Second, large numbers of animals that are genetically similar increases the validity of results. Third, mice have been extensively used by other researchers to examine the effects of bacterial toxins. Fourth, the mode of action, symptoms and pathophysiological changes caused by CPE in the murine model strongly mimic those observed in human cases of CPE toxicosis. Taking all these factors into consideration, we believe that the mouse is an excellent model, and its use can be justified. C. perfringem Enterotoxin (CPE) Purified, freeze dried CPE was gratefully obtained from Dr. Bruce McClane, University of Pittsburgh. For administration, toxin was resuspended in phosphate buffered saline-Tween (PBS-TW) (0.15 M NaCl, 0.01 M Na2HP04, 0.01 M

PAGE 45

36 NaH2P04, and 0.2% Tween 20, pH 7.2) and standardized to a final concentration of 1 fig/^l of protein. Protein determinations were made by the method of Lowry et al. (1951) with bovine serum albumin as the standard. Preparation of Antisera to CPE Three milligrams of purified CPE were supplied as antigens. Individual 4-5 kg New Zealand White rabbits were used to generate antisera. One milliliter of Freud's Complete adjuvant was emulsified with 1 mg (300 /xl) of the antigen and 5 injections (3 intradermal, 1 subcutaneous, 1 intramuscular) were administered to the rabbits. The injections were repeated using 1 mg of antigen and 1 ml of Freud's Incomplete adjuvant 30 days later and again 14 days after that. To monitor the titre, test bleeds from the ear vein were done every 7 days after an injection. After the third series of injections, the rabbit was anesthetized with Ketamine and Rompun and bled by cardiac puncture. Serum was separated by centriftigation and stored at -20C. Biotinvlation of Antisera Antisera to CPE was biotinylated using NHS-LC-Biotin (Pierce, Rockford, IL). One hundred microliters of 0.05 M bicarbonate buffer, pH 8.5, was added to 2 mg of antisera. NHS-LC-biotin (0.04 mg) was added to die antisera and incubated on ice with mixing every 20 minutes for 2 hours. Unreacted biotin was removed by the addition of PBS to a total volume of 1 ml followed by dialysis in 6-8,000 wt cut off

PAGE 46

37 dialysis tubing (Spectrum Medical Industries, Inc., Los Angeles, CA) against two 500 ml volumes of PBS for 16 hours. The biotin labeled antisera was aliquoted and stored at -20C. Animals Three week old (12-13g) male Swiss Webster (SW) mice used in this study were obtained from Harlan-Sprague Dawley, through the University's Department of Animal Resources. The Department of Animal Resources is an lACUC Veterinary controlled facility. Animal Resources (AR) order, install, raise and maintain the animals within the Food Science and Human Nutrition (FSHN) Departmental Animal Facility as prescribed by the University of Florida lACUC. Animals are kept on a 12/12 light-dark cycle at 25C and are examined on a daily basis by AR who also change bedding and provide food and water. All animal studies were performed within the FSHN Animal Facility. Animal Tissue Preparation Mice were allowed to reach body weights of between 15-18 g before experimental use. This was usually 4-6 days after delivery, which allowed time for the animals to become accustomed to their surroundings and being handled, thus reducing any contributing stress factors. Animals were euthanized by COj asphyxiation followed by cervical dislocation. Mice were dissected and all organs

PAGE 47

38 (brain, thymus, heart, lung, liver, kidney and small bowel) were isolated, washed in PBS-TW, weighed and transferred to sterile 15 ml polypropylene mbes. Three milliliters of PBS-TW were immediately added and each organ was homogenized and disrupted on ice using a Polytron (Brinkmann Instruments) three times for 15 seconds. All individual organs from each mouse were kept separately in polypropylene mbes, to which CPE resuspended in PBS-TW was added at a standardized concentration, vortexed and allowed to bind. CPE:Tissue Binding To determine the level of CPE binding to tissue from each organ, five separate experiments (E1-E5) were performed, each requiring 8 mice. In four experiments the toxin concentration varied from 40 ng (El) to 80 ng (E2), 1 /xg (E3) and 2.5 fig (E4) of CPE/ml of tissue homogenate, with incubation for 1 hour at 27C (lab controlled to this temperamre). In the fifth experiment (E5) the CPE concentration was 2.5 /ig/ml tissue homogenate, with incubation at 4C for 4 hours. During the incubation period, samples were gently vortexed every 10 minutes to ensure complete distribution of toxin and tissue. Each experiment contained an equivalent number of negative controls, that is, tissues treated with only PBS-TW and no (zero) CPE. After incubation, a 1 ml sample from each individual organ was removed, and the remaining homogenate was stored at -70''C. The 1 ml isolated supernatant sample

PAGE 48

39 was centrifuged at 12.000 x g for 30 minutes at 4C, and the supernatant collected and stored at -70C. Enzyme Linked Immunosorbent Assay (ELISA) The ELISAs were conducted using the methodology described by Crowther (1995). Before the binding experiments were performed, extensive preliminary studies (concentration vs concentration analysis) were undertaken to determine the optimum levels and incubation times for each of the ELISA components and steps. A 200 fil supernatant sample (At: defined as soluble protein-antigens) from each of the tissues was thoroughly mixed with 200 /xl of coaling buffer (CB: 1.5 mM NazCOj, 3.3 mM NaHCOj, 3 mM NaNj). Triplicate 100 jul aliquots were added to individual wells of a polystyrene 96 well Inmiunolon™ 2 flat bottom plate (Dynatech Laboratories, Chantilly, VA), and allowed to passively adsorb for 16 hours at 4C. Any unbound sample was then aspirated, and each well was washed five times (X 5) with 100 [il PBS-TW. Primary polyclonal antibody (lAb) to CPE previously diluted to 1/10-^ in PBS-TW was added to each well (100 /il/well) and allowed to adsorb for 16 hours at 4C. Any unbound TAb was then aspirated, and each well was washed X 5 with 100 /xl PBS-TW. Secondary (2Ab) goat-anti rabbit IgG antibody labeled with the enzyme alkaline phosphatase (AP) (Sigma Chemical Co., St. Louis, MO) diluted to 1/10 ^ in PBS-TW was added to each well (100 ixl/weW) and allowed to adsorb for 2 hours at 27C. Wells were then aspirated and washed as described

PAGE 49

40 previously. The At:rAb:2Ab-AP complex was detected by the addition of 100 jul of alkaline phosphatase buffer (50 mM Na2C03, 50 mM NaCO,, 0.5 mM MgCy containing 100 /Ag of p-nitrophenyl phosphate (Sigma Chemical Co.). The enzyme reaction was stopped after 1 hour by the addition of 100 /xl of 1.0 N NaOH. Presumptive CPE positive samples are indicated by a sample color change from clear to yellow, which was quantitated by spectrophotometric analysis at 405 nm (Bio-Rad ELISA reader, model 2550). Western Immunoblot Analysis Tissue proteins and any CPE contained in the supernatant were separated by polyacrylamide gel electrophoresis (PAGE) with sodium dodecyl sulfate (SDS) using Bio-Rad Mini-Protean II slab gels and the buffered system of Laemmli (1970). Stacking and separating gels were 4.0% and 10% polyacrylamide, respectively. Prestained protein molecular weight markers (27-180 kDa) were obtained from Sigma Chemical Co. Gels were stained with 0.15% Coomassie Brilliant Blue R-250 overnight after being used for Western immunoblots. After soaking in transfer buffer (0.192 M glycine, 0.025 M Tris, 20% methanol) for 30 minutes, gels were immunoblotted using a Bio-Rad mini Trans-blot electrophoretic transfer cell and nitrocellulose membranes (Stratagene, La Jolla, CA) according to manufacturers directions. After transfer, the membrane was blocked with PBS containing 0.1% Tween 20 and 7.0% casein for 1 hour. The membrane was drained and soaked for

PAGE 50

41 16 hours in a 10"^ dilution of biotinylated antisera in blocking buffer. The membrane was then washed and soaked in a 1 mg/ml solution of strepavidin (Sigma Chemical Co.) for 2 hours. After a second washing step with PBS-TW the membrane was soaked and rinsed 3 times for 5 minutes with substrate buffer (100 mM Tris, 100 mM NaCl, 5 mM MgClj, pH 9.5). The protein bands were made visible by soaking the membrane in 10 ml of substrate buffer containing 0.1 mg/ml of nitroblue tetrazolium and 0.05 mg/ml of 5-bromo-4-chloro-3-indolyl phosphate for 30 minutes at 4''C. Statistical Analysis and Interpretation Statistical analysis of ELISA readings was performed using an one-way analysis of variance (ANOVA) with the STATISTICA for Windows software program, release 4.5 (Copyright^ StatSoft, Inc. 1993). When a significant difference was found {p < 0.05), Post hoc comparisons were done by using Tukey's honest significant difference (HSD) test, choosing an alpha level for critical ranges set at (a = 0.05). Results and Discussion ELISA Data Interpretation and Difficulties The aim of these experiments was to determine, using in vitro techniques, to which tissues (organs) within the mouse the CPE molecule had the highest affinity. Previous studies (Wnek and McClane, 1989; McClane and Wnek, 1990) have

PAGE 51

42 indicated that one CPE molecule binds to a single 50 kDa receptor (Rl) on the cell surface to form the CPE-Rl "small complex". Formation of small complex occurs at botii temperatures 4C and 2TC. Large complex formation occurs when the hydrophobic portion (N-terminal: toxic) of CPE is internalized and interacts with a 70 kDa protein (receptor R2) to form CPE-R1-R2. This interaction (large complex formation) is irreversible and does not occur at 4''C. Several definitions were made for the ELISA results. First, an ELISA was considered statistically positive after performing an ANOVA (p < 0.05) and a Posthoc comparison (HSD test, a = 0.05). The positive ELISA (a = 0.05) indicated that after the tissue-CPE interaction was completed, the isolated supernatant still contained free CPE molecules. Second, an ELISA was considered statistically negative after performing an ANOVA and a Post-hoc comparison (a = 0.05) where no significant difference was detected between control and CPE treated tissues. Thus it appears that a given tissue supernatant contained a greater number of Rl receptors than toxin which bound all available CPE molecules. Third, that if the experiments were conducted with fresh organ tissues at 27''C, then CPE binding to receptor Rl would form the small complex (CPE-Rl). This small complex would, within the 1 hour incubation period conformationally change and interact with receptor R2 to form the large complex (CPE-Rl -R2). Since we were using polyclonal antibodies to CPE, a number of epitopes on the CPE molecule were available for interaction (binding). We simply did not know whedier the interaction of CPE with Rl on the membrane

PAGE 52

43 left any regions of the toxin molecule available for interaction with the TAb. This is not an unlikely situation since the N-terminal hydrophobic portion of CPE does not appear to bind to Rl, and thus could be available. However, since the binding of CPE to Rl is temperature dependent and rapid, we believed that if CPE did bind to Rl to form the small complex, then large complex formation would be relatively rapid, and thus no portion of the CPE molecule would be available to bind to the TAb. Additionally, if the amount of tissue from a specific organ was kept constant and the amount of toxin varied, it might be possible to approximate CPE saturation levels for a specific tissue. That is, since the molecular weight of CPE is known, it was possible to calculate the number of CPE molecules/mg protein, and thus the number of Rl receptors/organ. Initially there were some difficulties in conducting the ELISAs, particularly for liver and kidney tissues which had a higher degree of non-specific binding. This was possibly a function of some tissues having both a higher amount and more varied number of tissue proteins, which may have caused some stearic hindrance. Alternately some proteins may have had a higher non-specific affinity for the l''Ab, caused by conserved protein regions common to both CPE and tissue proteins, resulting in stearic interference. Fortunately, most of the non-specific binding problems were resolved by using purified TAb, and additional blocking steps which significantly reduced background interference. Despite these problems, we believe the ELISA results are valid for all tissues.

PAGE 53

44 Table 3.1 Detection of unbound CPE by ELISA in murine organ tissues after in vitro interaction. CPE concentration and binding conditions Organ 40 ng 80 ng 1.0 us. 2.5 iidlTC 2.5ug/4C Brain ++ + + Thymus — — — ++ + + Heart ++ ++ + + Lung ++ ++ +-H Liver Kidney — Bowel — __ Binding conditions: 40 ng/ml CPE at 2TC for 1 hour 80 ng/ml CPE at 2TC for 1 hour 1.0 /ig/ml CPE at 27C for 1 hour 2.5 /xg/ml CPE at 27C for 1 hour 2.5 /ig/ml CPE at 4<'C for 4 hours (-) = no free CPE detected [a = 0.05] (++) = free CPE detected [a = 0.05]

PAGE 54

45 CPE Binding A summary of the five experiments described in the materials and methods is shown in Table 3.1. In experiments El (40 ng/ml) and E2 (80 ng/ml) no free CPE was observed in the tissue supernatant from any mouse organ. This suggested that neither of these toxin levels saturated the Rl sites in any organ tissue. The molecular weight (MW) of a compound is the sum of the atomic weights of the atoms in the molecule. It is the number of grams containing Avagadro's number of molecules (6.022 x 10^') (a mole is Avagadro's number of molecules). Thus, since the MW of the CPE molecule is 35,317 (Czeczulin et al 1993) then the minimum number of Rl receptors in each of the organ tissues examined in this study was > 8.5 x 10' /mg. Experiment E3 examined a 12.5 fold increase in CPE to 1,000 ng/ml. Results indicated that at this level the heart and lung showed free CPE at a significance level of a = 0.05. This suggested that the receptor-saturation threshold for CPE in the these organs was between 80-1,000 ng/ml, and the number of Rl receptors per mg of tissue was between > 8.5 x 10' /mg and < 1.5 x 10" /mg. No free CPE was detected with the brain, thymus, liver, kidney and bowel. Experiment 4 examined a 2.5 fold increase in CPE from E3, and a 32 fold increase over E2. The results obtained were consistent with E3 indicating diat 2,500 ng/ml of CPE saturated Rl receptors in thymus, heart, lung and brain tissue. Receptor-saturation threshold for CPE in the brain and thymus was between 1,000-2,500 ng/ml, and the number of Rl receptors per mg of tissue was between > 4.3 x 10* /mg and < 1.5 x 10" /mg for

PAGE 55

46 the brain and between > 2.0 x 10^' /mg and < 5.0 x 10" /mg for the thymus. No free CPE was found with liver, kidney and bowel tissues indicating that the number of CPE Rl-receptors in these tissues was > 1.5 x 10" /mg tissue. These results confirm previous conclusions by McDonel (1980): first, that the receptor saturation threshold for liver, kidney and bowel tissues was > 1.5 x 10"/mg of tissue protein, and there appeared to be no difference between the number of CPE-receptors between these organ tissues. Second, he also considered that competitive or inhibitory factor(s) released from the brain homogenate into the reaction buffer prevented significant CPE-receptor binding to this specific tissue. In this present study significant binding of CPE to brain tissue was found, which contradicts McDonel (1980). However, it should be noted that in this current study the methods used were far more sensitive and the binding technique was different. McDonel manipulated the tissue through a greater number of steps which may have caused the release of the "inhibitory" CPE-binding factor he suggested. One of the initial assumptions was that even if CPE bound to Rl, the small complex formed might still allow additional binding of lAb to any available regions of the CPE molecule. We argued however, that under our experimental conditions any small complex formed would conformationally change and become large complex very rapidly. Thus PAb binding to small complex was not an issue. Small complex formation is temperature independent, however, large complex formation is temperamre dependent. In E5 we examined whether saturating levels of CPE (2,500

PAGE 56

47 ng/ml) gave the same results at 4C when compared to using 2,500 ng/ml CPE at 27C (E4). Table 3.1 shows that indeed the results from E4 and E5 are directly comparable, where in both experiments free toxin is only found in the supematants of heart, lung, thymus and brain tissues, but not in liver, kidney and bowel. There are several possible alternatives. If McClane and Wnek (1990) are correct and small complex formation is temperature independent and large complex formation temperature dependent, it simply makes no difference, since no free toxin is found in some tissues. That is, in liver, kidney and bowel tissues small complex formation does not appear to allow any residual TAb binding to CPE. The small complex must simply be in an unavailable conformational state. Although it cannot be equivocally stated that this situation also occurs with heart, lung, thymus and brain tissues, it would seem reasonable to assume that CPE binding in these tissues is no different than in liver, kidney and bowel. Western Immunoblots were performed to determine if the technique could distinguish between control and CPE treated tissue. Only lung tissue was examined since this organ bound a significant amount of toxin, yet had a low level of interfering soluble proteins compared to liver and kidney. Preliminary smdies indicated that with the Western immunoblot technique, the lower limit of CPE detection either as pure toxin or unbound CPE was 50 ng/ml. Results shown in Figure 3.1 lane C indicated that control (untreated) tissue showed no positive bands, indicating that contaminating tissue proteins from the lung did not bind non-specifically to the TAb. When CPE

PAGE 57

48 Figure 3.1 Western immunoblot of lung tissue: CPE, lung tissue without CPE, and pure CPE detected with a 10"^ dilution of antisera to CPE. The amount of protein (CPE) in each sample is noted in parenthesis. Lane A: purified CPE (28 fig); lane C: lung tissue supernatant without CPE (0 /ig); lane L: lung tissue supernatant with CPE (100 ng).

PAGE 58

49 was in excess to the receptors within the lung (see Table 3.1 treatment 1.0 ^g, and Figure 3.1 lane L) free toxin was clearly detected as a 35,000 band (lane L) exactly at the same position as pure CPE (lane A). Strong antibody positive bands were also detected at approximately 87,000 (small complex) and 160 kDa (large complex) (lane L). These in vitro binding studies suggested that the liver, kidney and bowel had the highest number of CPE receptors within their tissue (> 4.5 x 10'^ receptors/ organ), followed by brain and thymus (< 4.5 x 10'^ receptors/organ), and heart and lung (< 1.7 X 10'^ receptors/organ). These results are not unexpected since introduction of CPE to the body occurs via the gastrointestinal tract, and metabolism and detoxification of CPE likely occurs via the kidneys and liver. Indeed Skjelkvale et al. (1980) in his studies with '^'I-labeled CPE suggested that the kidney was a target organ for CPE binding and that this organ contained the highest level of CPE receptors. The studies conducted herein, confirm that the kidney is an organ that contains a large number of CPE receptors, however, there is no indication that this organ is the main target. Skjelkvale et al. (1980) suggested that CPE is metabolized and expelled from the body within urine, which would make the kidney an organ of focus. However, further in vivo studies are required to confirm this. The liver and small bowel would also be organs of focus since cells from these tissues are highly susceptible to the action of CPE (Lindsay and Dennison, 1986a, 1986b; Kokai-Kuhn and McClane, 1996; Mach and Lindsay 1997). Although our data strongly suggest

PAGE 59

50 that liver, kidney and bowel tissues contain the highest number of receptors, and require very high levels of CPE to reach receptor saturation, the results do not necessarily translate to the in vivo model. That is, these organs may be the main focus of CPE distribution but not necessarily CPE activity in vivo. It is possible that low levels of CPE within a specific tissue have a more dramatic effect on the host, than high levels of CPE in another. For example, high levels of CPE in the bowel may be enterotoxigenic and cytotoxic causing diarrhea, fluid loss and tissue desquamation. However, low level CPE binding and activity in brain and lung tissue may affect neurologic and respiratory stams. Additionally, CPE binding to heart tissue may have cardiotrophic effects. Thus, enterotoxigenicity may be transient, but alteration in neurologic, respkatory or cardiac status may be lethal. Chapters 4 and 5 detail our smdies to determine the in vivo distribution of CPE after intraperitoneal and intragastric administration and the consequential effects.

PAGE 60

CHAPTER 4 DETECTION AND DISTRIBUTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN AFTER IN VIVO INTRAPERITONEAL ADMINISTRATION INTO SWISS WEBSTER MICE Introduction McDonel (1980) indicated that in vitro, CPE binds with different specificities to various organ tissues. Results suggested that Hver and kidney tissues contained the largest number of CPE receptors, and that both organs were specific sites of CPE attack that required > 0.5 ^g of CPE/mg of tissue to reach receptor saturation. There is however, an anomaly in McDonel's supposition. The data suggested that to reach complete samration of liver and kidney tissues in vivo required lethal levels of CPE to be administered and absorbed, an unlikely circumstance in most CPE induced foodbome illnesses. Skjelkvale et al. (1980) using the murine model attempted to determine the distribution and levels of CPE binding after intravenous (IV) administration of radioiodinated CPE. Results suggested that the liver and kidneys were specific organs of CPE accumulation and attack, and that a major fraction of CPE was rapidly metabolized and excreted in urine. This apparently confirmed McDonel's findings, however, alternative explanations are possible. First, decay of radiolabeled toxin could have spread radioactivity to organs where CPE was not apparent, leading to false positives. Second, CPE was not administered through its 51

PAGE 61

52 natural port of entry, that is intragastrically (IG). It is not known wliether CPE administered both IP and IG give similar results. It could be argued however, that this second point is moot, since Tsai and Riemann (1975a) showed that CPE orally administered to mice was present in blood witiiin minutes. Thus, IG and IP administration may be similar, however, CPE is known to be structurally altered by proteolytic enzymes in the small bowel before systemic absorption (Granum and Richardson, 1991). Third, CPE disuribution in mice was only monitored during the two hours after IV administration, and only at a single CPE dose. It is possible that there are various CPE distribution patterns which are toxin dose and time dependent. This chapter describes the approaches to determine the organ/tissue distribution of varying doses of non-labelled CPE toxin after IP administration into Swiss Webster mice. Materials and Methods CPE-Toxin and Antisera Freeze dried CPE obtained from Dr. Bruce McCIane was prepared for IP administration as described in Chapter 3. The biological activity (specific activity) of preparations was examined before use by Vero cell analysis (Mach and Lindsay, 1994) and was standardized to 4,000 EUZ/xl toxin. The methods to produce antisera to CPE are described in Chapter 3.

PAGE 62

53 Animal preparation and IP enterotoxin administration Animals were purchased and maintained as described in Chapter 3. As all animals were bred from the same line and obtained from the same source it was assumed that they were genetically similar. Gross examination indicated no apparent phenotypic differences. When animals weighed 15-18 g (25-30 days old), they were randomized and grouped six/cage. To reduce any administration differences, one person (Keller) held the mice in the correct alignment, while another (Lindsay) measured the correct volume and performed the administration. Animals were administered CPE IP in the left side of the peritoneal cavity, using a 1 ml tuberculin syringe and a 27 gauge needle. After injection animals were remmed to their respective cages and monitored every 15 minutes. Sftidv la: CPE Concentration This smdy was performed to determine the following: i. the CPE concentration at a time which caused death within 72 hours; ii. die CPE concentration at which the pathophysiological changes observed during the toxicosis could be strongly predicted without interference from cases of random deatti within the treated mouse population. These data would then allow animal necropsy and tissue sampling with reliability. CPE was administered IP at various levels ranging from 0.1 to 5 mouse lethal doses (MLD): nonlethal 0.1 MLD (0.5 Mg/250 fil) to 0.25 MLD (1.25 ng/250 fil); sublethal 0.5 MLD (2.5 /ig/250 ^1) to 1.0 MLD (5.0 Mg/250 ^1); lethal 2.0 MLD

PAGE 63

54 (10.0 iig/250 fil) to 5.0 MLD (50.0 /Ag/250 pil); and control: 250 ^1 of PBS. Twelve mice were used at each treatment level, and 3 control mice were administered PBS. Any pathophysiological changes were monitored every 15 minutes from (immediately after administration) to T72 hours, or time to death, which ever came first. Mice were necropsied and all organs (brain, thymus, lung, heart, liver, kidney, and small bowel) were isolated, washed in PBS-TW, weighed and transferred to sterile 15 ml polypropylene mbes and stored at -10C for later analysis. Blood and urine (where possible) were also collected. Before storage, blood samples were gently centrifuged at 200 x g to collect serum. Both serum and urine were stored at -70C for analysis. Study lb: CPE Distribution vs Time Based on data obtained in Smdy la (see Results and Discussion for detailed data) a CPE concentration of 10 /xg CPE/250 fil PBS was chosen for IP administration in the disu-ibution versus time smdy. Twenty four animals were prepared, and administered CPE IP in a single dose as described above. Six similar weighted animals were administered PBS as controls. At six time intervals 0.25, 1, 2, 3, 4, and 5 hours, four animals were randomly chosen from the toxin administered group and one from the control group and euthanized. Mice were necropsied and all organs (brain, thymus, lung, heart, liver, kidney, and small bowel) were collected as described above. Blood and urine (where possible) were also collected. Before

PAGE 64

55 storage, blood samples were gently centrifuged at 200 x g to remove red blood cells, and the serum collected and stored at -TO^C, as was any urine sample. Immunological Methods To prepare organ tissue samples for ELISA, 3 ml of PBS-TW was added and each organ was homogenized and disrupted using a Polytron as described previously. Serum and urine were examined without PBS-TW dilution. The materials and methods to conduct the ELISA and Western immunoblot are described previously in Chapter 3. Interpretation of ELISA Assays and Statistical Analvsis Several assumptions made with regards to the interpretation of the ELISA assays, based on data presented in Chapter 3, and on recently published studies by Kokai-Kuhn and McClane (1996). First, that the ELISA assay could detect both free CPE and CPE-Rl receptor bound small complex noting that the small complex moiety had to be in the soluble fraction. Second, that CPE polyclonal antibodies (CPE-pAb) used in the ELISA assay had a higher affinity for free CPE, and thus would preferentially bind to this moiety. This is also stoichiometrically logical since CPE bmding to Rl would reduce the number of epitopes available for CPE-CPEpAb binding. Third, low (sublethal) levels of CPE would not saturate Rl receptors in most organs, while high (lethal) levels of CPE would saturate Rl receptors in most

PAGE 65

56 organs. Thus, when administering CPE IP at high levels, unbound CPE would preferentially bind to the CPE-pAb resulting in a positive ELISA (a = 0.05). See Chapter 3 for discussion. If the ELISA was positive at a = 0.2, CPE may be detected as a mixture of unbound, small complexed and large complexed CPE. Alternately, at sublethal CPE levels (without receptor saturation) CPE detection might be predominantly as a mixmre of small complex and large complex forms. Statistical analysis was performed as described in Chapter 3. Results and Discussion Symptoms During Toxicosis Immediately after CPE-A administration the animals presented an accelerated heart rate and a spasmatic breathing pattern (hyperpnea). Animals exhibited arched backs, ruffled-opaque fur, opaque eyes, disorientation, loss of appetite for both food and water, and a requirement for group association in the comers of the cage (gathering). The CPE toxin appeared to express apparent neurotoxicity where animals exhibited flaccid paralysis, suggesting that the vagus nerve had possibly been compromised. At no time did animals present any apparent signs of pain, and until recovery or death the animals would cuddle and rest. When animals recovered from the toxicosis they showed no signs of any long term pathophysiological damage, and ate and drank as before CPE administration. Similar symptoms in animals due to CPE toxicosis were observed by Tsai and Riemann (1975a, 1975b) and Skjelkvale et

PAGE 66

57 al. (1980) although not in as much detail. The toxicosis presented differences in onset time and recovery dependent on the route of administration and the purity of CPE preparations. Pathological Findings at Necropsy At necropsy the liver, thymus and kidneys showed petechiae. Some animals had blood in the bladder, while others had an enlarged liver and/or kidneys. At high toxin levels the animals had a mushy bowel with distinct signs of proteolysis of the intestinal walls. The integrity of the blood vessels seemed to be compromised as blood was found in the peritoneal and thoracic cavity. It was extremely difficult to collect blood, and the average blood volume collected was about 1 ml. This was likely due to two factors: first, the animals were very small, and only had a total blood volume of 1.5-3.0 ml. Second, blood coagulation seemed to be accelerated in CPE treated animals compared to control animals. This observation has been confirmed in smdies by Wallace et al. (1997). The liver in CPE treated animals was heavier than in controls, suggesting that the toxicosis causes a higher metabolic rate (detoxifying mechanism) which is an attempt to clear CPE from the body. Thus more blood and/or fluids were present in die organ(s), and either upon organ failure or death tlie fluids were trapped.

PAGE 67

58 Table 4.1 ELISA detection of unbound CPE in murine organ tissues after in vivo IP administration. CPE dose (MLD*) administered Organ OA 025 M LO LQ. Brain ++ + + Thymus ++ ++ ++ + + Heart Lung ++ ++ ++ + + Liver ~ ~ — — — Kidney — — ~ — Bowel — + + Blood ++ ++ + + Urine -++ ++ ++ + + CPE dose administered: 0.1 MLD (0.5 /ig/250 fxl) 0.25 MLD (1.25 /xg/250 ixl) 0.5 MLD (2.5 Mg/250 pil) 1.0 MLD (5.0 /ig/250 /il) 2.0 MLD (10.0 ^g/250 fil) 5.0 MLD (50.0 /ig/250 ^1) control: 250 ixl of PBS Mouse Lethal Dose. (-) = no free CPE detected [a = 0.05] (++) = free CPE detected [a = 0.05]

PAGE 68

59 CPE Concentration Study When CPE was administered at sublethal doses (0.1 and 0.25 MLD) none of the ELISA assays for any of the seven organs examined or blood (as serum) and urine was statistically significant at a = 0.05 level (Table 4.1). This suggested two alternatives: first, CPE at this low level was being rapidly metabolized by the liver and kidney, and excreted in the urine before the animals were necropsied and screened by ELISA. Second, the low CPE levels did not samrate cellular receptors as suggested in the section Interpretation of ELISA Assays. Thus neither unbound CPE nor small complex was available for detection. These results are consistent with the pathological findings as none of the animals died within the 72 hour time period after CPE administration. Indeed, while animals presented clear symptoms of CPE induced toxicosis they were able to recover within 8 to 12 hours. CPE administered at sublethal-lethal levels (0.5 to 1.0 MLD) was detected as free toxin in thymus and lung tissue, and urine (a = 0.05). All other organ tissues (brain, heart, liver, kidney, and bowel) were not statistically significant (a = 0.05). Serum was found to contain free toxin (a = 0.05) at 1.0 MLD but not at 0.5 MLD. Animal sensitivity and resistance to CPE between 0.5-1.0 MLD showed varied responses. Two patterns of sensitivity and resistance were apparent. First, 30-50 % of animals appeared resistant since they effectively recovered from the toxicosis within 8-12 hours of CPE administration. Animals sensitive to CPE died predominantly within the first 24 hours. This bimodal diso-ibution of effects was

PAGE 69

60 consistent through several studies, and was in part not unexpected, since the toxin concentration was near the LD50. Normally within a homogeneous population of mice (same species, similar age and weight) two subgroups are apparent that differ in their ability to metabolize a toxin. The nature or reason for this variability remains unknown, although it is likely that detoxifying mechanism(s) of an individual animal may be less effective and compromised. All mice administered CPE at 2.0 MLD died within 12 hours, with about 80% of mice dying between 8 and 9 hours. At 5 MLD all mice died within 15 minutes. Free CPE was detected (a = 0.05) in brain, thymus, lung and bowel tissues, and blood and urine. No free toxin was apparent in heart, liver and kidney tissues, which was in part unexpected. Several alternatives were suggested: first, none of these organs were saturated with toxin at these levels, and any toxin present in these tissues was present in a form (possibly large complex ?) that was undetectable. Second, due to multiple organ failure, any damaged tissue containing CPE was removed from the organ via the circulation. Third, the animals died from the CPE toxicosis before detectable levels of free toxin were reached. This latter alternative is applicable for CPE at 5 MLD where the toxin likely acted as a neurotoxin.

PAGE 70

61 Table 4.2 Detection of unbound CPE by ELISA in murine organ tissues after in vivo IP administration: time study. CPE binding time in hours Organ 0.25 LO M M M IM Brain -++ ++ + + Thymus ~ — ~ — — — Heart -++ ++ + + Lung .. .. + Liver -++ ++ + + Kidney — — — — Bowel — ._ Blood ++ + +/_ Treatments: administration of 10 /ig/250 ^1 (2 MLD*) CPE, animals necropsied at 0.25, 1.0, 2.0, 3.0, 4.0, and 5.0 hours; control: 250 ^1 of PBS. Mouse Lethal Dose. (-) = no free CPE detected [a = 0.2] (+/-) = free CPE detected [a = 0.2] (+) = free CPE detected [cc = 0.1] (++) = free CPE detected [a = 0.05]

PAGE 71

62 Time Study Based on data obtained from the concentration study a toxin level of 2 MLD (50 /ig/250 /il) was chosen for the time study. This level was chosen for two reasons: first, CPE could be detected with statistical significance and reliability by ELISA, and second, after toxin administration, most mice died within a narrow time-frame 8-9 hours and there was no random death during first 5 hours. Results from the time smdy at the single CPE level are shown in Table 4.2. Free toxin was detected in serum within 15 minutes of CPE administration, however, toxin was apparently cleared within three hours. Toxin was detected (a = 0.05) in brain, heart, and liver tissues 3 hours after administration, and in die lungs (a = 0.1) after 5 hours. CPE was not detected in diymus, kidney or bowel tissue (a = 0.2). Since CPE is systemically absorbed and distributed, its presence in brain, heart, liver and lung tissue was not unexpected. However, apparent lack of presence in thymus and kidney was unexpected. Lack of presence in bowel was not unexpected since, this tissue contains a high number of receptors. More importantly however, for toxin to be present in ileal tissue would require translocation of toxin from the peritoneal cavity to the lumen, against the normal translocation gradient. Alternately, lack of free toxin in thymus and kidney may also suggest that in these tissues, that if CPE is present, it is in the large complex form and therefore inaccessible to CPE-pAb. Another consideration can be made. It is possible that toxin distribution and binding does not follow a set disfribution pattern. That is distribution and binding may

PAGE 72

63 Table 4.3 ELISA and Western immunoblot detection of CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after in vivo IP administration. CPE concentration and binding conditions Organ ELISA Brain Thymus Heart Lung Liver Kidney Bowel CPE dose administered: 0.1 MLD (0.5 ^g/250 /xl) control: 250 fil of PBS ELISA (-) = no free CPE detected [a = 0.2] Western immunoblot 35: M, 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1); 160: M, 160 kDa (CPE:R1:R2). (~) = no band detected () = band detected Western immunoblot 35 85 160

PAGE 73

64 depend on how CPE was systemically absorbed, the time needed for toxin to be absorbed, exact location of IP administration, or other factors inherent to the animal itself, for example the metabolism of the individual animal. It is interesting to compare the results obtained for the concentration and time studies at the same CPE levels. It appears that there are differences in toxin distribution patterns regarding thymus, heart, and liver tissues. CPE was detected in the free form in the time study, but not in the concentration study in hean tissue, while the liver and thymus were negative (a = 0.1) One possible explanation for these differences was that the sample size of the two smdies was different. Also, the time allowed for CPE distribution and subsequent binding to receptors was different. In the time smdy the samples were collected within the first 5 hours, while in the concentration smdies samples were collected between 8-12 hours. Since it was not known how time affected CPE distribution and binding, we chose not to directly compare the ELISA data from the time and concentration smdies, but to further analyze each separately. Western Immunoblot From the in vitro studies it was presupposed that CPE bound to various organ tissues within the mouse, albeit with different affinities. It was also assumed that CPE bound initially to receptor Rl forming small complex, and then to R2 forming large complex. The ELISA data predicts the presence of CPE on various organs.

PAGE 74

65 However, in order to confirm CPE presence and determine the conformation and configuration state (small or large complex) of the toxin it was necessary to conduct analysis by Western immunoblot. It should be noted that sample preparation for each method was significantly different. ELISA samples required no special preparation steps, and simply screened the supernatant for soluble protein (antigen). Western's required tissue samples to be boiled in the presence of fi-mercaptoethanol and SDS, allowing the release of tissue associated proteins. In order to initially determine the sensitivity of the Western immunoblot assay, all tissue samples from the 0.1 MLD concentration smdy were screened for the presence of CPE. These samples were chosen since they represented the lowest level of CPE administration, and the ELISA assay was not able to detect free CPE in any samples. Previous work had mdicated that the sensitivity of the ELISA assay in diis smdy was > 100 pg CPE/ml, whereas the sensitivity for the Western assay > 1 pg CPE/ml. It was assumed that if CPE was present in tissue it had to be found either as the small (85 kDa) or more likely the large (160 kDa) complex. A summary of the analyses is shown in Table 4.3. As previously noted, at 0. 1 MLD the ELISA did not detect any unbound toxin as the 35 kDa protein in any tissue. However, Western analysis of CPE treated animals showed a 35 kDa protein band in the heart and lung which was not present in control (untreated animals). Whether toxin was present in these organs as free (unbound) toxin not detectable by ELISA, or as loosely bound small complex which dissociated upon sample preparation is unknown. A 85 kDa band, presumably small complex

PAGE 75

66 was detected in the brain, thymus, heart, kidney and bowel tissues. This band was not present in control (untreated samples). A 160 kDa band presumably large complex was present in brain and kidney tissues, and again this band was not present in control tissues. The Western analyses suggest that eidier the ELISA assay was less sensitive to CPE detection when toxin was administered at nonlethal levels or that at nonlethal levels CPE was bound as small and large complex, and thus not available for detection. The latter is more likely since the number of Rl receptors exceeds the number of CPE molecules, and thus receptor saturation was not reached. The Western assay was very informative since it allowed determination of the state and type of CPE association to specific receptors (either small or large complex). Both the ELISA and Western assays compliment each other since the ELISA determines receptor samration threshold, while the Western determines the type and distribution of CPE-receptor interaction. The type of CPE-receptor complex formed appeared to be specific within individual organs, that is, different organs showed different CPEreceptor complex patterns. These patterns may suggest that toxin distribution within an individual tissue is both CPE-concentration dependent and cell type specific. The intraperitoneal studies showed that CPE is systemically distributed throughout the body within minutes of administration. The pathophysiological effects of this rapid distribution were both concentration and time dependent. At nonlethal levels CPE was not detected in blood, however, it is likely that at these concentrations

PAGE 76

67 the toxin was rapidly bound to receptors within the liver and kidneys, and metabolized and excreted via urine. The basis for this assumption is first, that both the liver and kidney contain the highest number of CPE-receptors. Second, the receptors within these organs appear to have the highest affinity for CPE (McDonel, 1980). Third, these two organs are normally associated with the removal of toxic products from the body. At sublethal levels CPE appears to have several modes of action. First, as a cytotoxic-enterotoxin causing diarrhea accompanied by gross tissue damage to the bowel. Second, as a neurotoxin since animals presented symptoms usually termed parasympathomimetic (cholinergic). This neurotoxic effect, confirms previous descriptions associated with CPE toxicosis in animals and humans (Murrell et al. 1987; Lindsay, 1996). At lethal levels CPE appears to act as both a neurotoxin and a superantigen. Death is rapid, and symptoms of bradycardia, hyperpenia and flaccid paralysis were apparent, leading to multiple organ failure, shock and death. These symptoms suggest a complete collapse of the sympathetic nervous system accompanied by the massive induction of cytokines at the cellular level. While these findings apply to intraperitoneal administration of CPE it remains to be determined whether similar whole body distribution patterns and responses occur after intragastric administration. This is critical since CPE is normally synthesized, activated and systemically absorbed in the small bowel. This is the focus of Chapter 5.

PAGE 77

CHAPTER 5 DETECTION AND DISTRIBUTION OF CLOSTRIDIUM PERFRINGENS TYPE A ENTEROTOXIN AFTER IN VIVO INTRAGASTRIC ADMINISTRATION INTO SWISS WEBSTER MICE Introduction The effects of in vivo oral administration of C. perfringens type A vegetative ceUs, spores suspensions, sterile culture filtrates, and crude cell preparations containing type A enterotoxin (CPE) into various animal models has been studied previously by several groups (Canada and Strong, 1965; Weiss et al., 1965; Hauschild et al., 1971; Uemura et al., 1975; Tsai and Riemann, 1975a, 1975b). Unforttinately, either the basis for the studies were questionable or the results were inconclusive. Canada and Strong (1965) force-fed germ-free mice with Uve C. perfringens type A suspensions, and measured the heat resistance of the spores recovered after passage through the animals GI tract. As would be expected the recovered spores had the same heat resistance as those initially fed the animals. What this smdy showed or proved is moot. In other studies it appeared that either the administration or absorption levels were too low, thus enterotoxin could not be detected in serum, and the only signs of toxicosis were emesis and diarrhea (Weiss etal., 1965; Hauschild et al., 1971; Uemura et al., 1975). Tsai and Riemaim (1975a, 1975b) examined the responses of mice to oral challenge with C. perfringens type A vegetative cells, spore suspensions, and a crude 68

PAGE 78

69 C. perfringens cell supernatant containing unquantified levels of CPE. Challenge with vegetative cells and spore suspensions showed the bacterium to be present in the gastrointestinal (GI) tract, heart, lung, liver, kidney, spleen, and blood. The presence in various organs is interesting, however, there was no indication that the bacterium was actually bound within tissue. Presence could merely have been a function of systemic distribution via blood, and lack of washing prior to analysis. Unfortunately there was no determination of the level of CPE in any of these tissues or within blood. Administration of crude supematants containing CPE, and for that matter any other C. perfringens toxins, showed CPE present in serum. Mice also exhibited diarrhea, and mucosal enteritis with edema and hyperemia of the small intestine. Whether these pathophysiological responses were due to CPE or any of the other C. perfringens toxins was not determined. The work of Tsai and Riemaim (1975a) is perhaps die best of the previous studies, since it did show that CPE can be found in blood after oral administration. However, several criticisms of their work can be made. First, it is impossible to determine the level of toxin absorption let alone the amount of CPE produced in the small intestine simply based on feeding live cultures. There is significant variation in sporulation between C. perfringens strains in vitro, and the level of sponilation and CPE production in vivo is simply unknown, except for the 8239 lab strain. Even for this strain there are many variables that interplay in vivo. Second, were there any other C. perfringens toxins produced and absorbed, since this bacterium may produce

PAGE 79

70 up to 15 different toxins, many of which appear to act synergistically? Third, the use of crude toxin containing supematants is fraught with problems, since as noted, many toxins act synergistically. In a crude extract how can any one toxin be shown to cause a specific pathological or histopathological response. Thus pure preparations of a specific toxin have to be used to determine cause and effect. Fourth, the time frame used to study any pathological effects needs to be extended. Distribution stodies should at least mimic FBI symptoms which usually take 1-24 hours, or possibly longer to elicit. Further, a range of toxin doses should be administered to determine if there are dose and time dependent responses. Fifth, toxin detection or effect should not be investigated by a single type of assay, since a single assay may provide either or both, false positives and/or false negatives. Thus, several methods of analyses must be used, one of which must be biological. Finally, any method used, has to have the ability to be statistically vaUdated. This chapter reports studies to determine the whole body distribution of CPE toxin after intragastric administration through the namral port of entry, that is intragastrically, using die murine model. The criticisms and requirements discussed above have been incorporated into the study, in order to specifically define the response.

PAGE 80

71 Materials and Methods CPE-Toxin and Antisera Freeze dried CPE obtained from Dr. Bruce McClane was prepared for IG administration as described in Chapter 3. The specific activity of preparations was examined before use by Vero cell analysis (Mach and Lindsay, 1994) and was standardized to 4,000 EU/pil toxin. The methods to produce antisera to CPE are described in Chapter 3. Animal Pr eparation and IG Enter otoxin Admini stration Animals were purchased and maintained as described in Chapter 3. All mice were bred from the same line and obtained from the same source. It was assumed that they were genetically sunilar, and after gross examination no phenotypic differences were apparent. When animals weighed 15-18 g (25-30 days old), they were randomized and grouped six/cage. To reduce any administration differences, one person (Keller) performed all toxin preparation and oral administration (gavage). Animals were IG administered CPE via a 250 ixl volume into die oesphagous and stomach, using a 1 ml mberculin syringe and a 27 gauge, 2 mm ball-ended Popper needle. PreUminary smdies indicated that it was not necessary to neutrahze stomach acid using 50 mM bicarbonate buffer prior to CPE administration. Prior administration was observed to upset the mice and vary the biological response after

PAGE 81

72 CPE administration. Additionally, not using bicarbonate buffer more closely mimicked the true toxicosis, since it allowed protease activation of the toxin in the small intestine. CPE retained biological activity in the presence of PBS, and since the volimie administered was quite small, delivery was relatively straight-forward. After CPE administration mice were remmed to their respective cages and monitored every 15 minutes. Stodv la: CPE Concentration This smdy was performed to determine several factors. First, the CPE concentration that caused death within 72 hours, which would allow a direct comparison between the IP MLD obtained in Chapter 4, and the IG MLD. Second, the CPE concentration at which any pathophysiological changes observed during toxicosis could be strongly predicted witiiout interference from cases of random death within the treated mouse population. This would then allow an IG time concentration study to be conducted similar to tiiat performed in Chapter 4. Thus CPE administration, necropsies and tissue sampling could be done within a specific time frame with reUabihty. CPE was administered IG at various levels ranging from 25-75 fxg CPE (0.1-0.25 MLD) in 250 /xl PBS. Twelve mice were used at each treatment level, and 3 control mice were administered PBS alone. Mice were monitored for pathophysiological changes every 15 minutes from T„ (immediately after

PAGE 82

73 administration) to T72 hours, or time to death, which ever came first. Mice were necropsied and all organs (brain, thymus, lung, heart, liver, kidney and small bowel) were isolated, washed in PBS-TW, weighed and transferred to sterile 15 ml polypropylene tubes and stored at -70C for analysis. Smdv lb: CPE Distribution vs Time In retrospect the data obtained in Smdy la (see Results and Discussion for details) proved to be of limited value, since toxin concentrations within this range were nonlethal. Rather than conduct additional studies using large numbers of animals to determine the IG MLD, two mice were each administered CPE IG at concentrations of 1(X), 125, 150, 250 and 500 )ug/250 ^1 and the time to death determined. This two mouse method has been used previously to determine the MLD of C. botulinum neurotoxin with good statistical reliability (Sugiyama, 1986). Based on these additional data two CPE concentrations, 150 ixg CPE/250 ix\ PBS (0.5 MLD) and 500 Mg CPE/250 /xl PBS (2.0 MLD) were chosen for the distribution vs time smdies. For each time smdy, ten mice were administered CPE IG in a single dose, nine similar weighted control mice were administered PBS alone. Each toxin level was initially designed as a distribution vs time study, however, for reasons discussed in the Results and Discussion section each study presented its own unique problem. Unexpectedly at each of the CPE concentrations, two distinct time to death patterns were apparent. That is, despite the genotypic and

PAGE 83

74 phenotypic similarity of the mouse populations used for each study, half the population died within one time period, the other half within a different time period. Therefore the data were analyzed taking mio account the time to death variable. Mice administered CPE were necropsied immediately upon death. At the same time one matched control mouse was euthanized and necropsied. All organs (brain, thymus, lung, heart, liver, kidney, spleen, stomach, and small bowel) were collected and stored as described previously. Tmmiinnlo gical Methods and Statistical Analysis To prepare organ tissue samples for ELISA, 3 ml of PBS-TW was added, and each organ was homogenized and disrupted using a Polytron as described previously. Serum and urine were not exammed. The materials and methods required to conduct the ELISA and Western immunoblot are described previously in Chapter 3. The same assumptions for the interpretation of the ELISA assay and statistical analysis as described in Chapter 4 were used in the IG studies. Vero Cell Assay Tissue suspensions of all organs were tested for biologically active CPE using the Vero ceU assay (African Green Monkey kidney cells :ATCC CCL81) (Mach and Lindsay, 1994). CeUs were culuired in Sarstedt (Newton, NC) 75 cm^ flasks and maintained in modified Eagles medium (MEM) containing 10% fetal bovine serum

PAGE 84

75 (FBS), 2 mM L-Glutamine, and antibiotic/antimycotic (Sigma Chemical Co.). Cells were incubated as an adherent monolayer in a humidified incubator with 5% C02/95% air at 37C. When confluent, cells were removed with trypsin-EDTA (Sigma Chemical Co.), diluted to 10 ml in MEM and plated 0.2 ml/well (10* cells/ml) in Costar 48 well cluster dishes. Cluster dishes were incubated as described above for 24 hours, the media removed and each well washed twice with sterile 15 mM PBS. Earl's balanced salts solution (free of FBS) containing 0.28 mM phenol red (BSS-PR) was added to die wells (final well volume of 0.2 ml). To test the biological activity (cytotoxicity) of any CPE in the high-speed centrifuged tissue supernatant, 100 /il of BSS-PR was replaced with 100 /xl of tissue supernatant and incubated for 2 hours. Cells were examined for morphological damage and viability with a Nikon phase contrast microscope. A sample was considered biologically active if it caused > 50% cell lysis in duplicate wells. To determine whether cell death was caused by CPE, all suspected CPE positive tissues as determined by a positive ELISA, positive Western and positive Vero cell assay, were re-assayed in the Vero cell assay using CPE neutraUzing antibody. Tissue supematants were treated with a 1;100 dilution of CPE primary antibody, thus neutraUzing any available toxin. Any sample which was ELISA, Western and Vero cell assay positive, and negative in the Vero cell assay after neutraUzation with anti-CPE antibody, was considered a confirmed positive.

PAGE 85

76 Results and Discussion Symptoms During Toxicosis To induce a toxicosis by IG administration required a 10-15 fold higher CPE level. Upon IG CPE administration, mice presented similar symptoms and pathophysiological changes as observed for IP CPE administration, see Chapter 4 for detailed description. Mice administered nonlethal CPE levels fully recovered from the toxicosis within 6-8 hours. Animals administered sublethal levels of CPE, presented two response modes. Approximately 25 % of the mice died within 4 hours, while the remaining 75% recovered within 6-8 hours. Although the animals used were genetically and phenotypically similar, this response was not imexpected since the amount of toxin administered was near the 0.5 MLD. Administration of lethal levels of CPE caused the death of all animals, however, two modes of death were apparent. First, a "sudden" death induced within minutes of toxin administration, and second, a nonabrupt death that required several hours for manifestation. Pathological Findings at Necropsy At nonlethal CPE levels there was no apparent gross histopathological tissue damage to organs, other than the bowel. At sublethal and lethal CPE levels the liver, thymus, and kidneys showed petechiae, and the liver and kidneys were enlarged. Enlarged organs suggest hyperactivity which leads to organ failure. Additionally, the GI tract exhibited signs of extensive proteolysis (mushy and spongy) to the

PAGE 86

77 submucosa, and without care upon dissection the tissue would disintegrate. It was apparent that the damage to the digestive tract would significantly compromise metabolism to the point where animals had very little chance of recovery. This type of bowel tissue damage has been observed for other C. perfringens infections, for example pigbel, where recovery from the disease requires surgery. Collection of blood was not possible since the blood vessels in die peritoneal cavity showed signs of hemorrhage. Coagulated blood was found in the peritoneal cavity, lungs, liver, kidneys and digestive tract. This may have contributed to a decrease in blood volume within the animal. Wallace et al. (1997) observed similar responses in their work, and preliminary data indicated that the coagulation factors appeared to be affected during CPE-toxicosis. No urine could be collected during necropsy likely due to kidney organ failure, or possibly loss of involuntary control of the urinary system. CPE Concentration Studv The ELISA assay did not detect free CPE administered at 0. 1 MLD (nonlethal) in any organ-tissue sample (Table 5.1). This was expected since nonlethal levels would not have saturated tissue receptors, and thus toxin was not available for detection. At 0.25 MLD the brain and kidney had free CPE at a significance level of a = 0.2, but all the other organs showed no free CPE (Table 5.1). This may suggest that at 0.25 MLD the CPE threshold for receptor saturation in brain and kidney tissues was reached. However, since the kidney is the main organ of CPE

PAGE 87

IB Table 5.1 Detection of unbound CPE by ELISA in murine organ tissues after in vivo IG administration. CPE dose administered Organ 0.1 MLD 0.25 MLD Brain — +/_ Thymus Heart Lung Liver Kidney +/. Bowel CPE dose administered: 0.1 MLD (25 /^g/250 ^1) 0.25 MLD (75 /ig/250 /xl) controls 250 pil of PBS Statistics (--) = no free CPE detected [a = 0.2] (+/-) = free CPE detected [a = 0.2]

PAGE 88

79 Table 5.2 ELISA and Western immunoblot of CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after IG adminisn-ation. CPE concentration and binding condition 0.25 MLD Organs ELISA Western 35 85 160 Brain +/Thymus „ Heart ~ ~ Lung __ _. Liver -_. a Kidney +/.. Bowel ~ CPE binding conditions: administration of 0.25 MLD (75 /ig/250 fil) controls were administered 250 /nl of PBS Statistics : (-) no free CPE detected [a = 0.2] (+/-) free CPE detected [a = 0.2] Western: 35: M, 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1) 160: 160 kDa (CPE:R1:R2); () band; (-) no band

PAGE 89

80 detoxification and excretion, free CPE miglit be expected. Whether the CPEreceptors in brain tissue were saturated is unknown, however, based on the in vitro studies this would appear hkely. Western immunoblots of tissues from animals administered 0.25 MLD showed CPE not as free CPE, but rather as a combination of small (CPE-Rl) and large (CPER1-R2) complex (Table 5.2). These results confirm Kokai-Kun and McClane's (1996) studies showing that the ELISA assay could detect CPE when it is exposed but sequestered in small or large complex. The statistical significance of the ELISA results are lower than perhaps expected (a = 0.2) however, the binding coefficient of free CPE to the primary antibody is higher than bound toxin. Thus, as noted previously in Chapters 3 and 4, the sensitivity of the ELISA is lower when it is detecting CPE as a mixture of complexes and free toxin. An interesting and as yet unexplained finding was that lung and liver tissues contained CPE predominantly as large complex. Sensitivity differences between the ELISA assay and the Western immunoblot have been discussed previously in Chapter 4. However, it is important to stress that while the assays are different, that is, the immunoblot specifically discerns between free CPE and CPE-complexed forms, the ELISA predominantly detects free CPE in tissue supematants (due to a higher binding coefficient). Thus the methods become complimentary.

PAGE 90

81 Time Study Since there was negligible tissue damage upon administration of nonlethal amounts of CPE no time studies were conducted at this level. ELISA results of administration of sublethal CPE levels (0.5 MLD) showed free toxin (significant at a = 0.05 level) in diymus, heart, and bowel tissues, and (significant at a = 0.1 level) in kidneys (Table 5.3). Brain, lung, liver, stomach and spleen as assayed by ELISA contained no free CPE. As discussed previously two patterns of responses were found at sublethal levels, recovery or death. Comparison of CPE timedistribution patterns between these two groups showed remarkable similarities except for the kidneys. Animals that died within 4 hours after toxin administration had fi-ee CPE, while kidneys of animals that recovered from the toxicosis had no significant fi-ee CPE. This is not unexpected since animals that recovered from the toxicosis obviously metabolized and excreted toxin via urine. Western immunoblots confirmed the presence of free CPE in all organs except in bowel and spleen tissues (Table 5.4). Brain, thymus, heart, lung, liver, and kidney tissues also had CPE in the small and large complex form. The stomach tissue contained only free CPE strongly suggesting that this organ has no CPE receptors. Free CPE in stomach tissue confirmed that intragastric administration was successful, and that the presence of CPE in lung tissue was the result of systemic distribution not faulty administration. The bowel had no CPE in any form (free or complexes) as detected by Western immunoblot. Possibly, the gross destruction of the small bowel

PAGE 91

82 Table 5.3 ELISA detection of unbound CPE in murine organ tissues after in vivo IG administration: time study. CPE binding time in hours Organ 4 hr. 72 hr. Brain Thymus + + + + Heart + + + + Lung Liver Kidney + Bowel + + + + Stomach Spleen — CPE binding conditions: administration of 0.5 MLD (150 /ig/250 fil) animals necropsied either at 4 or 72 hours controls were administered 250 /il of PBS Statistics: (-) = no free CPE detected [a = 0.1] (+) = free CPE detected [a = 0.1] (++) = free CPE detected [a = 0.05]

PAGE 92

83 Table 5.4 ELISA and Western immunoblot of CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after IG administration. CPE concentration and binding condition 0.5 MLD Organs Brain Thymus Heart Lung Liver Kidney Bowel Stomach Spleen ELISA Western + + + + + + + 35 85 160 CPE binding conditions: administration of 0.5 MLD (150 /ig/250 fil) animals necropsied at either 4 and 72 hours controls were administered 250 /il of PBS Statistics : (-) no free CPE detected [a = 0.2] (+) free CPE detected [a = 0.1] (++) free CPE detected [a = 0.05] Western: 35: M, 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1); 160: M, 160 kDa (CPE:R1:R2); () band; (-) no band

PAGE 93

84 resulted in the complete elimination of any toxin from this tissue. Desquamated tissue was simply either excreted or proteolytically digested before necropsy. Lethal CPE administration of 2 MLD also showed a bimodal time response where approximately 50 % of the animals died within the 15 minutes, referred to as "sudden" death, while the remaining 50% died a nonabrupt death requiring several hours for lethal outcome. ELISA results showed free CPE in thymus, liver, kidney, stomach, and bowel tissue (a = 0.05), and at a = 0.1 heart tissue (Table 5.5). Brain, lung, and spleen tissue had no detectable levels of free CPE. Higher free CPE levels were expected in liver and kidney tissues since these are the organs of detoxification and excretion. Another hypothesis could be proposed that, immediately after absorption in the small intestine CPE reaches the liver and kidney via portal vein compromising these organs. This in turn causes multiple organ failure, shock, and death. Liver and kidney failure could reduce blood pressure and blood flow to other organs. As blood is entrapped in failing organs less CPE reaches other tissues, thus not saturating their receptors. Detection of higher levels of free CPE in more organs than previously observed for lower toxin levels could simply be a function of toxin concentration. Simply stated, as more CPE becomes available a greater number of organ-tissues would reach their CPE receptor-saturation threshold. Western immunoblot results showed CPE both as free toxin, small and large complex in all organs except the stomach and spleen (Figures 5.1, 5.2 and Table 5.6).

PAGE 94

85 Table 5.5 ELISA detection of unbound CPE in murine organ tissues after in vivo IG administration: time study. Organ Brain Thymus Heart Lung Liver Kidney Bowel Stomach Spleen Sudden + + +/+ + + + Nonabrupt + + + + + + + + + + CPE binding conditions: administration of 2.0 MLD (500 /xg/250 fil) controls were administered 250 fil of PBS Sudden death: death within 15 minutes Nonabrupt death: several hours for lethal outcome Statistics : (-) = no free CPE detected [a = 0.2] (+/-) = free CPE detected [a = 0.2] (++) = free CPE detected [a = 0.05]

PAGE 95

86 Figure 5.1 SDS-PAGE of tissue supernatants: tissue protein analysis of mice administered 2 MLD IG of CPE. Tissue supernatants from: (C) control tissue homogenate (no CPE), (U) lung, (H) heart, (B) brain, (M) high molecular weight standard, (T) thymus, (V) liver, (K) kidney (S) spleen, (A) CPE.

PAGE 96

87 Figure 5.2 Western immunoblot of tissue supematants: mice administered 2 MLD IG of CPE. Toxin detected with polyclonal antisera to CPE. Tissue supematants from: (C) control tissue homogenate (no CPE), (U) lung, (H) heart, (B) brain, (M) high molecular weight standard, (T) thymus! (V) liver, (K) kidney, (S) spleen, (A) purified CPE.

PAGE 97

88 Table 5.6 ELISA, Western immunoblot and Vero cell detection of: CPE (free toxin) CPE:R1 (small complex) and CPE:R1:R2 (large complex) in murine organ tissues after in vivo IG administration. CPE concentration and binding conditions for 2MLD Organs Brain ELISA Western 35 85 160 Cell death Thymus + + Heart + /Lung Liver + + 0 Kidney + + 0 Bowel Stomach + + Spleen 0 CPE binding conditions: administration of 2.0 MLD (500 /ig/250 /il) controls were administered 250 n\ of PBS Statistics : (-) no free CPE detected [a = 0.2] ( + /-) free CPE detected [a = 0.2] (++) free CPE detected [a = 0.05] Western: 35: M, 35 kDa (free CPE); 85: M, 85 kDa (CPE:R1); 160: M, 160 kDa (CPE:R1:R2); () band; (-) no band Cell death: () in Vero cell assay, sample supernatant caused > 50% cell lysis and cell lysis was completely inhibited by neutralization with primary antibody to CPE. (0) no cell lysis

PAGE 98

89 Figure 5.3 Vero cell assay of murine tissues: (A) liver, (B) brain, and (C) bowel tissue supernatants. Animals were administered 2 MLD IG of CPE.

PAGE 99

90 The stomach had only free CPE confirming the ELISA results, and the spleen had no detectable CPE in any form (free or complexed). The Vero cell analyses concurred with both the ELISA and Western immunoblot results. Free and biologically active CPE was detected in thymus, heart, bowel, and stomach tissue (Table 5.6 and Figure 5.3). Brain, and lung tissue had free biologically active CPE, which confirmed the Western results. The Vero cell assay did not detect free biologically active CPE in liver, kidney or spleen tissues, and these negative results were consistent and reproducible for all samples tested. It would appear that unlike other organ tissues except the stomach, the spleen has no receptors for CPE. In summary, these findings confirmed the need to use all three assays (ELISA, Western and Vero cell) to elucidate the state of the CPE molecule within any tissue. Although free CPE (35 kDa) appeared to be present within a specific tissue based on the ELISA and Western assays, it did not necessarily mean that the toxin was biologically active. Indeed, free toxin was detected m Uver and kidney tissues by ELISA and Western but neither of these tissues contained biologically active CPE.

PAGE 100

CHAPTER 6 CONCLUSIONS AND SUMMARY Clostridium perfringens type A enterotoxin is a known causal agent of foodborne illness in man, and has also recently been implicated as a trigger in the sudden infant death syndrome (SIDS). The pathophysiological changes associated with either disease state appear very complex. Diarrhea and histopathological tissue damage to many organs within the body are a function of the cytotoxigenic and enterotoxigenic namre of the CPE molecule. However, other symptoms observed during the toxicosis suggest both neurotoxigenic (cholinergic) and superantigenic activities. The goal of this research was to better understand the nature of these CPE induced pathophysiological disturbances, by determining where CPE was specifically distributed during toxicosis. The murine model was chosen for this smdy since as a model it offered many unique genotypic and phenotypic advantages. In vitro tissue smdies showed CPE binding to cellular membrane protein receptors in brain, thymus, heart, lung, liver, kidney, and bowel. Liver, kidney and bowel tissues contained the highest number of CPE-receptors. Intraperitoneal CPE administration revealed that toxin was systemically and rapidly absorbed, being distributed throughout the body within minutes of administration. CPE induced pathophysiological effects were both concentration and time dependent. All organ tissues within the body except the spleen and stomach 91

PAGE 101

92 bound CPE, with the liver and kidneys binding the highest amount of toxin. Nonlethai levels of CPE were metabolized by liver and kidneys and excreted via urine. Sublethal CPE levels were both cytotoxic and enterotoxic, causing diarrhea and presenting symptoms associated with neurotoxicity. Lethal CPE levels acted as a neurotoxin and apparently as a superantigen, causing multiple organ failure, shock and rapid death. Intragastric administration of CPE induced responses that paralleled the intraperitoneal administration findings, except that higher levels of toxin were required to induce similar responses. Results showed that CPE again had three different modes of distribution, but four types of activity which were time and concentration dependent. Nonlethai CPE levels induced enterotoxigenic symptoms, and toxin was metabolized and cleared from the body via urine. Sublethal levels of CPE again induced enterotoxigenic, cytotoxigenic and neurotoxigenic symptoms. Lethal levels of CPE however, induced two patterns of death. First, a "sudden" death induced within minutes of toxin administration, and second, a nonabrupt death that required several hours for manifestation. Animals expressed symptoms of respiratory distress, shock and multiple organ failure, similar to the action of a neurotoxicsuperantigen. Based on the various distribution, time and lethality patterns observed in these smdies, the following models were developed (Figures 6.1a-d).

PAGE 102

93 Figure 6.1 Murine model for CPE distribution after IG administration (a) Nonlethai murine model. Animals recover from sequelae within 6-8 hours after IG CPE administration. (+) organs/tissues containing CPE. (b) Sublethal murine model. Animals may either recover from sequelae or die within 72 hours after IG CPE administration. (+) organs/tissues containing CPE.

PAGE 103

94 CPE Distribution Model: CPE Threshold Saturation Level ^ Nonietha) : 0.1 -0.5 MLD -Animals recover from sequelae. 4 spleen kidney lung + + I i 1 Recovery urine< + + + + Stomach bowel liver lieart blood CPE CPE Distribution Model: CPE Threshold Saturation Level Sublethal : 0.51.0 MLD Animals may recover from sequelae or die within 72 hours. + + + + stomach bowel liver heart blood

PAGE 104

95 Figure 6.1 continued. (c) Sudden death murine model. Animal dies within 15 minutes after IG CPE administration. (+) organs/tissues containing CPE. (d) Nonabrupt death murine model. Animal dies hours after IG CPE administration. (+) organs/tissues containing CPE.

PAGE 105

96 CPE Distribution IViodel: CPE Threshold Saturation Level Lethal dose : > 2.0 MLD Immediate outcome: shock and death within minutes. spleen kidney lung thymus + + + + i I I Multiple organ fellure and shock CPE + + Stomach bowel liver heart blood LI CPE Distribution Model: CPE Threshold Saturation Level Lethal dose : > 2.0 MLD Nonabrupt outcome: multiple organ failure, shock and death requiring several hours for manifestation. spleen kidney lung thymus + + 4, bfain ) i i CPE Multiple organ failure and shock urinei + + + + Stomach bowel liver heart blood

PAGE 106

97 The murine models may be extrapolated to the human response to CPE toxicosis. Nonlethal levels of CPE may cause a toxicosis similar to a self-limiting foodborne illness. Systemic absorption of sublethal levels of CPE may induce a neurotoxicosis, from which a healthy individual would likely recover, however, death could occur with this toxin concentration in an immunocompromised person, the young or elderly. Systemic absorption of lethal CPE levels will cause death in all persons. The observation of a "sudden" pattern of death after lethal ingestion was remarkable, and pivotal. This pattern mimics the suggested response of some at-risk infants to a cytotoxic-superantigen toxicosis, and supports a role for CPE as a trigger in some cases of SIDS. The studies outlined in this dissertation suggest that CPE also induces pathophysiological responses associated with neurotoxicity and superantigenicity. To confirm this hypothesis, fuhire work should be focussed on determining whether cytokines are induced during CPE toxicosis. Specifically, the induction of interferongamma, mmor necrosis factor-alpha, and interleukins IL-1, IL-2 and IL-6. Studies should also focus on which cytokines are induced in relation to specific tissues, whether induction is localized, and if the cytokine induction is CPE concentration and time dependent. Additionally, if CPE has a role as a trigger for death in SIDS which mimics the "sudden" death murine model (Figure 6. Ic) then, increased cytokine levels may be apparent in infants who have died from SIDS.

PAGE 107

REFERENCES Arnon SS (1983) Breast-feeding and toxigenic intestinal infections: missing links in SIDS. In: Tildon JT, Roeder LM, Steinschneider A (eds) Sudden Infant Death Syndrome. New York, Academic Press, pp 539-555 Baron S, Fleishmann WR, Coppenhaver DH, Nielsen DW, Klimpel GR, Stanton GJ, Hughes TK (1991) The interferons: mechanisms of action and clinical applications. JAMA 266:1375-1383 Bartholomew BA, Stringer MF, Watson GN, Gilbert RJ (1985) Development and application of an enzyme-linked immunosorbent assay for Clostridium perfringens type A enterotoxin. J Clin Pathol 38:222-228 Bartoszcze M, Nowakowska M, Roszkowski J, Martas J, Palec S, Wysmp (1990) Chinchilla deaths due to Clostridium perfringens A enterotoxin. Vet Rec 126:342-343 Bean NH, Griffm PM (1990) Foodbome disease outbreaks in the United States, 19731987; pathogens, vehicles and trends. J Food Protect 53:804-817 Beckwith JB (1988) Intrathoractic petechial hemorrhages: a clue to the mechanism of death in sudden infant death syndrome? Ann NY Acad Sci 533:37-47 Bettelheim LA, Golkwater PN, Dwyer BW, Bourne AJ, Smith DL (1990) Toxigenic Escherichia coli associated with sudden infant death syndrome. Scand J Infect Dis 22:467-476 Birkhead G, Vogt RL, Heun EM, Snyder JT, McClane BA (1988) Characterization of an outbreak of Clostridium perfringens food poisoning by quantitative fecal culture and fecal enterotoxin measurement, J Clin Microbiol 26:471-474 Blackwell CC, Weir DM, Busuttil A, Saadi AT, Essery DE, Raza MW, James VS, Mackenzie DA (1994) The role of infectious agents in sudden infant death syndrome! FEMS Immunol Med Microbiol 9:91-100 98

PAGE 108

99 Blackwell CC, Weir DM, Busuttil A (1995) Infectious agents, the inflammatory responses of infants and sudden infant death syndrome (SIDS). Mol Med Today 1:7278 Bouvier-Colle MH. Flahaut A, Messiah A, Jougla E, Hatton F (1989) Sudden infant death and immunization: an extensive epidemiological approach to the problem in France-winter 1986. Int J Epidemiol 18:121-126 Bowness P, Moss PAH, Tranter H, Bell JI, McMichael AJ (1992) Clostridium perfringens enterotoxin is a superatigen reactive with human T cell receptors V/36.9 and V^22. J Exp Med 176:893-896 Bruce K, Becker L (1992) Quantitation of medullary astrogliosis in SIDS. Pediatr Neurosurg 17:74-79 Brynestad S, Iwanejko LA, Stewart GSAB, Granum PE (1994) A complex array of Hpr consensus DNA recognition sequences proximal to the enterotoxin gene in Clostridium perfringens type A. Microbiol 140:97-104 Buck GM, Cookfair DL, Michalek AM, Nasca PC, Standfast SJ, Sever LE, Kramer AA (1988) Intrauterine growth retardation and risk of sudden infant death syndrome. Am J Epidemiol 129:874-884 Canada JC, Dorothy HS (1965) Effects of animal alimentary passage on the heat resistance of Clostridium perfringens. Appl Microbiol 13:788-792 Canard B, Saint-Joanis B, Cole ST (1992) Genomic diversity and organization of virulence genes in the pathogenic anaerobe Clostridium perfringens. Molec Microbiol 6:1421-1429 Centers for Disease Control (1994) Clostridium perfringens gastroenteritis associated with corned beef served at St. Patrick's Day meals; Ohio and Virginia 1993. Morbidity and Mortality Weekly Reports 43:137-144 Cerami A (1992) Inflammatory cytokines. Clin Immunol Immunopath 62:S2-S10 Chonmaitree T, Baron S (1991) Bacteria and viruses induced production of interferon in the cerebrospinal fluid of children with acute meningitis: a smdy of 57 cases and the review. Rev Infect Dis 13:1061-1065 Collier RJ, Kaplan DA (1985) Immunotoxins. Sci Amer 9:56-64

PAGE 109

100 Comillot E, Saint-Joanis B, Daube G, Datayama S, Granum PE, Canard B, Cole ST (1995) The enterotoxin gene {cpe) of Clostridium perfringens can be chromosomal or plasmid-borne. Molec Microbiol 15:639-647 Crowther JR (1995) ELISA: theory and practice. Tontowa, NJ, Humana Press Czeczulin JR, Collie RC, McClane BA (1996) Regulated expression of Clostridium perfringens enterotoxin in naturally cpe-negative type A, B and C isolates of Clostridium perfringens. Infect Immunol 64:3301-3309 Czeczulin JR, Hanna PC, McClane BA (1993) Cloning, nucleotide sequencing and expression of the Clostridium perfringens enterotoxin gene in Escherichia coli. Infect Immun 61:3429-3439 Daube G, Simmon P, and Kaeckeenbeeck A (1993) IS1151, an IS-like element of Clostridium perfringens. Nucleic Acids Res 21:352 Duncan CL, Strong DH (1968) Improved medium for sporulation of Clostridium perfringens. Appl Microbiol 16:82-89 Duncan CL, Strong DH (1969) Ileal loop fluid accumulation and production of diarrhea in rabits by cell-free products of Clostridium perfringens. J Bacteriol 100-8694 Duncan CL, Strong DH, Sebald M (1972) Sporulation and enterotoxin production by mutants of Clostridium perfringens. J Bacteriol 110:378-391 Fekety R, Kim K, Gatts DH, Browne RA, Cudmore MA, Silva J, Toshniwal R, Wilson K (1980) Studies on die epidemiology of antibiotic-associated Clostridium difficile colitis. Am J Clm Nutr 33:2527-2532 Fleischer B (1994) Superantigens. APMIS 102:3-12 Gaskill ME, Khan SA (1986) Regulation of the enterotoxin B gene in Staphylococcus aureus. J Biol Chem 263:6276-6280 Gillan JE, Curran C, O'Reilly E, Cahalane SF, Unwin AR (1989) Abnormal patterns of pulmonary neuroendocrine cells in victims of sudden infant death syndrome Pediatrics 84:828-834 Granum PE, Harbitz O (1985) A circular dichroism study of the enterotoxin from Clostridium perfringens type A. J Food Biochem 9:137-146

PAGE 110

101 Granum PE, Richardson M (1991) Chymotrypsin treatment increases the activity of Clostridium perfringens enterotoxin. Toxicon 29:896-900 Granum PE, Shitaker JR, Skjelkvale R (1981) Trypsin activation of enterotoxin from Clostridium perfringens type A. Biochim Biophys Acta 668:325-332 Granum PE, Stewart G (1992) Molecular biology of Clostridium perfringens enterotoxin. In: Sebald M (ed) Genetics and Molecular Biology of Anaerobic Bacteria. New York, SpringierVerlag, pp 235-247 Granum PE, Whitaker JR (1980) Improved method for the purification of enterotoxin Clostridium perfringens type A. Appl Environment Microbiol 39:1120-1122 Granum PE, Whitaker JR, Skjelkvale R.(1980) Trypsin activation of enterotoxin from Clostridium perfringens type A. Biochim Biophys Acta 668:325-332 Grether JK, Schulman J (1989) Sudden infant death syndrome and birthweight. J Pediatr 114:561-567 Guilian GG, Gilbert EF, Moss RL (1987) Elevated fetal hemoglobin levels in sudden infant death syndrome. New Engl J Med 316:1122-1126 Gundieroth WG (1989) Theories of cardiovascular causes in sudden infant death syndrome. JACC 14:443-447 Hanna PC, McClane BA (1991) A recombinant C-terminal toxin fragment provides evidence that membrane insertion is important for Clostridium perfringens enterotoxin cytotoxicity. Molec Microbiol 5:225-230 Hanna PC, Mietzner TA, Schoolnik GK, McClane Ba (1991) Localization of the receptor-binding region of Clostridium perfringens enterotoxin utilizing cloned toxin fragments and synthetic peptides. J Biol Chem 266:11037-11043 Hanna PC, Wnek AP, McClane BA (1989) Molecular cloning of the 3'half of the Clostridium perfringens enterotoxin gene and demonstration that this region encodes receptor-binding activity. J Bacteriol 171:6815-6820 Hanna PC, Wiekowski EU, Mietzner TA, McClane BA (1992) Mapping functional regions of Clostridium perfringens type A enterotoxin. Infect Immun 60:2110-2114 Hauschild AH, Niilo L, Dorward WJ (1971) The role of enterotoxin in Clostridium perfringens type A enteritis. Can J Microbiol 17:987-991

PAGE 111

102 Hauser D, Eklund MW, Boquet P, Popoff MR (1994) Organization of the botulinum neurotoxin CI gene and its associated non-toxin protein genes in Clostridium botulinum C468. Mol Gen Genet 243:631-640 Hobbs BC (1979) Clostridium perfringens gastroenteritis. In: Riemann H and Bryan FL (eds) Food Borne Infections and Intoxications. New Yorlc, Academic Press, pp 131-167 Hobbs BC, Smith SF, Oakley CT, Warrack GH, Cruickshank JF (1953) Clostridium welchii food poisoning. J Hyg 51:74-101 Hoffman JH, Damns K, Hillman L, Krongrad P (1988) Risk factors for SIDS: results of the National Institute of Child Health an Human Development Cooperative Epidemiological Study of Sudden Infant Death Syndrome. Arch N Y Acad Sci 533:13-30 Hoffman JH, Jehy C, Damns K, Pakter J, Peterson D, van Belle G, Hasselmayer E (1987) Diphtheria-tetanus-pertussis immunization and sudden infant death: results of the National Institute of Child Health and Human Development Cooperative Epidemiological Study of Sudden Infant Death Syndrome. Pediatrics 79:598-611 Hollander N (1988) Beta-endorphin in the brainstem, pimitary, and spinal fluid of infants at autopsy: relation to sudden infant death syndrome. Forensic Sci Int 38:6774 Horiguchi Y, Akai T, Sakaguchi G (1987) Isolation and function of a Clostridium perfringens enterotoxin fragment. Infect Immun 55:2912-2915 Howatson AG (1992) Viral infection and a interferon in SIDS. J Clin Pathol 45:25-28 Hulkower KI, Wnek AP, McClane BA (1989) Evidence that alterations in small molecule permeability are involved in the Clostridium perfringens type-A enterotoxininduced inhibition of macromolecular synthesis in Vero cells. J Cell Physiol 140-498504 Ijzermans JNM, Maquet RL (1989) Interferon-gamma: a review. Immunobiol 179:456-473 Jakeman K, Rushton D, Smith H, Sweet C (1991) Exacerbation of bacterial toxicity to infant ferrets by influenza virus: possible role in SIDS. J Infect Dis 163:35-40

PAGE 112

103 Janssen MT, Put HM, Nout MJ (1996) Natural toxins. In: de Vries J (ed) Food Safety and Toxicity. Boca Raton, FL, CRC Press, pp 23-24 Johnson HM, Russell JK, Ponzer CH (1992) Superantigens in human disease. Set Am 4:92-101 Jones AM, Weston JT (1976) The examination of the sudden infant death syndrome infant: investigative and autopsy protocols. J Forensic Sci 21:833-841 Jones CL, Khan SA (1986) Nucleotide sequence of the enterotoxin B gene from Staphylococcus aureus. J Bacteriol 162:29-33 Kariks J (1989) Diaphragmatic muscle fibre necrosis in SIDS. Forensic Sci Int 43:281-291 Katayama S, Dupuy B, Daube G, China B, Cole ST (1996) Genome mapping of Clostridium perfringens strains widi \-Ceu\ shows many virulence genes to be plasmid borne. Mol Gen Genet 251:720-726 Klein E (1895) Ueber einen pathogenen anaeroben Darmbacillus, Bacillus enteritidis sporogenes. Zentralbl Bakteriol Parasitenkd Infektionskr Hyg 18:737 Knox R, McDonald EK (1943) Outbreaks of food poisoning in certain Leicester institutions. Med Officer 69:21-22 Kokai-Kun JF, Songer JG, Czeczulin JR, Chen F, McClane BA (1994) Comparison of western immunoblots and gene detection assays for identification of potentially enterotoxigenic isolates of Clostridium perfringens. J Clin Microbiol 32:2533-2539 Kokai-Kun JF, McClane BA (1996) Evidence that a region(s) of the Clostridium perfringens enterotoxin molecule remains exposed on the external surface of the mammalian plasma membrane when the toxin is sequestered in small or large complexes. Infect Immun 64:1020-1025 Krous HF (1984) The microscopic distribution of intrathoracic petechiae in sudden infant death syndrome. Arch Pathol Lab Med 108:77-79 Labbe RG (1989) Clostridium perfringens. In: Doyle MP (ed) Foodbome Bacterial Pathogens. New York, Marcel Dekker, Inc., pp 192-234

PAGE 113

104 Labbe RG, Duncan CL (1977) Evidence for stable messenger ribonucleic acid during sporulation and enteroioxin synthesis by Clostridium perfringens type A. J Bacteriol 129:843-849 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227:680-685 Lindsay J A (1988) The effect of a Clostridium perfringens 8-6 enterotoxin on viability and macromolecular synthesis in Vero cells. Biochem Biophys Res Comm 151:13711377 Lindsay J A (1996) Clostridium perfringens type A enterotoxin (CPE): more than just explosive diarrhea. Crit Rev Microbiol 22 (4): 257-277 Lindsay JA, Dennison JD (1986a) Histopathological effect of Clostridium perfringens 8-6 enterotoxin on rabbit intestine. Curr Microbiol 13:61-66 Lindsay JA, Dennison JD (1986b) A scanning electron microscope study of the effect of an enterotoxin from Clostridium perfringens 8-6 on mice of different ages. J Gen Microbiol 132:2839-2898 Lindsay JA, Johnson H, Wallace FM, Soos J (1994) Can superatingens trigger sudden infant death? Med Hypoth 43:81-85 Lindsay JA, Mach A, Wilkinson M, Martin L, Wallace L, Keller A, Wojciechowski L (1992) A role for Clostridium perfringens and its enterotoxin(s) in SIDS. A microbiological, toxicological and histopathological study. 2nd International SIDS Conference, Sydney, Australia, 65a, 120 Lindsay JA, Mach A, Wilkinson M, Martin L, Wallace L, Keller A, Wojciechowski L (1993) Clostridium perfringens type A cytotoxic-enterotoxin(s) as triggers for death in Sudden Infant Death Syndrome: Development of a toxico-infection hypothesis. Curr Microbiol 27:51-59 Lindsay JA, Sleigh RW, Ghitgas C, Davenport JB (1985) Purification and properties of an enterotoxin from a coatless spore mutant of Clostridium perfringens type A Eur J Biochem 149:287-293 Lowry OH, Rosebrough NJ, Parr AL, Randall RJ (1951) Protein measurement with the folin phenol reagent. J Biol Chem 193: 265-275

PAGE 114

105 Mach AS, Lindsay J A (1994) Activation of Clostridium perfringens cytotoxic enterotoxins in vivo and in vitro: role in triggers for SIDS. Curr Microbiol 28:261267 Mach AS, Lindsay J A (1997) Clostridium perfringens enterotoxin stimulates in vitro nitric oxide production in the presence of cytokines. 97th ASM General Meeting, Miami Beach, Florida, Abstract P-025, pp 441 McClane BA (1992) Clostridium perfringens enterotoxin: strucmre, action and detection. J Food Safety 12:237-252 McClane BA (1994) Clostridium perfringens enterotoxin acts by producing small molecule permeability alterations in plasma membranes. Toxicology 87:43-67 McClane BA (1997) Clostridium perfringens. In: Doyle M, Montville T, Beuchart L (eds) Food Safety. Washington, DC, ASM Press, pp 305-326 McClane BA, Hanna PC, Wnek A (1988a) Clostridium perfringens type A enterotoxin. Microb Pathog 4:317-323 McClane BA, McDonel JL (1979) The effects of Clostridium perfringens enterotoxin on morphology, viability and macromolecular synthesis in Vero cells. J Cell Physiol 99:191-200 McClane BA, Wnek AP (1990) Studies of Clostridium perfringens enterotoxin action at different temperatures demonstrate a correlation between complex formation and cytotoxicity. Infect Immun 58:3109-3115 McClane BA, Wnek AP, Hulkower KI, Hanna PC (1988b) Divalent cation involvement in the action of Clostridium perfringens type A enterotoxin. J Biol Chem 263:2423-2435 McClane B, Wnek A, Whitaker-Dowling P (1987) Interferon pretreatment enhances the sensitivity of Vero cells to Clostridium perfringens type A enterotoxin. Microbial Pathog 3:195-206 McClung LS (1945) Human food poisoning due to growth of Clostridium perfringens (C. welchii) in freshly cooked chickens. J Bacteriol 50:229-231 McDonel JL (1980) Binding of Clostridium perfringens '"I-enterotoxin to rabbit intestinal cells. Biochem 21:4801-4807

PAGE 115

106 McDonel JL (1986) Toxins of Clostridium perfringens types A, B, C, D and E. In: Doraer F, Drews H (eds) Pharmacology of Bacterial Toxins. Oxford, Pergamon Press, pp 477-517 McDonel JL, Duncan CL (1975) Histopathological effects of Clostridium perfringens enterotoxin in the rabbit ileum. Infect Immun 12:1214-1218 McDonel JL, McClane BA (1979) Binding vs. biological activity of Clostridium perfringens enterotoxin in Vero cells. Biochem Biophys Res Commun 87:497-504 MacLennan JD (1962) The histotoxic clostridial infections of man. Bact Rev 26(2): 177-276 Melville SB, Labbe RG, Sonenshein AL (1994) Expression from the Clostridium perfringens cpe promoter in C. perfringens and Bacillus subtilis. Infect Immun 62:5550-5558 Mietzner RA, Kokai-Kun JF, Hanna PC, McClane BA (1992) A conjugated synthetic peptide corresponding to the C-terminal region of Clostridium perfringens type A enterotoxin elicits an enterotoxin-neutralizing antibody response in mice. Infect Immun 60:3947-3951 Morris JA (1987) Hypothesis: common bacterial toxins are a possible cause of the sudden infant deadi syndrome. Med Hypoth 22:211-222 Murrell TGC, Ingham B, Moss J, Taylor W (1987) A hypothesis concerning Clostridium perfringens type A enterotoxin and SIDS. Med Hypoth 22:401-413 Murrell TGC, Murrell WG, Lindsay J A (1994) Sudden infant death syndrome (SIDS): are common bacterial toxins responsible, and do they have a vaccine potential. Vaccine 12:365-368 Murrell TGC, Walker PD (1991) The pigbel story Papua New Guinea. Trends R Soc Trop Med 85:119-122 Murrell WG (1989) Clostridium perfringens. In: Buckle K (ed) Foodbome Microorganisms of Public Health Significance. Sydney, Australia, AIFST, pp 209-232 Murrell WG, Stewart B, O'Neill C, Siarakas S, Kariks S (1992) The role of enterotoxigenic bacteria in SIDS. Abstract of the 2nd International SIDS Conference, Sydney, Australia 1992:52

PAGE 116

107 Murrell WG, Stewart B, O'Neill C, Siarakas S, Kariks S (1993) Enterotoxigenic bacteria in SIDS. J Med Micro 39:114-127 Niilo L (1975) Measurement of biological activities of purified and crude enterotoxin of Clostridium perfringens. Infec Immun 12:440-442 Niilo L (1980) Clostridium perfringens in animal disease: a review of current knowledge. Can Vet J 21:141-148 Pastan IH, Pai LH, Brinkman U, Fitzgerald DJ (1995) Recombinant toxins: new therapeutic agents for cancer. Ann NY Acad Sci 745:345-354 Pestka S, Langer J, Zoon K, Samuel C (1987) Interferons and their actions. Ann Rev Biochem 56:27-72 Ray CG, Hebestreit NM (1971) Studies of the sudden infant death syndrome in King county, Washington, II. Attempts to demonstrate evidence of viremia. Pediatrics 49:79-82 Rood JI, Cole ST (1991) Molecular genetics and pathogenesis of Clostridium perfringens. Microbiol Rev 55:621-648 Saito M (1990) Production of enterotoxin by Clostridium perfringens derived from humans, animals, foods and the natural environment in Japan. J Food Protection 53:115-118 Seto DS, Carver DH (1978) Circulating interferon in sudden infant death syndrome. Proc Soc Exp Biol Med 157:378-380 Sherman S, Klein E, McClane BA (1994) Clostridium perfringens type A enterotoxin induces tissue damage and fluid accumulation in rabbit ileum. J Diarrheal Dis Res 12:200-207 Skjelkvale R, Tolleshaug H, Jarmund T (1980) Binding of enterotoxin from Clostridium pefringens type A to liver cells in vivo and in vitro. Acta Path Microbiol Scan 88:95-102 Skjelkvale R, Uemura T (1977) Experimental diarrhea in human volunteers following oral administration of Clostridium perfringens enterotoxin. J Appl Bacteriol 46-281286

PAGE 117

108 Smith LDS (1979) Virulence factors of Clostridium perfringens. Rev Infect Dis 1:254260 Smith WP, McDonei JL (1980) Clostridium perfringens type A: in vitro systems for sporulation and enierotoxin synthesis. J Bacteriol 144:306-311 Songer JG (1996) Clostridial enteric diseases of domestic animals. Clin Microbiol Rev 9:216-234 Spika JS, Shaffer N, Hargrett-Bean N, Collin S, MacDonald KL, Blake PA (1989) Risk factors for infant bomlism in the United States. Am J Dis Child 143:828-832 Staton AN (1980) Is overheating a factor in some unexpected infant deaths? Lancet 3:1054-1057 Stephens T (1990) Animal models give clues to SIDS. J NIH Res 1:73-80 Sterne M (1981) Clostridial infections. Br Vet J 137:443-454 Stringer MF (1985) Clostridium perfringens type A food poisoning. In: Poriello SP (ed) Clostridia in Gastrointestinal Disease. Boca Raton, FL, CRC Press, pp 117-144 Sugii S, Horiguchi Y (1988) Identification and isolation of die binding substance for Clostridium perfringens enterotoxin on Vero cells. FEMS Microbiol Let 52:85-90 Sugimoto W, Takagi M, Ozutsumi K, Harada S, Matsuda M (1988) Enterotoxin of Clostridium perfringens type A forms ion-permeable channels in a Upid bilayer membrane. Biochem Biophys Res Commun 156:551-556 Sugiyama H (1986) Mouse models for infant bomlism. In: Zak O, Sande MA (eds) Experimental Models in Antimicrobial Chemotherapy. New York, NY, Academic Press, pp 73-91 Thach BT, Davies AM, Koenig JS (1988) Pathophysiology of sudden upper airway obstruction in sleeping infants and its relevance for SIDS. Ann N Y Acad Sci 533:314-328 Todd ECD (1989) Cost of acute bacterial foodborne disease in Canada and the United States. Int J Food Microbiol 9:313-326 Tracey KJ, Cerami A (1993) Tumor necrosis factor, other cytokines and disease Annu Rev Cell Biol 9:317-343

PAGE 118

109 Tsai CC, Riemann HP (1975a) Food poisoning signs in mice induced orally by Clostridium perfringens type A enteroioxin. J Formosan Med Assoc 74:310-315 Tsai CC, Riemann HP (1975b) Oral infection and food poisoning in mice by enterotoxigenic Clostridium perfringens type A. J Formosan Med Assoc 74:361-371 Uemura T, Sakaguchi G, Itoh T, Okazawa K, Sakai S (1975) Experimental diarrhea in cynomolgus monkeys by oral administration with Clostridium perfringens type A viable cells or enterotoxin. Japan J Med Sci Biol 28:165-177 Valdes-Dapena M (1980) Sudden infant death syndrome: a review of the medical literature 1974-79. Pediatrics 66:597-614 Valdes-Dapena M (1983) The morphology of the sudden infant death syndrome: an overview. In: Tildon JT, Roeder LM, Steinschneider A (eds) Sudden Infant Death Syndrome. New York, Academic Press, pp 169-182 Van DammeJongsten M, Weners K, Notermans S (1989) Cloning and sequencing of the Clostridium perfringens enterotoxin gene. Antonie von Leeuwnhoek J Microbiol 56:181-190 Verrier RL, Kirby DA (1988) Sleep and cardiac arrhythmias. Ann N Y Acad Sci 533:238-251 Wallace FM, Keller AM, Lindsay J A (1997) Cytokine response after administration of Clostridium perfringens enterotoxin in the murine model. 97th ASM General Meeting, Miami Beach, Florida, Abstract P-028, pp 441 Weiss KF, Strong DH, Groom RA (1965) Mice and monkeys as assays for Clostridium perfringens food poisoning. Appl Microbiol 14:479-485 Whitaker JR, Granum PE (1980) The role of amino groups in the biological and antigenic activities of Clostridium perfringens type A enterotoxin. J Food Biochem 4:201-217 Wieckowski EU, Wnek AP, McCIane BA (1994) Evidence that an ~50-kDa mammalian plasma membrane protein with receptor-like properties mediates the amphiphilicity of specifically bound Clostridium perfringens enterotoxin. J Biol Chem 269:10838-10848

PAGE 119

110 Wilkinson MA (1992) The sudden infant death syndrome in florida: an epidemiological, pathological and microbiological study. M.S. Thesis, University of Florida Willinger M (1989) SIDS: a challenge. J NIH Res 1:73-80 Wnek AP, McClane BA (1983) Identification of a 50,000 Mr protein from rabbit brush border membranes that bind Clostridium perfringens enterotoxin. Biochem Biophys Res Com 112:1099-1105 Wnek AP, McClane BA (1986) Comparison of receptors for Clostridium perfringens type A and cholera enterotoxins in isolated rabbit intestinal brush border membranes. Microb Pathog 1:89-100 Wnek AP, McClane BA (1989) Preliminary evidence that Clostridium perfringens type A enterotoxin is present in a 160,000-Mr complex in mammalian membranes. Infect Immun 57:574-581 Wnek AP, Stouse RJ, McClane BA (1985) Production and characterization of monoclonal antibodies against Clostridium perfringens type A enterotoxin. Infect Immun 50:442-448 Wojciechowski LM (1995) The cloning and characterization of the Clostridium perfringens type A 8-6 enterotoxin gene. Ph.D. Dissertation, University of Florida Zylke J (1989) Sudden infant death syndrome: resurgent research offers hope. JAMA 262:1565-1566

PAGE 120

BIOGRAPHICAL SKETCH Andreas M. Keller was bom in Numberg, Germany. In 1967 his parents immigrated to Brazil, and he was raised in Sao Paulo and Rio de Janeiro. He graduated with a B.S. degree in agronomy in 1988 from the Federal Rural University of Rio de Janeiro, majoring in food technology. He received his M.S. degree in foods and human nutrition from Eastern Michigan University in 1990. In fall of 1990 he enrolled in the Food Science and Human Nutrition Department's doctoral program at the University of Florida. During his tenure at UF he worked for the University Division of Housing as Resident Manager for Maguire Village, University Village South and Tanglewood. He was an active member of Student Government servmg the UF students as a Family Housing Senator, Chairman of Mayor's Council and as Smdent Senate President, and was awarded two University of Florida Presidential Recognitions, and an Outstanding Student Senator award for his outstanding achievements. Upon completion of his degree he plans to pursue a career as a food microbiologist with an international food company. Ill

PAGE 121

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate. ^ scope and quality, as a dissertation for the degree of Doctoi) of Philo§ James A. Lindsay, Chair y Professor of Food Science and Human Nutrition I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ULQi lelas L. Archer Dougl Professor of Food Science and Human Nutrition I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quahty, as a dissertation for the degree of Doctor of Philosophy. in P 0'l^PP.fp ( Sean F. O'Keefe Associate Professor of Food Science and Human Nutrition I certify that I have read this study and diat in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quaUty, as a dissertation for the degree of Doctor of Philosophy. Associate Professor of Family Youth and Community Sciences

PAGE 122

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fiilly adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ramon C. Littell Professor of Statistics This dissertation was submitted to the Graduate Faculty of the College of Agriculmre and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. — J May, 1997 /Clc^ > .^ Dean, College of Agriciriture Dean, Graduate School