‘ . I. ‘0
u I. |..u 1. o 9 r 00.. ‘II. . J1: illlxll ‘IILF
I. II. I. II I
I II‘ - I. n.' . t . II. . .
'. c. I}. . - Iv III bl- ... I
Q I D I. ‘0. “I O O 5-. .11 ‘Il‘ I ' I. 1.
I ‘- II ' i . |
u .. : -. u . a. -- - u u .r - .- -nUn»...
I I1
. -o \ a. A
.l .l.

I o affinih I!”
u I ol 3"

 

 

1

 

31.13::

(”mm

4 EMSMQ
1 Unhmy

\

M»

k_“_ _____ __ -5 ,

 

This is to certify that the

dissertation entitled

THE RELATIONSHIP BETWEEN TISSUE DISTRIBUTION
AND SELECTED TOXIC EFFECTS OF ENDOTOXIN IN
NORMAL AND GALACTOSAMINE SENSITIZED MICE

presented by

Bryon W. Petschow

has been accepted towards fulfillment

of the requirements for

Ph.D. Microbiology & Public Health

degree in

 

 

@th 07/4“, W/ D

Maid/protessor

 

Dam September 12, 1983

 

M.S’L"15em 'IUINVIuIH'r‘.‘h'nun Aqua! Opportunity lmu’lutmn 0 12771

 

 

MSU

LIBRARIES
“3—...

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES wiII
be charged if book is
returned after the date
stamped below.

 

 

 

 

 

 

 

— [new

x ,-
., .1

 

 

 

 

THE RELATIONSHIP BETWEEN TISSUE DISTRIBUTION
AND SELECTED TOXIC EFFECTS OF ENDOTOXIN IN

NORMAL AND GALACTOSAMINE SENSITIZED MICE

By

Bryon W. Petschow

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Microbiology and Public Health

1983

ABSTRACT

THE RELATIONSHIP BETWEEN TISSUE DISTRIBUTION
AND SELECTED TOXIC EFFECTS OF ENDOTOXIN IN
NORMAL AND GALACTOSAMINE SENSITIZED MICE

By

Bryon W. Petschow

The relationship between the quantitative localization of endotoxin
among host organs and selected toxic effects of endotoxin was studied in
normal and galactosamine sensitized mice.

The majority of intravenously (iv) or intraperitoneally (ip) injected
endotoxin became associated with the liver and spleen of normal mice by one
hour. By contrast, subcutaneously (sc) administered endotoxin was about
50—fold less concentrated in liver while only about 2—fold higher in mean
50% lethal dose when compared to iv or ip injected endotoxin. Although less
than 15% of so injected endotoxin became associated with the deep tissues
of the host, sc injected mice exhibited a variety of symptoms typically
associated with endotoxin poisoning.

Treatment of mice with galactosamine (galN) caused a 1000-fold decrease
in the LD50 of endotoxin. In addition, mice receiving galN and as little as
0.1ug of endotoxin exhibited marked liver damage as evidenced by increased
serum ornithine carbamyltransferase (OCT) activity. Galactosamine treatment
did not alter the organ distribution of iv injected endotoxin or enhance the
uptake of endotoxin by liver parenchymal cells. Furthermore, endotoxin
localized in livers of normal mice was primarily associated with Kupffer cells.

Serum from mice that were injected with the LD50 of endotoxin caused
significant increases in serum OCT activity when transferred to GalN—treated

mice. The activity of endotoxemic serum was amplified by pretreating donor

Bryon W. Petschow

mice with Corynebacterium parvum before injecting endotoxin. C.parvum/

 

endotoxin serum decreased survival and increased serum OCT levels when
injected into galN-treated but not normal mice. Serum from mice that
received either C.parvum or endotoxin (0.1 LDSO) alone was inactive when
injected into galN—treated mice.

C.parvum-primed mice were sensitized to the lethal effects of endotoxin
to about the same extent as galN—treated mice. Whereas C.parvum/endotoxin
serum was lethal and capable of eliciting elevated serum OCT levels in galN—
treated mice, neither response occurred following transfer of C.parvum/
endotoxin serum to C.parvum-primed mice. Likewise, donor serum obtained
from galN-treated mice given the same dose of endotoxin (0.1 LDSO) was not
lethal or capable of elevating serum OCT levels when transferred to C.parvum—
primed or galN—treated mice, thereby implying different mechanisms of

endotoxin sensitization.

DEDICATION

This dissertation is dedicated to Debbie
for her constant encouragement, unending support,
and patient understanding. Your presence and
willingness to listen throughout these years will
always be remembered for making the difficult times

bearable and the good times unforgettable.

ii

ACKNOWLEDGMENTS

I would like to thank Dr. Robert J. Moon for serving as my research
advisor, for review of this manuscript, and for your guidance throughout
this study and concern for my professional growth.

Special thanks is extended to Dr. Robert Leunk for his friendship
and assistance throughout this study. Thanks also to Ruth Vrable for
her friendship, excellent technical assistance, and ability to keep the
lab operating smoothly.

I also wish to thank Dr. E. Werner, Dr. H. Tavakoli, Paul Rota,
Dace Valduss, Ellen Keitelman, and Lee Schwocho for their friendship,
discussion, and encouragement.

Finally, I would like to thank my brother, Kevin W. Petschow, for
his support, friendship, and encouragement and for providing a refuge

on occasional weekends during the past several years.

iii

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . .
LIST OF FIGURES O O O I O O O O O O O O O O
INTRODUCTI ON C O O O 0 I O O O O O O O O 0

LITERATURE REVIEW. . . . . . . . . . . . .
Significance of Endotoxin Research .
Chemical Structure of Endotoxin. . .
Radioisotopic Labeling of Endotoxin.

Distribution of Endotoxin in vivo. .

Pathophysiologic Changes Induced by Endotoxin.

Mediation of Endotoxin Activity. . .

Galactosamine Induced Enhancement of Endotoxin

Susceptibility.

Experimental Models for Enhancing Endotoxin Responsiveness . .

Correlation Between Endotoxemia and Clinical Liver Disease . .

LITERATURE CITED . o . . . . . . . . . . .

ARTICLE I.

Lack Of Correlation Between Tissue Localization

And Biological Responses Of Mice To Endotoxin. . . . . .

Article II.

Potentiation Of Endotoxin Sensitivity By Galactosamine

Is Not Associated With Increased Uptake Of Endotoxin

By Hepatocytes . . . . . . . .

ARTICLE III. Involvement Of Humoral Factors In The Potentiation

Of Endotoxin Sensitivity By Galactosamine. . . . . . . .

APPENDIX 0 O O O O O O O O O O O O O O O O

SWARY O O O O O O O O O O O O O O O O O 0

iv

. vii

O
mO‘UIJ-‘b

. 15
. 20
. 25
. 29
. 32

.102
.130

.135

LIST OF TABLES

TABLE

1.

ARTICLE I.

Tissue distribution of subcutaneously injected 51Cr-labeled

endotoxin in normal mice . . . . . . . . . . . . . . .

Detection of endotoxin by Limulus assay in livers of normal

mice or mice injected sc or iv with 100ug of endotoxin .

Effect of subcutaneously administered endotoxin on induction

of liver PEPCK by hydrocortisone . . . . . . . . . . .
ARTICLE II.

Organ distribution of intravenously injected 51Cr—endotoxin

in normal and galactosamine-treated mice . . . . . . .

In vivo localization of 51Cr-endotoxin in liver parenchymal

and nonparenchymal cells of normal and galactosamine-

treated mice 0 o o o o o o o o o a a o o o o o o o o o

In vivo localization of 51Cr-endotoxin in nonparenchymal

cells enriched for specific cell types by centrifugal

elutriation. o o o o o o o o o o o o o o o o o o o o 0

Effect of galactosamine on the uptake of 51Cr-endotoxin

by normal mouse liver parenchymal cells in vitro . . .

ARTICLE III.

 

Comparison of the LD50 of endotoxin in normal and

galactosamine-treated mice . . . . . . . o . . . . . .

Serum OCT activity in galN-treated mice after intravenous

injection with various doses of endotoxin. o o o o o .

PAGE

57

60

64

88

89

91

92

109

110

LIST OF TABLES (cont'd.)

TABLE PAGE

ARTICLE III. (cont'd.)

 

3. Elevation of serum OCT activity in galN-treated mice

given serum from endotoxin challenged mice . . . . . . . . . 112

4. Ability of C.parvum/endotoxin serum to elicit increased

serum OCT activity in galN-treated mice. . . . . . . . . . . 114

5. Survival and serum OCT activity in galN-treated mice given

various dilutions of C.parvum/endotoxin serum. . . . . . . . 115

6. Stability of C.parvum/endotoxin serum after heating

or enzyme digestion. . . . . . . . . . . . . . . . . . . . . 116

7. Comparison of serum OCT activity in galN- or C.parvum-
treated mice after receiving C.parvum/endotoxin

or galN/endotoxin serum. . . . . . . . . . . . . . . . . . . 118
APPENDIX

1. Cellular composition of nonparenchymal cell fractions

produced by centrifugal elutriation. . . . . . . . . . . . . 133

vi

LIST OF FIGURES

FIGURE PAGE
ARTICLE I.

1. Whole animal organ distribution of 51Cr-labeled

endotoxin by 1 hour after injection . . . . . . . . . . . . . 56

2. Total radioactivity recovered from tissues of mice
following subcutaneous injection of 51Cr-endotoxin

51

or N32 Cr04 o o o o o o o o o o o o o o o o o o o o o o o 0 59

3. Liver glycogen levels in fasted mice following

subcutaneous injection with endotoxin or saline . . . . . . . 62

4. Mean decrease in body temperature of mice following

subcutaneous injection with various doses of endotoxin. . . . 66

ARTICLE II.

 

1. Lethal effects of endotoxin in galactosamine-

treated mice 0 I O O O O O O O O O I O O O O O O O O O O O O 85

2. Increase in serum OCT activity in mice following injection

with galactosamine and endotoxin. . . . . . . . . . . . . . . 86

ARTICLE III.

 

1. Proposed mechanisms of endotoxin sensitization and their

relationship to endotoxin toxicity. . . . . . . . . . . . . . 122

vii

INTRODUCTION

The biologic alterations elicited by gram-negative bacterial endotoxin
in susceptible animals are numerous and diverse. These include alterations
in metabolic processes, pyrogenicity, cardiovascular changes, development
of tolerance, immunogenicity, adjuvanticity, tumor necrosis, enhancement
of nonspecific resistance, shock, and lethality. In spite of many years
of intensive investigation the mechanisms underlying the toxic and lethal
responses to endotoxin remain unclear.

Several investigators have attempted to link the pathophysiologic
consequences of endotoxin-poisoning with its ability to induce a number
of host metabolic alterations in liver parenchymal cells. Such changes
include depletion of carbohydrate reserves (7,81), inhibition of hormonal
induction of certain hepatic enzymes (9,76,77,106), elevations in serum
transaminase levels (61,138), inhibition of gluconeogenesis (63,112),
labilization of lysosomal membranes (54,135), and impairment of mitochondrial
energy production (45,48,80). Although previous reports from this
laboratory (139) as well as others (102,103,108,137) indicate that
endotoxin becomes associated with liver parenchymal cells both in vitro
and in vivo, whether endotoxin affects these cells directly or via humoral
mediators has not been satisfactorily established.

Endotoxin research during the past decade has focused primarily
on the ability of endotoxin to activate a number of cell types including
tissue macrophages, lymphocytes, and neutrophils. Evidence has been

assembled to suggest that endotoxin acts to stimulate specific cell types

to produce and/or release a variety of soluble mediators which are
believed to be responsible for many of the biological effects of endotoxin.
Host cascade systems such as the complement, kinin, and coagulation
pathways are also activated by endotoxin. Because most studies describing
these mediated events employ in vitro systems to circumvent the
complexities of in vivo studies, it has been difficult to delineate how
these various mediators and cells interact in vivo to produce the toxic
and lethal effects of endotoxin.

Galactosamine (galN) is a liver-specific toxin that causes liver
cell necrosis (57,69) and increases the susceptibility of a variety of
experimental animals to the lethal effects of endotoxin (40). The well-
characterized biochemical changes induced by galN in hepatocytes leads
to structural and functional alterations in hepatocyte plasma membranes
(4,28,31), making this an ideal model with which to study the mechanisms
of endotoxin sensitization.

The intent of the present study is to gain insight into the
relationship between selected toxic effects of endotoxin and the local-
ization of endotoxin among host tissues, especially the liver. This
investigation was initiated by using a subcutaneous route of injection
to determine whether a correlation exists between the organ distribution
of endotoxin and various biologic responses to endotoxin, including
lethality. A second objective is to examine the distribution of intra-
venously administered endotoxin among various liver cell types and to
' determine whether the increased sensitivity of galN-treated mice to the
toxic effects of endotoxin is associated with enhanced uptake of endotoxin
by liver parenchymal cells both in vivo and in vitro. A third objective
of this study is to determine whether serum borne mediators elicited by

endotoxin are involved in endotoxin sensitization by galN. This objective

was approached experimentally by comparing lethality and serum ornithine
carbamyltransferase (OCT) levels in normal and galN-treated mice following
transfer of serum from mice treated with Corynebacterium parvum and
endotoxin. Observations made in other serum transfer experiments

distinguish between different mechanisms of endotoxin sensitization.

LITERATURE REVIEW

 

Significance of Endotoxin Research.

 

Bacteremia with gram-negative organisms is consistently associated
with fever, hypotension, hypoglycemia, and disseminated intravascular
coagulation. (47,140). These manifestations have been attributed to the
presence of endotoxin in the blood (101) which may lead to shock and
death. These observations have formed the basis for literally hundreds
of investigations into the chemical structure and biologic activities of
bacterial endotoxins.

The precise role of endotoxin in gram-negative bacteremic shock
remains a subject of controversy. One reason for this is the difficulty
of directly measuring endotoxin levels in the blood or tissues. The
discovery by Levin and Bang (65) that toxic endotoxin induces gelation
of a lysate of Limulus polyphemus amebocytes has led to the development
of a bioassay for endotoxin and has facilitated more controlled studies
of clinical endotoxemia. The clinical and experimental usefulness of
the Limulus assay is limited by the nonspecific and frequent
false—positive and false-negative test results (30).

According to McCabe et a1. (75) the incidence of septic shock has
increased significantly in the United States over the past 25 years and
is responsible for approximately 130,000 deaths per year. Other reports
on the mortality rate of gram-negative shock range from 11 percent (22)
to 82 percent (134), with a national average of 50 percent (71). The
majority of such clinical cases originate in patients with impaired

immune defenses or when gram-negative bacilli are introduced directly

into the urinary, respiratory, or vascular system by therapeutic vectors
such as urinary or intravascular catheters or ventilatory equipment.
Increased susceptibility to gram-negative shock may be a frequent
complication of administering large doses of antibiotics to the
bacteremic patient which may lead to shock, presumably by liberating

large amounts of endotoxin (140).

Chemical Structure of Endotoxins.

 

Endotoxins are normal constituents of the outer membrane of
gram—negative bacteria and consist primarily of lipopolysaccharides with
variable amounts of protein. Many of the biologic activities normally
attributed to endotoxins can be elicited with chemically pure
lipopolysaccharide. Both the bacterial source and the method of
extraction influence the biologic activity and extent of aggregation of
the extracted preparation. The molecular weight of experimental
preparations of lipopolysaccharides range from 1 to 20 million daltons.
While numerous investigators use the terms "endotoxin" and
"lipopolysaccharide. LPS" interchangeably, it must be recognized that
these two bacterial products may be experimentally different in both
biologic activity and chemical composition based on the presence or
absence of the protein component which serves to distinguish endotoxins
from purified lipopolysaccharides.

Lipopolysaccharides extracted from wild-type gram-negative bacteria
are composed of a repeating polysaccharide region, a central
polysaccharide core, and a lipid-rich hydrophobic region. The innermost
hydrophobic region, lipid A, consists of B-1,6-disaccharide units of

glucosamine containing both ester- and amide-linked long chain fatty

acids. It is generally acknowledged that lipid A represents the toxic
moiety of endotoxin (39). The core polysaccharide contains heptose,
several hexoses, and 0-phosphorylethanolamine, as well as a unique
sugar, 2—keto, 3—deoxyoctulosonate (KDO), ‘which. covalently ‘binds. the
core region to the lipid A moiety. The subunits of the outermost
polysaccharide region are made up of 3 to 4 repeating hexoses and
determine the O-antigenic specificity of the bacterial cell. The number
of repeating saccharide subunits can range from 0 (rough) to as few as 2
(semirough) or as many as 10 (smooth), possibly even within the same
bacterial cell (43,53). The basic structure of the lipid A and core
polysaccharide regions of lipopolysaccharides purified from diverse
groups of bacteria is remarkably similar while the O-antigenic region is
unique for each type of organism.

Variable amounts of protein are known to be associated with
endotoxin. preparations extracted. by the Boivin. TCA. procedure. This
protein is termed lipid A-associated protein (LAP) and is itself
biologically active (88,120). It should be recognized that highly
purified, well-defined lipopolysaccharide preparations are essential for
critical and consistent evaluation of the biologic effects of endotoxin

(74).

Radioisotopic Labeling of Endotoxin.

 

The precise cells or tissues that interact with endotoxin to
produce the toxic and lethal effects of endotoxin following its
intravenous injection are unknown. Many investigations have addressed
this problem experimentally by examining the blood clearance and tissue

localization of radiolabeled endotoxin. Labeling endotoxin with

different radioisotopes has presented investigators with 21 variety of
problems in the use of these radiolabeled products in studying the fate
of endotoxin in vivo. Studies with 32P—labeled endotoxin have shown
that a substantial portion of the isotope is liberated from the labeled
conjugate in vivo (51) or when. mixed with serum in vitro (107).
Internal labeling of endotoxin with 3H or 14C by growing the bacteria in
the presence of radiolabeled sugars requires time-consuming and costly
extraction procedures as well as additional steps to solubilize tissues
before determining their radioactive content in organ distribution
studies. Labeling of endotoxin with 3H by tritium gas—exposure results
in significant losses in toxicity as measured by lethality in mice and
ability to induce local Schwatrzman reactions in rabbits (109,110).
Ulevitch (125) devised a method for covalently binding radioiodine
to endotoxin extracted from a variety of bacteria without altering the
biOphysical, immunologic or biologic properties of the preparation.
This procedure is based on the initial reaction of endotoxin with p-OH,
methylbenzimidate. which is believed to render primary amino groups
contained within the endotoxin molecules, most likely ethanolamine
residues, susceptible to labeling with radioactive iodine. The author
stressed the need to use protein—free lipOpolysaccharide preparations
for radioiodination by this method in order to insure that iodine will
not covalently bind to protein which can be dissociated from endotoxin
in vivo. Specific activities obtained by this method are considerably
higher than those reported previously for a variety of other isotopes.
Braude et al. (14) were the first to describe a method for

radiolabeling endotoxin from Escherichia coli with 51Cr in the form of

 

sodium chromate or chromium chloride. Although the mechanism by which

51Cr is bound to endotoxin has not been defined, Braude and coworkers

maintain that 51Cr-labeled endotoxin is highly toxic and stable with
less than 0.1% loss of label from conjugated product. per' day"when
subjected to continuous dialysis. In a subsequent study Braude et al.
(15) found that following parental administration of 51Cr-labeled
endotoxin, up to 60% of the radioactivity was found in the liver by 45
min. In contrast, injection of NaZSICrO4 showed significant differences
in tissue distribution when compared to 51Cr--labeled endotoxin, with
very little association of radioactivity with liver and spleen. Further
studies on the character of s'ICr-labeled endotoxin by Chedid et al.
(19) showed that centrifugation of a fresh preparation of 51Cr-labeled
endotoxin yielded a pellet containing highly labeled, toxic endotoxin.
Furthermore, endotoxin recovered from plasma of mice that were
previously injected with 51Cr—endotoxin was labeled and lethal for
adrenalectomized mice. These results coupled with those of Braude et
a1. support the stable nature of 51Cr-endotoxin and demonstrate that
51C

r does not dissociate from the conjugated product following

injection.

Distribution of Endotoxin In Vivo.

 

Numerous experimental studies into the fate of endotoxin in vivo
indicate that the majority of intravenously injected endotoxin cleared
from the blood is localized in the liver, spleen, and lungs. These
organs contain the largest reserve of fixed tissue macrophages of the
mononuclear ‘phagocyte system. and represent the ‘major organs of the
reticuloendothelial system (RES) as described by Aschoff (3). Following

intravenous injection, the largest amount of endotoxin accumulates in

the liver (13,77,139). The resident macrophage pOpulation found in the
liver consists of Kupffer cells and constitutes the largest reserve of
fixed tissue macrophages of the RES.

Ruiter et al. (108) recently undertook a series of experiments that
showed a slow but progressive increase in liver uptake of 51Cr-labeled
endotoxin. With a 10-fold lower dose, liver uptake was more rapid but
transient, suggesting that liver uptake 'may depend on ‘blood
concentration or occur by different mechanisms. They also found
radioactive content of samples of blood and liver correlate
quantitatively with the amount of endotoxin in these samples as
estimated by Limulus assay. These observations effectively illustrate
that isotopic labeling of endotoxin with 51chromium is a viable means of
determining the fate of endotoxin in vivo.

Noyes et a1. (95) found that 51Cr-labeled endotoxin localized
primarily in livers of mice by 2 hrs after intravenous administration.
Following intramuscular injection they found that 35-50% of the label
remained at the site of injection. with as little as 2% becoming
associated with the liver. About 35% of the radioactivity was
unaccounted for. Despite the contrast in tissue distribution,
intramuscular injection of endotoxin was still lethal. Musson et an”
(89) compared the tissue distribution and rates of disappearance of
lipopolysaccharides in endotoxin responsive (C3H/St) and unresponsive
(C3H/HeJ) mice. The C3H/HeJ strain is refractory to most of the
biological effects of endotoxin including mitogenicity (119),
immunogenicity (133), lethality (118), and enhancement of nonspecific
resistance (20). This defect in endotoxin responsiveness is due to a

mutation in a single autosomal gene locus thought to be responsible for

10

the regulation of expression of various endotoxin activities (133).
Although no significant differences were observed between responsive and
unresponsive mice in rate of LPS clearance, more LPS accumulated in
lymph nodes, adrenals, lungs, and kidneys of C3H/St mice when compared
to C3H/HeJ mice. This observation led Musson et al. to suggest that
these tissues may be involved in the toxic response to endotoxin.

Uncertainty still exists as to whether liver parenchymal cells are
adversely affected by endotoxin directly or via endogenous factors
present in the circulation. Several investigators have attempted to
gain insight into this problem by determining whether endotoxin can be
detected in hepatocytes following parenteral administration.

The liver is strategically located along the venous circulatory
system between the gastrointestinal tract and the heart. Venous blood
passing through the splenic vein, small mesenteric veins, and gastric
veins, merge into a large portal vein and eventually circulates through
the liver sinusoids. Blood percolates through the liver sinusoids,
eventually collects in the central veins and is passed to the large
hepatic vein. The liver consists predominantly of parenchymal and
sinusoidal lining cells. The sinusoidal cell population is composed of
endothelial cells and macrophages (Kupffer cells). The liver sinusoidal
lumen is lined with a double layer of fenestrated endothelial cells.
Beyond this layer of endothelial cells lie the parenchymal cells with
numerous microvilli projecting into the space of Disse. Kupffer cells
are situated in the sinusoidal lumen in a manner which physically
maximizes their exposure to circulating blood. The Kupffer cell body is
anchored to the fenestrated endothelium by numerous, delicate,

cytOplasmic dendritic processes.

11

Although hepatic accumulation of endotoxin is usually attributed to
Kupffer cell function, radiolabeled endotoxin has also been detected in
liver parenchymal cells (108,137,139). Willerson. et al. (137) used
autoradiography to demonstrate association of 14C-labeled endotoxin from

S. enteritidis with liver parenchymal cell nuclei as well as Kupffer

 

cells. Zlydaszyk and Moon (139) isolated liver cells from mice 1 hour
after injection of 51Cr-labeled endotoxin. Susceptibility of the
parenchymal cells to lysis by pronase allowed these workers to
demonstrate 75% association of radiolabel with parenchymal cells and
only 25% with non—parenchymal cells. Subcellular fractionation of liver
homogenates by differential centrifugation suggested that approximately
45% of the 51Cr—endotoxin was associated with liver cell nuclei, 20%
with the mitochondrial-lysosomal fraction and 30% with the cytosol.
Tavakoli and Moon (123) attended these organ distribution studies of
51Cr-labeled endotoxin. Their results indicate that about 50% of an
intravenously injected dose of endotoxin is sequestered by the liver by
1 hour. The organ distribution of endotoxin did not change
significantly by 5 hours. In contrast to the earlier results of
Zlydaszyk and Moon (139), the more rigorous purification procedures used
in this study yielded data showing that only about 12% of the
liver-associated radiolabel was associated with a crude nuclear fraction
while less than 1% was associated with purified nuclei. These
discrepancies might be explained by contamination of crude nuclear
fractions in the earlier study' with 51Cr-endotoxin associated with
cellular debris or intact Kupffer cells.

Ramadori et al. (103) used immunofluorescence to detect endotoxins

in vivo and found that smooth-foam LPS is localized predominantly in

12

nonparenchymal cells while rough-form LPS and lipid A can be found in
both hepatocytes and nonparenchymal cells. In contrast, other
investigators found only Kupffer cell—associated endotoxin by
immunofluorescence (26) or autoradiography (74,109). These
discrepancies might be explained by the presence of endotoxin in
hepatocytes in concentrations below the limits of detection afforded by
autoradiography or immunofluorescence.

Ruiter et al. (108) recently employed differential centrifugation
and countercurrent centrifugal elutriation to study the distribution of
51Cr-endotoxin among rat liver cells 30 min after intravenous injection.
Their study also detected low levels of endotoxin in rat liver
parenchymal cells (17ng/106 cells). Separation of liver sinusoidal
cells by centrifugal elutriation allowed these investigators to show
that endotoxin localized primarily in Kupffer cells (340ng/106 cells) as
compared to endothelial cells (24ng/106 cells).

Another study by Ramadori et al. (102) demonstrated binding sites
for endotoxin on plasma membranes of isolated hepatocytes maintained in
culture. In vitro binding of smooth-form and rough-form LPS as well as
free lipid .A to mouse hepatocytes was demonstrated by direct
immunofluorescence. Maximal binding of LPS to hepatocytes occurred at
an LPS concentration of 500ug/ml. No binding was detected with LPS
concentrations below 10ug/ml. Both forms of LPS (500ug/ml) also bound
to nonparenchymal cells whereas free lipid A (SOOug/ml) bound only to
hepatocytes.

Additional work by Tavakoli and Moon (123) investigated the

51

association of Cr—labeled toxic and alkaline—treated detoxified

endotoxin with hepatoma tissue culture (HTC) cells. Their data show

13

that HTC-cells bind very little toxic endotoxin by 5 hours after
treatment. Exposure of toxic 51Cr-endotoxin to plasma in vivo did not
enhance binding to HTC-cells. Binding of alkaline-treated detoxified
endotoxin to HTC—cells was 3-8 fold higher than toxic endotoxin.
Conversely, organ distribution studies showed that only 25% of
intravenously injected detoxified endotoxin became associated with the
liver by 5 hours compared with 50% of toxic endotoxin. Furthermore,
nuclear association of toxic endotoxin gradually increased with time
while no such increases were observed in vivo. Since binding of
endotoxin to HTC-cells does not compare favorably with in vivo results,
caution should be exercised when extrapolating in vitro data to the
localization and action of endotoxin in vivo.

Mathison and Ulevitch (74) studied tissue distribution of
radioiodinated LPS injected intravenously into rabbits. Using both a
rough and smooth-form LPS, they observed a biphasic clearance of LPS
from the circulation. One-half of the injected LPS was removed from
blood within 30 min, localizing primarily in liver, spleen, and lung.
The detection of high concentrations of 125I-labeled LPS in gall bladder
of injected rabbits led these workers to suggest that endotoxin might
be detoxified in vivo by hepatocytes and subsequently passed into
the bile canalicular system. Examination of tissue sections by
autoradiography also showed LPS concentrated in phagocytic vacuoles of
hepatic Kupffer cells, splenic. macrophages, and leukocytes, but ‘not
hepatocytes.

A continuing interest of these authors has been the significance of
LPS interaction with plasma. Radioiodinated LPS remaining in plasma

beyond 30 min after injection into rabbits disappeared from the blood

14

with a half-life of about 12 hours and exhibited a decrease in buoyant
density from about 1.44 g/ml (native LPS) to less than 1.2 g/ml. When
the low density form of LPS was isolated from the blood and reinjected
into previously untreated rabbits, the initial rapid clearance phase did
not occur suggesting that reduction of LPS buoyant density was
associated with a slower rate of clearance from blood. In a previous
report (126) these investigators showed that LPS interaction with plasma
lipoproteins in vitro and in vivo was associated with a reduction in
buoyant density as well as inhibition of a number of endotoxic
activities including lethality in adrenalectomized mice,
anticomplementary activity, pyrogenicity in rabbits, and ability to
cause neutropenia in rabbits. More recent work by this group has shown
that delipidation of plasma by extraction with n-butanol/diisopropyl
ether prevents the reduction in LPS buoyant density, an effect which is
reversed by addition of purified high density lipoproteins (127).
Addition of low density lipoprotein or very low density lipoprotein had
no effect on the buoyant density of endotoxin. The authors suggest that
high density lipoproteins function as transport molecules for endotoxin
in plasma to aid in reaching organs of clearance or target organs.
Alternatively, Skarnes (116) suggests that plasma components play a
major role in detoxification of circulating endotoxin. By fractionating
serum Skarnes separated two alpha globulins which interacted with
endotoxin. and. ‘possess 'nonspecific carboxylic esterase activity.
Interaction of endotoxin with a heat—stable alpha lipoprotein led to
disaggregation of endotoxin while detoxification of exposed groups was

effected by interaction with another heat-labile alpha globulin. These

15

findings are in agreement with those of Ulevitch (126) in that endotoxin
forms complexes with plasma lipoproteins.

Cumulatively, these results suggest that the interaction of LPS
with plasma factors may be a major consideration in endotoxin

detoxification and clearance.

Pathophysiologic Changes Induced by Endotoxin.

 

Endotoxin elicits an impressive array of biological activities in
the susceptible host. Unfortunately, it is still not clear which
activities are important in the lethal response to endotoxin. Several
histologic and metabolic alterations are known to occur in liver
parenchymal cells following endotoxin challenge (9,24,54,66,77,79).
Although it appears that at least low levels of endotoxin become
associated with hepatocytes in vivo, whether hepatocytes serve as direct
targets of endotoxin toxicity or merely function in the processing of
endotoxin remains an open question.

The histopathologic alterations in liver tissue following a single
injection of endotoxin were investigated by Levy et al (66,67). Using
light microscopy and histochemical techniques (66), the earliest changes
noted were swelling of Kupffer cells and the appearance of greater
numbers of polymorphonuclear leukocytes (PMN's) and platelet thrombi.
Decreased glycogen content, increased fat deposition, and elevated acid
phosphatase activity of Kupffer cells was evident by 4 hrs. after
injection of endotoxin. In a subsequent report, Levy et al. (67) used
electron microscopy to confirm many of their previous findings with
light. microscopy. Hepatocytes showed extensive ‘vacuolation, swollen

mitochondria and endoplasmic reticulum, and depletion of glycogen

16

content. The number of lysosomes in the perinuclear region appeared to
increase with time over a 24 hr period. Areas of hepatocellular
necrosis were also observed. The sinusoidal region showed platelet
thrombi, leukocytic infiltration by PMN's, lymphocytes, and eosinophils,
and swollen Kupffer cells containing increased numbers of lysosomes and
vacuoles. The authors stressed the need for an accurate determination
of the topologic distribution of endotoxin in the liver.

The development of hypoglycemia and concurrent depletion of
glycogen. reserves are some of the first endotoxin-induced. metabolic
alterations to be reported in the literature (81). Studies by Berry et

al. (7) demonstrated that a lethal dose of killed Salmonella typhimurium

 

injected into mice resulted in an almost complete loss in total body
carbohydrate. The simplest explanation for endotoxin-induced depletion
of carbohydrate stores (77) is an increase in metabolic rate above
normal levels, as commonly seen accompanying fever. This appears not to
be the case, however, since endotoxin causes hypothermia in mice kept at
temperatures commonly encountered in the laboratory. Furthermore,
Shands et al. (112) were not able to show an increase in metabolic rate,
as judged by oxygen consumption, in endotoxin challenged mice.
Cumulatively, these data do not support the notion that an elevation in
metabolic rate is responsible for the severe depletion of carbohydrate
reserves. Alternatively, the inhibition of hepatic gluconeogenesis
might be responsible for depletion of liver glycogen content and
contribute to the pathOphysiologic consequences of endotoxemia. LaNoue
et al. (63) observed impaired gluconeogenesis in liver slices obtained
from endotoxin-treated rats. Shands et al. (112) have also shown that

endotoxin-poisoning impairs gluconeogenesis in mice made hyperreactive

17

to the toxic effects of endotoxin by prior infection with BCG (see
below). Consistent with this hypothesis is the ability of endotoxin to
inhibit the corticosteroid induction of several key regulatory enzymes
of gluconeogenesis, including phosphoenolpyruvate carboxykinase (PEPCK)
(9,106), glycogen synthase (77), and glucose-6-phosphatase (76).

Corticosteroid hormones antagonize many of the host responses to
endotoxin including lethality (10,42). Moreover, the importance of
endogenous adrenocorticoids to an animal's response to endotoxin is
clearly evident in the 1000—fold decrease in LDSO of endotoxin which
occurs following adrenalectomy. Elucidation of the precise mechanism of
corticosteroid protection from other simply chronologically' parallel
protective events has not been possible. Cremer and Watson (26) and
Ribble et al. (104) were unable to demonstrate a difference in the
tissue distribution of radiolabeled endotoxin in cortisone pretreated
animals when compared tn) untreated controls. Although various
investigators have postulated the involvement of the antiinflammatory
(114) and membrane stabilizing (41,52,117) prOperties of
corticosteroids, the preponderance of evidence supporting steroid
protection from the toxic effects of endotoxin poisoning suggests a
metabolic action.

Corticosteroid protection as regards changes in carbohydrate
metabolism following endotoxin administration is a problem which has
been thoroughly investigated by Berry et al. (7,9,10,76,77,85,86,106).
The possibility that endotoxin directly inhibits hepatic enzyme activity
is unlikely since endotoxin added directly to liver homogenates (9) or
isolated rat hepatocytes (36,78) has no direct effect on gluconeogenesis

or PEPCK activity. In an attempt to determine whether endotoxin caused

18

a generalized inhibition of hepatic protein synthesis, Shtasel and Berry
(115) followed ribonucleic acid (RNA) synthesis by measuring the
incorporation of either 32F or 14C-orotic acid in liver RNA of mice
receiving endotoxin, cortisone, or both. Endotoxin not only stimulated
whole liver RNA synthesis but also increased protein synthesis as
evidenced by the incorporation of 14C-leucine into acid-insoluble
peptides. These results led Shtasel and Berry to suggest that endotoxin
inhibits the hormonal induction of certain hepatic enzymes in a very
selective fashion and may involve the participation of mediating factors
(115).

Elevations in serum levels of a variety of enzymes occurs following
endotoxin administration. For example, Woodward. et al. (138) found
serum levels of aldolase, glutamic—oxaloacetic transaminase, and
isocitrate dehydrogenase were all elevated above normal levels in rats
within 4 hours after injection of a lethal dose of endotoxin. Konttinen
et al. (61) observed increases in serum levels of lactate dehydrogenase
and glutamic-oxaloacetic transaminase in rabbits given endotoxin
intravenously, and suggested the liver as a likely source.

Changes in serum enzyme activities is assumed to reflect cellular
injury (49). Although cellular' necrosis is one condition. known to
result in elevated serum levels of certain transaminase, similar changes
occur in the absence of necrosis because of increased cell permeability
(96). In order to correlate changes in serum enzyme activities with
specific effects of endotoxin, it would be valuable to know the identity
of the cells that release specific enzymes into the circulation.

Ornithine carbamyltransferase (OCT) is a mitochondrial enzyme which

catalyzes the condensation of carbamylphosphate and ornithine to form

19

citrulline in the urea cycle and is found almost exclusively in liver
parenchymal cells (18). Accordingly, elevations in serum OCT activity
is considered to be a specific indicator of hepatic injury due to its
high concentration in liver (18,41). Sobocinski et al. (117) reported a
seven-fold increase in plasma OCT activity over normal levels in rats by
II) hours after receiving the. LD90 of endotoxin. Cook. et al. (24)
measured plasma levels of OCT to demonstrate deterioration of hepatic
integrity in rats made hyperreactive to the toxic effects of endotoxin
by pretreatment with glucan. Rats given an LD80 of endotoxin exhibited
13 and 35—fold elevations in plasma OCT at 3 and 6 hours respectively,
when compared to untreated controls. Glucan pretreated rats receiving a
dose of endotoxin that caused 100% lethality only showed a 5-fold
increase in plasma OCT activity by 4 hours after treatment. Cook et al.
(24) suggest that loss of hepatocyte integrity may be important in
lethal endotoxemia but not in the mechanisms underlying glucan-induced
sensitization to lethal endotoxemia.

Numerous studies have attempted to link impairment of mitochondrial
function to the pathogenesis of endotoxin shock. Several investigators
have reported mitochondrial swelling and extensive disruption of
mitochondrial membranes in livers of endotoxin poisoned animals. These
structural manifestations are accompanied by impairment of mitochondrial
function. For example, Mela et al. (79) observed that isolated liver
mitochondria obtained from. endotoxin poisoned rats showed decreased
ATPase activity, loss of ability to maintain membrane-bound Mg++, and

reduced rates of oxidative phosphorylation. Several investigators have

also described uncoupling of oxidative phosphorylation, loss of

20

respiratory control, and ‘morphological changes in. mitochondria. when
treated with endotoxin in vitro (45,48).

Additional work by Mela et al. (80) examined the role of lysosomal
enzymes in impairment of mitochondrial function. By monitoring acid
phosphatase activity in lysosomal fractions isolated from livers of rats
treated with endotoxin 5 hours earlier, they showed a 50% decrease in
total liver acid phosphatase activity along with a 70% decrease in
lysosomal acid phosphatase activity when compared to controls.
Impairment of mitochondrial function accompanied the loss in liver
lysosomal enzyme activity. Increases in serum acid hydrolase activity
(54,55) as well as increased fragility of hepatic lysosomes (54,135)
during endotoxemia have been reported. As a second objective, Mela et
al. (80) were able to reproduce the impairment of mitochondrial
function, as evidenced by inhibition of ATPase activity and Ca++
transport, by simply adding lysosomal enzymes to normal mitochondria 33.
33533. Mela et al. (80) believe that extrusion of lysosomal contents
could. be partially responsible: for ‘mitochondrial. impairment seen. in
endotoxemic animals. Although it is generally believed that endotoxin
does not affect mitochondria directly, the precise role of lysosomal

disruption and mitochondrial impairment in the pathogenesis of endotoxin

remains unclear.

Mediation of Endotoxin Activity.

 

Recent research lends support to the hypothesis that a variety of
mediators released into the blood of a host after challenge with
endotoxin are responsible for many of the biological effects of

endotoxin.

21

Fever resulting from endotoxin administration is the subject of a
recent review by Westphal et al. (136). From this review it becomes
apparent that the febrile manifestation of endotoxin poisoning is
believed to be caused by an endogenous pyrogen (EP) released from one or
more of a variety of cells including monocytes, Kupffer cells, alveolar
macrophages, and polymorphonuclear leukocytes. Serum borne EP is
believed to influence the thermoregulatory center of the hypothalamus,
causing disturbances in temperature regulation.

Considerable evidence exists implicating prostaglandins as
contributing factors in the febrile response to endotoxin.
Phillip-Dormston (99) used a radioimmunoassay to determine the
concentration of prostaglandins in the cerebrospinal fluid of rabbits
following injection of endotoxin. While prostaglandins of the F series
remained the same, those of the E series were twice as high in endotoxin
challenged rabbits compared with untreated controls. Feldberg and
Saxena (35) recently showed that injection of either of two 13 series
prostaglandins into rats caused fever of short duration followed by
hypothermia. Intravenous injection of LPS or lipid A caused a prolonged
drop in body temperature only. Agents such as aspirin or indomethacin
are capable of interfering with prostaglandin biosynthesis and inhibit a
variety of endotoxin effects such as release of glucocorticoid
antagonizing factor (see 'below), lethality, and fever (128). These
findings suggest that impaired prostaglandin synthesis may be associated
with endotoxin refractoriness and increased release of puostaglandins
may be involved in endotoxin activity.

Berry (8) found that endotoxin poisoned mice housed at an ambient

temperature of 30°C or below, became hypothermic while those kept at

22

temperatures above 30°C became febrile. As a second objective he found
that the LD50 of endotoxin decreased when mice were maintained at
temperatures above or below 30°C.

Besides endogenous pyrogen and prostaglandins, a variety of other
soluble mediators have received considerable attention with respect to
their role in various biologic responses to endotoxin. For example,
leukocytic endogenous mediator (LEM) causes fever, lowers serum iron and
zinc, and increases nonspecific resistance to infection (56).

Moore et al. (85) first demonstrated the participation of a plasma
factor called glucocorticoid-antagonizing factor (GAF) in the inhibition
of hormonal induction of liver PEPCK by endotoxin. Mice made tolerant
to the biologic effects of endotoxin by repeated injection of increasing
doses of endotoxin were injected with endotoxin or serum from endotoxin
challenged mice. The ability of cortisone acetate to induce liver PEPCK
was determined 4 hrs later. Serum from endotoxin—poisoned mice but not
endotoxin itself or normal serum inhibited the hormonal induction of
PEPCK in the tolerant recipients. These results suggest that endotoxin
tolerant mice respond to exogenously administered GAF but fail to
produce GAF in response to endotoxin. Moore et al. (85) showed that
macrophages are involved in the production of GAF. GAF activity is
trypsin sensitive and is totally abolished by heating to 75°C for 1 hour
(87). The molecular weight of GAF was estimated by elution through
Sephadex G-200 to be approximately 150,000 daltons.

LipOpolysaccharides have been shown to induce lymphocyte activating
factor (LAF, Interleukin. 1, 1L1) production. and secretirux by' human
(124), murine (29), and murine cell line (83) macrophages. Although

originally defined as a murine thymocyte mitogenic factor, ILl also

23

enhances peripheral T cell proliferation, alloantigen-specific
cytotoxic T cell responses, and in vitro antibody responses to
thymic-dependent antigens, all of which depend upon the presence of the
priming antigen. Biochemical studies indicate that murine 1L1 is a
single polypeptide chain with a nwdecular weight of about 15,000 that
appears to be resistant to a ‘number of endOpeptidases and
protein-denaturing agents such as urea and sodium dodecyl sulfate (84).
Ample evidence exists in the literature supporting the importance of ILl
as a mediator of LPS adjuvanticity.

In recent years the combined treatment of mice with an attenuated

strain of Mycobacterium tuberculosis (bacillus Calmette-Guerin, BCG) and

 

endotoxin has been utilized to amplify the production of serum borne
factors which mediate the antitumor properties of endotoxin. Carswell
et al. (16) were the first to describe the presence of tumor necrosis
factor (TNF) in the serum of BCG-infected mice challenged with endotoxin
which caused hemorrhagic necrosis of a variety of transplanted tumors in

recipients by 24 hrs. Zymosan and Corynebacterium parvum, also capable

 

of inducing RES hyperplasia, were equally effective in priming mice for
the production of TNF in response to endotoxin. Manuel et al. (73)
subsequently develOped a sensitive and quantitative in vitro assay for
TNF activity by monitoring the release of 3I-l-thymidine into culture
supernatants of prelabeled tumor cells. Tumor cells of dissimilar
origins were inhibited or killed by TNF-containing serum while normal
embryonic fibroblasts were unaffected. Maximal titers of serum TNF were
obtained 14 days after BCG infection and 2 hr after challenge with as
little as 1.0 ug of endotoxin. Serum from mice treated with a similar

dose of endotoxin or infected with BCG alone were inactive in this assay

24

system (73). Furthermore, evidence has been assembled to distinguish
between the active principle of TNF-containing serum and endotoxin
itself (44,50). It is interesting to note that Parant et al. (98) were
able to show that TNF-serum enhanced the resistance of normally

susceptible mice to challenge with either Klebsiella pneumonia or

 

Listeria monocytogenes. Enhanced resistance to infection was not due to

 

endotoxin itself since nonspecific resistance to bacterial infection in
C3H/HeJ (LPS low responder) and 8 day old (LPS responder) mice used in
their experiments is unaffected by endotoxin directly (20,97). Whether
enhanced resistance to infection and tumor necrosis can be attributed to
the same serum factor remains an open question.

A number of authors have contributed to an excellent review of
endogenous ‘mediators elicited during the ‘host response. to ‘bacterial
endotoxin (82). From this review it becomes evident that the majority
of studies characterizing the biologic activities of individual soluble
mediators utilize in vitro systems to circumvent the complexities on in
E9. studies. The majority of these factors also appear to possess
biologic activities that, when considered individually, would appear to
be beneficial in the whole animal. For example, 1L1 and IL2 appear to
enhance both humoral and cell-mediated immunity while TNF causes
hemorrhagic necrosis of implanted tumors and increases nonspecific
resistance to infection. Similarly, there are as yet no definitive
studies, with the possible exception (ME EP and GAF, clearly
demonstrating the involvement of any endotoxin-induced, endogenous

mediator in the toxic manifestations of endotoxemia.

25

Galactosamine-induced Enhancement of Endotoxin Susceptibility.

 

Galactosamine hepatitis is a unique experimental model used to
study the biochemical mechanisms responsible for liver cell organelle
injury and ultimately, cell death. The injection of sufficient
quantities of D-galactosamine (galN) into experimental animals results
in liver cell necrosis and inflammatory reactions with striking
similarity to human viral hepatitis (57). Susceptibility to a given
dose of galN varies among species but most species react in a similar
manner. Liver cell necrosis induced by intraperitoneal administration
of galN occurs by a multi—step mechanism. The 'primary biochemical
response' following galN injection is characterized by the accumulation
of metabolites of galN in liver tissue, leading to the depletion of
uracil nucleotides. The 'secondary biochemical response' resulting from
these initial metabolic alterations appears to involve the inhibition of
specific enzymes and decreased synthesis of RNA, glycoproteins, and
glycolipids (28,59). These metabolic alterations culminate in the
development of galactosamine hepatitis.

Galactosamine is metabolized by enzymes of the galactose pathway
which are found predominately in liver tissue (27,28). Most amino
sugars, such as glucosamine, are N-acetylated prior to uridylation and
subsequently metabolized as N-acetylhexosamines. Within 30 min of galN
injection, UDP-galactosamine, UDP-glucosamine, and a variety of other
related uridylated galactosamine derivatives accumulate in livers of
galN-treated rats. Consequently, the liver content of metabolically
available UTP drops to approximately 7% of control levels by 2 hrs with
an equal depression in UDP and UMP content (59). On balance,

Consumption of uridine nucleotides greatly exceeds the capacity of

26

uridylate dg.ngyg synthesis (58). The decrease in uridine nucleotides
is associated. with a similar decrease in. normal levels of uridine
hexoses and hexosamines. This galN—induced deficiency in uridine
nucleotide content is not associated with an equivalent reduction in
pools of ATP, CTP, or CTP (28,59). This primary biochemical response to
galN injection is thought to impair plasma membrane protein and
glycoprotein biosynthesis resulting in functional and structural
alterations of hepatocyte plasma membranes (4,28,31). Because of the
metabolic selectivity of the liver for galN, when sufficient doses of
this amino sugar are administered to experimental animals, liver injury
ensues while other cells and organs appear to be umuphologically and
physiologically unaffected (28,57). Approximately one-half of an
injected dose of galN can be recovered from all tissues (62), two-thirds
of which is found in the liver (5).

A role for endotoxin in the pathogenesis of galactosamine hepatitis
has been suggested since many of the clinical symptoms experienced by
galN—treated animals are known to be caused by endotoxin (ie.,
inflammation, systemic hypotension, complement consumption, generalized
edema, and disseminated intravascular coagulation). Using the Limulus
assay, Grun et al. (46) investigated the development of endotoxemia
following injection of rats with galN (lg/kg). Endotoxemia was detected
by 6 hrs along with hypoglycemia, hypotension, and fever in galN-treated
rats but not controls. This response was followed by an hepatic
inflammatory reaction and increased serum transaminase levels occurred
by 24 hrs. Grun et a1. (46) also demonstrated that colectomy performed
three to six days prior to galN administration in order to eliminate the

intestinal flora as a source of endotoxin, prevented the development of

27

endotoxemia, inflammation, and histologically evident liver cell
necrosis seen to occur in sham-operated control rats. Both colectomized
and control rats responded metabolically to galN administration with a
typical decrease in hepatic uridine nucleotide content (46,69). Rats
made resistant to the biological effects of endotoxin by repeated
injections of increasing doses of endotoxin were also found to be
refractory to the hepatic deterioration induced by galN, despite the
development of endotoxemia (46). These observations led Grun et al.
(46) to conclude that the primary biochemical response of hepatocytes to
galN-injection cannot be solely responsible for the subsequent increases
in serum levels of liver-specific enzymes and histologically evident
liver cell necrosis. In another study, Farber et al. (32) showed that
administration of high (400mg/kg) or low (200mg/kg) doses of galN to
rats resulted in quantitatively and qualitatively similar biochemical
responses while only the higher dose of galN induced liver cell
necrosis.

Galanos et al. (40), using phenol extracted lipopolysaccharides,
showed that treatment of a variety of experimental animals with galN
enhanced their sensitivity to the lethal effects of concomitantly
administered endotoxin by a factor of at least several thousand.
Lethality usually occurred within 9 hrs of treatment with galN and
endotoxin. The state of hyperreactivity to endotoxin was greatest when
endotoxin was injected intravenously either concurrent with or 1 hr
after galN. Lethality did not occur when endotoxin was injected 1 hr
before or 4 hrs after galN nor when galN was injected alone.
Galactosamine-treated mice were susceptible to the lethal effects of

lipid A but not polysaccharide preparations purified from endotoxin.

28

Galanos et al. (40) made note of the potential toxic reactivity of the
liver to endotoxin in addition to its functional role in blood clearance
of endotoxin. In addition, they concluded that temporary"metabolic
changes of the host may be significant in rendering an animal more
susceptible to circulating levels of endotoxin.

A role for the complement system has been implicated in the
mechanism of liver cell necrosis induced by galN administration (69).
Liehr et al. (70) investigated liver cell necrosis and complement
activation in rats that received low doses of galN (200mg/kg) with or
without an additional injection of sublethal amounts of endotoxin
(1.5mg/kg). Endotoxin was administered in order to activate the
complement system. Fulminant hepatic necrosis, as shown histologically
and by dramatic increases in serum transaminase levels, accompanied by a
rapid decline in hemolytic complement activity was observed only when
endotoxin was administered with galN. Liehr et al. (70) also used
immunohistology to detect the accumulation of the third component of
complement on necrotic liver cells of rats receiving galN (1mg/kg) and
endotoxin. The authors of this study concluded that alterations in liver
parenchymal cell. membranes following galN treatment ‘may’ prepare the
hepatocyte for complement-mediated cytolysis, leading to fulminant liver
necrosis. The significance of the complement system in galN-induced
liver cell necrosis was further supported by the demonstration that
complement-deficient mice (strain NMR 1) do not show signs of liver cell
necrosis or inflammation after galN administration despite the

development of endotoxemia (69).

29

Experimental Models for Enhancing Endotoxin Responsiveness.

 

An enigma which has existed in endotoxin research for many years is
the diverse nature of experimental models which exist whereby animals
can be rendered more sensitive to the lethal effects of endotoxin.
Formal et al. (37) were the first to show that experimental animals are

rendered more susceptible (235—fold) to Shigella flexneri endotoxin by

 

subcutaneous injection of a sublethal dose of carbon tetrachloride
(CCla). Animals treated 48 hrs prior to endotoxin. with CCl4 died
earlier than normal animals given a dose of endotoxin of comparable
lethality, while the organ distribution of chromium-labeled endotoxin
was not altered by the CCl4 pretreatment. Carbon tetrachloride-treated
guinea pigs become severely hypoglycemic following endotoxin challenge.
Hypoglycemia, however, cannot be the sole factor underlying the enhanced
sensitivity of pretreated guinea pigs to endotoxin since maintenance of
blood sugar levels by glucose infusion did not alter survival (33).
Farrar et al. (34) later reported that livers removed from CCla-treated
rabbits were less capable of detoxifying endotoxin in vitro when
compared to normal livers.

The effect of lead acetate administration on the susceptibility of
rats to endotoxin was first reported by Selye et al. (111). Their data
show that intravenous injection of a sublethal dose of lead acetate
increases the sensitivity of rats to endotoxin from. a variety of
gram-negative bacteria by about 100,000-fold. The lead-sensitizing
effect was greatest when the two agents were administered concurrently,
yet still present when the endotoxin was administered no sooner than 1

hr before or no later than. 7 hrs after lead acetate. While the

mechanism. underlying the lead-induced enhancement of endotoxin

30

susceptibility remains to be established, the ability of lead to
inactivate sulfhydryl groups (12) or block RE function (111) has been
suggested.

Silica selectively alters macrophage functions and is toxic for
macrophages in vitro and in vivo (2,23,60). Vogel et al. (131) recently
used silica to assess the role of the macrophage in the outcome of
endotoxin challenge. A single injection of silica 24 hrs prior to
endotoxin resulted in a four-fold decrease in the 50% lethal dose of
endotoxin. Silica pretreatment also enhanced the production of
interferon in vivo 2 hrs after challenge with endotoxin. Furthermore,
peritoneal macrophages derived from silica-treated mice produced
significantly more lymphocyte activating factor (interleukin 1, 1L1)
than normal macrophages in response to endotoxin stimulation in vitro.
These results corroborate those of Becker and Rudback (6) who
demonstrated potentiation of endotoxin sensitivity by another macrOphage
toxin, carageenan. Although the precise mechanisms by which silica and
carageenan affect. macrophage function. has. not been. elucidated, both
agents destabilize lysosomal. membranes (17,105). The importance of
lysosomal membrane destabilization in the pathogenesis of endotoxicity
remains to be determined. However, these data support and extend the
notion that the macrophage and macrophage-derived factors play a central
role in mediating the toxic effects of endotoxin.

As early as 1958, Suter et al. (121) described the increase in
sensitivity to endotoxin lethality caused by pretreatment of

experimental mice with an attenuated strain. of Mycobacterium

 

tuberculosis (bacillus Calmette- Guerin, BCG). Maximum sensitivity to

 

endotoxin usually occurred 10-14 days after BCG infection, with the

31

majority of deaths occurring earlier than in normal mice receiving only
endotoxin in a dose of comparable toxicity.

The severe hypoglycemia elicited by endotoxin in BCG-infected mice
is a problem which has been investigated by Shands et al. (112,113).
Depletion of muscle and liver glycogen as well as the rate of decline in
blood sugar levels was found to be enhanced in BCG-infected mice
poisoned with endotoxin. Administration of glucose to endotoxin
challenged, BCG—infected mice prolonged death but failed to increase
survival (113). These data led Shands et al. to suggest that endotoxin
lethality is a more complex process than simply severe hypoglycemia.

In a more recent report by Vogel et al. (130), BCG-infected C3H/HeN
mice were found to be approximately 8 times more susceptible to the
lethal effects of endotoxin and produced more interferon upon challenge
with endotoxin than uninfected controls. Macrophages obtained from
BCG-infected C3H/HeN mice also produced significantly more interleukin 1

(1L1) and prostaglandin. E 'when stimulated 'with endotoxin. in 'vitro

2
(132).

It is apparent that the response of experimental animals to
endotoxin is intensified by a multitude of vastly different agents. The
numerous and disparate effects elicited by those agents in vivo coupled
with the scarcity of data utilizing these models has prevented
investigators from drawing firm conclusions regarding the mechanisms by
which endotoxin exerts its toxic and lethal effects. The single common
feature of several of these models of endotoxin potentiation is the
ability to influence the state of the RES. Zymosan, BCG, C.parvum, and

glucan elicit reticuloendothelial. hyperplasia. ‘while silica. and

carageenan are reportedly macrophage—specific toxins. Endotoxin also

32

stimulates RES proliferation yet induces a state of endotoxin tolerance
to the biologic effects of subsequently injected endotoxin as Opposed to
an augmented state of endotoxin susceptibility seen with other RES
stimulants. This paradox as well as the mechanisms surrounding the

enhancement of endotoxin susceptibility remain to be elucidated.

Correlation Between Endotoxemia and Clinical Liver Disease.

 

In addition to its probable role in septic shock, endotoxin has
been demonstrated in plasma and ascites of patients with various liver
conditions. Liehr and Grun (68) used the Limulus assay to determine the
incidence of endotoxemia in 259 patients with various liver diseases.
Their results showed that :h1 patients with fulminant hepatic failure
(n=3), uncomplicated viral hepatitis (n=35), alcoholic hepatitis (n=6),
and chronic active hepatitis (n=12), endotoxemia was present in 100%,
37%, 100% and 50%, reSpectively. Endotoxemia was also found in 79% of
patients with stable and unstable cirrhosis (n=170) (68). Further, a
study by Clemente et al. (21) reported Limulus positivity in plasma of 9
out of 10 patients with alcoholic cirrhosis and functional renal
failure, while negative in 21 cirrhotic patients with normal renal
function. In another study, mortality among patients with alcoholic
cirrhosis and endotoxemia was 48%, while only 17% among patients that
were cirrhotic and Limulus negative, despite otherwise similar initial
clinical and biochemical features (122).

Fulenwider et al. (38) were unable to detect endotoxin in portal
blood, ascites, and peripheral blood from a large group of cirrhotic
patients. A correlation between endotoxemia and fever in patients with

alcoholic hepatitis without apparent infection has also been reported

33

(91). Five patients with Reye's syndrome, a condition that is typically
accompanied by intense fatty infiltration of the liver, were all
positive for endotoxin (25). Elevated serum GPT activities and positive
Limulus results were also noted in 7 of 17 patients with acute benign
viral hepatitis (129). Despite the overwhelming evidence supporting the
occurrence of endotoxemia in patients with a variety of liver
conditions, whether endotoxemia is simply a consequence of hepatic
dysfunction or a contributor to hepatic injury remains in question.
Evidence to support the role of endotoxin in the pathogenesis of
liver disease comes not only from clinical studies but from experimental
studies as well. The relationship between gut-derived endotoxin and
experimental hepatic disease was the subject of a review by Nolan and
Camera (94). Several studies have supported the involvement of
endotoxin. in CClA-induced liver injury. Rats ‘made tolerant to the

biologic effects of endotoxin were largely protected from the

hepatotoxic effects of CCl (90). Another study by Nolan et al. (92)

4
compared the effect of pretreatment with polymyxin B, an antibiotic
which disrupts and renders endotoxin nontoxic, and with gentamycin, an
antibiotic of similar antibacterial spectrum but without antiendotoxin
properties, on CCla-induced liver injury. Pretreatment with polymyxin B
but not gentamycin prevented the elevation in serum aspartate and
alanine aminotransferase activities and histologic liver damage induced
by CCl4 in controls, suggesting a key role for endotoxin.

Endotoxin has also been implicated in the pathogenesis of liver
injury by galactosamine. Grun et al. (46) detected endotoxemia by the

Limulus assay as early as 3 hrs after galN injection. Both the

induction of endotoxin tolerance and removal of the source of endotoxin

34

by experimental colectomy prevented the galN-induced elevation in serum
transaminase levels observed in sham-treated controls.

Toxins that cause liver injury also lead to early impairment of RES
function. Ali and Nolan (1) demonstrated significantly impaired RES
function in rats following alcohol ingestion by measuring blood
clearance of labeled microaggregated albumin. In subsequent studies
using a solid—phase immunoradiometric assay, these investigators showed
delayed clearance of E. coli 026 endotoxin as early as 1 hr after
alcohol ingestion or administration of a sublethal dose of CCl4 (64,93).
These findings corroborate those of Liu (72) in that RES function was
found to be significantly depressed in acute alcoholics. Nolan and
Camara (94) conclude that initial toxic liver injury may allow endotoxin
access to the hepatocyte through membrane changes or impair endotoxin
detoxification.

Cumulatively, these studies clearly show that bacterial endotoxin
may be involved in the pathophysiologic mechanisms involved in not only
septic shook but also clinical liver disease where hepatic endotoxin
present in much lower quantities is suspected of mediating liver cell
necrosis induced by a variety of apparently unrelated toxins. A
prerequisite to understanding these hepatotoxic mechanisms is the
determination of whether liver parenchymal cells are direct targets of
endotoxin or whether the toxic effects of endotoxin on parenchymal cells

are mediated by factors present in the circulation.

10.

11.

12.

13.

LITERATURE CITED

Ali, M.V., and J.P. Nolan. 1967. Alcohol induced depression of
reticuloendothelial function in the rat. J. Lab. Clin. Med.
70:295-301.

Allison, A.C., J.S. Harington, and M. Birbeck. 1966. An
examination of the cytotoxic effects of silica on macrophages.
J. Exp. Med. 124:141-153.

Aschoff, L. 1924. Reticuloendothelial system, p.1-33. lg
Lectures on Pathology. Paul B. Hoeber Inc., N.Y.

Bachmann, W., E. Harms, B. Hassels, H. Henninger, and W. Reutter.
1977. Studies on rat liver plasma membrane. Altered protein
and phospholipid metabolism after injection of
D—galactosamine. Biochem. J. 166:455-462.

Bauer, Ch., W. Bachmann, and W. Reutter. 1972. Studies on
galactosamine hepatitis: Determination cfi? galactosamine
metabolites in the developing rat liver. Hoppe-Seyler's Z.
Physiol. Chem. 353:1053-1058.

Becker, L.J., and J.A. Rudbach. 1978. Potentiation of
endotoxicity by carrageenan. Infect. Immun. 19:1099-1100.

Berry, L.J., D.S. Smythe, and L.G. Young. 1959. Effects of
bacterial endotoxin on metabolism. 1. Carbohydrate depletion

and the protective role of cortisone. J. Exp. Med.
110:389—405.

Berry, L.J. 1966. Effect of environmental temperature on
lethality of endotoxin and its' effect on body temperature in
mice. Fed. Proc. 25:1264-1270.

Berry, L.J., D.S. Smythe, and L.S. Colwell. 1968. Inhibition of
hepatic enzyme induction as a sensitive assay for endotoxin.
J. Bacteriol. 96:1191-1199.

Berry, L.J. 1975. Bacterial Toxins. Lloydia 38:8-20.

Berry, L.J. 1977. Bacterial toxins. Crit. Rev. Toxicol.
5:239-273.

Bertok, L. 1968. Effect of sulfhydryl compound on the lead
acetate-induced endotoxin hypersensitivity of rats. J.
Bacteriol. 95:1974-1975.

Bradley, S.G. 1979. Cellular and molecular mechanisms of action
of bacterial toxins. Ann. Rev. Microbiol. 33:67-94.

35

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

36

Braude, A.I., F.J. Carey, M. Zalesky, and D. Sutherland. 1955.
Studies with radioactive endotoxin. I. The use of Cr to
label endotoxin of Escherichia coli. J. Clin. Invest.
34:850-857.

 

Braude, A.I., F.J. Carey, and M. Zalesky. 1955. Studies with
radioactive endotoxin. II. Correlation of physiologic effects
of radioactivity in rabbits injected with lethal doses of §h_
coli endotoxin labeled with radioactive sodium chromate. J.
Clin. Invest. 34:858-866.

 

Carswell, E.A., L.J. Old, R.L. Kassel, S. Green, N. Fiore, and B.
Williamson. 1975. An endotoxin-induced serum factor that

causes necrosis of tumors. Proc. Nat. Acad. Sci.
72:3666-3670.

Catanzaro, F.J., H.J. Schwartz, and R.C. Graham. 1971. Spectrum
and possible mechanism of carrageenan cytotoxicity. Am. J.
Pathol. 64:387-403.

Cerriotti, G. 1974. Ornithine carbamoyltransferase:
determination by manual method. p. 691-695. 13_ H.U.
Bergmeyer (ed.), Methods of enzymatic analysis, vol. 2.
Academic Press, Inc., New York.

Chedid, E1, R.C. Skarnes, and M. Parant. 1963. Characterization
of Cr-labeled endotoxin and it's identification in plasma

and urine after parental administration. J. Exp. Med.
117:561—571.

Chedid, L., M. Parant, C. Damais, F. Parant, D. Juy, and A.
Galelli. 1976. Failure of endotoxin to increase nonspecific

resistance to infection of lipopolysaccharide low-responder
mice. Infect. Immun. 13:722-727.

Clemente, C., .1. Bosch, and .1. Rodes. 1977. Functional renal
failure and hemorrhagic gastritis associated with endotoxemia
in cirrhosis. Gut 18:556-560.

Coleman, B.D. 1964. Septic shock in pregnancy. Obstet. Gynecol.
24:895-902.

Comolli, R., and A. Perin. 1963. In vitro action of silicogenic
and non—silicogenic dusts on macrOphage metabolism. Proc.
Soc. Exp. Biol. Med. 113:289-293.

Cook, J.A., W.J. Dougherty, and T.M. Hold. 1980. Enhanced
sensitivity to endotoxin induced by the RE stimulant, glucan.
Circulatory Shock 7:225-238.

Cooperstock, M.S., R.P. Tucker, and J.F. Baublis. 1975. Possible
pathogenic role of endotoxin in Reye's syndrome. Lancet
1:1272-1274.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

37

Cremer, N. and D.W. Watson. 1957. Influence of stress on
distribution of endotoxin in RES determined by fluorescein
antibody technic. Proc. Soc. Exp. Biol. Med. 95:510-513.

Cuatrecasas, P., and S. Segal. 1965. Mammalian galactokinase,
development and adaptive characteristics in the rat liver. J.
Biol. Chem. 240:2382-2388.

Decker, K., and D. Keppler. 1974. Galactosamine hepatitis: Key
role of the nucleotide deficiency period in the pathogenesis
of cell injury and cell death. Rev. Physiol. Biochem.
Pharmacol. 71:77-103.

Economu, J.S., and H.S. Shin. 1978. Lymphocyte activating factor.
1. Generation and physiochemical characterization. J.
Immunol. 121:1446—1452.

Elin, R.J., and S.M. Wolff. 1973. Nonspecificity of the Limulus
Amebocyte Lysate test: positive reactions with
polynucleotides and proteins. J. Infect. Dis. 128:349-352.

El—Mofty, S.K., M.C. Scrutton, A. Serrioni, C. Nicolini and J.L.
Farber, 1975. Early, reversible plasma membrane injury in
galactosamine-induced liver cell death. Am. J. Path.
79:579-596.

Farber, J.L., G. Gill, and Y. Konishi. 1973. Prevention of
galactosamine-induced liver cell necrosis by uridine. Amer.
J. Pathol. 72:53-62.

Farrar, W.E., Jr., and J.G. Watson. 1964. Hypoglycemia following
endotoxin administration in animals with liver damage. Proc.
Soc. Exptl. Biol. Med. 115:833-837.

Farrar, W.E., Jr., M. Edison, and T.H. Kant. 1968. Susceptibility
of rabbits to pyrogenic and lethal effects of endotoxin after

acute liver injury. Proc. Soc. Exptl. Biol. Med.
128:711-715.

Feldberg, W., and P.N. Saxena. 1975. Prostaglandins, endotoxin
and lipid A on body temperature in rats. J. Physiol.
249:601-615.

Filkins, J.F., and R.P. Cornell. 1974. Depression of hepatic
gluconeogenesis and the hypoglycemia of endotoxin shock. Am.
J. Physiol. 227:778-781.

Formal, S.B., H.E. Noyes, and IL. Schneider. 1960. Experimental
Shigella infections. III. Sensitivity of normal, starved, and
carbon tetrachloride-treated guinea pigs to endotoxin. Proc.
Soc. Exptl. Biol. Med. 103:415-418.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

38

Fulenwider, J.T., C. Sibley, and S.F. Stein. 1980. Endotoxemia of
cirrhosis: an observation not substantiated. Gastroenterology
78:1001-1004.

Galanos, C., 0. Luderitz, E.T. Reitschel, and O. Westphal. 1977.
Newer aspects of the chemistry and biology of bacterial
lipopolysaccharides with special reference to their lipid A
component. p. 239-335. T.N. Goodwin (ed.), International
review of biochemistry, biochemistry of lipids II. University
Park Press, Baltimore.

Galanos, C., N.A. Freudenberg, and W. Reutter. 1979. Galactos-
amine-induced sensitization to the lethal effects of
endotoxin. Proc. Nat. Acad. Sci. 76:5939-5943.

Galvin, M.J., K. Shupe, and A.M. Lefer. 1978. Antiendotoxin
actions of methylprednisolone in isolated perfused cat liver.
Pharmacology 17:181-190.

Geller, P., E. Merril, and E. Jawetz. 1954. Effects of cortisone
and antibiotics on lethal action of endotoxins in mice. Proc.
Soc. Exp. Biol. Med. 86:716-719.

Gmeiner, J. 1975. The isolation of two different lipopolysac-
charide fractions from various Proteus mirabilis strains.
Eur. J. Biochem. 58:621-626.

 

Green, 5., A. Dobrjansky, E.A., Carswell, R.L. Kassell, L.J. Old,
N. Flore, and M.K. Schwartz. 1976. Partial purification of a

serum factor that causes necrosis of tumors. Proc. Nat. Acad.
Sci. 73:381-385.

Greer, G.G., N.A. Epps, and W.J. Vail. 1973. Interaction of
lipopolysaccharides with 'mitochondria. II. Effects of

magnesium ions on toxicity of a rough lipopolysaccharide. J.
Infect. Dis. 128:724—729.

Grun, M., H. Liehr, and U. Rasenack. 1976. Significance of
endotoxemia in experimental "Galactosamine-Hepatitis" in the
rat. Acta Hepato-gastroenterol. 23:64-81.

Harlan, W.R. 1969. Gram-negative sepsis: another piece of the
mosaic. New Eng. J. Med. 281:1127-1128.

Harris, R.A., D.L. Harris, and D.E. Green. 1968. Effect of
Bordetella endotoxin upon mitochondrial respiration and
energized processes. Arch. Biochem. Biophys. 128:219-230.

 

Henon, M., A. Delaunay, and S. Bazin. 1968. Reactions
metaboliques 'provoquees "in. vivo" et "in.‘vitro" pars less
endotoxines bacterinnes. Biol. Med. (Paris) 57:471-519.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

39

Hoffman, M.K., H.F. Oettgen, L.J. Old, R. Mittler, and U.
Hammerling. 1978. Induction and immunological prOperties of

tumor necrosis factor. RES J. Reticuloendothelial Soc.
23:307-319.

Howard, J.G., D. Rowley, and A.C. Wardlaw. 1958. Investigations
on the mechanism of stimulation of nonspecific immunity by
bacterial lipopolysaccharides. Immunology 1:181-203.

Ignarro, L.J. 1973. Preservation of structural integrity of liver
lysosomes and membrane stabilizing action of anti-inflammatory
drugs, catecholamines and cyclic adenosine monophosphate in
isotonic salt media. Biochem. Pharmacol. 22:1269-1282.

Jann, B., K. Reske, and K. Jann. 1975. Heterogeneity of
lipopolysaccharides. Analysis of polysaccharide chain lengths
by sodium dodecylsulfate-polyacrylamide gel electrophoresis.
Eur. J. Biochem. 60:239-246.

Janoff, A., G. Weissmann, B.W. Zweifach, and L. Thomas. 1962.
Pathogenesis of experimental shock. IV. Studies on lysosomes
in normal and tolerant animals subjected to lethal trauma and
endotoxemia. J. Exp. Med. 116:451-465.

Janson, P.M.C., S.H. Kuhn, and J.J. Geldenhuys. 1975. Lysosomal
disruption during the development of endotoxin shock in the
baboon. S. Afr. Med. J. 47:1041-1047.

Kampschmidt, R.P., and L.A. Pulliam. 1975. Stimulation of
antimicrobial activity in the rat with leukocytic endogenous
mediator. RES. J. Reticuloendothel. Soc. 17:162-169.

Keppler, D., R. Lesch, W. Reutter, and K. Decker. 1968.
Experimental hepatitis induced by D-galactosamine. Exptl.
Mol. Pathol. 9:279-290.

Keppler, D., J. Rudigier, E. Bischoff, and K. Decker. 1970. The
trapping of uridine phosphates by D-galactosamine,
D-glucosamine, and 2-deoxy-D-galactose. A study on the

mechanism of galactosamine hepatitis. Eur. J. Biochem.
17:246-253.

Keppler, D., and D.F. Smith. 1974. Nucleotide contents of ascites
hepatoma cells and their changes induced by D-galactosamine.
Cancer Res. 34:705-711.

Kessel, R.W., L. Monaco, and M.A. Marchisio. 1963. The
specificity of the cytotoxic action of silica - a study _:I_._1}_
vitro. Br. J. Exp. Pathol. 443351-364.

Konttinen, A., M. Rajasalmi, and J. Paloheimo. 1964. Serum enzyme
activities in endotoxin shock. Am. J. Physiol. 207:385-388.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

40

Kuhn, R., J.H. Leppelmann, and H. Fischer. 1959.
Stoffwechselversuche mit radioactiv markieterm D-Galaktosamin.
Leibigs Ann. Chem. 620:15-20.

LaNoue, K., A. Mason, Jr., and J. Daniels. 1968. The impairment

of gluconeogenesis by gram-negative infection. Metab. (Clin.
Exp.) 17:606-611.

Leibowitz, A.I., A.O. Vladutiu, and J.F. Nolan. 1979.
lmmunoradiometric assay of endotoxin in the serum. Clin.
Chem. 25:68-70.

Levin, J., and F.B. Bang. 1968. Clottable protein in Limulus:
its location and kinetics of its coagulation by endotoxin.
Thromb. Diath. Haemorrh. 19:186-197.

Levy, E., F.C. Path, and. B.Ha Ruebner. 1967. Hepatic. changes
produced by a single dose of endotoxin in the mouse. Light
microscopy and histochemistry. Am. J. Pathol. 51:269-284.

Levy, B., F.C. Path, R.J. Slusser, and B.H. Ruebner. 1968.
Hepatic changes produced by a single dose of endotoxin in the
mouse. Electron microscopy. Am. J. Pathol. 52:477-502.

Liehr, H., and M. Grun. 1977. Clinical aspects of Kupffer cell
failure in liver diseases. p. 427-436. E. Wisse and D.L.
Knook (ed.), 12' Kupffer Cells and Other Liver Sinusoidal
Cells. Elsevier/North- Holland Biomedical Press, Amsterdam.

Liehr, H., M. Grun, H.P. Seelig, R. Seelig, W. Reutter, and W.D.
Heine. 1978. On the pathogenesis of galactosamine hepatitis.
Indications of extrahepatocellular mechanisms responsible for
liver cell death. Virchows Arch. B. Cell Path. 26:331-344.

Lierh, H., M. Grun, H.P. Seeling, R. Seelig, and U. Rasenack.
1978. Synergistic action of hepatocyte membrane defect and
activated complement system in liver cell death. Acta Hepato-
gastroenterol. 25:105-110.

Lillehei, R.C., J.K. Longerbeam, J.H. Bloch and W.G. Manax. 1964.
The nature of experimental irreversible shock with its
clinical application. p. 139. IE S.G. Hershey (ed.), Shock.
Boston, Little, Brown, and Co., Inc.

Liu, Y.K. 1979. Phagocytic capacity of the reticuloendothelial
system 2h: alcoholics. RES. .1. Reticuloendothel. Soc.
25:605-613.

Mannel, D.N., M.S. Meltzer, and SE. Mergenhagen. 1980. Generation
and characterization of lipOpolysaccharide-induced and
serum-derived cytotoxic factor for tumor cells. Infect.
Immun. 28:204—211.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

41

Mathison, J.G., and R.J. Ulevitch. 1979. The clearance, tissue
distribution, and cellular localization of intravenously

injected lipopolysaccharide in rabbit. J. Immunol.
123:2133—2143.

McCabe, W.R. 1974. Gram-negative bacteremia. Adv. Intern. Med.
19:135-158.

McCallum, R.E., and L.J. Berry. 1972. Mouse liver fructose-1,
6-diphosphatase and glucose-6-phosphatase activities after
endotoxin poisoning. Infect. Immun. 6:883-885.

McCallum, R.E., and L.J. Berry. 1973. Effects of endotoxin on
gluconeogenesis, glycogen synthesis, and liver glycogen
synthesis in mice. Infect. Immun. 7:642-654.

McCallum, R.E. 1980. Mediated inhibition of hepatic
gluconeogenesis by endotoxin. p. 87-90. IE_D. Schlessinger
(ed.), Microbiology-1980. American Society for Microbiology,
Washington, D.C.

Mela, L., L.V. Bacalzo, and L.D. Miller. 1971. Defective
oxidative metabolism of rat liver mitochondria in hemorrhage
and endotoxic shock. Am. J. Physiol. 220:571-577.

Mela, L., L.D. Miller, L.V. Bacalzo, K. Olofsson, and R.R. White.
1973. Role of intracellular variations of lysosomal enzyme
activity and oxygen tension in mitochondrial impairment in

endotoxemia and hemorrhage in the rat. Annals of Surgery
178:727-735.

Menten, M.L., and H.M. Manning. 1924. Blood sugar studies on
rabbits infected with organisms of the enteritidioparatyphoid
B group. J. Med. Res. 44:675-693.

Mergenhagen, S.E. 1980. Endogenous mediators in host responses to
bacterial endotoxins. p. 3-166. 22; Microbiology-1980 D.
Schlessinger (ed.), American. Society for ‘Microbiology,
Washington, D.C.

Mizel, S.B., J.J. Oppenheim, and D.L. Rosenstreich. 1978.
Characterization of lymphocyte activating factor (LAF)
produced by the macrophage cell line, P388D . II. Biochemical
characterization of LAF induced by activated T cells and LPS.
J. Immunol. 120:1504-1508.

Mizel, S.B. 1979. Biochemical characteristics of macrophage-
derived lymphostimulatory factors. I_n J.H. Hadden and W.E.
Stewart II (ed.), Molecular mediators of cellular immunity.
Humana Press, Clifton, N.J.

Moore, R.N., K.J. Goodrum, and L.J. Berry. 1976. Mediation of an

endotoxic effect by macrophages. RES. J. Reticuloendothel.
Soc. 19:187-197.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

42

Moore, R.N. K.J. Goodrum, R. Couch, Jr., and L.J. Berry. 1978.
Factors affecting macrophage function: glucocorticoid

antagonizing factor. RES. J. Reticuloendothel. Soc.
23:321-332.

Moore, R.N., K.J. Goodurm, R.E. Couch, Jr., and L.J. Berry. 1978.
Elicitation of endotoxemic effects in C3H/HeJ mice with
glucocorticoid antagonizing factor and partial
characterization of the factors. Infect. Immun. 19:79-86.

Morrison, D.C., S.J. Betz, and D.M. Jacobs. 1976. Isolation of a
lipid A ‘bound polypeptide responsible for "LPS initiated"
mitogenesis of C3H/HeJ spleen cells. J. Exp. Med.
144:840-846.

Musson, R.A., D.C. Morrison, and R.J. Ulevitch. 1978.
Distribution of endotoxin (lipopolysaccharide) in the tissues
of 1ipopolysaccharide-responsive and -unresponsive mice.
Infect. Immun. 21:448-457.

Nolan, J.F., and M.V. Ali. 1973. Endotoxin and the liver. II.
Effect of tolerance on carbontetrachloride induced injury. J.
Med. 4:23-38.

Nolan, J.F., and M.V. Ali. 1974. Endotoxemia in liver disease.
Lancet 1:999.

Nolan, J.F., and A.I. Leibowitz. 1978. Endotoxin and the liver.
III. Modification of aucute carbon tetrachloride injury by
polymyxin B-an antiendotoxin. Gastroentorology. 75:445-449.

Nolan, J.F., A.I. Leibowitz, and A.O. Vladutiu. 1980. Influence
of carbon tetrachloride on circulating endotoxin after
exogenous administration of endotoxin in rats. Proc. Soc.
Exp. Biol. Med. 165:453-456.

Nolan, J.F. and D.S. Camara. 1982. Endotoxin, sinusoidal cells,
and liver injury. p. 361-376. IE_H. Popper and F. Schaffner
(ed.), Progress in liver diseases. Grune & Stratton, Inc.,
N.Y.

Noyes, H.E., C.R. McInturf, and C.J. Blahuta. 1959. Studies on
distribution of Escherichia coli endotoxin in mice. Proc .
Soc. Exp. Biol. Med. 100:65-68.

 

Nydick, J., F. Wroblewski, and J.S. LaDue. 1955. Evidence for
increased serum glutamic oxaloacetic transaminase (SGO-T)
activity following graded myocardial infarcts in dogs.
Circulation 12:161-168.

Parant, M., F. Parant, and L. Chedid. 1978. Enhancement of the
neonate's nonspecific immunity to Klebsiella infection by
muramyl dipeptide, :3 synthetic immunoadjuvant. Proc. Natl.
Acad. Sci. U.S.A. 75:3395-3399.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

43

Parant, M.A., F.J. Parant, and L.A. Chedid. 1980. Enhancement of
resistance to infections by endotoxin-induced serum. factor

from Mycobacterium bovis BCG-infected mice. Infect. Immun.
28:654—659.

 

Phillip-Dormston, W.K. and IL. Siegert. 1974. Prostaglandins of
the E and F series in rabbit cerebrospinal fluid during fever
produced by Newcastle disease virus, E. coli endotoxin, or
endogenous pyrogen. Med. Microbiol. Immunol. 159:279-284.

Phillips, L.J., and L.J. Berry. 1970. Circadian rhythm of mouse

liver phosphoenolpyruvate carboxykinase. Am. J. Physiol.
218:1440-1444.

Postel, J., P.R. Schloerb, and D. Furtado. 1975. Pathophysiologic
alterations during bacterial infusions for the study of
bacteremic shock. Surg. Gynecol. Obstet. 141:683-692.

Ramadori, G., U. Hopf, and K.H. Meyer zum Buschenfelde. 1979.
Binding sites for endotoxin lipopolysaccharide on the plasma
membrane of isolated rabbit hepatocytes. Acta
Hepatogastroenterol. 26:368-374.

Ramadori, G., U. Hopf, C. Galanos, M. Freudenberg, and K.H. Meyer
zum Buschenfelde. 1980. In vivo and in vitro reactivity of
lipopolysaccharide and lipid A with parenchymal and
nonparenchymal liver cells in mice. lfl.H° Liehr and M. Grun
(ed,), The reticuloendothelial system and the pathogenesis of
liver disease. Elsevier/North Holland Biomedical Press.

Ribblg1 J.C., M. Zalesky, and A.I. Braude. 1959. Distribution of
Cr-labeled endotoxin in cortisone treated mice. Bull. Johns
Hopkins Hosp. 105:272-283.

Rios, A., :nui R.L. Simmons. 1972. Poly-2-vinylpyridine-N-oxide
reverses the immunosuppressive effects of silica and
carrageenan. Transplantation 13:343-345.

Rippe, D.F., and L.J. Berry. 1972. Study of inhibition of
induction of phosphoenolpyruvate carboxykinase by endotoxin
with radial immunodiffusion. Infect. Immun. 6:766-772.

Rowlegzl'D., J.G. 'Howard, and. C.R. Jenkin. 1956. The. fate of
P-labeled bacterial lipopolysaccharide in laboratory
animals. Lancet 270:366-367.

Ruiter, D.J., J. Vander Meulen, A. Brouwer, M.J.R. Hummel, B.J.
Mauw, J.G.M. Van der Ploeg, and E. Wisse. 1981. Uptake by
liver cells of endotoxin following its intravenous injection.
Lab. Invest. 45:38-45.

Schrader, W.H., B.F. Woolfrey, and R.D. Brunning. 1964. Studies
with tritiated endotoxin. III. The local Schwartzman
reaction. Am. J. Pathol. 44:597-611.

110.

111.

112.

113.

114.

115.

116.

118.

119.

120.

121.

122.

123.

44

Schrader, W.H., and B.F. Woolfrey. 1963. Studies with tritiated
endotoxin. I. Preparation and analysis of tritiated
endotoxin. Am. J. Pathol. 42:225-233.

Selye, H., B. Tuchweber, and L. Bertok. 1966. The effect of lead
acetate on the susceptibility of rats to endotoxin. J.
Bacteriol. 91:884-890.

Shands, J.W., Jr., V. Miller, H. Martin, and V. Senterfitt. 1969.
Hypoglycemic activity of endotoxin. II. Mechanisms of the
phenomenon on BCG-infected mice. J. Bacteriol. 98:494-501.

Shands, J.W., Jr., V. Miller, and H. Martin. 1969. The
hypoglycemic activity of endotoxin. I. Occurrence in animals

hyperreactive to endotoxin. Proc. Soc. Exp. Biol. Med.
130:413-417.

Sheagren, J.N. 1981. Septic shock and corticosteriods. New Eng.
J. Med. 305:456-457.

Shtasel, T.F., and 'L.J. Berry. 1969. Effect: of endotoxin. and
cortisone on synthesis of ribonucleic acid and protein in
livers of mice. J. Bacteriol. 97:1018—1025.

Skarnes, R.C. 1970. Host defense against bacterial endotoxemia:
Mechanism in normal animals. J. Exp. Med. 132:300-316.

'. Sobocinski, P.Z., M.C. Powanda, W.J. Canterbury, S.V. Machota, R.I.

Walker, and S.L. Snyder. 1977. Role of zinc in the abatement
of hepatocellular damage and mortality incidence in endotoxic
rats. Infect. Immun. 15:950-957.

Sultzer, B.M. 1968. Genetic control of leukocyte responses to
endotoxin. Nature (London) 219:1253-1254.

Sultzer, B.M., and B.S. Nilsson. 1972. PPD tuberculin is a B-cell
mitogen. Nature (London) New Biol. 240:198-200.

Sultzer, B.M., and G.W. Goodman. 1976. Endotoxin protein: a
B-cell mitogen and polyclonal activator of C3H/HeJ lyphocytes.
J. Exp. Med. 144:821-827.

Suter, E., G.E. Ullman, and R.C. Hoffman. 1958. Sensitivity of
mice to endotoxin after vaccination with BCG (Bacillus
Calmette-Guerin). Proc. Soc. Exp. Biol. Med. 99:167-169.

Tarao, K., K. So, and T. Moroi. 1977. Detection of endotoxin in
plasma and ascites of patients with cirrhosis: Its clinical
significance. Gastroenterology 73:539-542.

Tavakoli, H., and R.J. Moon. 1982. In vitro and in vivo
association and subcellular distribution of toxic and

experimentally' modified endotoxin. Can” J. ‘Microbiol.
28:822-829.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

45

Togawa, A. , J.J. Oppenheim, and S.B. Mizel. 1979.
Characterization. of lymphocyte activating factor (LAF)
produced by human mononuclear cells: biochemical relationship
of high and low molecular weight forms of LAF. J. Immunol.
122:2112-2118.

Ulevitch, R.J. 1978. The preparation and characterization of a

radioiodinated bacterial lip0polysaccharide. Immunochemistry
15:157-164.

Ulevitch, R.J., and .A.R~ Johnston. 1978. The. modification of
biophysical and endotoxic properties of bacterial
lipopolysaccharides by serum. J. Clin. Invest. 62:1313-1324.

Ulevitch, R.J., A.R. Johnston, and D.B. Weinstein. 1979. A new
function for high density lipoproteins: their participation

in intravascular reactions of bacterial lipopolysaccharides.
J. Clin. Invest. 64:1516-1524.

Vane, J.R. 1971. Inhibition of prostaglandin synthesis as a
mechanism of action of aspirin-like drugs. Nature (London)
New Biol. 231:232-235.

Van Vliet, A.C.M., H.C.M. Maas, and J.H.P. Wilson. 1980. The
effect of portasystemic shunting and RES function on
endotoxemia in human liver disease. p. 159-164. $2.3: Liehr
and M. Grun (ed.), The reticuloendothelial system and the
pathogenesis of liver Disease. Amsterdam, Elsevier/North
Holland.

Vogel, S.N., R.N. Moore, J.D. Sipe, and D.L. Rosenstreich. 1980.
BCG-induced enhancement of endotoxin sensitivity in C3H/HeJ
mice. 1. In vivo studies. J. Immunol. 124:2004-2009.

Vogel, S.N., K.E. English, and A.D. O'Brien. 1982. Silica
enhancement of murine endotoxin sensitivity. Infect. Immun.
38:681-685.

Vogel, S.N., L.L. Weedon, L.M. Wahl, and D.L. Rosenstreich. 1982.
BCG-induced enhancement of endotoxin sensitivity in C3H/HeJ
mice. II. T cell modulation of macrophage sensitivity to LPS
in vitro. Immunobiology 160:470—493.

Watson, J., and R. Riblet. 1974. Genetic control of responses to
bacterial lip0polysaccharide in mice. I. Evidence for a
single gene that influences mitogenic and immunogenic
responses to lipopolysaccharides. J. Exp. Med.
140:1147-1161.

Weil, M.H., H,. Shubin, and M. Biddle. 1964. Shock caused by
gram-negative microorganisms: analysis of 169 cases. Ann.
Intern. Med. 60:384-400.

135.

136.

137.

138.

139.

140.

46

Weissman, C., and L. Thomas. 1962. Studies on lysosomes. I. The
effect of endotoxin, tolerance and cortisone on the release of
enzymes from the granular fraction of rabbit liver. J. Exp.
Med. 116:433—450.

Westphal, 0., and T. Sommer. 1977. The history of pyrogen
research. p. 221-235, 13. D. Schlessenger (ed.),
Microbiology-1977. American Society for Microbiology,
Washington, D.C.

Willerson, J.T., R.L. Trelstad, T. Pincus, S.B. Levy, and S.M.
Wolff. 1970. Subcellular localization of Salmonella
enteritidis endotoxin in liver and spleen of mice and rats.
Infect. Immun. 1:440-445.

 

Woodward, J.M, M.L. Camblin, and M.H. Jobe. 1969. Influence of
bacterial infection on serum enzymes of white rats. Appl.
Microbiol. 17:145-149.

Zlydaszyk, J.C., and R.J. Moon. 1976. Fate of 51Cr-labeled
lipopolysaccharide in tissue culture cells and livers of
normal mice. Infect. Immun. 14:100-105.

Zweifach, B.W., and A. Janoff. 1965. Bacterial endotoxemia. Ann.
Rev. Med. 16:201-220.

ARTICLE I

Lack Of Correlation Between Tissue Localization

And Biological Responses Of Mice To Endotoxin

47

ABSTRACT

One hour after injection, subcutaneously administered endotoxin is
about 50-fold less concentrated in liver than is intravenously or
intraperitoneally injected endotoxin. By 24 hours less than 15% of
subcutaneously injected endotoxin has left the site of injection as
estimated by 51Cr-LPS localization. Despite the differences in
distribution, the mean LDSO of subcutaneously injected endotoxin is only
about 2-fold higher than injection by the other routes. Furthermore,
subcutaneously injected mice exhibit numerous physical and metabolic
alterations typical of endotoxin poisoning including (1) inhibition of
induction of liver PEPCK by hydrocortisone, (ii) liver glycogen
depletion, and (iii) enhanced hypothermia at 5°C. Collectively, these
data indicate little correlation between the 32 v_i_v_c_> localization of
endotoxin among host tissues and biological responses typically

associated with endotoxin poisoning, including death.

48

INTRODUCTION

Previous studies from our laboratory (27) and others (9,12,15,26)
have demonstrated the importance of the liver in clearance of endotoxin
from blood. The liver is also the site of numerous endotoxin-induced
metabolic alterations including carbohydrate depletion (2,4,16),
decreased activity of a variety of liver enzymes (3,5,16), impairment of
mitochondrial function (17,19), and liver damage as judged by elevation
of transaminase levels in serum (20,25). In addition, changes in a
variety of cell types, including neutrophils. platelets, blood
monocytes, and tissue macrophages, as well as activation of host cascade
systems such as complement and coagulation are observed following
endotoxin poisoning. The number and complexity of the in_yiyg changes
has hindered determinations on exactly how these various cellular and
molecular changes interact to produce the toxic and lethal effects of
endotoxin. Furthermore, whether the pathophysiologic changes induced by
endotoxin result primarily from the direct interaction of endotoxin with
critical host tissues or are mediated by soluble factors remains in
question (6). Most recent i2_yiyg_evidence tends to support the notion
that mediators play the major role in responses to endotoxin.
Additional data which would support that concept would be the
demonstration of :a lack of correlation between the biological
consequences of endotoxin poisoning and the tissue localization of the
bacterial toxin.

The initial objective of this study was to determine eXperimentally
whether the lethal effects of endotoxin are dependent upon the
quantitative localization of the toxin among specific host organs.

Experiments were designed to compare the 50% lethal dose (LDSO) and the

49

50

organ distribution of intravenously (iv), intraperitoneally (ip), and
subcutaneously (sc) injected endotoxin in mice. Noting that the organ
distribution of radiolabeled endotoxin following subcutaneous injection
was markedly different from the other routes of injection despite
minimal differences in the 11350. we evaluated a selected number of

pathOphysiological responses to subcutaneously administered endotoxin.

MATERIALS AND METHODS

Animals. Female CD-1 mice (22-28g), obtained from Charles River
Labs, Wilmington, Del., were used in all experiments. Mice were housed
5 per cage under standard laboratory conditions with Purina laboratory
chow and water available ad libitum. Fasting animals had access to
water only.

Endotoxin. Lyophilized Salmonella typhimurium lipopolysaccharide

 

prepared by Westphal phenol-water extraction was purchased from Difco
Laboratories, Detroit, Michigan, and served as endotoxin in all
experiments. Appropriate dilutions of endotoxin were prepared in
sterile, isotonic saline and stored at -20°C. The mean 50% lethal dose
(LDso) of endotoxin was determined by the method of Reed and Muench (22)
using a 72 hour post-injection endpoint. Groups of at least 10 mice
were used to establish an endpoint for each dose of endotoxin.

Radiolabeled endotoxin. Sodium chromate (Na251CrOu, New England

 

Nuclear Corp., Boston, Mass.) was purchased in 2 ml volumes (1 mCi/ml)
with a specific activity of about 200-500 Ci/g. 51Cr-labeled endotoxin
was prepared by the method of Chedid et al (7), as outlined by Zlydaszyk
and Moon (27). The final preparation contained toxic, high molecular
weight endotoxin with a specific activity of approximately 9.5 uCi/mg.
Chromium-labeled endotoxin was diluted in appropriate volumes of sterile
saline just prior to injection.

Concentrations of both radiolabeled and unlabeled endotoxin were
determined by assay of 2—keto-3-deoxyoctulosonate (ICDO), following a
modification of the method of Cynkin and Ashwell (8). Aliquots of

endotoxin to be quantified were initially hydrolyzed in 0.2 M H280“ for

51

52

45 minutes in a boiling water bath. The hydrolyzed samples were then
subjected to the thiobarbituric acid assay (8). This procedure allowed
the quantitative determination of endotoxin solutions with
concentrations ranging from 4-300ug/m1.

Tissue distribution of 51Cr-labeled endotoxin. Groups of at least
6 mice ‘were injected either in. a lateral tail 'vein (iv),
intraperitoneally (ip), or subcutaneously (sc) into the interscapular
region with the appropriate LDSO of endotoxin containing approximately
0.5 uCi of radiolabeled endotoxin. At selected times individual mice
were heparinized, killed by intravenous injection of 1.625 mg of sodium
pentobarbital, and transferred to a platform which drained into a
collecting vessel. Parallel lateral incisions were made to expose the
thorax and abdomen. The inferior vena cava was severed and residual
blood was flushed from the various organs by injection of 15ml of saline
into the apex of the heart. This procedure provided sufficient flushing
to visibly clear blood from all organs and allow the controlled
collection of blood to determine blood-associated radioactivity. Skin,
carcass, blood and selected organs were individually placed into vials
and counted in a Packard gamma counter. The amount of radioactivity per
sample was expressed as a percentage of the total recovered and was
calculated as follows: (radioactivity of sample/total recovered
radioactivity) x 100 = Percent Uptake. To determine the amount of
radioactivity excreted, urine and feces were collected throughout the
examination period, homogenized in 100 ml of water and counted. The
total amount of radioactivity associated ‘with tissues of mice (not
excreted) expressed as percent recovery was calculated as follows:

tissue-associatedjgpm
tissue-associated cpm + excreted cpm

 

X 100 = Percent Recovery

53

Over 90% of the injected 51Cr-endotoxin was routinely recovered in all
experiments.

Limulus Assay. Whole liver homogenates were prepared in 9 mls of

 

pyrogen-free sterile saline and assayed for endotoxin by the Limulus
amebocyte lysate assay. Limulus assays were performed in duplicate on
lO-fold dilutions of liver homogenates by mixing 0.1 ml of homogenate
sample with 0.1 ml of limulus lysate (M.A. Bioproducts, Walkersville,
Md.) in 10x75 mm pyrogen-free glass test tubes. The reaction mixtures
were incubated for 60 min at 37°C and then observed and graded for
degree of gelation. Results were recorded as firm gel (+2), incomplete
gel with increased opacity and viscosity (+1), or negative reaction (-).
Activity of the lysate batch ‘was verified with the same endotoxin
preparation used throughout this study. Negative controls consisted of
tubes containing; 0.1 ad. of lysate and 0.1 ad. of pyrogen-free saline
diluent.

Liver glycogen determination. Groups of fed mice were injected
subcutaneously with sterile saline or with 285ug of endotoxin at 9 am
(0 hr) and killed at appropriate times thereafter by cervical
dislocation. The entire experimental period encompassed 1%) hr. Food
was withheld following initial injection. Liver glycogen was assayed by
the method of Kemp and Kits-van Heijningen (10). The percentage of
liver glycogen expressed in glucose equivalents was determined as
follows: [mg of glycogen/mg liver (wet wgt)] x 100 = % liver glycogen.

Hepatic PEPCK assay. Whole liver homogenates were assayed for

 

phosphoenolpyruvate carboxykinase (PEPCK) activity by the method of Lane
et a1 (11). Hepatic PEPCK was induced by subcutaneous injection of 5 mg

of hydrocortisone (hydrocortisone-21-sodium succinate; Sigma Chemical

54

Co., St. Louis, M0.) in 0.2 ml of sterile saline. PEPCK activity was
expressed as nanomoles of H1”C03 fixed/mg protein/min. Protein
concentrations were determined by the method of Lowry et a1 (13).

Body temperature determinations. Groups of mice were injected

 

subcutaneously with sterile saline or with various amounts of endotoxin
in a total volume of 0.1 m1 and housed separately in a refrigerated room
maintained at 5°C. Body temperatures were determined rectally over a 6
hr period using a Telethermometer probe (Yellow Springs Instrument Co.,
Yellow Springs, Ohio).

Statistical evaluation. Where appropriate, results were analyzed

 

for statistical significance (P<05) by the Mann-Whitney rank order test

(24).

RESULTS

Endotoxin toxicity. Initially, the LDSO of endotoxin for the three

 

routes of injection was determined. The routes of injection and
corresponding LD50 of endotoxin were: iv - 125ug; ip - 160ug; sc -
285ug.

Organ distribution of the LD50 of 51Cr-endotoxin after iv, ip, or

 

sc administration. Groups of at least 6 mice were injected iv, ip, or

 

sc with the appr0priate LD50 of 51Cr-endotoxin. After one hour, the
distribution of radioactivity among host organs was determined. Blood
was flushed from the vasculature of mice to distinguish between organ-
associated and blood-associated endotoxin. With either ip or iv
injection the majority of radioactive endotoxin (>35%) was associated
with the liver and spleen with substantially lesser quantities
distributed among the remaining tissues (Figure 1). By contrast, after
subcutaneous injection over 85% of the label remained at the site of
injection. Very little endotoxin reached the dry tissues, especially
the liver. Not included in Figure 1 are the amounts of radioactivity
associated with gall bladder. adrenals, intestines, heart and brain,
which collectively accounted for less than 8% of the total radioactivity
recovered after iv, ip, or sc injection. By 24 hrs, only slight
increases in the organ radioactivity were observed in sc injected mice
(Table 1). Irregardless of the route of injection, all mice receiving
endotoxin exhibited similar physical signs such as anorexia, diarrhea,
conjuctival exudate, lethargia, and ruffled fur.

To be certain that the isotope distribution of 51Cr-endotoxin did

not reflect dissociated chromium ip vivo, whole animal distribution of

55

 

56

 

   

 

x\\\\\\\\

W

   
   

 

 

    

 

 

2::-
er: E
2\
j
k\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
.‘g'.'.'.'g'.I
N ‘0 ID Q'n N ""

BXVldn .LNBOHBd

SPLEEN KIDNEYS LUNG SKIN CARCASS BLOOD

LIVER

Whole animal organ distribution of 51Cr-labeled endotoxin by 1 hour

Figure 1.

Mean of at least six experimental determinations.

after injection.

57

 

 

 

Table 1. Tissue distribution of subcutaneously
injected 51Cr-labeled endotoxin in normal mice.
a
Percent Uptake
Organ
1 hour 24 hour
Liver 0.5: 0.4 2.5: 2.2
Spleen 0.1: 0.1 0.3: 0.1
Kidneys 0.31 0.1 1.11 0.4
Lungs 0.1: 0.1 0.2: 0.1
Skin 42.8:28.7 38.3114.6
Carcass 46.4:28.4 48.2:12.0
Blood 1.5: 0.6 3.3: 0.6

 

aMean value i standard deviation from six
experimental determinations.

58

so injected free NaZSICrOI+ was compared with 51Cr-labeled endotoxin.
Figure 2 shows that after 1 hour 99% of the injected 51Cr-endotoxin
remained associated with host tissues compared with 75% of the free
chromium-51. By 24 hours, approximately 97% of the sc injected
51Cr-endotoxin remained tissue-associated compared with only 37% of the
free isotope. The remainder of injected Na251CrOu not recovered from
tissues had been excreted in feces and urine.

These ‘results indicate that following sc injection, free
chromium-51 is absorbed into the bloodstream and excreted more rapidly
than 51Cr associated with endotoxin.

The presence of significantly lower amounts of endotoxin in livers
of ac injected mice was further confirmed by the Limulus amebocyte
lysate assay. The Limulus test was performed on IO-fold dilutions of
whole liver homogenates prepared 1 hr after sc or iv injection with
100ug of endotoxin. Liver homogenates prepared from uninjected mouse
livers served as negative controls. The final dilution of each liver
homogenate resulting in a positive Limulus reaction consistently
indicated that there. is at least a 100-fold lower concentration of
endotoxin in livers of sc injected mice (Table 2).

The ability of subcutaneously injected endotoxin to elicit selected

 

biological responses in mice. It has been indicated above that the LDSO

 

of endotoxin by sc injection is about twice that of ip or iv injection.
It was of some interest to determine whether selected biological
responses to ac injected endotoxin were similar to those previously
reported (2,4,16) for iv or ip injected animals. Firstly, liver
glycogen levels were determined in mice following sc injection with the

LD50 of endotoxin. As shown. in Figure 3, liver glycogen steadily

59

 

 

 

7% Ndz'CrO,

fl

/ x

24 HOUR

 

/
4!

 

 

V / am

 

If

 

. 5 O
7 5

>mw>oomm o\o

100
25

Figure 2. Total radioactivity recovered from tissues of

Cr-endotoxin or

51

mice following subcutaneous injection of

NaZSICrOA. Mean of at least five experimental determinations.

60

Table 2. Detection of endotoxin by Limulus assay in livers

of normal mice or mice injected sc or iv with 100ug of endotoxin.8

 

Limulus Reactionb

 

 

Dilution
Normal iv endotoxin sc endotoxin
—2
10 (-) (+2) (+2)
10’3 H (+2) (+1)
10’4 NDC (+2) (-)
10'S NDC (+1) NDC

 

3Represents the results from 2 separate experiments performed
on duplicate samples of each dilution of liver homogenate.

b(+2), firm gel; (+1), incomplete gel with increased opacity
and viscosity; (-), negative reaction.

C 0
Not determined.

61

.moHE xfim ummma
um 50pm .z.m.m fl moam> some mcu mucmmmuamu ucwoa summ
.ADV mafiamm no A.V :HxOBCmcm 5:3 :oEom-cw msomcmuzuonm

mewonHOL wows mmummm cw mam>oa :mwcoxaw uw>wg .m muowwm

62

20:0wfiZTPmOn. wmnox

m.

L

N.

 

 

1L

- v.0

. m6

. N.-

.m._

- QN

- QM

. Od

. o.m

. Qm

. ON

 

5w) Naooomo Ham-i

( °/o

63

declined in both groups during the first 10 hours of fasting. By 8-10
hrs, liver glycogen in control mice had stabilized at about 1.0-1.5mg%,
while in the endotoxin-poisoned mice it continued to decline to an
average of 0.3mg% by 12 hrs. Glycogen levels remained significantly
lower than in control mice for the remainder of the experimental period
(P<.05).

The effect of subcutaneously administered endotoxin on the
induction of hepatic PEPCK by hydrocortisone is shown in Table 3. Mice
were injected with the LD50 of endotoxin and 5mg of hydrocortisone
individually and in combination. Four hours later mice were killed by
cervical. dislocation. and liver' PEPCK. activity' ‘was determined.
Hydrocortisone administration significantly increased hepatic PEPCK
activity' 4 ‘hrs after injection (P<.001). By contrast, simultaneous
injection of endotoxin prevented the induction of liver PEPCK by
hydrocortisone (P<.01).

Mean body temperature changes after sc injection of either sterile
saline or 2.85, 11.4, 28.5, or 285ug of endotoxin in mice housed at 5°C
are shown in Figure 4. Body temperatures were determined rectally at
one hour intervals for 6 hours. For the mice receiving saline,
temperature decreased approximately 4°C after 4 hrs. Among mice
receiving endotoxin, body temperatures decreased more rapidly and in
proportion to the increasing doses of endotoxin. The mean decrease in
body temperature of mice receiving 285ug of endotoxin was significantly
greater than control mice at 2, 3 and 4 hrs post-injection (P<.01).
Mice injected with 11.4 and 28.5 micrograms of endotoxin were also
significantly more hypothermic than controls at 3 hrs, and 3 and 4 hrs,

respectively. By six. hours after administration of 285, 28.5, and

64

Table 3. Effect of subcutaneously administered endotoxin

on induction of liver PEPCK by hydrocortisone.

 

PEPCK Activitya

 

Experimental (nmoles/mg protein/min)
0.1 m1 saline (sc.) 19.12 1 6.5
5 mg Hydrocortisone 36.56 : 9.29b
5 mg Hydrocortisone +
285ug endotoxin (sc.) 25.13 1 5.94c
285ug endotoxin (sc.) 21.02 1 9.64

 

aMean : standard deviation of at least six experimental
determinations.
b

P<.001 versus saline control.

c
P<.01 versus hydrocortisone control.

65

.95 msmwm .0 an: mama“... .U 5.3

”$3.: .. ”was wsmw.m .. “mcwamm . 0 "3035.3 .moHE me ummma
um mo mpoumpmaswu moon cw mmmmuomp some ecu mucmmmuemu ucfioo
comm .cfixouopcm «0 ounce moofium> cufla coauommofi msomcmusonom

wCfiBOHHOM moHE mo whammHmQEou zoos cw mwmmpomm :mmz .q muswflm

66

 

 

° “3 ‘I “3 °? 9 a

(Go) BMOIVMEdWBJ. NI BONVHO

' Fl

 

HOURS

67

11.4ug of endotoxin, fatalities were 66%, 16% and 16% respectively. No
deaths were recorded among mice receiving saline or 2.85ug of endotoxin

over the experimental period employed.

DISCUSSION

Knowledge of the fate of lipopolysaccharide in the endotoxemic host
is essential to the understanding of the mechanism(s) by which these
complex macromolecules induce their biological effects. The organ
distribution data presented in Figurel confirms previous reports
showing the accumulation of major quantities of intravenously or
intraperitoneally injected endotoxin in RES organs, especially the liver
and spleen (12,15,27). Further, the localization of substantial amounts
of this endotoxin in Kupffer cells is well established (23). Zlydaszyk
and Moon presented data supporting the concept that over 75% of
51Cr—labeled endotoxin concentrated within the liver after iv injection
is associated with parenchymal cells (27). Interpretation of results
concerning the presence of endotoxin in gall bladder contents of rabbits
led Mathison and Ulevitch to suggest that endotoxin may be processed by
hepatocytes and passed into the bile canalicular system (15). Whether
hepatic localization of endotoxin is associated with direct
hepatotoxicity has not been completely defined. A major purpose of this
particular study was to address this question directly.

The organ distribution of endotoxin after sc administration is
markedly different than after iv or ip injection (Figure 1).
Significantly lower amounts of sc administered endotoxin reach various
organs, particularly the liver. For example, 1 hr after sc injection
less than 15% of the toxin was absorbed into the vascular compartment
and less then 1.0% became associated with the liver. Over 85% of the
radioactivity remained at the injection site for at least 24 hrs

(Table 1). Varying the dose of 51Cr-endotoxin did not substantially

68

69

alter this distribution (data not shown). It is equally apparent from
the results of our LD50 studies that the quantitative doubling in the
LD50 between sc and either ip or iv routes in no way approximates the
dramatic differences seen in the tissue localization data which reflect
upwards of 100-fold changes in concentration. Collectively, these
results suggest that the lethal effects of endotoxin are independent of
the amounts of toxin localized within the deep tissues of the host.
Some years ago, Noyes et al (21) also observed marked differences in the
patterns of distribution of 51Cr-endotoxin following intramuscular vs
intravenous injection. Although greater than 35% of the injected toxin
could not be accounted for at the time of sacrifice, they reported that
from 1/3 to 1/2 of intramuscularly injected endotoxin remained at the
site of injection.

A consistent concern of this type of study is that the isotope
might be liberated from the conjugated complex during host
detoxification. To address this concern experimentally, we compared the
distribution and excretion of NaZSICrou with the radiolabeled endotoxin.
Collectively, our results suggest that if the 51Cr-label had dissociated
from the toxin at the site of injection, the radioactivity would have
been absorbed into the blood stream and excreted much faster than was
observed after the sc injection of 51Cr-endotoxin (Figure 2). The
stability of the “Cr-endotoxin conjugate following sc injection was
further confirmed by detection of at least 100-fold lower levels of
endotoxin in livers of sc injected mice when compared to iv injected
mice (Table 2). Taken together, these data indicate that the
radioactive content of individual organs of mice injected with

51Cr-endotoxin represents endotoxin associated with that tissue.

 

70

The marked contrast in the _i_n__ y__i_y_<_3_ distribution of endotoxin after
iv and sc injection, led us to examine selected metabolic and
physiologic responses in mice receiving endotoxin subcutaneously to
determine whether the biologic reaponses were consistent with those
observed from the literature. Various reports have appeared concerning
the disruption of carbohydrate metabolism in the endotoxemic host
(4,16). Consistent with the rapid and prolonged depression of hepatic
glycogen levels in mice after iv or ip injection of endotoxin, sc
injection of endotoxin also impairs the ability of fasted mice to
maintain the hepatic glycogen reserves compared with control mice
(Figure 3). The impaired ability to carry out gluconeogenesis, perhaps
related to decreased PEPCK activity, has been suggested as an integral
component in liver carbohydrate depletion in endotoxemic animals (16).
Subcutaneously injected endotoxin also significantly inhibited the
induction of liver PEPCK by hydrocortisone (Table 3). When correlated
with data in Figure 1, these data indicate that hepatic metabolic
responses to endotoxin are not dependent upon the amount of endotoxin
associated with the liver or other tissues. Besides the hepatic
metabolic dysfunctions described, subcutaneous injection of increasing
doses of endotoxin caused dose dependent hypothermic responses and
enhanced lethality in mice subjected to an additional stress of 5°C
ambient temperature (Figure 4), and a transient leukopenia in mice kept
at room temperature (data not shown).

Recent efforts have been made by various workers to distinguish
between direct effects of endotoxin and indirect effects mediated by
soluble substances released from affected cells, especially macrophages

and lymphocytes (6). Because our results show that at most 10% of a

71

lethal dose of endotoxin administered subcutaneously ever reaches deep
tissues and yet the host exhibits typical biological responses to
endotoxin, our data strongly favor the thesis that soluble mediators
play a significant role in the pathophysiological responses to
endotoxin. Better understanding of the role of mediators in endotoxin
poisoning is essential to a clearer understanding of the basis of the

pathophysiologic changes observed during endotoxin poisoning and gram

negative septic shock .

 

 

 

10.

ll.

12.

LITERATURE CITED

ApUL R.N., C.F. Hertogs, and D.H. Pluznik. 1979. Generation of
colony-stimulating factor by purified macrophages and lymphocytes.

J. Reticuloendothel. Soc. 26:491-500.

Berry, L.J. 1971. Metabolic effects of endotoxin, p. 165-198. 1p;
S. Kadis, G. Weinbaurer, and S.J. Ajl, (ed.), Bacterial Toxins,

Vol. 5. Academic Press, New York.

Some metabolic aspects of

Berry, L.J., and D.S. Smythe. 1960.
Ann.

host-parasite interactions in the mouse typhoid model.
N.Y. Acad. Sci. 88:1278-1286.

Berry, L.J., D.S. Smythe, and L.G. Young. 1959. Effects of bacterial
endotoxin on metabolism. I. Carbohydrate depletion and the
protective role of cortisone. J. Exp. Med. 110:389-405.

Berry, L.J., D.S. Smythe, and L.S. Colwell. 1968. Inhibition of
hepatic enzyme induction as a sensitive assay for endotoxin.

J. Bacteriol. 96:1191-1199.

Cellular and molecular mechanisms of action

Bradley, S.G. 1979.
Ann. Rev. Microbiol. 33:67-94.

of bacterial toxins.

Chedid, g1, R.C. Skarnes, and M. Parant. 1963. Characterization
f

Cr-labeled endotoxin and its identification in plasma and

o
J. Exp. Med. 117:561-571.

urine after parenteral administration.

Cynkin, M.A., and G. Ashwell. 1960. Estimation of 3—deoxy-sugars
by means of the malonaldehyde-thiobarbituric acid reaction.

Nature (London) 186:155-156.

Eger, W., H. Jungmichel, and G. Kordon. 1958. Effect of the
lipopolysaccharide pyrexal on allyl alcohol damage to the liver
as an expression of a change in resistance of the organism.
Virchows Arch. Pathol. Anat. Physiol. Klin. Med. 331:154-164.

Kenn), A., and A.J.M. Kits van Heijningen. 1954. A colorimetric

micro-method for the determination of glycogen in tissues.
Biochem. J. 56:646-648.

1965. Phosphoenolpyruvate

Lane, M.D., H.C. Chang, and R.S. Miller.
Methods in Enzymology

carboxykinase from pig liver mitochondria.
23:270-277.

1967. Hepatic changes produced

Levy, E., F.C. Path, and B.H. Ruebner.
Am. J. Pathol.

by a single dose of endotoxin in the mouse.
51:269-284.

72

73

13. Lowry, O.H., N.J. Rosenbrough, A.L. Farr, and R.J. Randall. 1951.
Protein measurement with the Folin phenol reagent.
J. Biol. Chem. 193:265-275.

14. Manuel, D.N., M.S. Meltzer, and S.E. Mergenhagen. 1980. Generation
and characterization of a lipopolysaccharide—induced and
serum-derived cytotoxic factor for tumor cells.

Infect. Immun. 28:204-211.

15. Mathison, J.C., and R.J. Ulevitch. 1979. The clearance, tissue
distribution, and cellular localization of intravenously
injected lipopolysaccharide in rabbits. J. Immunol. 123:2133—2143.

16. McCallum, R.E., and L.J. Berry. 1973. Effects of endotoxin on
gluconeogenesis, glycogen synthesis, and liver glycogen synthase
in mice. Infect. Immun. 7:642-654.

17. Mela, L., L.V. Bacalzo, Jr., and L.D. Miller. 1969. Defective
oxidative metabolism of rat liver mitochondria in hemorrhagic
and endotoxin shock. Am. J. Physiol. 220:571-577.

18. Moore, R.N., K.J. Goodrum, and L.J. Berry. 1976. Mediation of
endotoxic effects by macrophages. J. Reticuloendothel. Soc.
19:187-197.

19. Moss, G.S., P.O. Erve, and W. Schumer. 1969. Effect of endotoxin
on mitochondrial respiration. Surg. Forum. 20:24—25.

20. Munson, A.E., D.C. Drummond, A.C. Adams, and S.G. Bradley. 1976.
Enhanced toxicity for mice of combinations of bacterial
lipopolysaccharide and vincristine. Antimicrob. Agents
Chemother. 9:840-847.

21. Noyes, H.E., C.R. McInturf, and C.J. Blahuta. 1959. Studies on
distribution of Eschericia coli endotoxin in mice. Proc. Soc.
Exp. Biol. Med. 100:65-68.

 

22. Reed, L.J., and H.A. Muench. 1938. A simple method for estimating
fifty-percent endpoints. Am. J. Hyg. 27:493-499.

23. Ruiter, D.J., J. Van der Muelen, A. Brouwer, M.J.R. Hummel, B.J. Mauw,
J.C.M. Van der Ploeg, and E. Wisse. 1981. Uptake by liver
cells of endotoxin following its intravenous injection. Lab
Invest. 45:38-45.

24. Siegel, S. 1956. The Mann-Whitney U test, p. 116, Ip_ Nonparametric
Statistics for the Behavioral Sciences. McGraw-Hill Book Co.

25.

26.

27.

74

Snyder, S.L., and R.I. Walker. 1975. Inhibition of lethality in
endotoxin-challenged mice treated with zinc chloride. Infect.
Immun. 13:998-1000.

Walker, R.I. 1979. p. 993-946, Ip_P. Rosenberg (ed.), Toxins:
Animal, Plant, and Microbial. New York: Pergamon.

Zlydaszyk, J.C., and R.J. Moon. 1976. Fate of 51Cr-labeled
lipopolysaccharide in tissue culture cells and livers of
normal mice. Infect. Immun. 14:100—105.

ARTICLE II

Potentiation Of Endotoxin Sensitivity By
Galactosamine Is Not Associated With Increased

Uptake of Endotoxin By Hepatocytes

75

 

 

Abstract

'The distribution. of intravenously administered 51Cr-labeled
endotoxin among various liver cell types was compared in normal mice and
ndtma made hyperreactive to endotoxin by treatment with galactosamine.
Galactosamine treatment of endotoxin challenged mice decreased survival
and caused marked liver damage by 8 hrs as reflected by an increasing
serum. ornithine carbamyltranferase (OCT) activity. In normal mice,
there was approximately a 5-fold lower uptake of 51Cr-endotoxin by liver
parenchymal cells than nonparenchymal cells. Further, fractionation of
nonparenchymal cell preparations by centrifugal elutriation showed that
51Cr-endotoxin was primarily associated with Kupffer cells. Treatment
of mice with galactosamine did not alter the organ distribution of
endotoxin or enhance liver parenchymal cell uptake of endotoxin in vivo.
Furthermore, treatment of 'normal liver' parenchymal cells ‘with
galactosamine in vitro did not increase cell uptake of 51Cr-endotoxin.
These results suggest that the augmented lethal and hepatotoxic effects
of combined exposure to galactosamine and endotoxin are independent of

the quantitative localization of endotoxin in liver.

76

 

 

 

Introduction

Gram-negative bacterial endotoxins are capable of inducing a number
of histologic and metabolic alterations in livers of experimental
animals. Such changes include depletion of liver glycogen (22,24),
inhibition of enzyme induction and gluconeogenesis (2,22,31) elevations
in serum transaminase levels (6,19,39), labilization of lysosomal
membranes (16,37), and impairment of mitochondrial energy production
(13,14,23).

Although numerous reports on the distribution of endotoxin in vivo
have emphasized the importance of the reticuloendothelial system (RES)
in the clearance of endotoxin from the blood, certain studies indicate
that liver parenchymal cells may also be involved (32,38,40).
Furthermore, binding sites for endotoxin. ‘were demonstrated by
immunofluorescence on rat hepatocyte plasma. membranes (29,30). The
precise importance of liver parenchymal cells in the uptake and
processing of endotoxin remains uncertain.

Galactosamine (galN) is a.liver-Specific toxin that causes liver
cell necrosis (17,20) and increases the sensitivity of a variety of
experimental animals to the lethal effects of endotoxin (12). The
sequence of biochemical changes elicited by galN in hepatocytes results
in impaired biosynthesis of macromolecules in liver cells leading to
alterations in the structure and function of parenchymal cell plasma
membranes (1,9,11).

The increased sensitivity to endotoxin coupled with the hepatocyte
plasma membrane damage induced by galN has led to a suggestion that

liver injury elicited by galN ‘may facilitate endotoxin. entry into

77

78

hepatocytes (27). The initial objective of this study was to
investigate the distribution of endotoxin among various liver cell types
following its intravenous injection in mice. As a second objective we
sought to determine whether potentiation of endotoxin lethality by galN
was accompanied by a shift in the distribution of 51Cr-endotoxin among
the various organs as well as among various liver cell types. We
reasoned that if endotoxin exerts its effects directly on hepatocytes
then one. would predict that sensitization to the toxic effects of
endotoxin by galN should be accompanied by enhanced uptake of endotoxin
by liver parenchymal cells. To accomplish these objectives we separated
liver cells by differential centrifugation and centrifugal elutriation
and experimentally determined the in ‘vivo localization. of endotoxin
among various liver cell types. The same techniques were also used to
study the effect of galN on uptake of endotoxin by primary cultures of

mouse liver parenchymal cells in vitro.

Materials and Methods

Animals. Female CD-1 mice (Charles River Labs, Wilmington, Del.)
weighing 20-26 gm were used in all experiments. Mice were housed 5 per
cage under standard laboratory conditions and received water and food ad
libitum.

Materials. Lyophilized Salmonella typhimurium lipopolysaccharide (LPS)

 

extracted by the Westphal phenol-water procedure was purchased from
Difco Laboratories, Detroit, MI., and served as endotoxin in all
experiments. Radiolabeled endotoxin was prepared by incubating LPS with
NaZSICrO4 (New England Nuclear Corp., Boston, Mass.) according to the
method of Chedid et al. (5), as outlined by Zlydaszyk and Moon (40).
The final preparation contained toxic, high-molecular weight endotoxin
with a specific activity of about lluCi/mg. Appropriate dilutions of
radiolabeled or unlabeled endotoxin were prepared in sterile, isotonic
saline and stored at -20°C. Concentrations of both labeled and
unlabeled endotoxin were determined by assay of 2-keto-3-deoxyoctuloso-
nate (KDO), following a modification of the method of Cynkin and Ashwell
(8). Aliquots of endotoxin were hydrolyzed in 0.2M H2804 for 45 min in
a boiling water bath and then subjected to the thiobarbituric acid assay
(8). This procedure ‘was quantitatively accurate over“ a. range from
4-300ug/m1 of endotoxin.

Neutralized solutions of D-(+)-galactosamine HCl (Sigma Chemical
Co., St. Louis, M0.) were prepared daily in pyrogen-free phosphate-

buffered saline (PBS) and always administered intraperitoneally (0.5mg/g

body weight) in 0.2ml aliquots.

79

80

Serum OCT determination.

 

Serum ornithine carbamyltransferase (OCT) activity was determined
spectrophotometrically by the manual method of Cerriotti (4), which
measures citrulline generated from the condensation of ornithine and

carbamylphosphate.

Tissue distribution of 51Cr-labeled endotoxin

 

Groups of at least 5 mice were injected intravenously (iv) with
approximately 0.1uCi of 51Cr—labeled endotoxin (25ug) with or without
galN. Three hours later individual mice were heparinized, killed by iv
injection of 1.625mg of sodium. pentobarbital, and transferred to a
platform which drained into a collecting vessel. Incisions were made to
expose the thorax and abdomen. The inferior vena cava was severed and
residual blood was flushed from the various organs by injecting 15ml of
saline into the apex of the heart. This procedure provided sufficient
flushing to visibly clear blood from all organs and allowed the
controlled collection of blood to determine blood-associated
radioactivity. Skin, blood, carcass, and selected organs were
individually placed in vials and counted in a Packard gamma counter.
The amount of radioactivity per sample was calculated as a percentage of
the total recovered as follows:

[radioactivity of sample/total recovered radioactivity] x 100 = % Uptake

Liver cell isolation.

 

Liver cells were isolated from normal mice or mice injected 90 min
earlier with 0.3uCi of 51Cr-endotoxin (25ug) according to a modification

of the method of Seglen (34). lp_situ liver perfusion procedures have

81

been described in detail previously (25,33). Briefly, mice were
initially heparinized and then anesthetized with pentobarbital within
10 min. Parallel lateral incisions were made and the viscera displaced
to one side to expose the portal vein. The portal vein and the thoracic
portion of the inferior vena cava were immediately cannulated. The
liver was then washed with about 50ml of Ca++-free wash buffer (34) at a
rate of about 3ml/min with the aid of a polystaltic pump (Buckler
Instruments, Inc., Fort Lee, N.J.) to remove gross blood. Perfusion
medium was prewarmed to 37°C by passing through a water bath-jacketed
coil of tubing just before entering the portal vein. Following the
preperfusion, the liver was perfused with 0.075% collagenase and 0.1%
hyaluronidase in Hank's Balanced salt solution (HBSS;GIBCO) without
NaHCO3 containing 75mM HEPES, 0.05% bovine serum albumin, 5.5mM glucose,
penicillin (200U/ml) and streptomycin (0.148mg/ml). The liver was then
removed and placed in ice-cold Leibovitz's L-15 medium (GIBCO
Laboratories, Grand Island, N.Y.) supplemented with 20mM HEPES, 10%
newborn calf serum, 0.4% bovine serum albumin, penicillin (200 U/ml),
and streptomycin (0.148mg/ml). The liver was immediately teased apart
and the resulting cell suspension filtered through 250nm nylon mesh.
Parenchymal cells were separated from erythrocytes, sinusoidal liver
cells, and cell debris by differential centrifugation (50 x G, 1min)
and washed twice with L-15 medium. Preformed gradients of Percoll
(Sigma Chemical Co., St. Louis, M0.) were used to enrich for viable
parenchymal cells (28). Density gradients of Percoll were generated by
diluting stock solutions of Percoll with L-15 medium to a final density
of 1.07 g/ml followed by centrifugation at 20,000 x G for 45 min in a

Beckman fixed-angle rotor. Preformed gradients of Percoll were

82

calibrated with density marker beads (Pharmacia Fine Chemicals,
Uppsala). Aliquots of washed parenchymal cells (3 ml) were layered on
30 ml gradients and centrifuged in a swinging bucket rotor at 650 x G
for 30 min. The enriched parenchymal cell band, found at a density of
approximately 1.10 g/ml, was then removed, washed 3 times and
resuspended in L-15 medium.

Liver nonparenchymal cells contained in the initial cell suspension
were sedimented and washed with L-15 medium at 300 x G for 10 min.
Enriched fractions of both Kupffer and endothelial cells were obtained
by centrifugal elutriation in a Beckman JE-6 elutriator rotor according
to the method of Ruiter et al. (32). Three fractions (100ml each) were
collected at a constant rotor speed of 2500 rpms: (i) contaminating
lymphocytes and cell debris (14 ml/min), (ii) enriched for endothelial
cells (22ml/min), and (iii) enriched for Kupffer cells (40ml/min).
Cells contained in each fraction were concentrated by centrifugation at
1500 rpms for 20 min and resuspended in L-15 medium.

To determine the amount of endotoxin associated with the various
liver cell fractions, samples of the final suspensions of partially
purified cells were removed and (i) counted in a gamma counter and
(ii) examined with a hemocytometer to determine cell numbers and

capacity to exclude trypan blue dye.

Association of 51Cr-endotoxin with hepatocytes in vitro.

Freshly isolated liver parenchymal cells (see above) were resuspended in
1.9 ml aliquot of L-15 medium (2.75x105/ml) in 6-well Costar plates and
incubated at 37°C in a humidified atmosphere containing 5% C02 to allow

attachment. Percoll-enriched parenchymal cells obtained from. normal

83

mice were greater than 94% viable, as estimated by trypan blue dye
exclusion, and at least 98% pure by hemocytometer count. After 30 min,
0.12uCi of 51Cr-endotoxin (50 ug) was added to each well and the total
volume brought to 2.0 ml with PBS or PBS containing galN. In control
wells, 51Cr-endotoxin was suspended in 2.0 ml of L-15 medium in the
absence of cells. For comparison, parallel wells containing
nonparenchymal cells (2.0x106/ml) were treated the same as control and
experimental. wells. At designated times, culture supernatants from
duplicate wells were decanted and attached cells were gently washed
twice with L-15 medium. Culture supernatants and wash medium were
combined and centrifuged at 2000 rpm for 2 min to pellet unattached
cells. Supernatants were counted in a Packard gamma counter to
determine unassociated 51Cr-endotoxin. Cells were detached from wells
and solubilized with 2 ml of 0.5% deoxycholate and 0.5% Triton X-100 in
0.1M potassium phosphate buffer, pH=7.5, 1mM EDTA, 1mM dithiothreitol,
and 5mg/ml bovine serum albumin (7). Solubilized cells were decanted
and added to unattached cell pellets (above). Wells were washed with
1ml of water which was added to respective solubilized cell tubes.
Solubilized cell fractions were then counted to determine cell-
associated 51Cr-endotoxin. Counts remaining Zhl control. wells after
removal of culture ‘medium. were subtracted from. experimental counts
before calculating the amounts of endotoxin associated with respective

cell cultures.

Statistical evaluation.

 

Where appropriate, results were analyzed for statistical significance

(p.<05) by the Mann-Whitney rank order test (35).

Results

Potentiation of endotoxin-induced lethality by galactosamine.

 

In preliminary experiments it was determined that the 50% lethal

dose of endotoxin (LD ) was ca. 125ug for CD-1 mice, with the majority

50
of deaths occurring between 18 and 24 hours. By contrast, concurrent
injection of mice with galN (0.5mg/g b.w.) and as little as 1.0ug of
endotoxin resulted in 67% lethality by 24 hrs, with the majority of
deaths occurring by 9 hrs (Fig. 1). A dose of 25ug of endotoxin was
lethal for 50% and 67% of galN-treated mice by 6 hrs and 24 hrs,
respectively. No deaths were recorded among mice receiving either galN
or 25ug of endotoxin. The onset of symptoms classically associated with
endotoxin poisoning (i.e., lethargia, diarrhea, conjunctival exudate)
was also markedly accelerated in galN-treated mice given endotoxin.
Release of OCT in galN-treated mice glyen endotoxin.

Serum OCT activity was determined in mice at various intervals
after treatment with galN and 1.0ug of endotoxin, either separately or
together. As shown in Figure 2, mice receiving either galN or endotoxin
alone had negligible increases in serum. OCT by' 8 ‘hrs. Concurrent
injection of the same doses of both endotoxin and galN resulted in a
slight increase in serum OCT activity by 6 hrs and almost a 100-fold
increase by 8 hrs (p$.01) when compared to mice given endotoxin or galN
alone.

Effect of galactosamine on uptake of iv injected endotoxin by the liver.

Reasoning that if galN sensitizes mice to endotoxin by enhancing
uptake by liver cells then one would predict that the differences in

uptake should be evident with a dose of 25ug of endotoxin which caused

84

85

"maonexm .moHE vmummuulmcfiemmOOUmamw a“ :fix0oomcm mo muommmm Hmnumq

mmDOI

 

 

vw N F

_ y“? _
T\\
191146. T

.CHxOqucm wamm .. newaqucm we wamm + szw . O “ejacuoccm mo mu: + szw .0 “21mm .‘

.H muswfim

tom

tom

 

 

Ioow

‘lVAIAHOS J.NEIOHSd

86

.Awno.~v cwauomcw mam mCHEmmOOUmHmw LuH3

:ofiuomHCfi wCHBOHH0w muse :fl zufi>fluom Boo Eduwm CH mmmouocH .N wuswflm

 

 

 

 

 

 

mEDOI
m w v m o
b _ — _
d I o
T 00?
r
e Exoaoocm +z_wm I room
Exoaoocm 4
2.8 . .
I com

S
has
won
ww
mm
W...
WW
mu

m

Ii

IA

87

rapid death in galN-treated mice but was well-tolerated by normal mice
(Fig. 1). Consistent with previous reports (3,22,40), the majority of
intravenously injected endotoxin was sequestered by livers of normal
mice (Table 1). This preportion was not significantly altered by
treatment with galN.

To examine whether galN might change the distribution of endotoxin
among liver cells, we investigated the distribution of 51Cr-endotoxin
among liver parenchymal and nonparenchymal cells in normal and galN-
treated ‘mice. Separation of liver cells ‘was started. 90 ‘min after
endotoxin injection. Final preparations of Percoll-enriched parenchymal
cells were an: least 96% pure and contained from 8-14x106 total cells
(Table 2). More than 85% of the isolated parenchymal cells were viable
as judged by exclusion of trypan blue dye. Nonparenchymal cell
preparations were obtained by differential centrifugation and contained
considerable numbers of parenchymal cells. For this reason the measured
uptake of 51Cr-endotoxin by nonparenchymal cells in individual
experiments was corrected for counts associated with contaminating
parenchymal cells before calculating the amount of endotoxin associated
with nonparenchymal cells (ng/106 cells). Table 2 shows that
approximately 5-fold higher amounts of 5'lCr-endotoxi‘n were associated
with nonparenchymal cells compared to parenchymal cells. Administration
of galN did not alter the amount of parenchymal cell-associated
51Cr-endotoxin.

The amounts of 51Cr-endotoxin associated with 3 different cell
types in the nonparenchymal cell population. was estimated in cells

isolated by utilizing centrifugal elutriation. Nonparenchymal cells

obtained from 3 normal mice 90 min after iv injection of 51Cr-endotoxin

88

.meu new upmm: .mmcwummucw mmoaaucHu

.mcowuwcfisumumv Houcmsaumaxo wumummom xwm Scum .x.m.m H w=Hm> cmoz

 

 

 

 

 

n
.cwROuovcoiuudm mo cowuuomcw momma mu: m omwosum cowuaowuumwnm
m.wm o.wa wwwm can“ deuce
N.OHE.~ n.oud.o cmm Esq uumnuo
m.auu.q m.ouo.m amm mmm oooam
m.0um.m H.0um.~ «mm sum cwxm
m.oum.m m.cuo.c ooc mom mmwuumu
ions .o flowed E 8 manual
~.owc.o ~.oun.o am cm was;
Ross «535 a? m3 8.3....
L.~wc.mo ~.oum.o~ mien Omen um>HA
szw HmEuoc szw finance
cmwuo
Acmmuo\Emuv

mxmuas ucmuuom

n

mmxmuas consume:

 

:waquchpu

.mqu vmummuulmcwammOuumamw was MmEno: cw

venom-cw zamsocm>muuCM we :owuaofiuumwc compo

.# maomH

89

.moHE ameoc 50pm ucmumMMHc %HHmoHumHumum uOZp

.wHHoo Hmewzocmumo wcfiumsHEmusoo
nuwz moumfioommm Emu pow wouomwwoo exams: Hams HashnocmumacozU

.A.3.n w\wEm.oV m:HEmm0uomHmw usonuwz no cows

cwxouoocmlwufim sows muse mo coauomnca pmumm GHE om omuumum cowumummwm Haoun

.coaumcflEumump Hmucwswumaxw mumumamm xwm ummma um Eouw .Z.m.m H msam> cmmZm

 

 

om.moflw.mmm Aowlmanw m.~H~.m Nwflqmm maamo Hmexsocmwmoco:
a~.o «m.ms Aeolioavwa s.ssm.w aaasm maamu ameazocmuma
omumwuulw:HEmw0uumeo
m.oanq.amm Assumsvaw m.~hm.mz mmsmNN maamu aaeaauamumaao:
o.-Hq.mq Aoofilomvwo q.o%m.mfi mHme mHHmo Anahnocmwmo
HmEuoz
ox oa\mcv A ones A oa\5auv
xmua: Ammcmuv Hoar wxmuo: some Hamo
pmuumuwou >ufiu=m N Hamo newsman: o

 

II

m.oowe vmummpuimCHEmmOBUmHmw pom Hmeo: mo mHHmo Hoaxnocmumasoc pom

Hmfizzucwuma um>fla ca cHx0uopcmleHm mo cowumuflamuoa o>fi> CH .N maomfi

90

were pooled and then separated into three fractions by centrifugal
elutriation according to the method of Ruiter et al. (32). While these
fractions were enriched for specific cell types (i.e., Kupffer cells,
endothelial cells, lymphocytes), the purity is not absolute (see
Appendix 1). Our results on cellular distribution (Table 3) indicate
that the uptake of endotoxin is much higher in the Kupffer cell enriched
fraction (496ng/106 cells) when compared to the endothelial (247ng/106
cells) or lymphocyte (94ng/106 cells) fractions.
Association of endotoxin with normal and_galN-treated hepatocytes in
The association of 51Cr-endotoxin with primary cultures of mouse
liver parenchymal cells was determined in the presence or absence of
galN at 4 and 8 hrs after treatment. Background counts measured in
control wells (without cells) were subtracted from experimental wells
before calculating the amount of 51Cr-endotoxin associated with cell
monolayers. Table 4 shows that uptake of 51Cr-endotoxin by parenchymal
cells was not increased significantly when incubated in the presence of
9mM galN by 4 or 8 hrs. Likewise, parenchymal cells incubated with 23mM
galN did not show an increased uptake of 51Cr-endotoxin by 8 hrs. In
addition, mouse liver nonparenchymal cells showed an approximately
2.5-fold higher uptake of radioactivity by 4hr when compared to

parenchymal cell uptake.

91

.mucwEHuwaxm ucmmchmpGH ozu

Boom mwcmu can some men ucmmmuomu muasmmm .GHxOuoosolwofim
spas wows mo coauuwhcfi wmuwm :HE om omuumum mm3 coauMHOmfi aamum

 

Amsm-0msvoms Aw.s-m.avm.m Amawaussnvwwma Aumomaaav
ma» COH uumhh

Aoamuoazvasm Am.m-a.zvs.m Aommuswsvanm Asmaamnuoeamv
Ne coauomum

Asaaumnvsa xa.m-a.svo.m Aows-smmeos AmmasooaQERHV

He sowuumpm

 

Amaamo oH\w:v A ofixv Ego coauomun
mambo: whommwz pamww Hamo Hmuoe coauMHuusam

 

m.c0Humfiuu:Hm Howswfiwucmo kn momma Hams uwmwomam wow omzofiucm

maamo HmE%£ucmumaco: aw :Hx0uoocwlpofim wo coaumswamuoa o>H> CH .m wanme

92

Table 4. Effect of galactosamine on the uptake of 51Cr-endotoxin

by normal mouse liver parenchymal cells in vitro.

 

Corrected Uptake (ng/well)a

 

 

Cell Type GalN
4hr 8hr
parenchymal cells none 199141 182:48
parenchymal cells 9mM 142:40b 183140b
parenchymal cells 23mM NDd 224:45b
nonparenchymal cells none 514165c NDd

 

aMean value : S.E.M. from seven separate experimental determinations.
bNot statistically different from untreated parenchymal cell cultures.
CP<.001 compared to untreated parenchymal cell cultures.

Not done.

Discussion

Administration of galactosamine (galN) increases the susceptibility
of mice at least 100-fold to the lethal effects of endotoxin (Fig. 1).
In addition serum levels of a liver-specific enzyme, ornithine
carbamyltransferase (OCT), are elevated in galN-treated mice given 1 ug
of endotoxin but not in mice given either endotoxin or galN alone (Fig.
2). These findings are consistent with those previously reported by
Galanos et al. (12) on the sensitization of a variety of experimental
animals to the lethal effects of endotoxin by galN administration.

The metabolic events following administration of galN have been
extensively characterized. Galactosamine is metabolized by enzymes of
the galactose pathway whiCh are found predominantly in the liver and
hence the effects of galN are primarily confined to the liver (17).
Within 30 min after galN injection, UDP-galactosamine and
UDP-glucosamine accumulate in the liver. This is accompanied by about a
90% drop in hepatic UTP content (18) with a resultant impairment in the
biosynthesis of RNA, proteins, glycogen, glycoproteins, and other
UTP-dependent reactions. These biochemical alterations affect the
structure and function of hepatocyte plasma membranes (1,9,11). By 24
hrs galN leads to leakage of liver enzymes into serum and to
histological liver damage.

The mechanisms by which galN exerts its potentiating effect on
endotoxin toxicity have not been elucidated.

The present study indicates that galN administration does not
markedly alter the organ distribution of a quantity of endotoxin that is

lethal for galN-treated mice but well-tolerated by normal mice

93

94

(Table 1). Approximately 70% of the injected endotoxin was localized in
livers of both normal and galN-treated mice by 3 hrs. These results
reaffirm observations made in previous studies which indicate that the
fate of endotoxin in vivo to a large extent follows the distribution of
the RES (21,32,40). The pattern of distribution of endotoxin among host
organs of normal mice remains constant for up to 8 hrs after injection
(36,40). Although the liver contains perhaps the largest reservoir of
reticuloendothelial cells (Kupffer cells), certain studies indicate that
endotoxin may also be sequestered by liver parenchymal cells
(29,30,38,40). Zlydaszyk and Moon (40) reported that following
injection of 51Cr-endotoxin in mice, up to 75% of the radioactivity
present in livers was found in parenchymal cells. Furthermore, they
demonstrated uptake of 51Cr-endotoxin In; nonphagocytic tissue culture
cells. Uncertainty still exists as to whether localization of endotoxin
in liver parenchymal cells is associated with direct toxicity.

The results of the present study further demonstrate that galN-
treatment does not increase the uptake of endotoxin by liver parenchymal
cells in vivo (Table 2). Parenchymal cells isolated from normal or
galN-treated mice given endotoxin were contaminated with, on the
average, only 2% nonparenchymal cells. Because the nonparenchymal cell
fractions contained considerably greater contaminating parenchymal
cells, the measured uptake of endotoxin by nonparenchymal cell fractions
was corrected for counts associated with contaminating parenchymal cells
based on the measured uptake of 51Cr-endotoxin by isolated parenchymal
cells. Although detectable in liver parenchymal cells in 5-fold lower
amounts, the majority of liver-associated endotoxin appears to be

sequestered by nonparenchymal cells.

95

It is becoming increasingly apparent that the ability of
centrifugal elutriation to separate nonparenchymal cells from mice is
substantially less than from rats. Hence the degree of purity of these
cell fractions used in this study is not optimal. Despite this
limitation, to our knowledge this is the first quantitative data
describing the distribution of intravenously injected endotoxin among
various liver cells in mice. Our observations are consistent with those
obtained by Ruiter et al. (32) who injected rats with a similar dose of
endotoxin per gram body weight and determined its distribution among
liver cell types by centrifugal elutriation. Their data indicate that
endotoxin is taken up almost exclusively by Kupffer cells (340ng/106
cells) but also by parenchymal (17ng/106 cells) and endothelial
(24ng/106 cells) cells In! 30 min after intravenous injection.
Collectively, these results suggest that liver localization of umjor
portions of intravenously injected endotoxin is largely a consequence of
blood clearance of endotoxin by the RES.

Many of the biochemical alterations observed in liver parenchymal
cells after injection of galN into the whole animal are reproduced by
treatment of hepatocytes with galN in vitro. Galactosamine is
metabolized by hepatocytes in vitro with the usual depletion of uridine
nucleotides and UDP-hexoses (15). Similarly, uridine will reverse the
uridine nucleotide deficiency induced by galN in cultured hepatocytes
(15) as well as in intact hepatocytes. In order to provide a more
controlled environment with which to study the uptake of endotoxin by
cells, we investigated the effect of galN on the uptake of
51

Cr-endotoxin by isolated liver parenchymal cells in vitro. Under the

conditions employed here, our experiments did not show enhanced uptake

96

of endotoxin by liver parenchymal cells when incubated in the presence
of galN. Furthermore, significantly greater amounts of endotoxin were
associated with nonparenchymal cells when compared to parenchymal cell
uptake in vitro.

Sensitization to the lethal effects of endotoxin by galN is maximal
within the first 2 hrs following galN administration (12) suggesting
that the biochemical alterations induced by galN are closely related to
the accompanying increase in endotoxin susceptibility. This observation
led Galanos et al. (12) to suggest that alterations in. metabolism
similar to those induced by galN might be involved in lowering the
threshold response to the toxic effects of endotoxin. While normal and
galN-treated mice show equivalent concentrations of endotoxin localized
in their livers by 3 hrs after injection (Tables 1&2), increased serum
OCT activity and lethality only occur in galN-treated mice, suggesting
that the lethal and hepatotoxic responses to endotoxin are not dependent
on the quantitative localization of endotoxin in liver parenchymal
cells. This notion coupled with significantly greater uptake of
endotoxin by Kupffer cells support the thesis that the liver functions
primarily in clearance of endotoxin from blood and not as a direct
target for endotoxin activity. However, these results do not completely
exclude a direct toxic effect of endotoxin on hepatocytes since galN may
simply decrease the threshold response to low levels of endotoxin
present in hepatocytes. It is unlikely that the concentration of
endotoxin by hepatocytes is strictly related to its lethal effects since
additional data obtained in this laboratory indicate that subcutaneously
administered endotoxin, when compared to intravenously injected

endotoxin, is about 50-fold less concentrated in livers of normal mice

97

while only about 2-fold higher in 50% lethal dose (data not shown).
Alternatively, the possibility that endotoxin sensitization by galN is a
mediated event must be considered. Indirect effects of endotoxin which
might contribute: to enhanced susceptibility of galN-treated. mice to
liver injury and lethality include, among others, ischemia (10), local
or disseminated intravascular coagulation (26), or release and
activation of injurious mediators into the circulation. Further
experiments are required to elucidate the mechanisms underlying the

sensitization of mice to the toxic effects of endotoxin by galN.

LITERATURE CITED

Bachmann, W., E. Harms, B. Hassels, H. Henninger, and W. Reutter.
1977. Studies on rat liver plasma membrane. Altered protein
and phospholipid metabolism after injection of D-galactosamine.
Biochem. J. 166:455-462.

Berry, L.J., D.S. Smyth, and L.S. Colwell. 1968. Inhibition of
hepatic enzyme induction as a sensitive assay for endotoxin.
J. Bacteriol. 96:1191-1199.

Bradley, S.G. 1979. Cellular and molecular mechanisms of action
of bacterial toxins. Ann. Rev. Microbiol. 33:67—94.

Cerriotti, G. 1974. Ornithine carbamyltransferase: determination
by' manual. method. p. 691-695. 22; H.U. Bergmeyer (ed.),
Methods of enzymatic analysis, vol. 2. Academic Press, Inc.,
New York.

Chedid, 5I’ R.C. Skarnes, and M. Parant. 1963. Characterization
of Cr-labeled endotoxin and its identification in plasma and

urine after parenteral administration. J. Exp. Med.
117:561-571.

Cook, J.A., W.J. Dougherty, and T.M. Holt. 1980. Enhanced
sensitivity to endotoxin induced by the RE stimulant, glucan.
Circulatory Shock 7:225-238.

Crane, L.J. and D.L. Miller. 1977. Plasma protein synthesis by
isolated rat hepatocytes. J. Cell. Biol. 72:11-25.

Cynkin, M.A. and G. Ashwell. 1960. Estimation of 3-deoxy sugars
by means of the malonaldehyde-thiobarbituric acid reaction.
Nature (London) 186:155-156.

Decker, K., and D. Keppler. 1974. Galactosamine hepatitis: Key
role of the nucleotide deficiency period in the pathogenesis
of cell injury and cell death. Rev. Physiol. Biochem.
Pharmacol. 71:77—103.

DePalma, R.C., J. Coil, J.H. Davis, and W.D. Holden. 1967.
Cellular and ultrastructural changes in endotoxemia: a light
and electron microscopic study. Surgery 62:505-515.

El-Mofty, S.K., M.C. Scrutton, A. Serroni, C. Nicolini and J.L.
Farber. 1975. Early, reversible plasma membrane injury in
galactosamine-induced liver cell death. Am. .L. Path.
79:579-596.

Galanos, C., M.A., Freudenberg, and W. Reutter. 1979.
Galactosamine—induced sensitization to the lethal effects of
endotoxin. Proc. Nat. Acad. Sci. 76:5939-5943.

98

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

99

Greer, G.G., N.A. Epps, and W.J. Vail. 1973. Interaction of
lipopolysaccharides with mitochondria. II. Effects of

magnesium ions on toxicity of a rough lipOpolysaccharide. J.
Infect. Dis. 128:724-729.

Harris, R.A., D.L. Harris, and D.B. Green. 1968. Effect of
Bordetella endotoxin upon mitochondrial respiration and
energized processes. Arch. Biochem. BiOphys. 128:219-230.

 

Hofmann, P., J. Wilkening, J. Nowack, and K. Decker. 1976.
Response of isolated rat hepatocytes to D—galactosamine and
uridine. Hoppe Seyler's Z. Physiol. Chem. 357:427-433.

Janoff, A., G. Weissmann, B.W. Zweifach, and L. Thomas. 1962.
Pathogenesis of experimental shock. IV. Studies on lysosomes
in normal and tolerant animals subjected to lethal trauma and
endotoxemia. J. Exp. Med. 116:451-465.

Keppler, D., R. Lesch, W. Reutter, and K. Decker. 1968.
Experimental hepatitis induced by D-galactosamine. Exptl.
Mol. Pathol. 9:279-290.

Keppler, D., and D.F. Smith. 1974. Nucleotide contents of ascites

hepatoma cells and their changes induced by D-galactosamine.
Cancer Res. 34:705-711.

Konttinen, A., M. Rajasalmi, and J. Paloheimo. 1964. Serum enzyme
activities in endotoxin shock. Am. J. Physiol. 207:385-388.

Liehr, H., M. Grun, R.P. Seelig, R. Seelig, W. Reutter and W.D.
Heine. 1978. On the pathogenesis of galactosamine hepatitis.

Indications of extrahepatocellular mechanisms responsible for
liver cell death. Virchows Arch. B. Cell Path. 26:331-344.

Mathison, J.C., and R.J. Ulevitch. 1979. The clearance, tissue
distribution, and cellular localization of intravenously
injected lipopolysaccharide in rabbits. J. Immunol.
123:2133-2143.

McCallum, R.E., and L.J. Berry. 1973. Effects of endotoxin on
gluconeogenesis, glycogen synthesis, and liver glycogen
synthase in mice. Infect. Immun. 7:642—654.

Mela, L., L.V. Bacalzo, and L.D. Miller. 1971. Defective
oxidative metabolism of rat liver mitochondria in hemorrhage
and endotoxic shock. Am. J. Physiol. 220:571-577.

Menten, M.L., and H.M. Manning. 1924. Blood sugar studies on
rabbits infected with organisms of the enteritidioparatyphoid
B group. J. Med. Res. 44:675-693.

Moon, R.J., R.A. Vrable, and J.A. Brocka. 1975. In situ
separation of bacterial trapping and killing functions of the
perfused liver. Infect. Immun. 12:411-418.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

100

Morrison, D.C., and C.G. Cochrane. 1974. Direct evidence for
Hageman factor (factor XII) activation by bacterial
lipopolysaccharide (endotoxin). J. Exp. Med. 140:797-811.

Noan, J.P., and D.S. Camara. 1982. Endotoxin, sinusoidal cells,
and liver injury. p. 361-376. Ip_H. POpper and F. Schaffner
(ed.), Progress in liver diseases. Grune & Stratton, Inc.,
New York, New York.

Pertoft, H., K. Rubin, L. Kjellen, T.C. Laurent, and B. Klingeborn.
1977. The viability of cells grown or centrifuged in a new
density gradient medium, Percoll (TM). Exp. Cell Res.
110:449-457.

Ramadori, C., U. Hopf, K.H. Meyer zum Buschenfelde. 1979. Binding
sites for endotoxin lipopolysaccharide on the plasma membrane

of isolated rabbit hepatocytes. Acta Hepatogastroenterol.
26:368-374.

Ramadori, G., U. Hopf, C. Galanos, M. Freudenberg, and K.H. Meyer
zum Buschenfeld. 1980. In vivo and in vitro reactivity of
lipOpolysaccharide and lipid A with parenchymal and
nonparenchymal liver cells in mice. 12 H. Liehr and M. Grun
(ed.), The reticuloendothelial system and the pathogenesis of
liver disease. Elsevier/North Holland Biomedical Press.

Rippe, D.F., and L.J. Berry. 1972. Study of inhibition of
induction of phosphoenolypyruvate carboxykinase by endotoxin
with radial immunodiffusion. Infect. Immun. 6:766-772.

Ruiter, D.J., J. Vander Meulen, A. Brouwer, M.J.R. Hummel, B.J.
Mauw, J.C.M. Van der Ploeg, and E. Wisse. 1981. Uptake by
liver cells of endotoxin following its intravenous injection.
Lab. Invest. 45:38-45.

Sawyer, R.T., R.J. Moon, and S.E. Beneke. 1976. Hepatic clearance
of Candida albicans in rats. Infect. Immun. 14:1348-1355.

Seglen, P.O. 1976. Preparation of isolated rat liver cells. p.
29-83. Ip_D.M. Prescott (ed.), Methods in Cell Biology, vol.
13. Academic Press, New York, San Francisco, London.

Siegel, S. 1956. The Mann-Whitney U Test, Ip_ Nonparametric
Statistics for the Behavioral Sciences. McGraw-Hill Book
Company.

Tavakoli, H., and R.J. Moon. 1982. In vitro and in vivo
association and subcellular distribution of toxic and

experimentally' modified endotoxin. Can" J. IMicrobiol.
28:822-829.

37.

38.

39.

40.

101

Weissmann, C., and L. Thomas. 1962. Studies on lysosomes. I. The
effect of endotoxin, tolerance and cortisone on the release of
enzymes from the granular fraction of rabbit liver. J. Exp.
Med. 116:433-450.

Willerson, J.T., R.L. Trelstad, T. Pincus, S.B. Levy, and S.M.
Wolff. 1970. Subcellular localization of Salmonella
enteritidis endotoxin in liver and spleen of mice and rats.
Infect. Immun. 1:440-445.

 

 

Woodward, J.M., M.L., Camblin, and M.H. Jobe. 1969. Influence of
bacterial infection on serum enzymes of white rats. Appl.
Microbiol. 17:145-149.

Zlydaszyk, J.C., and R.J. Moon. 1976. Fate of 51Cr-labeled
lipopolysaccharide in tissue culture cells and livers of
normal mice. Infect. Immun. 14:100-105.

ARTICLE III

Involvement Of Humoral Factors In The Potentiation

Of Endotoxin Sensitivity By Galactosamine

102

Abstract

Galactosamine (galN) sensitizes mice to endotoxin at least
1000-fold. Such mice experience considerable liver damage as evidenced
by increased serum ornithine carbamyltransferase (OCT) activity. Serum
transfer experiments were performed to determine whether humoral
mediators ‘were involved in the pathogenesis of endotoxemia in galN

sensitized mice. Normal mice given the ID of endotoxin possessed

50
serum mediators which caused symptoms of endotoxicity in galN-treated
mice. Further, serum from C. parvum - treated mice given low doses of
endoxin was even more toxic for galN-treated mice. The active factor(s)
in serum could be destroyed by incubation with trypsin or protease or by
heating at 75°C for 1 hr. Since serum from galN-treated mice given
endotoxin did not possess mediator which was toxic for C. parvum treated

mice, the data suggests that more than one mechanism of enhanced

sensitivity to endotoxin exists. Possible alternatives are discussed.

103

Introduction

Numerous studies in experimental animals suggest that the liver
plays a central role in the host response to endotoxin. For example,
(a) a major portion of intravenously injected endotoxin localizes in
liver tissue (8,24,46), (b) a variety of hepatic metabolic alterations
occur in response to endotoxin administration (2,5,25,26,31), (c) serum
levels of enzymes generally assumed to be of hepatic origin increase
following endotoxin injection (11,18,38), and (d) detoxification of
endotoxin occurs in vitro by cell-free liver homogenates (14). Whether
endotoxin acts directly on hepatocytes or indirectly through humoral
factors is essential to understanding the role of hepatic patho-
physiological responses to endotoxin. In the present study we have
utilized agents which amplify the toxic responses of endotoxin to
address this issue.

A variety of experimental agents or regimens alter sensitivity to
endotoxin. Almost all elicit a wide range of effects in vivo. A common
theme among endotoxin sensitizing agents is their ability to influence

the liver. Corynebacterium parvum and Mchobacterium bovis BCG induce

 

 

significant hyperplasia in major reticuloendothelial organs with
resultant increases in resistance to a variety of bacterial and
parasitic infections (6,7,22). Conversely, these agents also increase
susceptibility to the toxic effects of endotoxin. Galactosamine is a
liver-specific toxin (12,20) which also enhances sensitivity to
endotoxin (16) but ‘without hyperplastic effects on. RES cells. The
biochemical changes in hepatocytes following injection of galactosamine

have been described extensively (12).

104

105

The primary objective in this study is to investigate the
involvement of humoral factors in the potentiation of endotoxin
sensitivity by galactosamine. To accomplish this aim a variety of serum
transfer experiments have been performed with donor serum coming from
endotoxin-challenged. mice. Galactosamine-treated. ‘mice served as
recipients. Liver injury was assessed by measuring serum ornithine
carbamyltransferase activity.

Collectively, the data demonstrate the existence of a humoral
mediator of toxicity in serum of endotoxin-poisoned mice and the ability

of C. parvum treatment to enhance mediator production.

Materials and Methods

Animals. Female CD-1 mice weighing 20-26 gm were obtained from Charles
River Laboratories (Wilmington, DE) and used throughout this study. All
mice were housed conventionally and fed food and water ad_libitum.

Materials. A phenol-water extracted preparation of Salmonella

 

typhimurium lipopolysaccharide was obtained from. Difco Laboratories

 

(Detroit, MI), suspended at appropriate concentrations in pyrogen-free
saline, and stored at -20°C. Endotoxin. was injected intravenously

unless otherwise Specified. The mean 50% lethal dose (LD of

50)
endotoxin with or without galN was determined by the method of Reed and
Muench (29) using at least 10 mice for each dose of endotoxin.
Ornithine, carbamylphosphate (dilithium salt), and D-(+)-galactosamine
HCl were all obtained from Sigma Chemical Co. (St. Louis, MO).
Neutralized solutions of galactosamine were prepared in pyrogen-free
phosphate buffered saline and administered intraperitoneally (0.2ml) at

a standard dose of 0.5mg/g body weight.

Serum OCT determination. Serum ornithine carbamyltransferase (OCT)

 

activity was determined spectrophotometrically using the manual method
of Cerriotti (10), which measures citrulline generated from the
condensation of ornithine and carbamylphosphate. Serum OCT activity was
expressed as nanomoles of citrulline produced/ml of serum per minute.

Preparation of donor serum. Corynebacterium parvum was obtained from
Wellcome Research Laboratories (Beckenham, England; lot#CA763) as a
formalin-killed suspension supplied at a concentration of 7 mg dry
wgt/ml containing thiomersol. C.parvum was washed 3 times, suspended in

PBS at a concentration of 4.375 mg/ml, and stored at 4°C. Mice were

106

107

injected intraperitoneally with 875ug of killed C.parvum. Ten days
later, they received an intravenous injection of 10ug of endotoxin.
Mice were exsanguinated 2 hrs after endotoxin injection and serum was
prepared (C.parvum/endotoxin sera). As controls for C.parvum/endotoxin
serum, serum from normal, C.parvum-treated (C.parvum sera), or
endotoxin-treated (endotoxin sera) mice was used. All sera were stored
at -20°C until use. Aliquots (0.1ml) were injected intravenously.

Endotoxin levels were determined in serum from C.parvum-treated
mice using the Limulus lysate assay (M.A. Bioproducts, Walkersville, MD)
or by quantitation of 51Cr-labeled endotoxin serum obtained 2 hrs after
injecting C.parvum-treated mice with 10ug of endotoxin contained less
than 0.6ug of endotoxin per ml.

Sensitivity of serum mediators to heat and enzyme digestion.

Trypsin digestions (180U/mg, Worthington, Biochemical Corp.,
Freehold, N.J.) were performed by incubating 1 ml of serum (diluted 1:4
with pyrogen- free saline) with 500ug of enzyme at 37°C for 2 hrs.
Protease digestion (5.4U/mg, type XIV, Sigma Chemical Co., St. Louis,
MO) was performed with 1.85 mg of enzyme and serum diluted 1:4 in a
total volume of 1 ml at 37°C for 2 hrs. Serum samples incubated in the
absence of specified enzymes served as controls for enzyme digestion
experiments.

To determine the sensitivity of serum to heat, serum samples were
diluted 1:4 with saline and incubated at 56°C or 75°C for 1 hr.

Following treatment, 0.2 ml of serum was injected intravenously
into mice and serum OCT levels determined 6 hrs later.

Statistical Evaluation. Statistical significance was determined by the

 

Mann-Whitney rank order test (35).

Results

Potentiation of endotoxin toxicity by galactosamine.

 

Groups of normal or galN-treated mice were injected intravenously
with various amounts of endotoxin to establish the mean 50% lethal dose.
A standard dose of 0.5 mg of galN per gm body weight was used. This
dose markedly increased sensitivity to endotoxin and caused no visible
morbidity or mortality. The LD50 of endotoxin in normal mice was
approximately 125ug and in galN-treated mice about 0.1ug (Table 1).
This represents at least a 1000-fold increase in sensitivity to the
lethal effects of endotoxin. Table 1 also shows that galN treatment
decreased the LD50 of subcutaneously administered endotoxin by a factor
of at least 200. This observation was interesting in light of data
obtained in this laboratory which indicates that less than 1.0% of a
subcutaneously administered dose of endotoxin becomes associated with
the liver by 1 hr (Pe tchow and Moon, Article 1). Galactosamine not
only enhanced sensitivity but also accelerated the onset of lethality
after injection of endotoxin. The majority of deaths in galN-treated
mice occurred between 9 and 12 hrs post-injection of endotoxin. Mice
receiving endotoxin alone usually died between 18 and 24 hrs.

The hepatoxic effects of combined treatment with galN and endotoxin
was assessed by monitoring increases in serum OCT activity (10). Normal
serum OCT was 1.310.7nmoles/ml/min. By 8 hrs after galN alone, this
activity increased only slightly (Table 2). When 0.01 ug of endotoxin
was given to galN-treated mice, OCT activity was elevated to ca

45nmoles/ml/min by 6 and 8 hrs. When endotoxin was increased to 0.1 or

1.0ug, the six hr elevation was not changed; but by 8 hrs, OCT activity

108

109

Table 1. Comparison of the LD of endotoxin in

 

 

 

50
normal and galN-treated mice.
Route LD of endotoxin
50

of

Injection
Normal GalNa
iv 125ug 0.1ug
sc 285ug 1.2ug

 

aAll mice treated simultaneously with endotoxin
and galactosamine (0.5mg/g b.w.)

110

Table 2. Serum OCT activity in galN-treated mice after

intravenous injection of various doses of endotoxin.

 

 

 

Dose of Serum OCT Activity (nmoles/ml/min)a
Endotoxin 6 hr 8 hr
None NDb 6.7: 1.8 (6)
0.01ug 44.0:16.7 (5) 45.4: 12.4 (5)
0.10ug 40.3: 7.6 (4) 404.6:137.2 (6)
1.00ug 19.2: 4.5 (4) 365.5:122.1 (7)

aMean value : S.E.M. after concurrent injection of
endotoxin and galactosamine (0.5mg/g b.w.).

bNot determined.

111

had increased to 404 and 365 n moles/ml/min respectively. Control mice
receiving 1.0ug of endotoxin alone showed serum OCT activity of 5.6:1.1

nmoles/ml/min by 8 hrs (data not shown).

Abilityi of serum from endotoxin-poisoned mice to increase OCT in

galN-treated mice.

 

Serum (0.1 ml) obtained from normal mice or mice given the LD50 of
endotoxin (sc) was injected intravenously into either normal mice or
mice receiving a single concurrent injection of galN. Serum OCT
activity was measured in recipient mice after 8 hrs. As shown in Table
3, normal serum given to galN-treated mice stimulated about a 20—fold
increase in OCT level (53.6:30.0). Similar serum given to normal mice
did not alter OCT levels (2.8:0.8). Serum from endotoxin-poisoned mice
did not alter OCT in normal animals (1.9:1.0) but when given to
galN-treated mice increased OCT activity up to 558.6:38.5. This figure
represents a 10-fold higher response when compared to recipients given
normal serum. Serum OCT levels in galN-treated mice receiving endotoxin
serum.‘were almost 300-times higher than 'with normal recipients of

endotoxin serum demonstrating the ability of galN to amplify the effect

of an active factor(s) present in endotoxin serum.

Enhanced mediator response in serum from C. parvum treated mice.

C. parvum vaccine sensitizes mice to endotoxin at least lOOO-fold
(data not shown). It was of interest to determine whether serum from
C. parvum - treated mice given endotoxin (C. parvum/endotoxin) contained
the mediator of endotoxin toxicity observed in mice given endotoxin

alone (cf Table 3). If so, was it present in quantitatively greater

112

Table 3, Elevation of serum OCT activity in galN-treated

mice given serum from endotoxin challenged mice.

 

 

 

Donor Recipient Serum OCT Activity (8hr)a
Serum normal galN-treated
Normal serum 2.8:0.8 (5) 53.6: 30.0 (5)
Endotoxin serumb 1.9:1.0 (5) 558.6:138.5 (7)C

aMean value : S.E.M. by 8 hrs after receiving 0.1 ml of
donor serum with or without galN (0.5mg/g b.w.).

bDonor serum prepared 2 hrs after subcutaneous injection
of mice with 285ug of endotoxin.

CP<.005 versus galN-treated mice receiving normal serum.

113

amounts than in mice given only endotoxin. For these experiments only
IOug of endotoxin was administered. OCT remained the assay tool. Serum

from ‘mice given endotoxin alone (10ug), C. parvum, or C. parvum/

 

endotoxin had no effect on OCT activity in normal mice (Table 4).
Likewise, when endotoxin or C. parvum serum was given to galN-treated
mice, no significant changes were observed. When serum from mice given
both C.parvum and endotoxin was given to galN mice, serum OCT increased
approximately 140-fold.

Significant elevations in serum OCT could still be observed when
C.parvum/endotoxin serum was diluted as much as 4-fold with pyrogen-free
sterile saline (Cutter' Laboratories, Inc., Berkeley, CA) (Table 5).
Active serum also caused significant mortality in galN-treated
recipients. Survival of mice receiving galN and 2-fold or 4-fold
diluted C.parvum/endotoxin serum was 50% and 70%, respectively. A 1:8
dilution. of active seruni no longer caused significant mortality or
elevations of serum OCT activity in galN-treated recipients.
Collectively, these results suggest that humoral factors elicited by
endotoxin are involved in the galN-induced potentiation of endotoxin

lethality and hepatotoxicity.

Partial Characterization of C.parvum/endotoxin serum activiyy.

The sensitivity of C. parvum/endotoxin serum to heat is shown in
Table 6. One hr at 56°C had no effect on the activity whereas 1 hr at
75°C completely destroyed the ability of the serum to cause significant
increases in OCT in galN-treated mice. Further, incubation of .9;
parvum/endotoxin serum 'with trypsin (500ug/ml) for 2 hrs at 37°C

completely abrogated its ability to elevate OCT in galN-treated mice

114

Table 4. Ability of C.parvum/endotoxin serum to elicit

increased serum OCT activity in galN-treated mice.

 

Recipient Serum OCT Activity (6 hr)b

 

Donor Serum

 

normal galN—treated
Normal NDC 20.4: 6.5 (9)
C.parvum/endotoxina 4.1:1.8 (5) 596.2:96.0 (10)d’e
Endotoxina 2.0:o.5 (4) 18.0: 8.2 (5)
C.parvum 7.3:0.9 (4) 2.8: 0.5 (5)

 

aDonor serum prepared 2 hrs after injection of long of
endotoxin in normal or C.parvum-treated mice.

bMean value : S.E.M. in normal or galN-treated mice
by 6 hrs after receiving 0.1 ml of donor serum.

CNot determined.

dRepresents serum OCT activity in 7 of 10 surviving
mice by 6 hrs.

eP<0.001 versus controls.

115

Table 5. Survival and serum OCT activity in galN-treated mice

given various dilutions of C.parvum/endotoxin serum.

 

 

Dilution Percent Recipient
C.parvum/endogoxin Survival Serum OCT Actgvity

Donor Serum by 6 hrs. by 6 hrs.

1:2 50% (10) 966.9:118.0

1:4 70% (10) 502.3:103.4

1 100% (10) 12.6: 2.6

 

aDonor serum prepared 2 hrs after injection of 10ug of
endotoxin in C.parvum-treated mice and diluted with saline.

bMean value : S.E.M. of surviving mice after injection with
galN and 0.1 ml of diluted serum.

116

Table 6. Stability of C.parvum/endotoxin serum

after heating or enzyme digestion.

 

 

a Recipient Seru Percent of
Treatment OCT Activity Control
Control 311.7: 78.1 100
56'C, 1 hr 324.6:129.1 104
75'C, 1 hr 23.4: 2.9 7
Control 234.5:128.3 100
Trypsin 7.5: 3.5 3
Protease 32.0: 13.5 14

 

aC.parvum/endotoxin serum was diluted 1:4 and
incubated at selected temperatures for
1 hr or at 37°C with or without indicated
enzymes for 2 hrs.

bMean value : S.E.M. from at least 4 mice by
6 hrs after receiving galN and 0.2 ml
of test serum.

117

(Table 6). The activity of C. parvum/ endotoxin serum was also
destroyed by digestion with protease. Dialysis for 18 hrs against PBS

resulted in no loss in activity (data not shown).

Susceptibility of galN- or C.parvum-treated mice to the hepatotoxic
activity of C.parvum/endotoxin orygalN/endotoxin serum.

To further investigate the mechanisms at work in this experimental
model, we compared serum OCT in galN-treated and C.parvum stimulated
mice after transfer of either C.parvum/endotoxin serum or galN/endotoxin
serum. GalN/endotoxin serum was prepared 2 hrs after injecting mice
concurrently with long of endotoxin and 0.5mg/g b.w. of galN. As shown
in Table 7, no difference was observed in serum OCT activity of C.parvum-
treated mice by 6 hrs after injection with either C.parvum/endotoxin
serum or galN/endotoxin serum. By contrast, 4-fold diluted C.parvum/
endotoxin serum elicited a significant elevation of serum OCT in
galN-treated mice when compared to undiluted galN/endotoxin serum in
similar recipients. Serum OCT levels were not significantly higher in
galN-treated mice given galN/endotoxin serum (82.8:22.8) when compared
to C.parvum stimulated mice receiving either C.parvum/endotoxin serum
(50.9:11.8) or galN/endotoxin serum (41.2:11.5), suggesting that
galN-induced lesions and sufficient amounts of active serum factor(s)

must be present to cause significant increases in serum OCT levels.

 

118

Table 7. Comparison of serum OCT activity in galN- or C.parvum-treated

mice after receiving C.parvum/endotoxin or galN/endotoxin serum.

 

 

 

Donor Serum OCT Activity in Recipients Treated as Follows:b
Seruma Normal galN C.parvum
C.parvum/endotoxin NDC 229.3:7o.o (5)°'° 50.9:11.8 (5)
galN/endotoxin 11.2:3.0 (5) 82.8:22.8 (7) 41.2:11.5 (4)

 

aDonor serum prepared 2 hrs after injecting mice intravenously
with 10ug of endotoxin.

Mean value : S.E.M. determined 6 hrs after serum transfer (0.1 ml).
Not determined.
Serum transferred after diluted 1:4 with pyrogen-free saline.

I‘D an 0‘

P<.05 versus galN-treated mice receiving GalN/endotoxin serum.

Discussion

Although a variety of experimental approaches have been employed to
study the toxic manifestations of endotoxin in vivo, a clear
understanding of the actual mechanisms by which endotoxin exerts its
toxic effects is not yet available. The results of the present study as
well as those reported by Galanos et al. (16) demonstrate that
administration of galN enhances the sensitivity of mice to endotoxin.
Concurrent administration of galN reduces the mean 50% lethal dose of
intravenously injected endotoxin In! at least 1000-fold (Table 1). In
addition, combined treatment of mice with galN and as little as 0.1ug of
endotoxin caused liver injury by 8 hrs as reflected in the release of a
liver-specific enzyme, OCT, into the circulation (Table 2). Treatment
of mice with either galN or as much as 1.0ug of endotoxin separately did
not cause demonstrable increases in serum OCT activity by 8 hrs.

Furthermore, serum from mice challenged with the LD of endotoxin, but

50
not serum from normal mice, elicited significant elevations of serum OCT
in galN-treated recipients (Table 3). These data suggest that liver
injury induced by endotoxin in galN-treated mice is mediated by humoral
factors.

The indirect action of endotoxin on liver injury in galN-treated
mice was confirmed by our experiments using C.parvum to amplify the
activity of donor serum in response to smaller quantities of endotoxin
(Table 4). Only serum from C.parvum-primed mice given endotoxin was
capable of inducing significant elevations of serum OCT in galN-treated

recipients. Normal recipients given the same serum. also showed 'no

elevations in serum OCT activity, suggesting that the galN-induced

119

120

lesion must be present in order for the active principal of C.parvum/
endotoxin serum to cause liver injury.

A consistent concern of this type of study is that residual
endotoxin present in. donor serum is 'responsible for the ‘biological
activity observed in recipient mice. Three lines of evidence support
the hypothesis that soluble factors present in serum of C.parvum-primed
mice given endotoxin, and not endotoxin per se, induce liver injury in
galN-treated mice. These are:

(a) Transfer of serum from normal mice injected with the same dose
of endotoxin (10ug) to galN-treated mice resulted in neither lethality
or increased serum levels of OCT (Table 4).

(b) Doses of fully toxic endotoxin greater than that found in
0.1 ml of donor C.parvum/endotoxin serum, as determined by the Limulus
assay and quantifying serum levels of S’ICr-endotoxin from identically
treated mice, were unable to induce comparable increases in serum OCT
activity in galN-treated recipient mice by 6 hrs (Table 2).
Furthermore, dilution of C.parvum/endotoxin serum ‘with pyrogen-free
saline up to 4-fold resulted in full activity upon transfer to
galN-treated recipients (Table 5).

(c) Sensitivity of C.parvum/endotoxin serum to heat (75°C, 1 hr)
and digestion by trypsin or protease (Table 6) is inconsistent with the
heat resistance nature of endotoxin.

There is growing evidence implicating the macrophage as the
principal cell type involved in mediating the toxic manifestations of
endotoxin poisoning (32,43). Endotoxin has been reported to activate
macrophages in vivo (27) whereas it may cause cytotoxicity (34) or

activation (45) of macrophage in vitro. Consistent with this hypothesis

121

is the potentiation of endotoxin responsiveness following treatment of

experimental animals with RES stimulants, such as Mycobacterium bovis

 

BCG (39) and glucan (11). For example, preinfection with BCG markedly
enhances the sensitivity of animals to endotoxin in vivo (39,40) while
macrophages obtained from BCG-infected mice produce greater amounts of
Interleukin 1 and Prostaglandin E2 when stimulated with
lipopolysaccharides in vitro (42). In addition, BCG-infected animals
are more resistant to both transplantable tumors (28) and infection by a
variety of unrelated organisms (6,7). These observations are generally
attributed to nonspecific activation of macrOphages since macrophages
derived from BCG-infected animals possess greater ability to kill
unrelated organisms in vitro (22). The ability of C.parvum to markedly
amplify the endotoxin-induced serum activity upon transfer to
galN-treated recipients (Table 4) strengthens the hypothesis that
macrophages and their products play a major role in mediating the toxic
effects of endotoxin.

Although the nature of the mechanisms underlying both variation in
endotoxin responsiveness and the mode of action of endotoxin remain to
be established, several possible mechanisms are possible. The results
of this study suggest that endotoxin-sensitizing agents may be
differentiated into at least two general categories based. on their
mechanisms of sensitization as shown schematically in Figure 1. The
first group, typified. by’ C.parvum. in this study, ‘may intensify the
response of macrophages to endotoxin leading to enhanced production and
release of monokines or lysosomal contents in response to endotoxin

stimulation. Endotoxin has been reported to increase the fragility of

liver lysosomes (18,44) and elevate seruut levels. of *various enzymes

122

humoral mediators

++ (1)

 

endotoxin > death

 

ti (2) i

metabolic lesion

Sensitization Agents:

(1) C.parvum (2) galN
BCG CCl
zymosan lead acetate
Figure 1. Proposed mechanisms of endotoxin sensitization and their

relationship to endotoxin toxicity.

123

normally associated with lysosomes (18,19). Support for this hypothesis
resides in the ability of various membrane stabilizing agents, such as
cortisone or ZnClZ, to protect against the toxic effects of endotoxin
challenge (3,17,37,38) while agents with lysosomal membrane
destabilizing properties, such as silica and carrageenan (9,30),
increase endotoxin sensitivity (4,41). Furthermore, activated
macrophages are more susceptible to endotoxin (34) and release greater
amounts of various monokines (23,42) when stimulated with endotoxin ip_
yitgp, In addition to RES stimulants such as BCG or C.parvum, a number
of agents enhance endotoxin responsiveness, such as CCl4 (15,36), silica
(41), lead acetate (33), and galN (16).

Galactosamine enhancement of endotoxin responsiveness may represent
a second ‘mechanism of endotoxin sensitization. based on *well
characterized biochemical alterations induced by galN in liver
parenchymal cells. Administration of sufficient doses of galN to a
variety of experimental animals elicits a significant depletion of
hepatic uridine nucleotides (21) which leads to impaired biosynthesis of
RNA, proteins, glycogen, and glycoproteins in liver cells. These
biochemical changes are followed by structural and functional
alterations in the plasma membranes of hepatocytes (1,12,13) and may be
reversed by administration of uridine within 3 hrs after galN
administration (12). Galanos et a1 (16) found that. maximal
sensitization to endotoxin occurred during the first 4 hrs following
galN injection while only galN-induced biochemical alterations are in
progress. Furthermore, inhibition of the galN-induced biochemical
alterations by administration of uridine also prevented endotoxin

sensitization by galN. These results suggest that endotoxin

124

sensitization by galN administration is primarily associated with
metabolic alterations induced by this amino sugar.

The data in Table 7 support and extend the notion that multiple
mechanisms of endotoxin sensitization exist. Neither serum from
endotoxin challenged mice (Table 4) or from galN-treated mice given
endotoxin caused increased serum OCT activities in galN-treated
recipients. Serum from C.parvum-primed mice given endotoxin, however,
caused significant elevations in serum: OCT activity in galN-treated
recipients, suggesting that priming mice with C.parvum results in
enhanced serum levels of endotoxin-induced mediators involved in causing
liver injury in galN-treated mice. Because C.parvum/endotoxin serum did
not alter serum OCT levels in normal mice it is possible that galN
causes metabolic perturbations in hepatocytes which enhance host
susceptibility to endotoxin-induced factors found in serum. Conversely,
serum from galN-treated mice given endotoxin did not elicit elevated
serum OCT levels or lethality when transferred to galN-treated mice
(Table 7), suggesting that galN-sensitization is not associated with
enhanced production of humoral factors (Fig. 1). Likewise, C.parvum/
endotoxin serum did not elevate serum OCT levels in C.parvum-stimulated
mice, suggesting that endotoxin sensitization by RES stimulants is not
strictly associated with an amplified susceptibility to humoral factors
(Fig. 1). It remains to be determined whether a metabolic basis or
hyperreactivity to endotoxin exists for agents such as CCl4 and lead
acetate.

The use of agents to modulate endotoxin responsiveness may be
particularly useful in considering the events underlying endotoxin

toxicity. The findings in this report support the notion that both

125

humoral factors and metabolic components are involved in the lethal
response to endotoxin. Endotoxin is known to stimulate the production
of a variety of humoral factors with widely divergent activities and to
elicit a number of metabolic alterations in liver parenchymal cells.
These metabolic alterations may act in concert with injurious humoral
mediators to elicit cellular damage or, when of sufficient magnitude,
result in death of the host. Amplification of the humoral component
with reticuloendothelial stimulants such as C.parvum may enhance the
production and release of injurious mediators from particular cells in
response to endotoxin. Further experiments are required to characterize
the humoral factor(s) and the nature of the cells involved in the
production of this factor(s) which may be involved in the sensitization
of galN-treated mice to the lethal and hepatotoxic effects of endotoxin.
Amplification of the amiabolic component (Fig. 1) with agents such as
galactosamine may sensitize the host to endotoxin by increasing the
susceptibility of particular cells to injury by specific humoral
factors. The specific metabolic alterations induced by galN 'which

directly contribute to endotoxin sensitization remain to be determined.

10.

ll.

12.

LITERATURE CITED

Bachmann, W., E. Harms, B. Hassels, H. Henninger, and W. Reutter
1977. Studies on rat liver plasma membrane. Altered protein
and phospholipid metabolism after injection of
D-galactosamine. Biochem. J. 166:455-462.

Berry, L. J., D. S. Smythe, and L. S. Colwell. 1968. Inhibition
of hepatic enzyme induction as a sensitive assay for
endotoxin. J. Bacteriol. 96:1191-1199.

Berry, L. J. 1975. Bacterial Toxins. Lloydia 38:8-20.

Becker, L. J., and J. A. Rudbach. 1978. Potentiation of
endotoxicity by carrageenan. Infect. Immun. 19:1099-1100.

Berry, L. J. 1977. Bacterial toxins. Crit. Rev. Toxicol.
5:239-273.

Blaese, R. M. 1976. Macrophage function in the development if
immunocompetence and immunodeficiency, RES J.
Reticuloendothel. Soc.20:67-70.

Blanden, R. V., M. J. Lefford, and G. B. Mackaness. 1969 The host
response to Calmette-Guerin bacillus infection in mice. J.
Exp. Med. 129:1079-1107.

Bradley, S. G. 1979. Cellular and molecular mechanisms of action
of bacterial toxins. Ann. Rev. Microbiol. 33:67-94.

Catanzaro, P. J., H. J. Schwartz, and R. C. Graham. 1971. Spectrum
and possible mechanism of carrageenan cytotoxicity. Am. J.
Pathol. 64:387-403.

Cerriotti, G. 1974. Ornithine carbamyltransferase: determination
by manual method. p. 691-695. _I_r_1_ H. U. Bergmeyer (ed.),
Methods of Enzymatic Analysis, vol. 2. Academic Press, Inc.,
New York.

Cook, J. A., W. J. Dougherty, and T. M. Holt. 1980. Enhanced
sensitivity to endotoxin induced by the RE stimulant, glucan.
Circulatory Shock 7:225-238.

Decker, K., and D. Keppler. 1974. Galactosamine hepatitis: Key
role of the nucleotide deficiency period in the pathogenesis
of cell injury and cell death. Rev. Physiol. Biochem.
Pharmacol. 71:77-103.

126

l3.

14.

15.

l6.

l7.

l8.

19.

20.

21.

22.

23.

127

El-Mofty, S. K., M. C. Scrutton, A. Serroni, C. Nicolini and J. L.
Farber. 1975. Early, reversible plasma. membrane injury in
galactosamine-induced liver cell death. Am. J. Path. 79:579-596.

Farrar, W. E., Jr., M. Edison, and T. H. Kant. 1968.
Susceptibility of rabbits to pyrogenic and lethal effects of

endotoxin after acute liver injury. Proc. Soc. Exptl. Biol.
Med. 128:711-715.

Formal, S. B., H. E. Noyes, and H. Schneider. 1960. Experimental
Shigella infections. III. Sensitivity of normal, starved and
carbon tetrachloride-treated guinea pigs to endotoxin. Proc.
Soc. Exptl. Biol. Med. 103:415-418.

Galanos, C., M. A. Freudenberg, and W. Reutter. 1979.
Galactosamine-induced sensitization to the lethel effects of
endotoxin. Proc. Nat. Acad. Sci. 76:5939-5943.

Geller, P., E. Merril, and E. Jawetz. 1954. Effects of cortisone
and antibiotics on lethal action of endotoxins in mice. Proc.
Soc. Exp. Biol. Med. 86:716-719.

Janoff, A., G. Weissmann, B. W. Zweifach, and L. Thomas. 1962.
Pathogenesis of experimental shock. IV. Studies on lysosomes

in normal and tolerant animals subjected to lethal trauma and
endotoxemia. J. Exp. Med. 116:451-465.

Janson, P.M.C., S. H. Kuhn, and J. J. Geldenhuys. 1975. Lysosomal
disruption during the development of endotoxin shock in the
baboon. S. Afr. Med. J. 47:1041-1047.

Keppler, D., R. Lesch, W. Reutter, and K. Decker. 1968.
Experimental hepatitis induced by D-galactosamine. Exptl.
Mol. Pathol. 9:279-290.

Keppler, D., and D. F. Smith. 1974. Nucleotide contents of
ascites hepatoma cells and their changes induced by
D-galactosamine. Cancer Res. 34:705-711.

Mahmoud, A.A.F., P.A.S. Peters, K.H. Civil, and .L. S. Remington.
1979. In vitro killing of shistosomula of Shistosoma mansoni

by BCG and C. parvum-activated. macrophages. J. Immunol.
122:1655-1657.

Manuel, D. A., R. N. Moore, and S. E. Mergenhagen. 1980.
Macrophages as a source of tumoricidal activity
(tumor-necrotizing factor). Infect. Immun. 30:523-530.

 

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

128

McCallum, R. E., and L. J. Berry. 1973. Effects of endotoxin on
gluconeogenesis, glycogen synthesis, and liver glycogen
synthase in mice. Infect. Immun. 7:642-654.

Mela, L., L. V. Bacalzo, and. L. D. ‘Miller. 1971. Defective
oxidative metabolism of rat liver mitochondria in hemorrhage
and endotoxic shock. Am. J. Physiol. 220:571-577.

Mela, L., L. D. Miller, L. V. Bacalzo, K. Olofsson, and R. R.
White. 1973. Role of intracellular variations of lysosomal
enzyme activity and oxygen tension in mitochondrial impairment

in endotoxemia and hemorrhage in the rat. Annals of Surgery
178:727-735.

Neter, I. 1969. Endotoxins and the immune response. Curr. Top.
Microbiol. Immunol. 47:82-124.

Old, L. J., D. A. Clarke, and B. Benacerraf. 1959. Effect of
Bacillus Calmette-Guerin infection on transplanted tumors in
the mouse. Nature 184:291-292.

Reed, L. J., and H. A. Muench. 1938. A simple method for
estimating fifty percent endpoints. Am. J. Hyg. 27:493-499.

Rios, A., and R. L. Simmons. 1972. Poly-2-vinylpyridine-N-oxide
reverses the immunosuppresive effects of silica and
carrageenan. Transplantation 13:343-345.

Rippe, D. F., and L. J. Berry. 1972. Study of inhibition of
induction of phosphoenolpyruvate carboxykinase by endotoxin
with radial immunodiffusion. Infect. Immun. 6:766-772.

Rosenstreich, D. L., L. M. Glode, L. M. Wahl, A. L. Sandberg, and
S. E. Mergenhagen. 1977. Analysis of the cellular defects of
endotoxin-unresponsive C3H/HeJ mice. P. 314-320. 12_ D.
Schlessinger (ed.), Microbiology-1977. American Society for
Microbiology, Washington, D. C.

Selye, H., B. Tuchweber, and L. Bertok. 1966. The effect of lead
acetate on the susceptibility of rats to endotoxin. J.
Bacteriol. 91:884-890.

Shands, J. W., Jr., D. L. Peavy, B. J. Gormus, and J. McGraw.
1974. In vitro and in vivo effects of endotoxin on mouse
peritoneal cells. Infect. Immun. 9:106—112.

Siegel, S. 1956. The Mann-Whitney U Test, Ip_ Nonparametric
Statistics for the Behavioral Sciences. McGraw-Hill Book
Company.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

129

Snyder, I. S., M. K. Agarwal, and L. J. Berry. 1967 Influence of
carbon tetrachloride on inducible enzymes and response to
endotoxin in mice. J. Bacteriol 94:1817-1823.

Snyder, S. L., and R. 1. Walker. 1976. Inhibition of lethality in
endotoxin-challenged mice treated with zinc chloride. Infect.
Immun. 13:998-1000.

Sobocinski, P. 2., M. C. Powanda, W. J. Canterbury, 8. V. Machota,
R. 1. Walker, and S. L. Snyder. 1977. Role of zinc in the
abatement of hepatocellular damage and mortality incidence in
endotoxic rats. Infect. Immun. 15:950-957.

Suter, E., G. E. Ullman, and R. G. Hoffman. 1958. Sensitivity of
mice to endotoxin after vaccination with BCG (Bacillus
Calmette-Guerin). Proc. Soc. Exp. Biol. Med. 99:167-169.

Vogel, S. M., R. N. Moore, J. D. Sipe, and D. L. Rosenstreich.
1980. BCG-induced enhancement of endotoxin sensitivity in
C3H/HeJ mice. I. In vivo studies. J. Immunol. 124:2004-2009.

Vogel, S. M., K. E. English, and A. D. O'Brien. 1982. Silica
enhancement of murine endotoxin sensitivity. Infect. Immun.
38: 681-685.

Vogel, S. M., L. L. Weedon, L. M. Wahl, and D. L. Rosenstreich.
1982. BCG-induced enhancement of endotoxin sensitivity in
C3H/HeJ’ mice. II. T cell. ‘modulation. of ‘macrophage
sensitivity to LPS in vitro. Immunobiology 160:479-493.

Vogel, S. M., and S. E. Mergenhagen. 1982. Cellular basis of
endotoxin susceptibility. p. 160-168. Ip_R. Genco and S. E.
Mergenhagen (ed.), Hostparasite interactions in. periodontal
diseases. American Society for Microbiology, Washington, D.C.

Weissmann, C., and L. thomas. 1962. Studies on Lysosomes. I. The
effect of endotoxin, tolerance and cortisone on the release of
enzymes from the granular fraction of rabbit liver. J. Exp. Med.
116:433-450.

Wilton, J. M. A., D. L. Rosenstreich, and J. J. Oppenheim. 1975.
Requirement of bone-marrow derived lymphocytes for macrophage

activation In! bacterial lipopolysaccharid .J. Immunol.
114:388-393.

Zlydaszyk, J. C., and R. J. Moon. 1976. Fate of Cr-labeled
lipopolysaccharide in tissue culture cells and livers of
normal mice. Infect. Immun. 14:100-105.

APPENDIX

Cellular Composition of Nonparenchymal Cell Fractions Produced By

Centrifugal Elutriation.

 

Normal mouse liver nonparenchymal cell preparations were fractionated
using centrifugal elutriation following the method of Ruiter et a1. (6) as
outlined in Article II. Three fractions of 100 ml volume were collected at
flow rates of 14, 22, and 40 ml/min at a constant rotor speed of 2500 rpm.
Cells were harvested from each fraction by centrifugation at 1500 rpm for
20 minutes and resuspended in L-15 medium. Cell number and viability were
determined by hemocytometer count in 0.2% Trypan blue in saline. Cell
types in the three fractions were characterized by Mr. Robert Leunk (Ph.D.
Dissertation, Michigan State University, 1983). Cytocentifuge preparations
of cell suspensions were made using a Cytospin 2 (Shandon Instruments).
Slides were air dried, fixed in methanol for 10 min, and stained by the May-
Grunwald—Giemsa technique. Nonparenchymal cell morphology was observed by
light microscopy. At least 200 cells of each suspension were observed.

Cytochemical staining for nonspecific esterase was performed according
to to the method of Li et al. (3,5) using Naphthyl-AS-acetate as substrate.

The ability of cells to phagocytose latex beads was assessed in vitro
by mixing 1x106 cells with 2 drops of a 1:10 dilution of latex spheres
(0.81 micron diameter, Difco) in medium 199 with 10% newborn calf serum.
The mixtures were incubated under mild agitation at 37'C for 1 hour in a

humidified atmosphere containing 5% CO After incubation, the mixtures

2.
were filtered through Nucleopore membranes of 5 micron pore size and rinsed
with 5 ml of medium. Cells were washed off the filters and cytocentrifuge
preparations were made and stained as described above. At least 200 cells

of each suspension were examined for the presence of intracellular latex

spheres at 1000X magnification.

130

131

About 40% of the initial nonparenchymal cell suspension could be
recovered after centrifugal elutriation and they were usually more than
90% viable. The various cell types contained in each fraction were
identified by light microscopic observation of (i) cellular morphology,
(ii) in vitro phagocytosis of latex beads, and (iii) cytochemical staining
for nonspecific esterase.

Kupffer cells exhibited an elongated or oval, often indented, nucleus
which was usually located eccentrically. Their ropy chromatin stained
only lightly. The cytoplasm was lightly basophilic and highly vacuolated.
Endothelial cells had a more darkly staining, round or oval nucleus with
densely packed chromatin. The slightly granular cytoplasm was basophilic.
Lymphocytes exhibited a round, deeply staining nucleus, often filling
almost all of the cell. The cytoplasm was deeply basophilic and stained
homogenously. Other cell types present included polymorphonuclear leukocytes
(PMN's) and eosinophils.

Latex positive cells exhibited several, often clumped, refractile latex
spheres contained within the cell cytoplasm. Cellular morphology could
also be observed in these cytocentrifuge preparations. Kupffer cells are
the only nonparenchymal cell type which internalize latex spheres of 0.81
micron diameter (1,2,7). A low number of cells exhibited typical Kupffer
cell morphology yet contained no latex spheres. Some PMN's were also
observed to have phagocytosed latex spheres but were excluded from latex
positive counts.

Esterase positive cells appeared blue while esterase negative cells were
clear when viewed by light microscopy. Both Kupffer and endothelial cells
are positive for nonspecific esterase (1,3) and these cells exhibit diffuse

cytoplasmic staining. Lymphocytes are esterase negative, except for a

132

subclass of human T lymphocytes which reportedly contain a single perinuclear
spot of esterase activity (4).

The cellular composition of three fractions of liver nonparenchymal
cells produced by centrifugal elutriation are shown in Table 1. Fraction
1 was enriched for lymphocytes (54%) but also contained considerable
endothelial cells and small amounts of subcellular debris and cell nuclei.
The second fraction contained primarily endothelial cells (72%) with lesser
amounts of Kupffer cells (19%) and lymphocytes (10%). Fraction 3 was
enriched for Kupffer cells (62%) but also contained considerable endothelial
cells (31%). Results obtained by the three methods of cellular identification
were in fairly close agreement.

Nonparenchymal cells isolated from endotoxin-poisoned mice and
fractionated by centrifugal elutriation were identified by observation of
cellular morphology. No major differences were observed in the cellular
composition of the three fractions of cells produced by centrifugal

elutriation.

133

.mCOHum=HEumumw kucmEeuwaxw mumumame a mo cowumfi>mo vuwocmum « ammucmuuma some may we cacao m=Hm>

a

.Eau oomw mo woman acuou
unnumcoo m up .>Hw>euuwemwu .cfiE\HE as one .NN .qfi «0 women 30am on vmuomaaoo Mia mcoauomumm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m

Am>fiumwwcv Am>fiUfimcav
m dflm aflmm wcacwmuw wwmumumm
. .I I
W Am>fiumwmcv Am>fiufimoav
a name mflmm maneumoowmna xmumq
f, 0 ~40 mfifim mflmo zwoaosapoz me coauomum
MI I
. Am>aomumcv Am>aBamoav
mfim~ mfiww wcacwmum mmmuwumm
Am>fiumwmcv Ao>wufimoav
mfimw mama maneumoowmsa xmumq
{mmw-ili;li -mmfiiy.illiillr II Nan .t- Nflme swoaonauoz we coauomum
Ao>fiummocv Ao>eufimoav
~«Om Neon wcficamum mmmumumm
Am>fiumamcv Am>eufim0dv
fifinm dfim maneuxuowmsa xmuma
fiflfi whom mawm omflw awoaosduoz fie cowuowum
it. 2:. .IL. -i:.: - --. -:tliiiiiliilu.
mHHmu wHHmu maemo
umnuo mwuzocso6>m Hmfifimcuoocm someday

 

 

 

 

 

 

.coaumfipusam HmwSwaucmu so vmosvoua mcofiuumuu Hams Hmezsocmwmacoc

mo coeufimoeeoo umHSHHmU

.H manna

LITERATURE CITED

Crofton, R.W., M.M.C. Diesselhoff-den Dulk, and R. Van Furth. 1978.
The origin, kinetics, and characteristics of the Kupffer cells
in the normal steady state. J. Exp. Med. 148:1-17.

De Leeuw, A.M., E.C. Sleyster, M.J. Quiet, and D.L. Knook. 1982.
Rat and mouse Kupffer and endothelial cells: a comparison of
structural and functional characteristics. Kupffer Cell
Bull. 5:4—8.

Emeis, J.J., and B. Planque. 1976. Heterogeneity of cells isolated
from the rat liver by pronase digestion: Ultrastructure,
cytochemistry, and cell structure. J. Reticuloendothel. Soc.
20:11-29.

Knowles, D.M., T. Hoffman, M. Ferrarini, and H.C. Kunkel. 1978.
The demonstration of acid alpha-naphthyl acetate esterase
activity in human lymphocytes: Usefulness as a T cell marker.
Eur. J. Immunol. 35:112-123.

Li, C.Y.,L.T. Yam, and W.H. Crosby. 1972. Histochemical
characterization of cellular and structural elements of the
human spleen. J. Histochem. Cytochem. 20:1049-1057.

Ruiter, D.J., J. Van der Meulen, A. Brouwer, M.J.R. Hummel, B.J. Mauw,
J.C.M. Van der Ploeg, and E. Wisse. 1981. Uptake by liver cells
of endotoxin following its intravenous injection. Lab. Invest.
45:38-45.

Widman, J, R.S. Cotran, and H.D. Fahimi. 1972. Mononuclear
phagocytes (Kupffer cells) and endothelial cells. Identification
of two functional cell types in rat liver sinusoids by
endogenous peroxidase activity. J. Cell Biol. 52:159-170.

134

SUMMARY

It is clearly evident that attempts to determine the primary cellular
and molecular basis of the toxicity of gram-negative bacterial endotoxin
are impeded by a variety of complications. Knowledge of the fate of
endotoxin in vivo is essential to the understanding of the biological
actions of endotoxin. Whether the pathophysiologic changes induced by
endotoxin result primarily from the direct interaction of endotoxin with
critical host tissues or are mediated by humoral factors present in the
circulation remains in question.

There is considerable evidence illustrating the role of the liver in
the clearance of endotoxin from blood (10,14,17). The liver is also the
site of numerous endotoxin-induced metabolic alterations (1,2,11,12). The
results presented in Article I illustrate that the organ distribution of
sc administered endotoxin is dramatically different from iv or ip injected
endotoxin. Significantly lower amounts of sc administered endotoxin reached
various organs, particularly the liver. For example, 1 hour after sc
injection less than 15% of the toxin was absorbed into the vascular compart-
ment and less than 1% became associated with the liver. Despite the dramatic
differences seen in the tissue distribution the LD50 of endotoxin following
sc injection was only twice that of either iv or ip injection. Taken
together, these results suggest that the lethal effects of endotoxin are
independent of the amounts of endotoxin localized within the deep tissues
of the host. Furthermore, subcutaneously injected endotoxin inhibited the
hormonal induction of hepatic PEPCK and depleted liver glycogen. When
correlated with the distribution data, these results indicate that hepatic

metabolic responses to endotoxin are not dependent on the amounts of

135

136

endotoxin associated with the liver or other tissues.

The conclusions drawn from the results of Article I were further
tested by determining whether potentiation of endotoxin sensitivity by galN
is associated with enhanced uptake of endotoxin by liver parenchymal cells
(Article II). Galactosamine is a liver-specific toxin that causes liver
cell necrosis (8) and increases the susceptibility of a variety of experimental
animals to the lethal effects of endotoxin (6). The findings of Galanos
et al. (6) were confirmed by our observation that galN-treatment causes a
1000-fold decrease in the LD50 of endotoxin. Furthermore, a dose of endotoxin
that was well-tolerated by normal mice caused marked increases in serum
OCT activity in galN-treated mice. Galactosamine-treatment did not alter
the organ distribution of iv injected endotoxin nor enhance the uptake of
endotoxin by liver parenchymal cells. Taken together, these results suggest
that the increased susceptibility of galN-treated mice to the toxic effects
of endotoxin is independent of the quantitative localization of endotoxin
in host liver parenchymal cells. Furthermore, the majority of iv injected
endotoxin found in livers of normal mice was associated with Kupffer cells,
supporting the thesis that the liver acts primarily to clear endotoxin from
blood and not as a direct target for endotoxin action.

Until recently it had been difficult to understand how endotoxin acts
to produce such a diverse array of host responses. Recent evidence indicates
that endotoxin activates specific cells to release a variety of mediators
into the circulation which are believed to be responsible for many of the
biological effects of endotoxin in vivo. Whether these factors are involved
in mediating the toxic and lethal effects of endotoxin remains uncertain.

The results presented herein further illustrate that humoral factors

mediate the increased susceptibility of galN-treated mice to the toxic effects

137

of endotoxin. Serum from mice that were injected with the LD50 of endotoxin
elicited marked increases in serum OCT activity when transferred to galN-
treated mice. The activity of endotoxemic serum was amplified by combined
exposure of donor mice to C.parvum and 0.1 LD50 of endotoxin. C.parvum/
endotoxin serum decreased survival and increased serum OCT levels in galN-
treated but not normal mice. Serum from mice that received C.parvum or

0.1 LD50 of endotoxin alone was inactive.

C.parvum-primed mice were sensitized to the lethal effects of endotoxin

to approximately the same extent (LD =.13ug) as galN-treated mice (LDSO=

50
.O9ug). Whereas C.parvum/endotoxin serum caused lethality and marked
increases in serum OCT levels in galN-treated mice, neither response occurred
following transfer of C.parvum/endotoxin serum to C.parvum-primed mice.
Likewise, donor serum obtained from galN-treated mice given the same dose of
endotoxin (0.1 LDSO) was not lethal or capable of elevating serum OCT levels
when transferred to C.parvum-primed or galN-treated mice.

Considerable evidence exists implicating the macrophage as the principal
cell type involved in mediating the toxic effects of endotoxin (16).
Activation of the RES with agents such as BCG or C.parvum is associated
with enhanced resistance to infection (3,4) and enhanced sensitivity to the
lethal effects of endotoxin (5,13). It is generally accepted that these
phenomena are directly related to an enhanced state of macrophage activation
in vivo (7) since macrophages derived from BCG-infected mice produce greater
amounts of monokines and lysosomal enzymes in response to LPS in vitro (9,15).
The demonstration that combined exposure of mice to C.parvum and endotoxin
amplifies the appearance of humoral mediators which are toxic for galN-treated
mice strengthens the hypothesis that macrophages and their products play a
major role in mediating the toxic effects of endotoxin.

Collectively, the results presented herein demonstrate that humoral

138

factors mediate the increased susceptibility of galactosamine-treated mice

to the toxic effects of endotoxin and further suggest that potentiation of
endotoxin sensitivity may occur as a result of either (1) specific alterations
in metabolism leading to enhanced susceptibility to mediator effects or

(ii) enhanced production of mediators in response to endotoxin due to an

elevated state of macrophage activation.

10.

11.

12.

LITERATURE CITED

Berry, L.J., D.S. Smythe, and L.S. Colwell. 1968. Inhibition of
hepatic enzyme induction as a sensitive assay for endotoxin.
J. Bacteriol. 96:1191-1199.

Berry, L.J., D.S. Smythe, and L.H. Young. 1959. Effects of bacterial
endotoxin on metabolism. I. Carbohydrate depletion and the
protective role of cortisone. J. Exp. Med. 110:389-405.

Blaese, R.M. 1976. Macrophage function in the development of
immunocompetence and immunodeficiency. RES J. Reticuloendothel.
Soc. 20:67-70.

Blanden, R.V., M.J. Lefford, and G.B. Mackaness. 1969. The host
response to C-lmette-Guerin bacillus infection in mice.
J. Exp. Med. 129:1079-1107.

Cook, J.A., W.J. Dougherty, and T.M. Holt. 1980. Enhanced sensitivity
to endotoxin induced by the RE stimulant, glucan. Circulatory
Shock 7:225-238.

Galanos, C., M.A. Freudenberg, and W. Reutter. 1979. Galactosamine-
induced sensitization to the lethal effects of endotoxin.
Proc. Nat. Acad. Sci. 76:5939-5943.

Karnovsky, M.L., and J.K. Lazdins. 1979. Biochemical criteria
for activated macrophages. J. Immunol. 121:809-814.

Keppler, D., R. Lesch, W. Reutter, and K. Decker. 1968. Experimental
hepatitis induced by D-galactosamine. Exptl. Mol. Pathol.
9:279—290.

Mannel, D.A., R.N. Moore, and S.E. Mergenhagen. 1980. Macrophages
as a source of tumoricidal activity (tumor-necrotizing factor).
Infect. Immun. 30:523-530.

Mathison, J.C., and R.J. Ulevitch. 1979. The clearance, tissue
distribution, and cellular localization of intravenously
injected lipopolysaccharide in rabbits. J. Immunol. 123:
2133-2143.

McCallum, R.E., and L.J. Berry. 1973. Effects of endotoxin on
gluconeogenesis, glycogen synthesis, and liver glycogen
synthase in mice. Infect. Immun. 7:642-654.

Mela, L., L.V. Bacalzo, and L.D. Miller. 1971. Defective oxidative

metabolism of rat liver mitochondria in hemorrhage and endotoxic
shock. Am. J. Physiol. 220:571-577.

139

13.

14.

15.

16.

17.

140

Suter, E., G.B. Ullman, and R.C. Hoffman. 1958. Sensitivity of mice
to endotoxin after vaccination with BCG (Bacillus Calmette-
Guerin). Proc. Soc. Exp. Biol. Med. 99:167-169.

Tavakoli, H., and R.J. Moon. 1982. In vitro and in vivo association
and subcellular distribution of toxic and experimentally
modified endotoxin. Can. J. Microbiol. 28:822-829.

Vogel, S.N., L.L. Weedon, L.M. wahl, and D.L. Rosenstreich. 1982.
BCG-induced enhancement of endotoxin sensitivity in C3H/HeJ
mice. II. T cell modulation of macrophage sensitivity to LPS
in vitro. Immunobiol. 160:479-493.

Vogel, S.N., and S.E. Mergenhagen. 1982. Cellular basis of endotoxin
susceptibility. pp. 160-168, 12 R. Genco and S.E. Mergenhagen
(ed.), Host-parasite interactions in periodontal diseases.
American Society for Microbiology, Washington, D.C.

Zlydaszyk, J.C., and R.J. Moon. 1976. Fate of 51Cr-labeled
lipopolysaccharide in tissue culture cells and livers of
normal mice. Infect. Immun. 14:100-105.