MUCUS SECRETAGOGUE ACTIVITY IN CECAL CONTENTS
OF RABBITS WITH EXPERIMENTALLY-INDUCED
CHARLOTTE EVANS HOTCHKISS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Charlotte Evans Hotchkiss
To Mark, Laura, and Arthur who never let me forget what
is really important.
First and foremost, I would like to thank my advisor, Dr.
Merritt for his continuous advice and support. Special thanks
also go to Dr. Moreland for support in the laboratory animal
training program that allowed me to work for this degree. In
addition, my committee members whose scepticism inspired me to
prove my point were instrumental in completing this work. I
would like to thank Dr. Jacobson and Sylvia Tucker for
allowing me to use their laboratory, computer,
photomicroscope, and supplies. Mark Hotchkiss provided help
with the photography. Finally, thanks go to Dr. Davis and the
Animal Resources Department for diagnostic laboratory support,
and particularly to Dr. Schoeb for his help with the
TABLE OF CONTENTS
ACKNOWLEDGMENTS . .
LIST OF FIGURES .
KEY TO ABBREVIATIONS .
1. INTRODUCTION .
2. LITERATURE REVIEW . .
Introduction . .
Characteristics of Mucus .
Physical Characteristics .
Production . .
Secretion . .
Degradation . .
Postulated Functions of Mucus .
Lubrication . .
Protection from Infection .
Nutrient for Flora .
Diffusion Barrier .
Creation of Microenvironment
Techniques Used to Study Mucus .
Model Systems .
Quantitation of Mucus .
Secretion of Mucus .
Composition of Mucus .
Diffusion Through Mucus .
Microbial Virulence .
Theoretical Considerations .
Lubrication . .
Protection from Infection .
Nutrient for Flora .
Diffusion Barrier .
Creation of Microenvironment
Conclusions . .
3. PILOT STUDIES .
Cecal Filtrate Collection
Intestinal Explants .
Enzyme-Linked Lectin Assay .
Data Analysis .
Cecal Ligation .
4. REFINEMENT OF METHODS . 73
Use of Soybean Agglutinin to Quantitate Mucus 73
Introduction . . 73
Materials and Methods . 73
Results . . 76
Discussion . . 78
Evaluation of Enzyme-Linked Lectin Assay 79
Introduction . . 79
Materials and Methods . 80
Results . . 83
Discussion . . 85
Harvesting of Mucus from Explants . 88
Introduction . . 88
Materials and Methods . 88
Results . . 89
Discussion . . 89
5. CECAL LIGATION AS A MODEL OF MUCOID ENTEROPATHY 95
Introduction . . .. 95
Materials and Methods . 96
Results . . 98
Discussion . . 102
6. MUCUS SECRETION FROM INTESTINAL EXPLANTS 120
Introduction . . 120
Materials and Methods . 120
Results . . 123
Discussion . . 124
7. COMPARISON OF ELLA AND IN VITRO LABELLING 128
Introduction . . 128
Materials and Methods . 128
Results . . 131
Discussion . . 132
8. PHYSICAL CHARACTERISTICS OF MUCUS SECRETAGOGUE 136
Introduction . . 136
Materials and Methods . 136
Results . . 139
Discussion . . 141
9. CONCLUSIONS . . 152
Future Directions . . 155
A. DATA TABLES FOR CHAPTER 3 . .
B. DATA TABLES FOR CHAPTER 5 . .
C. DATA TABLES FOR CHAPTER 6 . .
D. DATA TABLES FOR CHAPTER 7 (RADIOACTIVITY) .
E. DATA TABLES FOR CHAPTER 7 (ELLA) .
F. DATA TABLES FOR CHAPTER 8 (RADIOACTIVITY) .
G. DATA TABLES FOR CHAPTER 8, PART A (ELLA) .
H. DATA TABLES FOR CHAPTER 8, PART B (ELLA) .
REFERENCE LIST . . .
BIOGRAPHICAL SKETCH . .
LIST OF FIGURES
3.1. Sources of intestinal explants. .
3.2. Standard curves (ELLA). .
3.3. Serial dilutions of control filtrates. .
3.4. Serial dilutions of ME filtrates. .
3.5. Mucus in medium/filtrates (ELLA). .
3.6. Mucus secretion from explants. .
3.7. Site of cecal ligation. .
4.1. 6% SDS-polyacrylamide gel. .
4.2. Western blot (lectin). . ..
4.3. Explant treated with N-acetylcysteine. .
5.1. Excessive colonic mucus. .
5.2. Gross cecal necrosis. . .
5.3. Weight gain or loss. . .
5.4. Cecal necrosis and inflammation. .
5.5. Goblet cell hyperplasia. .
5.6. Depletion of acidic mucin. .
5.7. Colonic inflammation. . .
5.8. Characteristic of cecal contents. .
6.1. Mucus in medium/filtrates (ELLA). .
6.2. Mucus secretion from explants. .
7.1. Mucus in medium/filtrates (ELLA). .
Comparison of ELLA and tracer secretion.
Proportion of tracer in medium/filtrate.
Precipitable secreted tracer. .
Mucus secretion, part A (ELLA) .
Mucus secretion, part B (ELLA). .
Western blot (lectin). . ..
Western blot (lectin). . ..
SDS-polyacrylamide gel. .
KEY TO ABBREVIATIONS
Bovine serum albumin
Cyclic adenosine monophosphate
Enzyme-linked immunosorbent assay
Enzyme-linked lectin assay
Hemotoxylin and eosin
Hank's balanced salt solution
New Zealand White
Phosphate-buffered saline-tween 20
Soybean agglutinin-horseradish peroxidase
Standard error of the mean
Vasoactive intestinal peptide
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
MUCUS SECRETAGOGUE ACTIVITY IN CECAL CONTENTS
OF RABBITS WITH EXPERIMENTALLY-INDUCED
Charlotte Evans Hotchkiss
Chairperson: Dr Alfred Merritt II
Major Department: Veterinary Medicine
Mucoid enteropathy is a disease of rabbits, characterized
by copious mucus secretion in the intestinal tract. The
etiology of this disease is unknown. This study uses an
adaptation of the cecal ligation model developed by Toofanian
and Targowski to induce experimental mucoid enteropathy.
Filtrates of cecal contents from these rabbits and from
control animals were prepared. Explants from healthy rabbits
were incubated with these filtrates, and mucus secretion was
measured using an enzyme-linked lectin assay, or by measuring
secretion of radiolabelled mucus from explants that had been
preincubated with 3H-glucosamine. The results demonstrate the
presence of a secretagogue for colonic mucus in the cecal
contents of rabbits with experimentally-induced mucoid
enteropathy. Pretreatment of filtrates demonstrates that this
secretagogue can be precipitated with ammonium sulfate, and is
destroyed by heat (1000C for 30 minutes) or strong acid (pH 1
for 30 minutes).
Mucoid enteropathy (ME) is a disease that affects only
rabbits. It primarily affects young animals, but it can occur
at any age (105). Mortality from mucoid enteropathy has been
reported as 1-2% (238) to 10-20% (105, 251) of rabbits
kindled. Signs associated with the disease include anorexia,
dehydration, depression, subnormal temperature, distended
abdomen, and the defecation of clear to yellow gelatinous
mucus in the feces (238). Afflicted rabbits may also have
diarrhea, or may pass nothing but mucus (208) There is
disagreement as to whether or not affected rabbits drink
water. In our experience, and in some reports (235), they are
adipsic; but many reports describe polydipsia (238, 251).
Clinically, the rabbits are dehydrated. Afflicted rabbits may
die acutely, or after several days of illness. Affected
rabbits may recover, with or without intervention (105).
At necropsy, the consistent finding is copious amounts of
clear to yellow gelatinous mucus in the colon. The small
intestine may contain fluid and/or mucus. The cecum is often
impacted with dry digesta, which may contain gas pockets. The
gallbladder is often distended. Histologically, there may be
goblet cell hyperplasia, most noticeable in the ileum but may
occur in any part of the gastrointestinal tract. There is
minimal histologic evidence of inflammation, with mild mixed
infiltrates seen inconsistently (142, 238, 251).
Some reports of mucoidd enteritis" (142, 163) describe
gross paintbrush hemorrhages of the large intestine, and
microscopic edema and inflammation. Such lesions are more
characteristic of clostridial enterotoxemia, a more recently
defined disease of young rabbits (20), which is caused by an
overgrowth of toxin-producing clostridial organisms, C.
spiroforme or C. difficile most commonly. This disease is
usually initiated by antibiotic administration or overeating
of a high calorie, low fiber diet. Clinical signs of
enterotoxemia include anorexia and dehydration, as in mucoid
enteropathy. However, rabbits with enterotoxemia usually have
watery diarrhea and tend to die acutely. Interestingly,
animals with enterotoxemia can also pass large quantities of
mucus, suggesting that mucoid enteropathy may be a syndrome
with many possible etiologies, rather than a specific disease.
Studies of the cause and treatment of mucoid enteropathy
has thus far been unrewarding. Several studies have
concentrated on potential infectious causes, especially since
the incidence of the disease occurs in "outbreaks." Greenham
(64) found that tetracycline decreased diarrhea and delayed
death, but did not prevent or cure the disease. Van
Kruiningen (238) created an oral inoculum of macerated
intestine and contents in an attempt to transmit the disease.
Only 5 out of 17 of these transfaunated rabbits became ill,
while 2 out of 17 rabbits fed a control inoculum also
developed mucoid enteropathy. Intestinal coliform
concentrations are often increased in animals with mucoid
enteropathy (64, 113, 207); however inoculation with E. coli
obtained from rabbits exhibiting diarrhea, or E. coli heat
labile enterotoxin (207) failed to induce disease. Lelkes and
Chang (114) found several differences between the cecal flora
in normal rabbits and in those with mucoid enteropathy, but
could not identify a single organism associated with the
disease. They also transfaunated cecal contents from sick
rabbits directly into the ceca of healthy fistulated rabbits
and were unable to cause increased mucus secretion.
Research on ME has generally focused on determining the
inciting cause. There could very well be more than one
etiology. Individual cases have been seen in adult rabbits
with debilitating and/or stressful conditions, such as
neoplasia, surgical stress, and septicemia (Hotchkiss,
unpublished observations). The present study was therefore
designed to focus on the direct cause of the increased mucus
secretion, the one consistent finding in all afflicted
In the past, studies of mucoid enteropathy have been
hampered by the lack of viable models of the disease. There
is now one model available: based on the hypothesis that
mucoid enteropathy is primarily due to constipation,
Sinkovics (207) was able to induce colonic mucus
hypersecretion by surgical ligation of the colon. Toofanian
and Targowski (235) found that ligation of the blind pouch of
the cecum, permitting the flow of ingesta from ileum to colon,
was just as effective, causing mucus hypersecretion in
approximately 70% of rabbits. Partial cecectomy did not
induce a mucoid enteropathy-like syndrome, while injection of
tetracycline into the ligated cecum prevented the development
of disease (235). Little work has been done with the cecal
ligation model, however. Toofanian's group found decreased
small intestinal disaccharidase activity in ligation-induced
sick rabbits (233), and also minor alterations in cecal
volatile fatty acid concentrations (234). They incubated
colonic explants with cecal contents from rabbits with
experimentally-induced mucoid enteropathy and reported goblet
cell hyperplasia (236).
In the project described herein, methods have been
developed to test this finding in a quantitative manner.
Evidence is presented to support the hypothesis that cecal
contents from rabbits with mucoid enteropathy possess mucus
secretagogue activity. Preliminary tests have been carried
out to determine the nature of the secretagogue.
The importance of doing research on this disease is
twofold: First, little is known about it, considering that it
is one of the most important causes of morbidity and mortality
in rabbit production (113). Since rabbits are used in large
numbers in biomedical research, these losses can add up to
substantial costs, especially in situations where affected
animals have been injected for antibody production with very
precious and irreplaceable antigens. Second, this disease
provides a unique opportunity to study the control of mucus
secretion. Increased mucus secretion without inflammation can
be seen in enterotoxic diseases, such as cholera and E. coli
enterotoxemia, but there is no associated goblet cell
hyperplasia (96), and there are other confounding toxic
effects, notably a secretary diarrhea. On the other hand,
goblet cell hyperplasia with mucus hypersecretion can be seen
with some diseases and parasitisms (96), but because of the
associated inflammatory changes it is difficult to examine the
mucus related aspects alone.
Although the secretagogue effects of specific chemicals
such as acetylcholine and PGE2 are well known (160, 258),
their physiological importance is not always clear. This
study is intended to demonstrate the presence of a
secretagogue that is pathologically relevant, with few effects
other than the stimulation of mucus production. It may be a
microbial product or a host secretion; in either case it could
prove to be useful for studying the processes controlling
differentiation of goblet cells, as well as production and
secretion of intestinal mucus.
In studying a disease in which the most consistent sign
is mucus hypersecretion, it is necessary to examine the role
of mucus in the normal physiological state. This review
presents information on what mucus is, what controls its
synthesis, secretion, and degradation, and what the normal
functions of gastrointestinal mucus are.
The function of mucus has been a subject of study since
ancient times. The ancient Greeks considered mucus, commonly
translated as "phlegm", to be one of the four humors, along
with blood, yellow bile, and black bile. In the second
century Galen (56) stated, "The best physicians concur in the
opinion that if a considerable amount of phlegm accumulates on
account of some bad conditions of the body, the most serious
abdominal disorders ensue, intestinal obstruction, lientery,
and tenesmus". Although it is now generally believed that
mucus is involved in lubrication rather than obstruction,
there is no question that increased mucus production occurs in
certain disease states.
Characteristics of Mucus
Initially, the term mucus was used to describe any
viscous material in the stomach and intestines. Primarily,
mucus was the jelly-like material which lined the
gastrointestinal tract. However, for a time, the material
which could be precipitated from gastric juice with suitable
agents was known as "dissolved mucin" (9). However, as it was
found that this material contained a heterogenous mixture of
substances, this terminology was dropped (59). Still, mucus
was variably described as clear and colorless, white and
cloudy, and occasionally yellow. Hollander et al. (81)
determined that these differences were due to the amount of
cells and debris dissolved in the mucus layer. They were able
to show that the irritants used to stimulate mucus secretion
resulted in desquamation of epithelial cells, associated with
Pilocarpine injected intravenously causes increased
secretion of mucus of egg-white consistency; however, an
enormous dose is required (44). In 1949, Morton and Stavraky
(153) found that injection of acetylcholine into the
mesenteric arteries caused secretion of mucus, while Janowitz
et al. (92) found that topical administration of acetylcholine
resulted in secretion of clear, cell-free mucus, facilitating
the study of "pure" mucus. It is now generally accepted
terminology that mucin is a specific glycoprotein, while mucus
is a heterogenous mixture containing mucin, protein, lipids,
nucleic acids, water, and electrolytes. Specifically, mucus
contains trefoil-type protein, kallikrein protease, lactose-
binding lectins, vitamin B, binding protein, and other
proteins which are secreted with mucin from the same granules
in the goblet cell (178). In the respiratory tract, most of
the lipid in the mucus is bound to the mucin glycoprotein
(157). Also, nucleic acids entwine other proteins with mucin,
making acidification and nuclease treatment necessary in the
purification of mucin (157). In the digestive tract, the
mucus layer also includes bacteria, food particles, and
products of digestion. In disease states, such as cystic
fibrosis, mucus also contains materials from degenerated
leukocytes, especially DNA and actin which contribute to its
viscosity, and interfere with clearance (239).
The most striking feature of mucus is its jelly-like
nature. Early studies revealed that unlike most proteins,
which show minimum viscosity at the isoelectric point, gastric
mucin shows a maximum viscosity at its isoelectric point of
4.98 (144). As rheological studies have become more
sophisticated, it has been shown that mucus is a visco-elastic
gel; and that the elastic component is greater than the
viscous component over the entire range of frequency and
strain studies with native and reconstituted mucus (4). The
elastic properties will be decreased, allowing the gel to
collapse if the mucin is treated with proteolytic enzymes or
if disulfide bonds are reduced with mercaptoethanol (4). In
addition, proper hydration is necessary for maintenance of the
gel. Hydration is controlled by small polyionic proteins, pH,
and electrolytes within the gel (242). Since mucin is a
negatively charged glycoprotein, a high concentration of a
positively-charged ion such as calcium can result in
condensation of the mucus gel, with an increase in viscosity
and lack of elasticity.
The chemical properties of mucin have been slowly
elucidated over the past century. Mucins are a heterogeneous
group of glycoproteins, with an average subunit molecular
weight of 2 x 106 (220). Carbohydrate comprises approximately
80% of this molecular weight, with the rest being protein.
The protein forms a core, with linear and branched
carbohydrate side chains covalently attached to form a "bottle
brush" structure (126). As well as heavily glycosylated
areas, there are believed to be "naked" areas on the protein
core, which are subject to cleavage by proteolytic enzymes
(4). These regions also contain cysteine residues, allowing
for polymerization via disulfide bonds (160), and are the
major antigenic determinants of mucin (131). The glycosylated
portion of the protein core contains a high proportion of
serine and threonine, which are involved in oligosaccharide
linkage via hydroxyl groups (220). In addition, there is a
high percentage of proline, which prevents a-helix formation,
allowing for increased glycosylation and flexibility.
The oligosaccharide chains of mucin are attached to the
protein core via an 0-glycosidic bond between serine or
threonine and N-acetylgalactosamine (220). These chains may
be linear or branched, and vary in length from 2 to 12
residues. Some chains will terminate in A, B, H, Lewisa, or
Lewisb antigens, reflecting the blood group genes of the
individual (41). The backbone of each chain contains N-
acetylgalactosamine, N-acetylglucosamine, and galactose, which
may comprise I or i antigens (126). The chains often
terminate in a fucose, sialic acid, or ester sulfate residue.
The glycosylation pattern depends on genetics, the region of
the gastrointestinal tract, the age of the animal, and can
vary in certain disease states, as determined by the
glycosyltransferase activity levels (228). There is also a
great deal of heterogeneity in a single preparation, even
within a single granule of a goblet cell (220).
The DNA sequences of several mucins have been determined
(116). These genes show a high degree of polymorphism due to
variable numbers of tandem repeats. The gene MUC1 encodes a
mucin-like cell surface glycoprotein which is present on many
epithelial cells, but MUC2 to MUC6 appear to encode secreted
mucins. MUC2 and MUC4 are associated with colonic mucins, and
MUC3 and MUC5 are associated with gastric mucins (116).
For several years it was debated whether or not mucin
contains a "link" glycopeptide, of approximately 118 kDa
(160). Roberton et al. (188) demonstrated that this
glycoprotein is different from other mucin subunits, in that
it contains only 50% carbohydrate, and it contains mannose,
which is generally associated with N-linked glycosylation
(220). Amino acid composition, carbohydrate composition, and
immunological studies have indicated that the "link"
glycopeptide is actually a 118 kDa fragment of fibronectin
(210). However, the cDNA for this peptide has recently been
sequenced in rats (257) and humans (256), and the 118 kDa
glycopeptide appears to represent the cysteine-rich carboxyl
terminal of a much larger mucin-like peptide. It is now
generally accepted that proteolytic cleavage of the mucin core
peptide occurs during purification, and that reduction and
alkylation allows for separation of the larger N-terminal
fragment and the 118 kDa C-terminal fragment (262).
In this chapter, there has been some generalization in
order to describe the overall functions of mucus. However, it
is important to recognize that mucin molecules are
heterogeneous, and some of the characteristics of mucus vary
along the length of the gastrointestinal tract, change during
development and maturation, and can be altered by disease
(160). In fact, a diurnal rhythm to mucus secretion has been
seen in rats and mice which were given access to food only in
the daytime: there was more cell proliferation, and more mucus
was made during the daytime (73).
The mucin produced in newborn rats is different from that
found in adults (160). Newborn mucin contains more protein
and less sugar, resulting in increased density (216). The
threonine content is increased, while serine and glutamic acid
content are decreased in newborn rats (160). Sulfation
decreases in the first 2 months of a rats life, with the major
change occurring at weaning; increased sulfation has also been
associated with mucus from "immature" cells (160).
Correspondingly, the pH within the mucus layer in the colon is
lower in suckling rats than in weanling or adult rats (193).
Aged (350 day old) rats also have increased sulfation of mucus
Newborn rat mucin has fewer fucose and N-
acetylgalactosamine residues, corresponding to altered lectin
binding (216). This likely relates to developmental changes
in glycosyl-transferase activities (12). During the suckling
period, N-acetylgalactosaminyl-transferase activity remains
constant, sialyl-transferase activity decreases just prior to
weaning, while fucosyl-transferase activity decreases
gradually throughout the suckling period. Following weaning,
the activities of all three enzymes increase dramatically. In
a parallel manner, more sialyl-transferase can be detected
immunocytochemically in the mucus of mature goblet cells, when
compared to immature goblet cells in the colonic crypt (226).
Sialylation of surface glycoproteins is replaced by
fucosylation at weaning, but sialylation of goblet cell mucus
The composition and structure of mucin varies along the
length of the gastrointestinal tract. The data on the actual
differences are sometimes contradictory, due to different
techniques, species differences, and the heterogeneous nature
of all mucins (160). In general, colonic mucin is more
aggregated, and contains more protein. It may not contain the
fragment that was formerly known as the "link" glycopeptide.
The concentration of acidic glycoproteins, containing sialic
acid and/or sulfate, relative to neutral glycoprotein
increases distally in the digestive tract. The gastric and
Brunner's gland secretions are largely neutral, lacking sialic
acid, but enriched in fucose, although in gastric mucin there
is a discrete subpopulation of highly sulfated mucin
associated with the mucous neck cells (90).
The viscosity of gastric mucus is greater than that of
duodenal mucus, while the elasticity is decreased (263). In
the stomach, there is a continuous layer of mucus, but the
thickness is variable, and tends to be thickest along the
lesser curvature (165) Mucus is present in the small
intestine, but mostly in the intervillous spaces, so that the
tips of the villi are sometimes visible. There is less mucus
in the ileum than in the duodenum and jejunum, so that most of
the villi are bare in vivo (165), although a layer 50-200g
thick can be seen in explants (67). There is no distinct
mucus layer in the cecum (165, 194). In the proximal colon
there is a discontinuous layer of mucus, varying in
composition and thickness, while in the distal colon there is
a thin, compact, continuous mucus layer present (194, 225).
Mucus from the healthy colon is more hydrophobic than ileal
mucus; however, induced colonic inflammation decreases the
Smectite is a cytoprotective agent that binds to mucus to
increase its barrier properties. It also causes gastric mucus
cells to produce mucus that contains more fucose and less
sulfate, as is seen with fasting (151) Dietary fiber has
been shown to increase the rate of turnover of intestinal
mucins, and also change their lectin-binding properties,
reflecting a change in composition (152). This is consistent
with the fact that alterations in glycosyltransferase
activities are seen with different diets (12).
Mucus can be abnormal in various disease states.
Although in cystic fibrosis the major mucus abnormalities are
likely decreased hydration (160, 242) and the presence of
degenerated leukocyte components (239), some investigators
have also found an increased fucose content (231). Celiac
disease (gluten-related enteropathy) results in a decreased
number of goblet cells, and a shift from neutral to acidic
mucin (160). Atrophic gastritis is associated with decreased
epithelial mucus, either due to decreased secretion or
increased degradation (135). In mucoid enteropathy of rabbits
there are increased numbers of immature colonic mucins with
heterogeneous lectin binding patterns, perhaps due to the
increased rate of turnover (89). Abnormal mucins are also
seen in infection of rats with Nippostrongylus brasiliensis
Goblet cell numbers are markedly reduced in patients with
ulcerative colitis (160). In addition, there is a change in
the composition, with decreased fucosylation and sulfation,
and exposure of galactose residues. This, along with deficient
0-acetylation of sialic acid (94), is suggestive of incomplete
mucin glycosylation. Intestinal malignancy is also often
associated with loss of terminal sugars, decreased 0-acylation
of sialic acid, and shortened oligosaccharide chains (160).
Consequently, the presence of small intestinal mucin antigen
(SIMA) in the colon, where there is normally masking of
antigens by sulfate, may reflect malignancy (43). SIMA can be
detected in the serum of 36% of colorectal cancer patients vs.
5% of controls, and may be useful as a screening test (179).
The abnormal glycosylation appears to be a clonal phenomenon,
reflective of somatic mutation (55). Gastric adenocarcinomas
are abnormally rich in sulfomucin and sialic acid, which may
reflect cellular dedifferentiation, or abnormalities in the
regulation of mucin biosynthesis (160).
Mucus is formed by several cells in the gastrointestinal
tract. In the stomach, mucus is secreted by surface cells,
the mucous neck cells of the funds, and the mucous cells of
the cardiac and pyloric glands. In the duodenum, the mucus-
secreting cells are the Brunner's glands. Throughout the rest
of the intestine the goblet cells are the source of mucus
(47). Goblet cells are named for their characteristic goblet
shape seen in histologic section, but electron microscopic
studies on cryofixed tissues show that these cells are
columnar in vivo (84, 242).
Duthie (38) reported that the mucus granules are first
formed near the nucleus, then are transported to the Golgi
apparatus, where stainable mucus first appears. This is
consistent with current knowledge that the peptide backbone,
which comprises less than 50% of each mucin molecule, is
formed in the endoplasmic reticulum. Sugar residues are then
added by the linkage of N-acetylgalactosamine to serine or
threonine (19). Unlike N-linked glycosylation that is
initiated in the endoplasmic reticulum, O-linked glycosylation
occurs entirely within the Golgi and condensing vacuoles (63).
The initial enzyme involved, al,3N-acetylgalactosaminyl-
transferase, is located in the Golgi apparatus (191). The
enzyme responsible for terminal sialylation, i-galactoside
u2,6-sialyltransferase, is present in the Golgi apparatus, but
also in post-Golgi apparatus structures, including the mucus
droplets and the plasma membrane (228). The mucin granules
are stored in condensed form in membrane bound vesicles within
the goblet cells until the mucus is secreted. During goblet
cell migration along the middle half of the villus, the mucin
granules are apparently renewed twice (23). The composition
of mucin in the granules is different along the crypt-villus
axis (23, 160); perhaps there is continued sialylation of
mucin within the storage granules.
The rate of mucin synthesis depends on the rate of
protein synthesis within the cell, the concentration and
positioning within the cell of glycosyltransferases, and the
presence of amino acid and sugar precursors (160). Glutamine
is known to stimulate glycoprotein synthesis as the major
metabolic substrate in the intestinal epithelium (51), but
little is known about the control of internal cellular
Mucus synthesis can be decreased by factors which
interfere with protein or oligosaccharide incorporation.
Factors which interfere with protein synthesis will also
decrease mucus production. These include malnutrition, and
metabolic inhibitors such as cycloheximide and puromycin
(160). Dietary fatty acids, non-steroidal antiinflammatory
drugs, and zinc diminish sulfation or interfere with
incorporation of individual sugars (51, 160) and may thereby
alter the composition of mucus. Cysteamine, which is used to
experimentally induce duodenal ulcers, decreases glycoprotein
production in Brunner's glands (104).
S-Adrenergic drugs, cyclic AMP, and theophylline have
been shown to increase overall glycoprotein synthesis in rat
and rabbit intestine (50, 110). Adrenergic compounds
(dopamine, epinephrine, isoproterenol, phenylephrine) have no
direct effect on mucus secretion, although they may increase
fluid secretion, making the mucus blanket appear thicker
(160). Cholinergic drugs and cholera toxin increase both
synthesis and secretion of mucus (198). Reserpine, a
hypotensive agent and carcinogen, increases glycoprotein
synthesis; the mucin that is released is more viscous than
normal, so that reserpine-treated rats have been used as a
model for cystic fibrosis (169). Epidermal growth factor has
been shown to increase glycosaminoglycan synthesis, and has a
protective effect on the gastric mucosa, but its relationship
to mucus secretion has not been studied (209).
Histamine increases mucin synthesis in the canine stomach
via H2 receptors, through activation of adenylate cyclase and
increases in cAMP (198). Gastrin increases mucus synthesis in
the corpus of the rat stomach, and this effect is not blocked
by H2 blockers (86). Conversely, the H2 blockers roxatidine
and FRG-8813, but not cimetidine or ranitidine, increase mucin
synthesis (87). Additionally, the H,K-ATPase inhibitor NC-
1300-0-3, but not omeprazole, stimulates mucin synthesis (87).
Anatomic alterations which affect the number of goblet
cells present and/or the rate of development will also change
the amount of mucus synthesized. Duodenal epithelial cells
from chick embryos show an increased number of goblet cells
when cultured in vitro compared to the number seen in vivo
(13). This increase in goblet cells is accelerated by
thyroxine, but prevented by hydrocortisone. Vitamin A
deficiency causes atrophy of goblet cells in salivary glands,
trachea, and small intestine (30). Salmonella infection in
mice decreases the number of goblet cells, apparently via
tumor necrosis factor a (8). Feeding of dietary fiber has
been associated with increased turnover of jejunal mucins
(152). Some substances which damage the intestinal
epithelium, such as methotrexate, will decrease the total
number of goblet cells, and therefore decrease mucin synthesis
(95). Following radiation damage (154), or the radiomimetic
disease caused by canine or feline parvovirus infection (96),
there is an initial increase in the number of immature goblet
cells, followed by a relative decrease, and then a second
increase as the tissue recovers. Corresponding changes in
mucus production take place following radiation exposure
(201). A decreased proportion of goblet cells is seen in
lesions of transmissible colonic murine hyperplasia caused by
Citrobacter freundii biotype 4280, and is followed by goblet
cell hyperplasia during recovery (10).
Several organisms have been associated with goblet cell
hyperplasia leading to mucus hypersecretion, including
Treponema hyodysenteriae in swine, and Ostertagia and
Oesophagostomum species in ruminants (36, 96). These
organisms are also associated with a varying degree of
inflammation, so the hyperplasia may be a response to
inflammation, repair, or the organism.
Nippostrongylus braziliensis infection in rats causes
goblet cell hyperplasia as the worms are being expelled (147).
This appears to be associated with the immune response, since
hyperplasia will be seen earlier in immunized rats, and does
not occur if the rats are treated with antihelmintics early in
the course of infection. Trichinella spiralis, Nematodirus
battus, Trichostrongylus tenuis, and Hymenolepis dimunuta have
all been associated with goblet cell hyperplasia (111, 139).
Yersinia enterocolitica has been shown to cause goblet
cell hyperplasia and increased mucin synthesis throughout the
intestinal tract of rabbits (129). There is also a great deal
of inflammation associated with this disease. The goblet cell
hyperplasia develops more rapidly and to a greater extent in
those areas of the intestine where mucosal injury is most
severe, and consequently may be associated with injury and or
inflammation. Alternatively, the hyperplasia may be part of
a repair mechanism, since it persists as the mucosa recovers
Intestinal coccidiosis caused by Eimeria species has been
associated with inflammation and goblet cell hyperplasia in
rabbits (114). Mucoid enteropathy, the subject of this study,
is characterized by copious production of intestinal mucus,
where the major histologic change reported is goblet cell
hyperplasia, and inflammation is minimal (238).
On the other hand, it is important to realize that goblet
cell hyperplasia in response to infection and/or inflammation
is not a universal phenomenon. Mucus secretion is decreased
in both small and large intestines during infection with
Isospora suis in piglets (106, 107), and numbers of colonic
goblet cells are decreased with Ehrlichia risticii in horses
(187). Clostridium difficile toxin A had no effect on goblet
cells in rabbits (118).
Mucus is secreted from intestinal goblet cells in two
ways. "Baseline", or constitutive, secretion, involving the
slow transport and secretion of glycoproteins that can be
documented by autoradiography, occurs in mucosal explants for
up to 24 hours in the absence of circulating factors or
enteric nerves. In fact, baseline secretion occurs even in a
cultured goblet cell line (173). While the bulk of the mucin
granules remain in the center of the goblet cell after
formation, there is constant formation of mucus at the
supranuclear region of the cell. This mucin migrates along
the periphery towards the apical border of the cell over the
course of 4 to 6 hours (184). Thus, some newly formed mucin
reaches the apical border and is secreted before older mucin
granules (23) A single granule will undergo exocytosis,
discharging mucus into the lumen. This process apparently
involves the cytoskeleton, as depolymerization of microtubules
with colchicine prevents this migration (219). However, once
the granule has reached the apical membrane, the actin
filaments localized there normally act as a barrier, since
depolymerization of actin filaments with cytochalasin D or
dihydrocytochalasin B results in increased baseline secretion
(167). Organelles are gradually shed from goblet cells as
mucin is secreted, so that the mean cell volume decreases
along the crypt-villus axis (184).
In the cases where it has been examined, stimulated
secretion has been shown to occur by compound exocytosis
(217). In this situation fusion of the initial mucin granule
with the plasma membrane is rapidly followed by tandem fusion
with the subjacent granules, allowing the contents of many
granules to exit through a single surface site. More recent
studies suggest that there is also fusion of granules within
the cell prior to exocytosis (84, 158) Cavitation in a
goblet cell that has just discharged its granules in this
manner can be recognized by electron, or even light,
microscopy (217). Following cholinergic stimulation,
cavitation is visible in crypt cells for 15-30 minutes; after
that it is difficult to distinguish mucin-depleted goblet
cells from epithelial cells (178). Cavitation may never be
apparent in villous goblet cells (178). In rabbit colonic
goblet cells stimulated by acetylcholine, compound exocytosis
is not inhibited by colchicine, implying that microtubules are
not necessary, since the granules are already at the site of
release (219). However, the protein synthesis inhibitor
cycloheximide, the microtubule inhibitor colchicine, and the
actin inhibitor cytochalasin B were all found to inhibit
cholera toxin stimulated secretion in rabbit ileal loops
(164). A jack-in-the-box mechanism for secretion has been
proposed (242), in which a small pore is opened over the
granule, water is allowed to enter and calcium can leave, with
the expansion of the granule contents causing release from the
It has been reported recently that mucus secretion is
stimulated independently by cAMP/protein kinase A and
increased intracellular calcium/protein kinase C mechanisms
(93), although previous reports state that cAMP does not
affect mucus secretion (161). In situ hybridization reveals
the calcium binding protein calcyclin preferentially expressed
in mucus-secreting cells, suggesting that calcyclin, in
conjunction with the p36 subunit of calpactin, is involved in
calcium-stimulated mucus secretion (232).
Several substances are known to stimulate mucus
secretion. Agents that disrupt the mucosal barrier, such as
mustard oil, alcohol (160), bile salts (117), and even
mechanical irritation (51) cause mucus release from surface
goblet cells. It has been shown that mustard oil causes
compound exocytosis (217). Florey (45) demonstrated that
secretion in response to mustard oil could be blocked with
cyanide, indicating that the process requires energy, and
secretion is not simply due to disruption of the plasma
membrane. Mechanical irritation has been shown to increase
levels of prostaglandins (26), which could be acting as
secondary messengers to stimulate secretion. Crypt goblet
cells are not affected, presumably because they do not come
into contact with the irritant. Proteolytic enzymes are known
to be mucus secretagogues in respiratory epithelium, but this
has not been shown in the intestine (22).
Stimulation of extrinsic autonomic nerves or electrical
field stimulation causes discharge of intestinal mucus in vivo
(175). This has been shown to be due to muscarinic
cholinergic innervation, as injection of pilocarpine
accelerates release of mucus by compound exocytosis. Vagotomy
or vagal stimulation does not affect mucus secretion in the
rabbit jejunum, implying that cholinergic control takes place
entirely within the enteric nervous system (65).
Acetylcholine-induced secretary events occur rapidly, and are
generally complete within 5 minutes (158, 178). It was long
thought that only crypt cells were susceptible to stimulation
(218), but it is now known that villous cells secrete a
significant amount of mucus (100), although they do not
demonstrate cavitation (178). In the rat, even crypt cells
are unresponsive to carbachol until 20 to 25 days of age,
corresponding to weaning (176). Acetylcholine (218),
pilocarpine (219), and carbachol (174) have also been shown to
stimulate mucus release in colonic explant systems.
Furthermore, it is fairly certain that the goblet cells
themselves possess muscarinic receptors, as secretion occurs
in a goblet cell line descended from a colonic adenocarcinoma
(173). The response is variable, but this may reflect the
variation of responsiveness in vivo between crypt cells and
the older surface or villous cells.
Prostaglandins, particularly PGE2, increase mucus
secretion, and perhaps synthesis (197, 198, 258). Non-
steroidal antiinflammatory drugs and indomethacin decrease
mucus synthesis and secretion, and it has been proposed that
inhibition of prostaglandin synthesis may be responsible (164,
206). Glucocorticoids also decrease mucus secretion (141).
Low doses of nicotine greatly decrease rectal prostaglandin
levels and decrease the thickness of the mucus layer, while
high doses cause an increase in mucus thickness and decrease
prostaglandin concentrations less dramatically (266). Other
arachidonic acid metabolites leukotrieness,
hydroxyeicosatetranoic acids) act as mucus secretagogues in
respiratory epithelium (171), but do not affect rabbit colon
in vitro (177).
Gastrointestinal peptide hormones have also been
suggested as potential mucus secretagogues. Neutra et al.
(161) found no increase in compound exocytosis in rabbit
intestinal explants when stimulated with caerulein,
cholecystokinin, pentagastrin, secretin, somatostatin,
substance P, or vasoactive intestinal peptide (VIP). However,
some studies indicate that secretin increases mucus secretion
in the stomach (99), and that both secretin and the related
hormone VIP weakly stimulate colonic mucus secretion in vivo
(40). Recently, VIP receptors have been found on mucus-
secreting cells in culture (108). VIP alone does not
stimulate mucus secretion from these goblet-like cells, but
both VIP and cAMP potentiate the secretagogue effects of
carbachol. This potentiating effect of VIP has also been seen
in tracheal submucosal glands (204).
The vasoactive amine histamine causes increased colonic
mucus secretion (161), but only under nonphysiologic culture
conditions (160). Histamine has also been shown to increase
PGE2 levels, and so may act indirectly (247). A greater
amount of gastric mucus was recovered after stimulation with
serotonin (141); however, there was no morphologic evidence of
increased secretion in rabbit colon following serotonin
Several inflammatory products of neutrophils,
macrophages, and mast cells have been shown to stimulate mucus
secretion from respiratory epithelium (70, 160), but they have
not been studied extensively in the intestine. There is now
increasing evidence that interleukin 1 (IL-1) increases
intestinal mucus release (27, 69) In addition, a macrophage
product, MMS-68, first isolated from the respiratory tract,
has been shown to stimulate intestinal mucus secretion (221).
Immune complexes have been shown to induce compound exocytosis
of mucin (246). Antigen challenge following oral, but not
intraperitoneal, inoculation increases mucus secretion,
suggesting involvement of mucosal immunity (109). When a
jejunal self-filling blind loop is created, causing bacterial
overgrowth, there is increased mucus secretion within the
loop, but it is decreased outside the loop (202). This
phenomenon could be the direct result of the organisms, or a
secondary response to inflammation.
Cryptosporidium parvum infection in mice resulted in an
increased amount of mucus in ileal washings (77). It is not
clear whether more mucus was actually secreted or if it was
simply dislodged more easily. If secretion was increased, it
is still not possible to tell if the effect is a direct result
of the organism, or secondary to some inflammatory mediator.
Virulent strains of Entamoeba histolytica have been shown
to increase secretion of preformed and newly synthesized mucus
glycoproteins, and also to increase mucin synthesis in rats,
in the absence of an inflammatory response (22). Intestinal
trematodes in dogs and cats have been associated with mucoid
inflammatory response, but no quantitative studies of mucus
secretion have been performed (96).
Cholera toxin and the heat labile toxin (LT) from
Escherichia coli both stimulate small intestinal mucus
secretion (160). This phenomenon can be separated from fluid
and electrolyte secretary effects, which are stimulated by
cyclic AMP (189), and blocked by tetrodotoxin (149). The
mechanisms for both increased fluid and mucus secretion are
complex. Lencer et al. (115) have shown that the B subunit of
cholera toxin binds to cloned human goblet cells in monolayer
culture, but there was no stimulation of mucus release,
suggesting that an indirect mechanism may be involved. It is
currently believed that binding of cholera toxin to
enterochromaffin cells results in increased intracellular cAMP
and secretion of serotonin, which in turn stimulates
cholinergic neurons (149). Both the fluid and mucus secretary
activities of cholera toxin are blocked by capsaicin,
supporting the involvement of the enteric nervous system
(149). However, fluid secretion is blocked by tetrodotoxin,
while mucus secretion is not, suggesting that mucus secretion
may be mediated by local effectors released by sensory
neurons. In addition, a portion of the capsaicin-sensitive
response is also atropine-sensitive, suggesting that there may
be a tetrodotoxin-insensitive interaction between cholinergic
and sensory nerve terminals occurring in the small intestine.
Although neither cyclic AMP or cyclic GMP alone affects
mucus secretion in explant systems (161), Jarry et al. (93)
found that cAMP directly stimulated both MUC2 gene expression
and mucus secretion from goblet cells in culture, in a protein
kinase A-dependent manner, which could account for the effects
of cholera toxin. This toxin has also been shown to increase
mucosal prostaglandin synthesis (183), and the effects of PGEI
and cholera toxin are qualitatively the same in in vivo rabbit
intestinal loops (164).
Mucus secretion has been shown to be increased in rats
with streptozotocin-induced diabetes, with less mucin present
in the tissue than in the luminal washings (134). Synthesis
of mucins relative to other glycoproteins is also increased.
Interestingly, the intestinal goblet cells from these diabetic
rats are no longer responsive to bethanechol or cholera toxin.
This phenomenon has not yet been confirmed, and may be an
artifact of collection, since fluid secretion is also
increased in diabetes, and may flush out the mucus present.
The antiulcer drugs zolimidine, carbenoxolone, quercetin,
and oral copper compounds have been reported to increase mucus
secretion, but the mechanisms are unknown (2, 3, 51).
Parathyroid hormone has been reported to increase gastric
mucus secretion (141). The carbonic anhydrase inhibitor,
acetazolamide, also is a mucin secretagogue (28) A new
antiinflammatory drug, SCH12223, which has protective effects
on the gastric epithelium increases gastric mucus content
(25). The amount of gastric mucus that can be aspirated in
vivo is increased after a meal (98) Feeding increases
intestinal mucus secretion, even when the duodenum has been
transplanted to a subcutaneous location and separated from its
nerve supply (47). Mucus secretion is increased during the
daytime in rats that are allowed food only during the daylight
The synthetic opioid anti-diarrheal, loperamide, has been
shown to slightly decrease baseline mucus secretion, and
drastically reduce secretion stimulated by PGE2 or deoxycholic
acid in the rat colon (121). This may be related to neural
effects or calcium channel blocking activity (186).
Cysteamine, which is used to induce duodenal ulcers
experimentally, also decreases mucus secretion (104).
Interferon-y does not affect baseline mucin synthesis or
secretion, but inhibits secretion stimulated by both cAMP and
calcium ionophores, apparently at the exocytotic step (93).
Little mucus is excreted in the feces. It was found that
crude mucus undergoes a spontaneous loss of viscosity when
incubated overnight at body temperature (72, 91). This
phenomenon was accelerated by the addition of certain
proteases; different investigators described different
sensitivities. It has been shown that pepsin, but not HC1,
can dissolve the gastric mucus layer, and cause epithelial
damage (126). That some degradation occurs in vivo is
supported by the fact that secreted mucin has a lower
molecular weight than stored mucin (220).
Although proteolytic enzymes can degrade mucin into
smaller glycosylated fragments, resulting in collapse of the
mucus gel, and loss of elasticity (4), they cannot completely
digest mucopolysaccharides (126). Both a- and 1-glycosidases
are required for the removal of sugar residues from mucin
(126), and it has been found in humans that these enzymes are
primarily derived from Bifidobacterium and Ruminococcus
species (82) in the colon.
Gastrointestinal mucin digestion is markedly decreased in
germ-free rodents, due to a lack of glycosidases capable of
removing the oligosaccharide sidechains from the peptide
backbone (120, 126). This results in an increased mass of
mucus in the cecum, increased mucus excretion in the feces,
and retention of mucus within a thickened cecal wall. When
normal enteric bacterial flora are administered to germ-free
rats, excretion of glycoproteins increases for 2-3 days, then
drops to the level of conventional animals, confirming that
bacterial enzymes play a major role in mucin digestion (120).
As mentioned previously, pepsin partially degrades
gastric mucin, resulting in solubilization, and dissolution of
the mucus layer. It has been found that patients with peptic
ulcer disease have a higher ratio of pepsin 1 to pepsin 3 than
normal individuals, and pepsin 1 is much more efficient at
dissolving mucus. Consequently, ulceration may be secondary
to the breakdown of the protective mucus (126).
It has been demonstrated in vivo that Helicobacter pylori
infection decreases the thickness of the gastric mucus layer
in humans (166, 196). This has been related to the production
of protease, lipase, phospholipase, and glycosulfatase by the
organisms which impair the protective mucus layer, and may
promote mucosal injury. Sulglycotide, a modified sulfomucin
gastroprotective agent, acts by inhibition of these enzymes,
along with aggregation of the organisms (156, 180). In
addition, Helicobacter produces urease, and the high levels of
ammonia and bicarbonate produced may impair the protection
afforded by mucus (35).
Vibrio cholera contains a virulence factor which
consists of a "mucinase complex" (168). Since mucus can bind
and inhibit cholera toxin, the destruction of mucus by a
metalloproteinase allows the toxin to retain activity, as well
as permitting better access to the epithelial cells (32). In
addition, a neuraminidase may increase the amount of GMl
ganglioside available in the membranes for toxin binding.
Postulated Functions of Mucus
In 1800, Glover postulated that the mucus secretion of
the intestinal tract was involved in lubrication (79). Florey
(46) saw that particles were removed from the intestine by
entrapment in mucus, which was then pushed downstream by the
intestinal villi. Nondigestible solids that are emptied from
the stomach are entrapped in mucus plugs (66). Mucus has been
experimentally shown to aid in ciliary propulsion of objects
in tubes (255). The need for lubrication varies with the
segment of the intestinal tract discussed. There is little
mucus secretion in the small intestine where the contents are
fluid, and there is little need for lubrication (47).
Conversely, there are many goblet cells within the colon where
lubrication is clearly needed for the passage of solid feces
(47). More recent studies have shown an incomplete mucus
layer in the proximal colon, where the contents still have a
high water content, but a thick, compact layer in the distal
colon, in order to facilitate the passage of feces (194, 225).
In the same thesis, Glover proposed that mucus "must
likewise defend the internal surface of the stomach and
intestines, from the action of the gastric juice, and from the
acritude of bile when regurgitated" (79). Several theories
were put forth for the mechanism of protection. The simplest
was that of a diffusion barrier, which will be discussed in a
later section. It was believed for a period of time that
mucus had direct acid-neutralizing capacity (79). This was
first discussed by Pavlov, and studied in detail by Hollander
(79). He found a definite buffering capacity in the mucous
secretions from Heidenhain pouches of dogs. However, it was
shown conclusively by Heatley that mucin itself has minimal
buffering power against hydrochloric acid; rather the
buffering activity is due to the content of bicarbonate (47).
On the other hand, the mucus is responsible for keeping that
bicarbonate next to the mucosal surface, as will be discussed
later in the section on creation of a microenvironment.
In 1855, Claude Bernard postulated that mucus has a
specific chemical or adsorbent action (79). At the beginning
of the twentieth century, the presence of a specific
antipepsin in mucus was reported (79). However, since 1914
there has been no further evidence of any enzymatic activity
of mucus. Bucher (17) showed adsorption of pepsin by the
mucus, and Zaus and Foskick (264) and Bradley and Hodges (15,
16) also documented antipeptic effects, however Heatley (74)
found that purified mucin does not directly inhibit peptic
Hydrogen peroxide is rapidly degraded in porcine gastric
mucus in vitro (34). Although purified mucin is subject to
attack by reactive oxygen intermediates, it has been shown
that the lipids associated with native mucus have a protective
effect against these radicals (62). Consequently, one
function of mucus may be to protect the mucosa from attack by
reactive oxygen intermediates released by the host when
killing bacteria or as a response to toxins. However, mucus
does not affect the cytotoxic activity of Clostridium
difficile toxin A in rabbits (118).
It has been shown that mucus plays a major role in
"adaptive cytoprotection" (21). Following a mild epithelial
injury with oleic acid, the thickness of the mucus layer is
increased. This prevents injury from a second exposure by
delaying the passage of the irritant through the mucus to the
epithelial cells. Increased mucus secretion stimulated by the
antiulcer drugs carbenoxolone and quercetin protects gastric
epithelium from damage by ethanol, 0.6N HC1, or 30% NaCI (2,
3). When acidified ethanol is used for challenge to negate
antacid effects, the H2-blocker FRG-8813 still offers
significant cytoprotection through increased mucus production
and secretion (85).
Protection from Infection
Cramer (30) found that vitamin A deficient rats are more
susceptible to bacterial infections, and related this to the
decreased production of intestinal mucosubstance, secondary to
decreased numbers of goblet cells. This suggests that mucus
can act as a barrier to bacteria. Goldsworthy and Florey (61)
then demonstrated that intestinal mucus contains lysozyme, a
non-specific antibacterial enzyme. Specific immunity in the
intestinal tract is dependent on secretary IgA. sIgA alone
does not prevent bacterial interaction with intestinal
epithelium (245), but rather appears to bind to mucin
glycoproteins through hydrogen or disulfide bonds (140),
allowing for aggregation and neutralization of bacteria.
Because mucus comprises a physical barrier, bacteria have
developed mechanisms to allow association with the underlying
epithelium. One of these mechanisms is motility. Flagella
allow motility of bacteria in aqueous solutions, but are
ineffective in viscous environments. However, Campylobacter
jejuni, although flagellated, acts in high viscosity solution
like a spirochete, relying on endocellular organelles for
locomotion, and shows an increase in motility related to
increased viscosity (42). Treponema hyodysenteriae, is
similarly highly motile in intestinal mucus (101). This may
provide a selective advantage to helical bacteria, such as
Campylobacter, Treponema, Vibrio, and Helicobacter species, in
penetrating mucus and colonizing the intestinal epithelium.
A lipopolysaccharide-deficient mutant of Salmonella
typhimurium is able to colonize mouse large intestine when
given alone, in combination with E. coli, or with low
concentrations of wild type Salmonella (162). However, if
high dose of wild type Salmonella is given concurrently, or if
both strains are allowed to multiply within the host for 8
days, the lipopolysaccharide-deficient mutant is eliminated.
This appears to be related to the fact that the mutant adhered
to cecal mucus far better, but penetrated mucus less well than
the wild-type, or even nonflagellated or nonchemotactic
transductants (137). This also supports the idea that
flagellar motility is not important in mobility within mucus.
The other major mechanism bacteria use to remain in the
intestinal tract is adhesion. Piliated Escherichia coli are
able to bind to sugar residues in the glycoproteins and
glycolipids of both epithelial cells and mucus. In
enterotoxigenic E. coli strains, K99 fimbriae bind to
galactose (155) and sialic acid residues (119), K88ab fimbriae
bind D-galactosamine residues (143), and the terminal subunit
of F17 fimbriae binds to undetermined carbohydrate sidechains
which are also present in cow plasma glycoproteins and hen egg
white (195). The F-18 colicin which is found in nonpathogenic
E. coli binds to mannose (243). Although E. coli lacking F-18
are able to colonize streptomycin-treated mouse large
intestine when given alone, they cannot compete with the
fimbriated bacteria in establishing infection when given
Both piliated and non-piliated strains of
enteropathogenic E. coli adhere to mucus (259). Denaturation,
trypsinization, or removal of carbohydrate from the mucin all
decrease binding (248). More mucin bound to both
enteropathogenic and enterotoxigenic E. coli strains at pH 5.7
than at pH 7.4 (248). In vitro, the presence of mucus
competitively inhibits binding of E. coli to colonic
epithelium (125), suggesting a protective effect. On the
other hand, adherence to mucus may allow for the initial
colonization by bacteria. Hydrophobicity of the mannose-
resistant AF/R1 pilus of the enteropathogenic E. coli strain
RDEC plays a major role in bacterial binding to mucus and
membranes (37), and RDEC preferentially colonizes colons of
weanling rabbits, where mucus is more hydrophobic than in
sucklings or in the ileum (124). Additionally, more mucus
binding is seen in weanlings than in adolescents (248), which
may reflect the increased susceptibility of weanlings to
Virulent strains of Yersinia enterocolitica are able to
colonize the intestine because the yadA gene on the virulence
plasmid codes for a high molecular weight outer membrane
protein which allows binding to mucin (170). However,
preincubation with mucus decreases the adherence of the
organism to brush border membranes (130). It appears that
coating with mucus changes the bacterial surface from
hydrophobic to hydrophilic, decreasing the interaction of the
organism with the epithelial surface (170). However, strains
containing the virulence plasmid are able to degrade mucin to
a greater extent than non-virulent strains (132), and thus
overcome the protection afforded by mucus.
Helicobacter pylori has also been shown to bind to mucin,
and the binding is decreased after the removal of sialic acid
(237). Electron microscopy shows that most strains of Vibrio
cholerae bind preferentially to mucus, rather than to the
epithelial surface (260, 261). Guinea pig mucus inhibits
invasion of epithelial cells by Shigella flexneri, but monkey
mucus does not (245).
Intestinal mucins inhibit replication of rotavirus in
vitro (24, 262). The rotavirus vp4 protein, which also
mediates binding to cells, is involved in the binding (24);
however it is not clear if sialic acid residues or short
oligosaccharide chains act as the mucin receptor (262).
Mucus also plays a significant role in parasitic
infections. Frick and Ackert (54) found that duodenal mucus
of adult chickens inhibits the growth of Ascaridia galli more
than that of young birds, and that this may play a role in age
resistance. A nematode which has been studied more
extensively is Nippostrongylus brasiliensis in rats.
Following a transient decrease, an increase in the number of
goblet cells is associated with an immune-mediated expulsion
of this worm. If the number of goblet cells is decreased
secondary to a protein deficient diet, the efficiency of
expulsion is decreased (249). Secretion of mucus is necessary
for the expulsion event: if mustard oil or a combination of
cysteine and papain is administered to immune rats to cause
secretion of stored mucin 1.5 hours prior to challenge, then
worm expulsion is inhibited (146). The role of prostaglandins
in this process is unclear. While some investigators have
found that administration of prostaglandins enhances
expulsion, others have been unable to confirm this (28). The
vehicles used (chloroform or alcohol), or the timing of
administration may be responsible for the discrepancy.
Some worms are entrapped by mucus prior to expulsion
(145). In addition, in a phenomenon called "immune
exclusion", worms are unable to penetrate the mucus of
immunized animals, while in naive animals, worms are able to
burrow through the mucus to the intervillous space (145).
Similarly, it has been shown that mucus from infected sheep
inhibits the motility of Trichostrongylus colubriformis in
agar gel (103) Nippostrongylus is able to move through
viscous gels by forming a tight corkscrew (111); the presence
of coating antibodies may prevent this motion, resulting in
immune exclusion. It has also been shown that the worms
ingest mucin, and morphologic damage of the adult worm gut is
associated with the development of immunity (145).
Trichinella spiralis is expelled from immunized rats in
a very similar manner, although Trichinellae enter the
epithelium, while Nippostrongylus remain between villi (145).
It has been shown that antibody or complement coating of
Trichinella larvae allows entrapment in mucus (18, 145). This
may be important in the expulsion of worms from the organisms,
although the injection of antibody allows for the expulsion of
worms from the infected epithelium, prior to association with
The expulsion of both Trichinella and Nippostrongylus is
an immune-mediated event. Transfer of thoracic duct
lymphocytes from immunized rats results in accelerated
expulsion of Nippostrongylus associated with early goblet cell
hyperplasia (147), and this was later shown to be T cell-
dependent (145). However, transfer of hyperimmune serum will
also result in rapid expulsion (145). Furthermore, at least
two steps are involved in worm expulsion (88). The first
involves T-cell dependent "damage" to the worms. These
damaged worms can then stimulate alterations in the terminal
sugar residues of mucus resulting in the selective expulsion
of damaged or healthy worms, even from athymic rats (88).
There is also synergistic interaction between immune serum and
thoracic duct lymphocytes in the rapid expulsion of
Trichinella (1). In both worms there are increased
concentrations of leukotrienes and rat mast cell protease II
(145, 150). Corticosteroids, which inhibit the immune
response in many ways and also decrease mucus secretion (141),
delay expulsion of Nippostrongylus (145) as well as the rat
tapeworm Hymenolepis diminuta (139). Reserpine also alters
mucus secretion (169) and delays expulsion of Nippostrongylus
On the other hand, mucus can apparently increase
pathogenicity of other parasites. In particular, it promotes
survival of Giardia lamblia in many ways. For example,
Giardia can be killed by the lipolytic products of milk, but
this killing is inhibited by the presence of intestinal mucus
(265). Mucus provides a nutrient source for the parasite
(57), and allows for its enhanced adherence to the intestinal
epithelium (145). This lectin-like adhesion to simple sugar
residues within mucus, especially sialic acid, is a
characteristic of several protozoa, including Entamoeba
histolytica (145) and Tritrichomonas mobilensis (33). The
yeast Candida albicans has also been shown to associate with
mucus, but the ability to colonize the intestinal mucosa is
dependent on the absence of normal flora (102).
Nutrient for Flora
Evidence that mucus may aid the growth of bacteria was
demonstrated by Smith et al. (213, 214), who found increased
virulence of intraperitoneally injected bacteria
(Staphylococcus aureus and Streptococcus species) when mucus
was administered with the organism. This effect was partially
due to sequestration of the bacteria from the immune system.
However, the chemical components of the mucus itself increased
the growth of the organisms, presumably by serving as a
nutrient source. It has been shown in vitro that jejunal
mucus stimulates the growth of Giardia lamblia (57). Mucin
can also enhance growth of both virulent and avirulent
Yersinia enterocolitica, as well as serve as a nutrient source
for E. coli, Salmonella typhimurium, Clostridium perfringens,
Bacteroides sp, Shigella flexneri, Rumenococcus, and
Bifidobacterium (8, 132).
By comparing the cecal mucus of conventional and germfree
rats, Lindstedt et al. (120) were able to show that normal
flora degrade a significant amount of mucus. There was less
mucus in the cecum of conventional animals, and a higher
proportion of what was there was of lower molecular weight,
indicating partial degradation. Other investigators have
confirmed these findings (52). However in the colon,
Szentkuti et al. (225) found a thinner mucus layer in germfree
rats, associated with decreased mucosal thickness, and
decreased numbers of goblet cells.
As discussed previously, Rumenococcus and Bifidobacterium
species possess glycosidases that are responsible for the
degradation of mucin in humans. The monosaccharides released
from mucin by these glycosidases support growth of Bacteroides
and other fecal bacteria that lack such glycosidases (82).
However, a strain of Bacteroides vulgatus isolated from
patients with Crohn's disease is capable of degrading mucus
glycoproteins (192). The enzymes produced by normal flora may
also directly attack pathogenic bacteria (160); normal flora
certainly inhibit colonization by pathogens, since antibiotics
increase infection with pathogenic organisms (245).
Much of the cytoprotective action of the mucus layer has
been attributed to the barrier properties of mucus (79).
However, when this property was experimentally examined,
Heatley illustrated that mucus is not a barrier to the passage
of solutes, such as H' ions and pepsin. Furthermore, when
studies were performed to determine the thickness of the
unstirredd water layer" in the intestine, the calculated value
for several solutes was similar, so it was proposed that mucus
merely acts as a support for the unstirred water layer (215,
250). However, more recent studies show that mucus does
retard diffusion of many molecules. For instance, it slows
the passage of H* ions relative to their rate of diffusion in
water (172, 253), and Lucas (122) demonstrated that this
retardation is dependent on the concentration of mucus. The
mobilities of sodium, potassium, and chloride are also greatly
reduced in mucus (68).
Mucus forms a polyanionic gel, and thus acts as an ion
exchange resin (48) In general, chloride is excluded,
whereas there is a high affinity for calcium and potassium
(67, 199). However, although the transport of chloride (and
other ions) is retarded, it is not prevented (68).
Sequestration of potassium by mucus and cell surface
glycoproteins may be important in recycling of potassium by
the Na/K-ATPase. Calcium may be important in the
condensation of mucus into compact granules prior to
Mucus retards diffusion of molecules as well as ions.
Smith (212) found a diffusion coefficient for butyrate in
mucus to be 50-60% of that in water. Desai et al. (34)
determined the diffusion coefficients for a wide variety of
molecules in mucus and concluded that no consistent effect of
molecular weight was evident with regard to barrier properties
for the weight range tested (34-660 Da). Using cultured
goblet cells, it has been shown that the overlying mucus layer
is a significant barrier to the passive absorption of the
lipophilic and uncharged drug testosterone (97). Mucus has
been shown to trap iron, but it is not clear if this aids
(181), or prevents excess, (252) absorption. This trapping
may make iron more available to bacteria (51).
Since mucus is a polyelectrolyte gel, it behaves as a
Donnan system (67). Consequently, the hydration of mucus is
dependent on the ionic concentrations in the bathing solution.
As ionic strength is decreased, hydration is increased, and
noncovalent interactions between mucin subunits are fewer,
decreasing viscosity. Conversely, at high ionic strength,
anionic charges within the gel will be shielded, and hydration
and volume decreased. It has been postulated that lack of
functional chloride channels resulting in decreased chloride
movement in cystic fibrosis could alter ionic composition
resulting in abnormal mucus secretions in intestinal,
pancreatic, and respiratory tissues (138, 242). Mucus from
cystic fibrosis patients is hyperpermeable to small ions and
water, associated with an increased calcium content (58).
Furthermore, glucose absorption was enhanced in afflicted
patients, coinciding with a decreased thickness of the
calculated unstirred water layer (53).
Mucus may even have a "waterproofing" function (159). If
the external mucus layer is removed from an eel, the weight of
the animal will increase when the eel is placed in distilled
water. This property has not been examined in the intestine,
and would be difficult to address in light of the complex
absorptive and secretary processes involved in digestion and
absorption. However, it is interesting that Westergaard and
Dietschy (250) saw swelling of intestinal villi inversely
correlated the thickness of the unstirred water layer.
Creation of Microenvironment
A microenvironment occurs when the concentration of
solutes next to a membrane is different from that of the bulk
phase on either side of the membrane, as has been shown to
occur in mucus (67) In the gastrointestinal tract, the
presence of a microenvironment is important in two basic
areas. In the stomach, it is necessary to prevent direct
contact between the highly acidic bulk contents and the
mucosal surface. When Heatley (75) found that mucus did not
act as a barrier for hydrogen ions, he proposed a pH gradient
within the mucus layer. In his dynamic model, mucus and
bicarbonate are secreted at the mucosal surface, and then
migrate outward. As the mucus progresses into the lumen, it
becomes more hydrated and accumulates more hydrogen ions. The
viscosity thus decreases, until the outermost layer is shed
into the lumen. Therefore, some of the hydrogen ions will be
brought back into the lumen with the dissolving mucus, while
others will be neutralized by the bicarbonate within.
This "mucus-bicarbonate barrier" model has stood the test
of time very well (5, 31, 148, 205). With the development of
microelectrode techniques, pH gradients have been demonstrated
in rat (190), rabbit (254), and human (182) gastric mucus in
vivo. The pH gradient can be decreased with the addition of
compounds that dissolve mucus, such as N-acetylcysteine (190).
Mucus is responsible for the creation, as well as the
support, of a microenvironment in the intestine. Hogben et
al. (78) proposed that an acidic microenvironment could
influence the absorption of drugs from the intestinal tract.
Acidic drugs are absorbed more rapidly than would be expected
at neutral pH; however at acidic pH these acids would be
undissociated, and would readily cross the epithelial membrane
by nonionic diffusion. He calculated that the pH at the rat
jejunal surface must be 5.3 to account for measured drug
absorption rates. Surface pH measurements with
microelectrodes have revealed an acidic microenvironment, but
with pH 6-7 (123, 185, 203).
Similarly, a slightly acidic microenvironment is present
in the colonic unstirred layer (185). This would allow for
the passive absorption of butyrate and other short chain fatty
acids in the undissociated form (7). Although there is
evidence for a bicarbonate gradient-dependent, carrier-
mediated anion exchange process for butyrate in the colon,
this cannot account for the rate of absorption measured
experimentally, nor could passive diffusion of the dissociated
ionized form of butyrate (136). In the rodent cecum, the
microenvironment is more basic than the luminal contents, and
the pH may have an effect on the virulence of Entamoeba
histolytica (112). This organism is subject to killing by
ammonia, and species with a higher cecal mucosal pH, such as
the rat, are less susceptible to infection than those with a
lower pH, such as the gerbil.
Techniques Used to Study Mucus
Intestinal loops. For in vivo experimentation of the
gastrointestinal tract, it is common to isolate the part of
the tract under study to minimize effects of the rest of the
system, and to allow for easy access. For the study of
gastric secretion, the first widely used technique was
formation of the Heidenhain pouch (76). Florey (44, 45, 46,
47) used a variety of intestinal pouches, blind loops, and
transplanted segments to examine movement of mucus within the
intestine, and the effects of feeding, neural stimulation, and
various chemicals on mucus secretion.
Explants. Mucosal samples (161) and full thickness
intestinal explants (50) have been used to measure
secretagogue activity in vitro. These tissue sections can be
maintained in culture medium for several hours, and mucus
production and secretion can be measured by several methods.
This technique allows for concurrent testing of several
different chemicals on tissues from a single animal. It has
even been used to demonstrate goblet cell secretion following
electrical field stimulation, due to intrinsic nerve
HT-29 cloned goblet cells. Until recently, it was
impossible to work with isolated goblet cells, because these
cells lose their polarity and ability to secrete when
separated from the mucosal epithelium (160). However, cell
culture systems are now available. The HT29 colon
adenocarcinoma cell line is undifferentiated under standard
culture conditions. However, substitution of glucose with
galactose (83), or long term treatment with butyrate (108)
results in differentiation of cells. Individual cells can be
selected for cloning. Several sublines, including HT29-18N2
and C1.16E, show characteristics of goblet cells, including
baseline mucus secretion, and the ability to respond to
cholinergic stimulation (108, 173).
Quantitation of Mucus
Biochemical techniques. Biochemical measurement of total
protein-bound hexose (258), acid-precipitable protein (50), or
glycoprotein (127) has been used to estimate mucus
concentration; however mucins are not the only glycoproteins
present in the intestine. Mucin can be separated from many
other glycoproteins by gel filtration on a Sepharose-4B
column, but many proteins remain associated with the sticky
Radioactivity incorporation. Glycoprotein synthesis has
been measured by determining the rate of "C- or 3H-glucosamine
incorporation into intestinal slices (50), and secretion is
measured by counting precipitable radioactivity released into
the media (49). By fractionating the cell sap on a Sepharose-
4B column, and collecting the void volume, it is possible to
estimate the amount of mucin synthesized. It has also been
shown that labelled butyrate and acetate are incorporated into
mucus glycoproteins (29). However, absorptive cells
synthesize and release glycoproteins at a faster rate than
goblet cells (160).
Morphometry. Goblet cell hyperplasia, which is
associated with increased mucus secretion, can be quantitated
morphometrically (147). Compound exocytosis can be
visualized, and the percentage of cavitated goblet cells is
reflective of the amount of mucus secreted (217).
Immunologic techniques. Standard radioimmunoassay (131)
and enzyme-linked immunosorbant assay (ELISA) techniques (188)
have been developed, using antibodies to mucin. Polyclonal
antibodies react predominantly to the "naked" regions of
mucin, and therefore do not measure degraded mucin (131).
Monoclonal antibodies have been made specific for both the
naked peptide backbone, and for intact mucin (210).
Antibodies have also been made which are specific for the
fragment known as the "link" glycopeptide (188).
Lectin binding techniques. An enzyme-linked lectin assay
(ELLA), very similar to a standard ELISA, has been developed
(27, 69). The lectin soybean agglutinin preferentially binds
to N-acetylgalactosamine residues, which are abundant in mucin
but less common in other glycoproteins.
Secretion of Mucus
Light microscopy. A monoclonal antibody has been
developed which specifically labels goblet cells of the human
colon, appendix, and small intestine (240). This may be
useful in the diagnosis of disease states, in which the
quantity or quality of mucus is altered. Histologic
examination of the thickness of the mucus coating can be used
to evaluate changes in secretion (225).
Morphometry has also been used to quantitate secretion by
comparing the amount of stained mucin in goblet cells before
and after addition of a secretagogue (100). Autoradiography
has been used to measure the rate of secretion (73).
Cavitation of goblet cells indicates that compound exocytosis
has taken place, and has been used to measure secretagogue
activity (176, 177, 217). Exocytosis has been directly
observed by video-enhanced light microscopy (230).
Electron microscopy. The steps of the secretary process
can best be followed by electron microscopy (217). Since
conventional fixation techniques can fragment the limiting
membranes, cryofixation may allow for better evaluation of the
mechanism of secretion (84) The thickness and composition of
the mucus layer can also be examined by this technique (165).
Composition of Mucus
Biochemistry. Standard biochemical techniques have been
used to determine the amino acid and sugar composition of
mucus (160). Glycosyl-transferase activities can be measured
by conventional biochemical assays (12, 226), providing clues
to changes in carbohydrate composition. Methods for measuring
the adhesiveness, plasticity, viscoelasticity, and
spinnability of mucus microsamples are also now available
Histochemistry. Differential staining techniques can be
used to distinguish mucin characteristics. When sections are
stained with combined alcian blue and periodic acid Schiff
reagent, acidic mucins will appear blue, and neutral mucins
will appear red (225). High iron diamine will stain sulfated
sialomucins (266). Positive staining with mild periodic acid
Schiff (mPAS) indicates deficiency of 0-acetylation of sialic
acid, as is seen in ulcerative colitis (94) Mucins of
different chemical composition can be distinguished by
differential lectin binding. Therefore, lectin histochemistry
can be used to identify abnormally glycosylated mucins in
goblet cells, which may be associated with disease (89, 241),
developmental changes (227), or diet (152). Finally, specific
antibodies can be used to distinguish small intestinal from
large intestinal mucins by indirect immunoperoxidase staining
Diffusion Through Mucus
Mucus can be immobilized on a filter, or between two
filters, and the rate of diffusion through this compared to
the rate of diffusion through the filter alone (75). The
mathematical modeling of diffusion through mucus in this type
of apparatus has been described (172). Diffusion through
mucus overlying HT29 monolayers can also be measured (97).
In vivo colonization assays, where the number of colony-
forming units of bacteria recovered in the feces after
experimental infection is measured over time, reflect a
combination of virulence factors (162). Microbial adherence
can be evaluated by coating plates with mucus or brush border
membranes, and measuring the percentage of radiolabelled
bacteria which will then adhere (130). Competition for
binding can also be measured with this assay, by adding the
competitor along with the organism. Adhesion of organisms to
cells in culture can also be evaluated microscopically (33).
Penetration of organisms through mucus can be evaluated by the
passage of radiolabelled bacteria through a mucus gel (170),
or by direct observation of the organism (42, 111).
The idea that mucus is necessary for lubrication to aid
in the passage of feces has been mostly intuitive, with little
direct experimental support. However, there is one disease
state which illustrates the need for such lubrication quite
clearly. In cystic fibrosis, where mucus is abnormally
viscous and lacks elasticity, it is common for infants to
develop meconium ileus, and for adults to develop meconium
ileus equivalent (160). As indirect evidence, the most
compact, continuous mucus layer is in the distal colon, where
there is greatest need for lubrication for the passage of
formed feces (194). In addition, the fact that nondigestible
solids are coated in mucus when they are expelled from the
stomach suggests that mucus aids in lubrication to facilitate
their passage through the pyloric sphincter and expel them
from the body (66).
Mucin itself has no direct buffering or antienzymatic
activities. However, its presence is necessary to protect
underlying cells. Its removal may result in increased gastric
or duodenal ulceration (104). Chronic inflammatory bowel
disease is associated with decreased mucus production, which
may help to perpetuate the disease (160). This protection is
partly due to the role of mucus as a diffusion barrier, and
the creation of a microenvironment, which will be discussed
below. In addition, it appears that mucus is able to detoxify
reactive oxygen intermediates (34, 62). Irritants of many
types stimulate the accelerated secretion of mucus; this may
be a protective mechanism to keep such irritants away from the
intestinal mucosa (21, 47).
Protection from Infection
The interactions of mucus and pathogenic organisms are
complex. It is perhaps easier to determine the protective
effects of mucus by looking at mechanisms that pathogens have
developed to overcome these effects. Mucus is a slippery
substance, and when the motility of the intestine is taken
into account, it is clear that attachment to mucus is
necessary if organisms are to avoid being flushed out. The
viscosity of mucus is also a barrier, and motility based on
internal structures rather than flagellae helps microorganisms
negotiate this barrier (42). That organisms have developed
mucinase enzymes indicates that mucus is a barrier that must
be negotiated (168).
Mucus also contains antibacterial substances, in
particular lysozyme (61) and secretary IgA (140). Mucin bound
to IgA allows for entrapment of parasites such as
Nippostrongylus, so that the parasite is immobilized and swept
out of the body (145). The goblet cell hyperplasia seen with
many parasitic infections emphasizes the importance of mucus
in their expulsion. Mucus coating can also prevent attachment
of pathogens to the intestinal epithelium by competition (125)
or by changing bacterial surface properties (170).
Pathogenic organisms have also been able to use mucus to
increase virulence. Within mucus, bacteria are "hidden" from
some of the defense mechanisms of the host (265) Also, mucus
can provide nutrients to these organisms, as discussed below.
Nutrient for Flora
Although the experimental evidence is incomplete, it
seems reasonable that degraded mucus provides nutrients for
bacterial flora, especially in times of fasting. The trapping
of iron by mucus (181, 252) allows the flora increased access
to this mineral which is so tightly controlled in the rest of
the body. It is important to remember that the normal flora
also provide a major defense against pathogenic bacteria (160,
245), so that in providing nutrients for flora, mucus is
actually protecting the host from infection.
The gastrointestinal mucus layer has been shown to delay
diffusion of a wide variety of molecules. In the intestine,
this layer is between 50 and 500 jim thick, depending on
species and site evaluated (34, 67). In general, the measured
unstirred water layer is slightly thicker. This may be due to
an unstirred layer overlying the mucus, but is more likely due
to the decreased diffusion of solutes in mucus leading to
erroneously large values for the unstirred water layer (215).
Consequently the barrier to nutrient absorption is formidable,
although the calculated maximum allowable water barrier which
would allow physiological absorption of glucose and lipid is
40 pm (223). Furthermore, the mucus layer is not static;
there is constant renewal at the mucosal surface and
degradation at the luminal surface, resulting in a net flow of
mucus away from the mucosa.
It would seem that this unstirred mucus layer would make
absorption of nutrients impossible. But on further
examination, the term unstirredd" is also inappropriate. The
mixing caused by segmental and peristaltic contractions of the
intestine, with three dimensional shear forces, as well as
expansion and contraction of the mucus layer itself, is not
comparable to mixing by a stir bar in a beaker. Furthermore,
the intestinal villi move in and out of the mucus layer,
cleansing themselves of particulate matter, and moving the
entire mucus layer aborad (46). Villi bare of mucin are
present in the small intestine, particularly the ileum, in
vivo (165). The villi may also cause mixing within the mucus
layer, as speculated by Strocchi and Levitt (223).
The fluxes of water in and out of the intestine must also
be considered. Large volumes (approximately 1 liter/20cm/hr
in a human, 53) are secreted from the crypts and absorbed from
the villi. This could aid in absorption by solvent drag
effects, as well as contribute to mixing within the mucus
layer. And finally, the mucus may actually "trap" certain
molecules, such that the concentration within the mucus can
exceed the luminal concentration, leading to increased
Creation of Microenvironment
The pH difference between the lumen and the mucosal
surface is most important in the stomach, where the presence
of a "mucus-bicarbonate" barrier is generally accepted. By
delaying the diffusion of H' toward the epithelial cells, and
the diffusion of bicarbonate away from them, mucus plays a
vital role. The microenvironment may also play a role in the
selective absorption of drugs and nutrients from the
intestinal tract. In particular, a slightly acidic
microenvironment would aid in colonic absorption of short
chain fatty acids (7).
The differences in the mucus layer throughout the
gastrointestinal tract reflect the differing functions
required. In the stomach, the major function is
cytoprotection, by creation of a diffusion barrier and a less
acidic microenvironment. For this, the mucus is neutral
rather than acidic, and is more viscous. In the small
intestine, mucus must act as a diffusion barrier to
destructive enzymes and pathogenic organisms, yet provide a
microenvironment to allow for absorption of nutrients. The
fact that absorption occurs at the tips of the villi, which
stick up through the mucus layer, may be a physiologic
necessity, rather than a random fact for students to memorize.
In the proximal colon, the loose, discontinuous mucous
layer provides for the nutrition of friendly flora, and
creates a microenvironment which allows for absorption of the
short chain fatty acids produced by these bacteria. Finally,
in the distal colon, a thin, compact mucus layer provides
lubrication for the passage of feces. All in all, mucus aids
the host in a number of ways, and its importance in the
maintenance of good health is often overlooked.
The overall purpose of this project was to determine if
there is a mucus secretagogue in the cecal contents of rabbits
with mucoid enteropathy, as proposed by Toofanian and
Targowski (236). Therefore pilot studies were performed to:
1) explore the presence of such a secretagogue in the cecal
contents of a rabbit with naturally-occurring mucoid
enteropathy (ME); and, 2) confirm that the cecal ligation
model developed by Toofanian and Targowski (229, 235) would
reproducibly cause a mucoid enteropathy-like syndrome.
Cecal Filtrate Collection
Cecal contents were collected from a juvenile female New
Zealand White (NZW) rabbit (Oryctalagus cuniculus) with
naturally-occurring mucoid enteropathy, and also from two
healthy adult NZW rabbits. The affected rabbit also had a
severe intestinal Eimeria infestation. For each rabbit, the
volume of the contents was estimated, and an equal volume of
Hank's Buffered Salt Solution (HBSS) was added. The slurry
was centrifuged at 3000 rpm (700 x g), 40C for 15 minutes.
The supernatant was collected, and recentrifuged at 9500 rpm
(10,000 x g), 40C for 30 minutes. The resulting supernatant
was then filtered through a Whatman #50 paper filter, followed
by a 0.8 Am membrane filter, and finally a 0.22 im sterilizing
membrane filter. The sterile filtrate was stored at -700C
The cecal filtrates that were collected were tested for
mucus secretagogue activity in an in vitro intestinal explant
system. Five healthy adult NZW rabbits were sedated with
ketamine/xylazine, and killed with an overdose of barbiturate.
Intestinal segments were collected, opened along the
mesenteric border, and rinsed in HBSS, containing 100 Ag/ml
gentamicin, to remove ingesta. A 6 mm diameter Baker's biopsy
punch (Baker Cummins Pharmaceuticals, Inc., Miami, FL) was
used to take explants from the ileum and proximal colon
(approximately 10 cm from the cecocolic junction, at the point
where the number of longitudinal bands (teniae) decreases from
3 to 1, fig. 3.1). Longitudinally paired punches were taken
for control and experimental samples, as the concentration of
goblet cells varies along the length of the colon (236).
Explants were incubated in 24 well polystyrene microtiter
plates containing 0.5 ml of Trowell's T8 medium (GIBCO BRL,
Gaithersburg, MD) with 100 Ag/ml gentamicin, and 0.5 ml of the
appropriate cecal filtrate in each well. The plates were
incubated for 1 hour at 370C in a 95%02/5%CO2 humidified
environment. Following incubation, the culture medium was
removed and frozen, while the explants were fixed in 10%
neutral buffered formalin. Following routine processing and
sectioning, the tissues were examined histologically. After
1, 2, and 3 hours of incubation, the tissues appeared viable
in every respect.
Enzyme-Linked Lectin Assay
Quantitation of mucus release by the explants was
accomplished using an enzyme-linked lectin assay (ELLA) as
described by Cohan et al. (27). Ninety-six well polystyrene
microtiter plates were coated overnight at 40C with 100 p1 of
serial dilutions of porcine gastric mucin standard or
medium/filtrate mixture following explant incubation in 0.5M
sodium carbonate buffer, pH 9.6. The plates were washed with
phosphate buffered saline containing 0.5% Tween-20 (PBS-T20),
and blocked with 5% fetal calf serum in PBS-T20 for 1 hour at
370C. The plates were again washed, and incubated for 1 hour
at 370C with 100 .l/well of a 10pg/ml solution of a soybean
agglutinin (Glycine max)-horseradish peroxidase conjugate,
which specifically binds N-acetylgalactosamine residues.
Plates were then washed a third time, and the substrate o-
phenyl diamine (OPD) was added. After 10 minutes, the
reaction was stopped with 100 .l of 4N H2S04, and the optical
density at 492 nm was read. The standard curves for porcine
gastric mucin were consistent from plate to plate (figure
After the filtrates were incubated with paired explants,
the mucus secreted into the culture medium was quantitated by
the ELLA described above. The effect of sample dilution on
the amount of mucus in the medium/filtrate measured by the
ELLA is illustrated in figure 3.3 for a control rabbit, and in
figure 3.4 for the rabbit with naturally-occurring mucoid
enteropathy. At the lower dilutions (higher concentrations)
there were too many proteins in each sample competing with the
mucus for a limited number of binding sites on the plate,
resulting in falsely low mucus readings. At the higher
dilutions, there seemed to be some appropriate dilutional
effect; however the signal-to-noise ratio is decreased,
creating a larger variance. The 1:64 dilution was selected
for studies requiring quantitation of mucus.
Using the porcine gastric mucin standard curve for the
same plate, the optical density readings were converted to
apparent Ag mucus per ml of medium/filtrate (figure 3.5). The
amount of mucus in the medium/filtrate incubated without an
explant was subtracted from each value to give the amount of
mucus secreted. Since each explant was incubated in 1 ml
medium/filtrate, this number was equivalent to ng mucus
secreted per explant. Each value was then divided by the
weight of the explant to obtain nanograms mucus secreted per
milligram of tissue (figure 3.6).
The amount of mucus in the medium/filtrates initially and
following explant incubation varied with the source of the
cecal filtrate (figure 3.5; Appendix A). Secretion from both
ileal and colonic explants was significantly increased
compared to control (P < 0.05) when incubated with the cecal
filtrate from the rabbit with mucoid enteropathy (figure 3.6).
To better study mucoid enteropathy, it is necessary to
have a reproducible experimental model of the disease. Cecal
ligation has been reported to cause a mucoid enteropathy-like
syndrome in rabbits (229, 235). Surgery was performed on one
adult, female, conventionally housed NZW rabbit (Oryctalagus
cuniculus). The procedure was approved by the University of
Florida Institutional Animal Care and Use Committee. The
rabbit was anesthetized with 35 mg/kg ketamine and 5 mg/kg
xylazine intramuscularly. A sterile field was prepared, and
the abdomen was opened. A window was made in the mesentery
adjacent to the cecum. Two adjacent ligatures of 2-0 silk
were placed around the cecum just distal to the sacculus
rotundus, preventing the flow of ingesta into and out of the
cecum, but not obstructing the flow from the ileum to the
colon (figure 3.7). The linea alba was closed with 3-0
chromic gut, and the skin was closed with 3-0 silk. Recovery
The rabbit was observed twice daily. By the first
postoperative day the rabbit was eating small amounts, and
passing dry feces. However, appetite and fecal output
decreased on the second day. On the third day the rabbit was
anorectic, appeared painful, and had mucoid diarrhea.
Euthanasia and necropsy were performed.
At necropsy, there was perineal staining with greenish,
liquid fecal material. There were mild subcutaneous
hemorrhages at the incision site, but the sutures were intact,
and there was no gross inflammation. The cecal contents and
the body of the cecum distal to the sutures appeared normal.
The ampulla coli and proximal 2-3 cm of the colon were grossly
dilated with gas and fluid, although there was no mechanical
obstruction. The colon contained a small amount of greenish
fluid, with mucus adhering to the mucosa. The contents of the
ileum were similar. The jejunum contained yellow fluid, with
pockets of accumulation of clear, gelatinous mucus. On
histologic examination there appeared to be mild goblet cell
hyperplasia in the ileum and colon. These results are
consistent with those described by Toofanian (229, 235), and
suggest that this method provides a workable model to study
Rabbit (Oryctologus cunicalus)
Figure 3.1. Sources of intestinal explants.
Explants were taken from the distal ileum (closed arrow) and
the single-banded proximal colon (open arrow). Adapted from
reference 6 (p. 270). Used by permission of the publisher,
Cornell University Press.
E 2.000 -
1.000 Plate 12
.2/ / ---Plate 13
a. -- --Plate 14
0 20 40 60 80 100 120 140
Porcine Gastric Mucin (ng/ml)
Figure 3.2. Standard curves (ELLA).
Serial dilutions of porcine gastric mucin analyzed by ELLA.
The standard curves for the 5 plates used to collect data for
this chapter are shown.
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Figure 3.3. Serial dilutions of control filtrates.
Serial dilutions of medium/filtrate samples from a control
rabbit (P8) incubated in the absence of an explant (P8B) or
the presence of ileal explants from 3 rabbits (RI, R2, and
R3), analyzed by ELLA.
I I I I I 1
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Figure 3.4. Serial dilutions of ME filtrates.
Serial dilutions of medium/filtrate samples from a rabbit with
mucoid enteropathy (P4) incubated in the absence of an explant
(P4B) or the presence of ileal explants from 3 rabbits (R1,
R2, and R3), analyzed by ELLA.
1.00 0 Colon
Control Control Mucoid
(P3) (P8) Enteropathy
Figure 3.5. Mucus in medium/filtrates (ELLA).
Calculated mucus (porcine gastric mucin equivalents) in
medium/filtrates, measured by ELLA. Bars represent mean +
SEM. N=1 (P8 blank), 5 (P3, P4 blank), 6 (P8 ileum, colon),
7 (P3 ileum, colon), 10 (P4 colon), or 14 (P4 ileum).
" 0.30 -
E Mucold Enteropathy
Figure 3.6. Mucus secretion from explants.
Mucus secretion was determined by subtracting the amount of
mucus in the medium/filtrate incubated without explants (as
measured by ELLA) from the amount incubated with explants.
Values shown are relative to the wet weight of the explant.
Bars represent mean + SEM. N = 8 (ME) or 11 (control). =
significantly different from control (P < 0.05).
Figure 3.7. Site of cecal ligation.
The cecum was ligated distal to the sacculus rotundus (arrow),
allowing the flow of ingesta from ileum to colon. Adapted
from reference 235, with permission.
REFINEMENT OF METHODS
Use of Soybean Agglutinin to Quantitate Mucus
Soybean agglutinin, the lectin from Glycine max, specifically
binds to N-acetylgalactosamine residues, which are a major
constituent of the O-linked mucin glycoprotein, but are rare
in N-linked glycoproteins. A direct enzyme-linked assay using
this lectin for the measurement of mouse intestinal mucus has
been described (27). Its suitability for the measurement of
rabbit intestinal and colonic mucin was investigated using
Western blot techniques.
Materials and Methods
Purification of rabbit colonic mucin. Colonic mucus from
a rabbit with experimentally-induced mucoid enteropathy
(rabbit H5, see chapter 5) was isolated by the method of
Mantle and Allen (128) with minor modifications. Mucus gel
was removed from the colonic mucosa with forceps, and was
homogenized by hand in a Kontes tissue grinder with an equal
volume of 5 mM EDTA and 1 mM PMSF. The suspension was
centrifuged at 30,000 x g for 30 minutes. 8.1 g of cesium
chloride (CsCI) was added to 13 ml of the supernatant for a
final concentration of 0.6 g/ml. This solution was
centrifuged for 24 hours at 1.5 x 105 x g at 4C in a Beckman
L7 Ultracentrifuge with a 70.1Ti rotor. Eight 1.6 ml
fractions were collected, and 100 gl of each was evaluated for
carbohydrate content. The four densest fractions were pooled,
and the volume was brought up to 13.5 ml with 0.6 g/ml CsCl,
and the centrifugation was repeated, this time for 48 hours.
Following carbohydrate quantitation, fractions 2 to 5 (from
heaviest to lightest) were pooled, then concentrated and
desalted using a Centricon-100 concentrator (Amicon, Inc,
Beverly, MA). This solution was applied to a 2.6 x 14 cm
Sepharose 4B column. The fractions containing the void volume
were pooled, dialyzed against distilled water, and frozen at -
70C. Porcine gastric mucin (Sigma Chemical Co., St. Louis,
MO), was used as a mucin standard for comparative purposes.
Periodic acid-Schiff assay for carbohydrate quantitation.
The method used was that of Mantle and Allen (127). Periodic
acid solution was prepared by dissolving 25 mg of periodic
acid in 7% acetic acid (3.5 ml glacial acetic acid in 50 ml
distilled water. Each sample was brought up to a volume of
2.0 ml. 0.2 ml periodic acid solution was added to each
sample and incubated for 2 hours at 370C. Then 0.2 ml
Modified Schiff reagent (Fisher Chemical, Pittsburgh, PA) was
added, and samples were incubated 30 minutes at room
temperature. The optical density at 555 nm was read.
Protein concentration. A kit (Sigma Chemical Co., St.
Louis, MO) employing Peterson's modification of the micro
Lowry method was used to measure total protein concentration.
The kit was used according to the manufacturer's instructions.
Direct enzyme-linked lectin assay (ELLA). Porcine
gastric mucin standards, and serial dilutions of purified
rabbit colonic mucin in 100 Al 0.5 M carbonate buffer (pH 9.6)
were coated onto duplicate 96-well polystyrene plates for 2
hours at room temperature. The plates were washed 4 times
with phosphate buffered saline containing 0.1% Tween-20 (PBS-
T20) using an automated plate washer. The plates were blocked
with 3% bovine serum albumin (BSA) in PBS-T20 for 1 hour at
370C. Following incubation, the plates were again washed, and
then incubated with 100 Al of a 1:200 dilution of a 1 mg/ml
solution of soybean agglutinin-horseradish peroxidase
conjugate (SBA-HRP) in PBS-T20 for 1 hour at 370C. Following
5 washes with PBS-T20, 85 Al of the substrate 0.4 mg/ml o-
phenylene diamine in 0.05 M phosphate-citrate buffer (OPD),
was added to each well. After 10 minutes at room temperature,
the reaction was stopped with 85 1l of 4 N H2SO4. The optical
density was read at 490 nm.
Western Blots. Six percent SDS-polyacrylamide gels were
prepared (71). Samples, up to 100 /i, were boiled 3 to 5
minutes in 0.25 M 2-mercaptoethanol and 0.1% SDS and applied
to gels. Following electrophoresis, gels were either stained
for protein with 0.25% Coomassie Blue, or samples were
transferred to nitrocellulose by wet electrophoretic transfer
for measurement of lectin-binding activity (mucus) of total
After transfer, blots to be labelled with lectin were
washed with PBS, and blocked in 2.5% casein or 3% BSA in PBS
for 1 hour at 370C. The blots were transferred to a solution
of 10 Ag/ml SBA-HRP in blocking solution. Following a 1 hour
incubation at 370C, blots were washed 4 times for 5 minutes
each in PBS. The substrate solution of 300 mg/ml 4-chloro-l-
naphthol was added, and color was allowed to develop for 30
Alternatively, transferred blots were stained for total
glycoprotein with periodic acid/Schiff (224). Membranes were
washed in distilled deionized water for 5 minutes, then
incubated in 1% periodic acid, 3% acetic acid for 15 minutes.
Following 15 minutes washing with several water changes,
modified Schiff's reagent (Fisher Chemical, Pittsburgh, PA)
was added, and the membrane was stained for 15 minutes in the
dark. The reaction was stopped by a 5 minute incubation in
0.5% sodium bisulfite solution, and the membrane was washed
Rabbit colonic mucin was successfully purified, as
evidenced by a lack of contaminating proteins on SDS-PAGE
(figure 4.1, lane 3), and a strong lectin-binding signal in
the stacking gel of a Western blot (figure 4.2, lane 3). No
band was seen at 118 kDa, corresponding to the "link"
glycopeptide (188), although the mucus was collected with
protease inhibitors. Using bovine serum albumin as a
standard, the concentration of protein in the purified mucus
solution was calculated to be 134 ig/ml by a modified Lowry
assay. Using commercially obtained porcine gastric mucin as
a standard, the glycoprotein concentration for the same sample
was calculated to be 73 Ag/ml by the PAS method, but only 17
ug/ml by ELLA.
Coomassie blue staining reveals a major band at 55 kDa in
all the cecal filtrates examined (figure 4.1, lanes 5-8),
along with a smear that may or may not contain minor bands.
This major band most likely represents serum albumin (65).
Mucin itself is not visible at these concentrations, in either
purified or crude samples. As would be expected, homogenates
of ileal and colonic explants demonstrate a large number of
For many cecal filtrate and medium/filtrate samples, the
only band detectable with soybean agglutinin was in the
stacking gel, and consistent with mucus (figure 4.2, lanes 5-
7). Other samples showed a smear of high molecular weight
material, with a major band at 144 kDa (lanes 8-11). PAS
staining of gels (data not shown) gave very similar results to
lectin staining. However, only the material in the stacking
gel representing mucus was visible in homogenates of ileal and
colonic explants (lanes 12-13), although there was a large
number of protein bands (figure 4.1, lanes 9-10).
Soybean agglutinin binds to both rabbit colonic mucin and
porcine gastric mucin. There appear to be fewer available
binding sites on the rabbit mucin, as the amount of purified
rabbit colonic mucin calculated from the ELLA was much less
than was calculated by more conventional means. This is not
surprising, since colonic mucins are more highly sialylated
and sulfated than gastric mucins (160), making the internal N-
acetylgalactosamine residues less accessible. Therefore,
although porcine gastric mucin is used as a standard
throughout this study because of availability and consistency,
the actual concentrations of rabbit mucus are probably higher
than reported. However, the relationships between rabbit
samples will remain the same.
It is worth noting that this mucin was collected from a
rabbit with experimental mucoid enteropathy. A recent paper
(89) demonstrates changes in lectin-binding capacities in
mucins from rabbits with mucoid enteropathy. It indicates an
increase in binding to soybean agglutinin, as well as other
lectins with affinities for internal sugar residues,
consistent with the incomplete glycosylation characteristic of
many disease states.
The soybean agglutinin is not as specific a marker for
mucus as had been expected. However, when non-mucin bands are
present in cecal filtrates, they do not appear to change after
explant incubation, and can therefore be subtracted out.
Furthermore, they are not present in the homogenates of the
explants themselves, and so cannot be released into the
medium/filtrates following incubation.
Evaluation of Enzyme-Linked Lectin Assay
Soybean agglutinin, the lectin used in the Western blots,
was also used in an ELLA (27). The mucus-containing sample
was used to coat polystyrene wells, and a SBA-HRP conjugate
was used to quantitate the amount of mucin bound.
However, a direct assay is generally not considered
appropriate for samples that contain only a small
concentration of the target molecule. A standard polystyrene
ELISA plate binds only 100 ng protein/well (71); the
medium/filtrate samples contain approximately 10 mg/ml.
Therefore, the majority of an undiluted sample does not bind
to the plate, and is removed in the first wash. Only when the
sample is diluted will a significant fraction of it bind.
Looking at the problem another way, if there is 1 Ag/ml mucus
in a sample that contains 10 mg/ml total protein, then mucus
constitutes 0.01% of that sample. If 100 ng protein binds to
the plate, then only 0.01 ng of mucus will bind, which is at
approximately the limit of detection (71).
Three types of modifications were made in an attempt to
develop a lectin-based assay that would be more sensitive.
One set of modifications was based on classic ELISA techniques
(71). Competition (indirect) assays and sandwich assays are
used to capture the antigen of interest out of contaminated
samples. Crude purification of the samples themselves was
also considered, and affinity chromatography and size
separation by ultrafiltration were performed. Finally, use of
a substrate that has a greater binding capacity (e.g.
nitrocellulose vs. polystyrene) allows for greater binding of
both contaminants and the molecule of interest, so that it can
be accurately quantitated without as great dilution of the
Materials and Methods
Unless otherwise noted, all reagents were obtained from
Sigma Chemical Co. (St. Louis, MO). The samples used for this
experiment were medium/filtrate samples following explant
incubation that had been obtained from pilot studies.
Competition assay. Plates were coated with porcine
gastric mucin (10 ng/well, 100 ng/well, or 1 gg/well) for 2
hours at room temperature, and then blocked with 2.5% casein
for 1 hour at 370C. 100 1l of serial dilutions of porcine
gastric mucin (100 ng/ml stock) or medium/filtrate sample were
added simultaneously with 100 A1 of 1, 2, 5, or 10 Ag/ml SBA-
HRP. Alternatively, mucin containing samples and SBA-HRP were
preincubated together for 1 hour at 37C, and then applied to
the plate. Plates were washed with PBS-T20, and 100 Al of the
substrate OPD was added. After 10 minutes, the reaction was
stopped with 100 [l of 4N H2SO4, and the optical density was
read at 492 nm.
Sandwich assay. Plates were coated with 2 Ag/well of
unconjugated soybean agglutinin (Glycine max lectin), Wisteria
floribunda lectin, or Maclura pomifera lectin for 2 hours at
room temperature, and then blocked with 2.5% casein for 1 hour
at 37C. Then 100 Al of serial dilutions of porcine gastric
mucin (100 ng/ml stock) or medium/filtrate sample were added,
and incubated 1-2 hours at 37C. After multiple washes with
PBS-T20, 100 Al of 10 Ag/ml SBA-HRP was added, and incubated
1 hour at 37C. Plates were washed with PBS-T20, and 100 jl
of the substrate OPD was added. After 10 minutes, the
reaction was stopped with 100 Al of 4N H2SO4, and the optical
density was read at 490 nm.
Centricon filtration. Centricon-100 (Amicon, Inc.,
Beverly, MA) filters were rinsed with distilled water, and
used according to the manufacturer's instructions to separate
1 ml samples. Both retentates, which should contain only
mucus and other molecules > 100,000 Daltons, and filtrates,
which should contain molecules < 100,000 Daltons, were assayed
in direct ELLA. In addition, these samples were applied to a
6% SDS-polyacrylamide gel to determine the actual protein
Affinity chromatography. One hundred pl of SBA-sepharose
beads were washed and resuspended in 100 Al PBS, and added to
50 [l of sample. After 15 minutes of incubation at room
temperature with some mixing, the beads were removed by
centrifugation for 1 minute at 1000 x g. The supernatant was
retained as "wash 0", an indication of how much mucus did not
stick to the beads. The beads were washed 3 times with 1.4 ml
PBS. Free N-acetylgalactosamine (200 pl of 10 mM, 0.1 M, 0.5
M, or 1.0 M) was added, and incubated for 15 minutes at room
temperature to elute the mucus. Following centrifugation, the
eluates were collected and dialyzed overnight vs. PBS. The
washes and eluates were analyzed for mucus concentration using
a direct ELLA.
Enzyme-linked lectin flow-through assay. An Easy-TiterT
ELIFA apparatus (Pierce, Rockford, IL) was assembled according
to the manufacturer's instructions, using either
nitrocellulose or a Biodyne B membrane. The membrane was
coated with 100 pl of serial dilutions of porcine gastric
mucin (10 [g/ml stock) or medium/filtrate sample, which was
pulled through the membrane over a 5 minute period. A 2.5%
casein or 3% BSA solution was pulled through over a 10-15
minute period to block the membrane. One hundred 1l of 0.1,
0.25, 1, 2.5, or 10 [g/ml SBA-HRP was added, and pulled
through in 5 minutes. Plates were washed 3 times with 200 pA
PBS, and 100 pl of the substrate OPD was pulled through as
quickly as possible. The reaction was stopped with 100 pl of
4N H2SO4, and the optical density was read at 490 nm.
Dot blots. Dot blots were performed using nitrocellulose
in a miniblot apparatus. Serial dilutions of mucus or samples
in PBS or 0.5 M carbonate buffer, pH 9.6, were used to coat
the nitrocellulose for 2-3 hours at room temperature. The
blots were removed from the apparatus, washed with PBS, and
blocked with 3% BSA for 1 hour at 250C. Five Ag/ml SBA-HRP in
PBS was added, and incubated for 1 hour at 250C. As a control
for endogenous peroxidase activity, some blots were not
incubated with SBA-HRP, but rather remained in blocking
solution. Following four 5 minute washes with PBS, blots were
developed by addition of 0.3 g/ml of the substrate 4-chloro-1-
naphthol in 50 mM tris, pH 7.6. After 5-10 minutes, the blot
was washed with distilled water and dried. Densitometry was
performed to quantitate color development.
In all competition assays the optical density reading
obtained was a direct reflection of the coating mucus
concentration, and SBA-HRP concentration. There was no
difference in optical density in the presence or absence of
competing mucus or filtrate sample.
In the sandwich assay, the optical density readings
depended solely on the type of lectin used for coating the
plate. With soybean agglutinin coating, the optical density
ranged from 1.8 to 2.5; for Wisteria floribunda the range was
2.1 to 2.7; for Maclura pomifera it was 1.5 to 2.3. The
presence or absence of mucus had no effect.
Direct ELLA of retentates following Centricon-100
ultrafiltration did not demonstrate appropriate decreases in
optical density with serial dilution. SDS-PAGE demonstrated
that the banding pattern for these retentates was identical to
that for unfractionated medium/filtrate samples (figure 4.1).
No protein was visible by Coomassie blue staining of the
ultrafiltrates on SDS-PAGE.
Following affinity chromatography with soybean
agglutinin, the material eluted from the beads was calculated
to contain approximately 25 ng/ml mucus by direct ELLA.
However, the material in "wash 0" (that which did not bind to
the beads) contained approximately 250 ng/ml mucus, as
measured by ELLA.
With the Easy-TiterT system, the optical density readings
were dependent on the SBA-HRP concentration used for
detection, but did not vary with mucus concentration.
Background readings (no mucus added) were 2.4-2.8 for 10 Ag/ml
SBA, 2.1-2.5 for 2.5 Ag/ml, 1.9-2.3 for 1 gg/ml, and 1.1-1.4
for 0.25 Ag/ml on Biodyne B membranes. The background values
on nitrocellulose were lower; however values for mucus coated
wells were also lower, and consequently meaningless.
In contrast, dot blots demonstrated appropriate
dilutional effects for both samples and mucus standards; wells
which did not contain mucus did not have color development.
However, dots containing medium/filtrate samples that were not
incubated with SBA-HRP demonstrated significant color
development, indicating the presence of endogenous peroxidase
activity within the samples. Subtraction of densitometry
values obtained in the absence of SBA-HRP from those obtained
in the presence of SBA-HRP led to inconsistent data that were
Attempts to increase sensitivity and specificity of mucus
quantitation over the direct ELLA described by Cohan (27), and
used in the pilot studies were uniformly unsuccessful.
Competitive assays for the measurement of mucus have been
successful using an antibody-based system (133). However,
mucin in solution was not able to compete with bound mucin for
lectin binding. It may be that the lectin has a greater
affinity for mucin that is bound to the plate as opposed to
mucin that is free in solution. Attachment to the plate may
allow greater exposure of N-acetylgalactosamine sites to which
the lectin may bind. Alternatively, mucin may bind to the
lectin in solution, but not block all of the multiple binding
sites, thereby allowing the lectin to bind to bound mucus.
Two sets of lectins were used for sandwich assays. At
first, the same soybean agglutinin was used to coat the plate
and for detection. Soybean agglutinin itself contains N-
acetylgalactosamine residues (60), so it is no surprise that
once the plate was coated with lectin, the lectin bound
maximally, even in the absence of mucin. The other lectins
used for coating were Wisteria floribunda lectin, a
glycoprotein which does not contain N-acetylgalactosamine, and
has an affinity for N-acetylglucosamine; and Maclura pomifera
lectin, a non-glycoprotein that has an affinity for N-
acetylgalactosamine (60). Unfortunately, the lectin used for
detection was again soybean agglutinin; so that each coating
lectin had a direct affinity for the detecting lectin. Use of
a different lectin for detection was not attempted, although
it might have been successful.
Ultrafiltration was performed to remove contaminants with
a molecular weight of < 100,000 Daltons. However, the
acrylamide gel demonstrates that essentially all proteins were
retained by the filter. Therefore, the contamination in the
sample was not reduced, and ultrafiltration was of little
benefit. It has been recently reported that nuclease
treatment and acidification are required to separate
extraneous proteins from mucin (157).
Affinity chromatography was successful in removing
contaminants from the final sample. Unfortunately, more mucus
was present in the initial wash than in the sample eluted from
the beads. This lack of affinity between the lectin and mucin
in solution may relate to the problems with the competition
Nitrocellulose was selected as a substrate with increased
binding capacity for the enzyme-linked lectin flow-through
assays and dot blots. This worked well for samples containing
just mucus, or even samples that had been "spiked" with a
known contaminant, such as BSA. Unfortunately, cecal contents
contain significant endogenous peroxidase activity.
Peroxidase was selected as an enzyme over alkaline
phosphatase, which is known to have significant endogenous
activity in the gut (71). In the polystyrene plate assay,
small amounts of endogenous peroxidase activity can be
detected, but these values are very small compared to the
mucus/lectin-associated activity. However, when measured by
densitometry, nitrocellulose blots incubated without
lectin/horseradish peroxidase conjugate often showed greater
activity than those incubated with the conjugate. It is not
clear why the endogenous peroxidase should bind with more
avidity to nitrocellulose than to polystyrene, but the fact
exists that it is not feasible to quantitate mucus bound to
Therefore, it was not possible to improve upon the direct
ELLA using the above techniques. In order to use the direct
assay, it is necessary to perform serial dilutions to
ascertain that the values used for analysis are not
artifactually low. It is also desirable to select the lowest
dilution (highest concentration) possible to obtain the best
Harvesting of Mucus from Explants
Mucus is a tenacious substance, and a problem in
quantitation is collection of the mucus from the tissue (160).
In fact, a specific gastric mucosal mucin receptor has been
described in the stomach (211). Different methods were
evaluated to see if an improvement could be made over simple
washing, without seriously damaging tissue and releasing
additional mucus. Since mucus is primarily held in
association by noncovalent mechanisms (160), acid and base
were added to alter pH. Because mucus contains mostly
negative charges, and relies on the presence of calcium,
magnesium, and other positive ions to maintain gel-forming
properties (67), EDTA was added in an attempt to chelate
cations, and disaggregate mucus. Finally, N-acetylcysteine,
a known mucolytic which breaks down disulfide bonds (200) was
Materials and Methods
Explants were collected with a 6 mm Baker's biopsy punch,
and incubated in 0.5 ml Trowell's T8 medium with gentamicin
and 0.5 ml of a filtrate of cecal contents at 370C in