LIBRARY

Michigan State
University

 

This is to certify that the
thesis entitled

DEVELOPMENT OF AN INTESTINAL CELL CULTURE METHOD
TO STUDY DIETARY EFFECTS ON INTESTINAL MUTAGENS

presented by

Wha Young Kim Lee

has been accepted towards fulfillment
of the requirements for
Ph. D. degree in Food Sc1ence and Human.
Nutr1t1on

 

 

Mama; fl meog'é

Major professor

Date March 15, 1978

 

0-7639

 

' ‘ Ill \\ \llll l \l ll \lllllllll

3 1293 01096 8968

«9:7 ‘
l~ 03

f; a
W”? 0 5 2002

DEVELOPMENT OF AN INTESTINAL CELL CULTURE METHOD TO STUDY
DIETARY EFFECTS ON INTESTINAL MUTAGENS

BY
Wha Young Kim Lee

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Food Science and Human Nutrition

1978

Cry/{.2 (Z, 47%;)

ABSTRACT

DEVELOPMENT OF AN INTESTINAL CELL CULTURE METHOD TO STUDY
DIETARY EFFECTS ON INTESTINAL MUTAGENS

BY
Wha Young Kim Lee

Intestinal fibroblasts were cultured to develop a
method to screen potential mutagenic and carcinogenic
agents found in the gastrointestinal tract. Cells were iso-
lated from fetal rat intestines and cultured in Eagle's
Minimum.Essentia1 Medium with 10% fetal calf serum and
antibiotics. Cell doubling time was 20 hours and the cells
were subcultured 2-5 times before experimental treatment.
Optimum conditions for determining mutation frequency were:
exposure time to mutagenic agents, 3 hours; expression time,
3 days; cloning in selective medium, 2 weeks; and 1.0 mM
ouabain in culture medium to select mutant cells.

Bile acids, often implicated as carcinogens in colon
cancer, were tested for mutagenesis in this system. Deoxy-
cholic, lithocholic and chenodeoxycholic acids were more
cytotoxic than cholic, hyodeoxycholic and 3,12-dione-58-
cholanic acids. Deoxycholic acid produced the highest muta-
tion frequency at 0.5 mM. Lithocholic and chenodeoxycholic

acids at 0.5 mM also were mutagenic. Cholic, hyodeoxycholic

Wha Young Kim.Lee
and 3,12—dione-58-cholanic acids at concentrations 5
1.0 mM were not mutagenic. Cholic acid at 2.0 mM was muta-
genic. N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) was
used to initiate cell mutation and bile acids were tested
as promoting agents in this system. When MNNG treated
cells were incubated in 1.0 mM ouabain medium containing
cholic acid at 0.1 mM or 0.2 mM or chenodeoxycholic acid at
0.1 mM, the mutation frequencies increased above the mutation
frequency induced by MNNG alone. When deoxycholic and litho-
cholic acids at 0.1 mM were used in ouabain medium, mutation
frequencies decreased. This decrease in mutation frequency
is probably due to the greater cytotoxicity of deoxycholic
and lithocholic acids rather than the inability of these acids
to promote MNNG induced mutation.

Organic and water extracts of feces and bile collected
from rats fed high-fat, low-fiber or low-fat, high-fiber diets
also induced higher mutation frequencies.

The effects of dietary fat and fiber on steroid meta-
bolism, bowel function, and anaerobic cecal bacteria were
studied in rats. Both wet and dry fecal mass were increased
when bran and agar were added to the diet. Intestinal tran-
sit time decreased as the level of bran in the diet increased.
There was no significant difference in daily bile acid excre-
tion either when bran was added to the diet or when the fat
and agar content of the diets were varied. However, the
greater fecal mass caused the fecal concentration of bile

acids to be decreased when dietary fiber was added to the diet.

Wha Young Kim Lee
Mere extensive bile acid degradation was found as the level
of bran in the diet increased. Mere cholic acid was de-
graded when cecal contents from rats fed bran diets were in—
cubated anaerobically than when cecal contents from rats fed
a bran-free diet were incubated anaerobically. The number of
anaerobic cecal bacteria was not altered by the bran content
of the diet. Therefore, bacterial type or activity rather
than the number of viable anaerobic bacteria in the large
bowel appears to be more important in bile acid degradation.

Neutral steroid excretion and concentration were
higher in rats fed a high-fat, low-fiber diet compared to rats
fed a low-fat, high-fiber diet. Neutral steroids were more
extensively degraded in the low-fat, high-fiber fed rats. Rats
fed the high-fat, low-fiber diet excreted more coprostanol in
the feces than coprostanone, whereas more caprostanone was
found in the feces of rats fed the low-fat, high-fiber diet.
This different ratio of coprostanol/coprostanone in the feces
of rats fed the two different diets implies that microbial
activity in the large bowels of the two groups of rats was
different.

Results of this study indicate that l) the intra-
1umina1 environment can be changed due to variations in the
diet, 2) it is feasible to use intestinal fibroblasts iso-
lated from rat fetuses to screen potential mutagens found in
the gastrointestinal tract. The second conclusion is based
on the observations that compounds which caused cell muta-
tions are either a known carcinogen or strongly implicated

in colon cancer.

To my mother

ii

ACKNOWLEDGEMENTS

The author wishes to express her sincere appreci-
ation to Dr. M.R. Bennink and Dr. W.L. Chenoweth for their
academic advising in both research and curricular programs.

Other guidance committee members, Drs. D.R.Romsos,
W;W. Wells, M.T. Yokoyama and R.B. Young, are also warmly
thanked for their concern and interest in this work. A
special note of thanks is due Dr. J.I. Goodman for his valu-
able suggestions on this work.

The author also acknowledges the Department of Food
Science and Human Nutrition at Michigan State University for
providing facilities and the National Institute of Health
for the financial support.

The understanding, patience and encouragement of the

author's husband, WOn Hong, is most sincerely appreciated.

iii

TABLE OF CONTENTS

INTRODUCTION
LITERATURE REVIEW
Cancer of Colon

Effect of Diet on Experimentally Induced
Cancer

Relationship of Cut Microflora to Colon
Cancer

Methods for Studying Carcinogenesis and
Mutagenesis

MATERIALS AND METHODS
PART I
Experiment 1
Experiment 2
PART II
Culture Media and Growth Conditions
Preparation of Reagents Used

Isolation of Intestinal Epithelial
Cells

Isolation of Intestinal Fibroblasts
Protocols of Experiments

Doubling Time Determination

RESULTS

iv

16

19

22
32
32
32
36
41
41
42

43
45
46
50
51

TABLE OF CONTENTS (Continued . . .)

PART I
Experiment 1
Experiment 2
PART II
Intestinal Epithelial Cell Isolation

Isolation of Fibroblasts and
Doubling Time

Mutagenesis Study
DISCUSSION
CONCLUSIONS AND RECOMMENDATIONS
BIBLIOGRAPHY

59
59
84
104
107

Table

10.

11.

12.

LIST OF TABLES

Diet and Cancer: Possible Relationships.

Effect of diet on the fecal steroid
concentration

Composition of diets for experiment 1.
Composition of diets for experiment 2.

Bowel function, steroid metabolism.and
bacterial content of the cecum in rats
fed graded levels of wheat bran.

Fecal mass and steroid metabolism in
feces and in bile from rats fed either
a high-fat, lowefiber diet or a low-fat,
high-fiber diet.

Cytotoxicity and ouabain resistant
mutants response to graded levels of
deoxycholic acid.

Variations in mutation frequency within
each experiment and within a treatment
from three experiments.

Mutagenic and comutagenic effect of
selected bile acid.

Mutagenic effect of chloroform extracts
of feces from rats fed a high-fat, low-
fiber diet or a low-fat, high-fiber diet.

Mutagenic effect of water extracts of
feces from rats fed a high-fat, low-fiber
diet or a low-fat, high-fiber diet.

Mutagenic effect of bile collected from

bile ducts of rats fed a high-fat, low-
fiber or low-fat, high-fiber diet.

vi

14
33
38

52

55

72

73

74

76

78

8O

 

LIST OF TABLES (Continued . . .)

Table Page
13. Mutagenic and comutagenic effects of

fecal extracts and bile from rats fed

a high-fat, low-fiber diet. 81
14. Promotor effect of selected bile acids. 82

vii

 

 

 

LIST OF FIGURES

Geographical distribution of colonic
cancer. 1968-1969.

Protocols for (A) plating efficiency
(B) mutagenesis and (C) cytotoxicity.

Phase micrographs of epithelial cells
isolated from rat intestine by tissue
culture/cell culture method.

Phase micrographs of fibroblasts isolated
from.intestine of rat fetus.

Cytotoxicity and ouabain resistant
mutant response to MNNG dose.

Effect of ouabain concentration on con-
trol and MNNG treated cells.

Bright field micrograph of Giemsa
stained cells.

Expression time of MNNG (1.5 ug/ml) -
induced ouabain resistant mutations.

Cytotoxicity of selected bile acids.

viii

49

58

58

61

63

66

68
7O

INTRODUCTION

Epidemiologic studies show that the incidence of
colon cancer varies with geographic area and socioeconomic
level. It is high in Western countries and low in African
and Asian countries. Based upon studies with migrant popu-
lations, many investigators agree that environmental factors
play a major role in the tumorigenesis of colon cancer, but
genetic factors play a minor role. It has been shown that
the risk rate of developing colon cancer among the immigrants
from.the low risk areas into the United States is high,whi1e
the risk rate of their counterparts in the native country
remains relatively low.

Among the environmental causes, dietary factors
have been suggested in the etiology of colon cancer. Several
hypotheses have been proposed how diet may play a role in
colon carcinogenesis. Dietary factors include fat, protein,
and fiber levels in diet. Diet influences intraluminal coma
ponents and intestinal microflora. Intestinal microflora may
produce carcinogens or cocarcinogens from the intraluminal
metabolites. Also, the diet can change gut function and the
mass of the material in the colon. Bowel function controls
the length of time that colonic cells are exposed to carcino-

genic agents, and the mass of indigestible material in the

1

2
colon affects the concentration of carcinogenic agents that
come in contact with colonic cells. Bile acids, cholesterol
and protein have been suggested as substrates which micro-
flora metabolize to produce carcinogens.

The relationship between diet, intraluminal compo-
nents, microflora, and gut function needs to be elucidated.
Through studying the foregoing interrelationships one might
be able to identify agents responsible for the etiology of
colon cancer. A great deal of research has been conducted
in an effort to elucidate the relationship between diet and
colon carcinogenesis in animal models. However, there is
not general agreement among investigators which component(s)
of the diet is (are) etiologically important in colon carcino-
genesis.

The existing animal bioassays are time consuming and
expensive. There is an urgent need to develop a relatively
rapid and reliable in_vi££g method to screen suspected
carcinogenic agents which are found in the gut. Positive com-
pounds can then be tested by the existing bioassays. Com-
pounds which are positive in the bioassays then can be re-
lated to diet.

The objectives of this research are: 1) to study
the effect of diet on steroid metabolism, gut function and
cecal bacteria, and 2) to develop an in_vi££g method to
screen suSpected carcinogens identified in the literature and

from objective 1.

 

 

LITERATURE REVIEW

Nutritional and dietary factors have been thought
to play a role in the development of cancer in man and
animals for many years. However, experimental evidence to
support this concept and to explain the mechanisms involved
is very limited at the present time.

First I will review the epidemiologic findings
which implicates a dietary relationship to cancer. The
relationships of specific dietary components and site of
cancer will be discussed also. Secondly, I will review ex-
perimental evidence which implicates dietary components and
cancer. Finally, I will review currently employed methods
to evaluate carcinogenicity and/or mutagenicity of compounds
which can be found in the gastrointestinal tract.

Food, food additives and contaminants may be involved
in the cause and/or the development of some of the common
cancers. Epidemiologic data show that there is an uneven
distribution of many major types of cancer in the world.
Race, culture, environmental factors and diet are factors
which could influence the incidence rate of cancer. The
question of which component(s) account for the difference in

cancer incidence remains unanswered.

4

Some observations have been reported. Cancer of
colon and breast are most prevalent in North.America and
Western European countries (1,2). The incidence of gas-
tric cancer is high in Japan, Chile, South Africa and
Finland. These epidemiologic data provide circumstantial
evidence that diet may be important in the cause of cancer,
but nothing about its etiology and pathogenesis.

Results of studies with immigrant pOpulation do not
support genetic and racial differences as a major cause of
cancer (3-6). Rather, environmental factors appear to
account for the geographic differences in cancer distribu-
tion. Generally the areas with low incidence of colon and
breast cancer have a low standard of living with limited

amount of industrialization, whereas the high incidence

areas are economically developed and industrialized countries.

Air and water pollutants common to the industrialized coun-
tries, food additives, customs and stress of Western society
and dietary pattern have been suggested as possible environ-
mental factors which cause the cancer (1).

Mbst investigators agree that diet is a major en-
vironmental factor which could cause cancer, especially
cancer of the gastrointestinal tract. Populations in high
risk areas consume diets containing large quantities of
animal fats and protein, refined carbohydrate (7-9) and
small quantities of dietary fiber (10), whereas people in
low risk areas eat foods low in meat and meat products but
high in fish and vegetables, unrefined cereals and fiber

rich foods such as legumes and tubers (1).

_—_

 

5

In addition to food consumed, nutritional status
may affect the incidence of tumors. High energy intake,
nutritional excesses or deficits, and consumption of
alcohol increase the risk of certain cancers. Specific
deficiencies of essential nutrients have been shown to en-
hance experimental chemical carcinogenesis. These include
the deficiencies of vitamin A (12,13), methionine and cho-
line (14), selenium (15), zinc (16), copper (16), vitamin C
(17) and protein (18).

Table l, cited from Lowenfels and Anderson (11),
summarizes possible relationships between diet and cancer.

Cancer of the colon. Cancer of the large bowel is a

 

leading cause of cancer related deaths in United States.
Mortality due to cancer of the large bowel is second only to
lung cancer in males, and third only to breast and lung
cancer in females. The incidence of colon cancer is high in
North America, Western Europe and Australia, moderate in
Eastern Europe, Israel, and Ibeiran Peninsula and low in
Asia, Africa, Central and South America (Figure 1). This
difference in the distribution could be attributed to genetic
or environmental factors. There is some evidence that gene-
tic factors may play a role in colon carcinogenesis (l9).
Relatives of colon cancer patients have a greater risk of
developing this cancer. Also, persons with familial poly-
posis coli, Gardner's syndrome, adematous polyps, and Pentz-
Jegher's syndrome, all of which are hereditary diseases,

have an increased risk of colon cancer.

6

TABLE 1. Diet and Cancer:

Possible Relationships

 

Factor

Organ or organ system
at risk

 

Energy Content

Type of Diet

Abrasive, irritant foods
High fat
Low residue

Toxic Plants
Cycad palm
Bracken fern
Safrole

Vitamins
A
Riboflavin

Carcinogens
Aflatoxins
Nitrosamines
Benzopyrene

Miscellaneous

Alcohol

Artificial sweeteners

Overall incidence of cancer
may be affected by total
caloric intake

Stomach
Colon, breast

Colon

Small, Large bowel
Small, Large bowel

Liver

Epithelial tissues
Epithelial tissues

Liver
Upper gastrointestinal tract

Stomach

Oropharynx, larynx, esophagus,
liver

Bladder

 

Adapted from Lowenfels and Anderson (11).

Figure 1. Geographical distribution of colonic cancer
1968-1969. Adapted from Kassira et al. (1).

 

 

8

Geographic Variance in Colonic Cancer Mortality Rates
(Age-Adjusted-per 100,000) 1968-1969

Scotland
Ireland
Austria
England
Denmark
Germany
France
Sweden
New Zealand
U.S.A.
Canada
Australia

Switzerland
Norway
Hungary
Portugal
Italy
Czechoslovakia
Spain
Malta
Israel
Finland
Greece

Poland
Hong Kong
Barbados
Japan
Romania
Singapore
Yugoslavia
Mauritius
Venezuela
Panama
Mexico
Costa Rica

Thailand
Egypt

El Salvador
Honduras

FIGURE 1.

H l-' N N
U1 0 U! C U!
j_14_141_1 141_1 L_LJ_L_1_LJ_1_LJ L1 l l

 

 

 

9

Studies with migrant populations, however, suggest
that the incidence of colon cancer is more associated with
environmental causes than with genetic factors. While the
incidence of colon cancer in Japan is low compared to the
united States, the risk rate increases sharply among the
immigrated Japanese in California (3) and Hawaii (4). The
incidence of colon cancer in the first generation of Japan-
ese immigrants is similar to the incidence of colon cancer
in Japan, but the incidence in the second generation is
essentially the same as that found in native American whites.
This is especially true for males. This implies that pro-
longed exposures to the American environment may be primar-
ily responsible for the higher incidence of colon cancer in
the second generation. Data from.other immigrant popula-
tions show similar trends. The colon cancer mortality rate
among Polish immigrants into the United States is rapidly
approaching the mortality rate for American whites, while
the mortality rate in Poland remains low (5). The import-
ance of environmental effects is also supported by the work
of Mass and Modan (6) who compared the incidence of coloro-
rectal tumors among Jews who immigrated into Israel from
different parts of the world. The incidence was high among
Europe-American born Jews, moderate among Israel born Jews,
and low among Asia-Africa born Jews. It is also reported
that American blacks have much higher incidence of colon

cancer than African blacks (10,20).

 

 

10

In view of the above epidemiological findings it
seems unlikely that the difference in the incidence of
large bowel cancer is due mainly to genetics. There is
good correlation between incidence of colon cancer and
economic development. The more affluent societies have
the higher incidence of the colon cancer.

Among the environmental causes, diet has been pro-
posed as a factor in the development of colon cancer.

Diets in countries with a high risk of colon cancer differ
from diets of countries with a low incidence of colon
cancer. The high risk countries consume more animal fat
and protein, processed sugar, and refined carbohydrate

(7-9) and less fiber (10) than countries with a low risk of
developing colon cancer. Haenszel et a1. (21) found that
Hawaiian Japanese, who ate "Western-style" meals more fre-
quently than typical Japanese meals, had a high incidence of
colon cancers. Seventh-Day Adventists, who are vegetarians,
are reported to have 20% less large bowel cancer than con-
trol American whites (22).

From these observations, many investigators have
developed a hypothesis relating diet and tumorigenesis in
the colon. Composition of the diet can alter the intralumi-
nal components and microflora of the gut. Also, the meta-
bolism of intestinal microflora can be altered by the diets.
It is hypothesized that carcinogenic or cocarcinogenic com-

pounds are produced from normal intraluminal components by

 

ll
altered microfloral populations or altered microfloral meta-
bolism. Cocarcinogenic compounds refer to compounds which
are not carcinogenic by themselves but which may stimulate
carcinogenesis in the presence of a carcinogen. Thus, the
diet may have its effect both by altering the supply of
substrates for carcinogen or cocarcinogen production, and by
changing the type and number of the intestinal bacteria
available to act on such substrates. A diet high in animal
fat, animal protein, and refined carbohydrate may lead to
different p0pulations of intestinal microflora than diets low
in animal products and high in dietary fiber (23). Also,
differences in diets can change endogenous secretions into
the gut which influences intraluminal environment.

The key question is what diet-dependent components,
modified by intestinal bacteria, could possibly account for
the development of colon cancer. Due to structural similar-
ities between steroids and polycyclic aromatic carcinogens
(23,24), bacterial metabolites of bile acid and cholesterol
have been suspected to have a role in carcinogenesis. As
the fat content of the diet increases, the amount of bile
secreted into the gut increases. (25) Also, the concentra-
tion of fecal bile acids and neutral steroids increase as
the fat level in diet increases. Moreover, the high fat
diet may change the population of intestinal flora which
could result in more extensive bacterial degradation of
steroids into products which may have carcinogenic or co—

carcinogenic activity.

12

A number of bacterial metabolites of amino acids
are also suspected as carcinogens or cocarcinogens.
Bacterial production of ammonia, indole and other trypto-
phan metabolites in the large bowel from.unabsorbed protein
are also mentioned in relation to colon carcinogenesis (7,
26).

Dietary fiber may be another component capable of
altering colon carcinogenesis through its effect on gut
function and microflora population and metabolism. It is
well agreed that as fiber in the diet decreased, fecal mass
becomes smaller and intestinal transit time gets longer
(27-29). The small stool mass and the longer transit time
might be very important factors in causing colon cancer.
Any carcinogen ingested or formed in the gut would not only
be present in a more concentrated form in small indigestible
masses, but would also be held in contact with cells for a
prolonged period. Transit time also has been proposed as
one of the factors determining the extent of microbial de-
gradation of compounds. (30) However, human studies indi-
cate that bowel transit time is not a crucial factor in deter-
mining the extent of bacterial metabolism in the gut (31).

There is some epidemiological evidence which shows
that high and low risk groups of people consuming different
dietary regimes have differencies in fecal constituents and
in gut bacteria. Hill and Aries (23,32) studied the excre-

tion and degradation of bile acids and neutral steroids in

13

feces among four populations with different risk rates.
English and Scots who consumed mixed "Western-type" diets
excreted more bile acids and neutral steroids in their feces
than Indians on a rice diet or Ugandans on a matoke diet.
Additionally, bile acids and neutral steroids were more con-
centrated and more degraded in the stools of the English
and Scottish subjects. A higher ratio of total anaerobic
bacteria to total aerobic bacteria also was found in high
risk populations. Anaerobic bacteria metabolize steroids
more actively than the aerobic bacteria (23). Various
studies show that vegetarians have a lower fecal concentra-
tion of bile acids and neutral steroids than omnivorous
people living in the same city (Table 2).

Reddy et a1. (33) compared fecal constituents among
Americans on a typical mixed "Western" diet, American vege-
tarians, American 7th-Day Adventists, Japanese Americans on
a Japanese diet, and Chinese Americans on a Chinese diet.
They found a higher fecal B-glucuronidase activity for
Americans eating a mixed "Western" diet than for the other
populations. Activity of B-glucuronidase, an inducible en-
zyme, is associated with many components of microflora pre-
sent in the feces and therefore used as an indicator of
metabolic activity of gut bacteria (34). Glucuronide forma-
tion is a major detoxification mechanism in animals. Many
endogenous and exogenous compounds that are excreted in bile

as glucuronide conjugates are deconjugated by bacterial

14

.muflo mean can SH mamooa woouo>ficao awaken sues oonmgaoo

.va Hahn scum emuamu<

n

.wonam> Houuaoo mo owmucoouom mm oomwoumxo manmomm

 

 

 

 

Ammo Assam mm SN aumu-wcoq SH nmuaauao>e< Amn-su~
Anny Reuse mm mm aumu-wcoq NH haemaumummo>
AmmV mmss< mm mm anou-waoq ma amammm>
Ammo sauce «we now mamas s A uses
0: cu umoa.£wwm
mouaom moflououm mowom oawm poem mo muoonnnm aowumasaom
Hmuudoz cowumuan mo .02
cowumuuamoaoo
owououm Hooch
Gowumnucoocoo vacuouw
Squaw was no umau mo powwow .N money

15
B-glucuronidase and further modified by intestinal bacteria.
Therefore the difference in B-glucuronidase activity indi-
cates that the intestinal microflora of Americans eating
"Western diet" are more able to hydrolyze glucuronide con-
jugates than microflora of other groups. Americans eating a
"Western diet" excreted more fecal bile acid and neutral
steroid and also these steroids were more extensively de-
graded.

Fecal steroid excretion for cancer and other disease
patients have been compared. Reddy and Wynder (35,36)
reported that there was a greater concentration of total
bile acid and neutral steroids in the feces of colon cancer
and adenomatous polyps patients when compared to patients
with other digestive diseases or to healthy controls. More-
over, the steroids were more extensively degraded in the
feces from cancer or adenomatous polyps patients. Feces
from colon cancer patients had a higher activity of fecal
bacterial 7a-dehydroxylase, the enzyme which converts cholic
and chenodeoxycholic acids to deoxycholic and lithocholic
acids, respectively. Hill et a1. (39) demonstrated that
there were higher levels of both fecal bile acid metabolites
and nuclear dehydrogenating clostridia (NDC) in large bowel
cancer patients than in patients with other diseases. NDC
are able to perform the nuclear dehydrogenation reaction.
It is hypothesized that nuclear dehydrogenation of bile

«Beid can result in the production of 20-methy1 cholanthrene,

a potent carcinogen.(23)

16

It appears that the fat and protein content of diet
can influence the incidence of naturally occurring or chem-
ically induced colon cancer.

Effect of diet on experimentally induced cancer.
Reddy and coworkers (40) investigated the effect of type and
quantity of dietary fat on colon tumor incidence in rats
treated with 1,2-dimethylhydrazine (DMH), an organospecific
carcinogen. Rats fed either a 20% lard or a 20% corn oil
diet were more susceptible to colon tumor induction by DMH
than rats fed 5% of either fats. There was no difference in
tumorigenesis beuwmn the lard and the corn oil fed groups at
the same level of fat. They also studied the biliary and
fecal bile acid and neutral steroid excretion in these rats
(25). Rats fed corn oil or lard at the 20% level showed
Inarkedly increased concentrations of biliary and fecal bile
lacids and neutral steroids compared to rats fed diets contain-
:ing 5% fat. Animals treated with DMH excreted more biliary
and fecal bile acids and neutral steroids than the rats on
Same level of dietary fat without DMH treatment. Since DMl-I
treatment increased bile acid excretion, this result may in-
<iicate that animals subject to tumor development may have
greater excretion of bile acids and neutral steroids.

When beef fat was added to a normal chow diet to 35%
by weight, rats develOped more azoxymethane (AOM) induced
jIntestinal tumors than rats fed the normal chow diet (41).
However, the results of this report should be viewed with

Caution. Since weight gain for animals fed the high fat

l7
diet was greater, the observed difference in tumor induc-
tion may have resulted from variation in calorie intake as
well as fat content of the diet. In addition, adding fat
to the chow diet would have resulted in dilution of all
nutrients, therefore the different tumor production may be
due to dietary deficiencies.

There is some evidence that bile acids may act as
carcinogens. As early as 1940, Cook et a1. (42) reported
that deoxycholic acid dissolved in sesame oil and injected
under the skin induced tumors in mice at the site of injec-
tion. However, Shear (43) could not find any tumors in mice
injected with dry crystals of deoxycholic acid. Since Cook
et a1. did not include a proper control group receiving only
solvent (sesame oil), questions can be posed regarding a
possible role of sesame oil in the carcinogenesis. Later
Lacassagne et a1. (44) showed that apocholic acid produced
tumors in mice at the site of injection. They also found
that 7-dehydrocholesterol and 38-acetOxy-bisnor-AS-cholenic
acid had carcinogenic activity (45). Their control groups
received solvent (olive oil) and did not develop any tumors.

Nigra and others (46) found an increased incidence of
tumor in the large intestine of rats treated with chemical
carcinogens when the animals were fed a 2% cholestyramine-
containing diet compared to rats fed a regular chow diet.
Since cholestyramine binds bile acids and causes more bile
acids to enter into the large intestine, it seems that bile

acids in the large bowel may have a tumor promoting effect.

18
In another experiment, these investigators transplanted the
bile duct in rats from the proximal half to the distal half
of the small intestine of rats to divert more bile acids to
the colon (47). They found more tumors in rats with divert-
ed bile ducts than in the control rats when all were given
the same amount of AOM.

Reddy, Wynder and coworkers demonstrated that deoxy-
cholic acid enhances the carcinogenic effect of N-methyl-N'-
nitro-N-nitrosoguanidine (MNNG) in the large intestine of
germ-free rats (48). When rats were given repeated intra-
rectal doses of sodium deoxycholic acid followed by intra-
rectal instillation of MNNG, they developed more tumors in
the large intestine than rats receiving MNNG alone. They al-
so demonstrated that lithocholic and taurodeoxycholic acids
promoted cancer in conventional rats given intrarectal in-
stillation of MNNG. (49) Sodium cholate and sodium chenode-
oxycholate also showed tumor promoting effects in MNNG induced
colon cancer in both germ-free and conventional rats, however,
these salts of bile acids had greater tumor promoting effect
in conventional rats than germ-free rats. These results in-
dicate that secondary bile acids produced by gut microflora
have greater tumor promoting effects than primary bile acids
(50). None of these bile acids, however, showed carcinogenic
activity when administered without MNNG.

In a recent report (51), cholecystectomized mice
showed an increased rate of colon carcinoma induced by DMH

compared to intact control mice. It was reported that

l9
cholecystectomized animals increase bile acid flow from the
liver to the large intestine without the gallbladder to
store the bile (52). Thus, there would be an increased
concentration of secondary bile acids, produced by bacteria,
in the colon leading to increased carcinogenic or cocarcino-
genic activity.

In a review of the distribution of small bowel
tumors, it was found that most tumors occur in the duodenum
or proximal jejunum.near the entrance of bile duct where the
concentration of bile is the highest. This result also
suggests that bile acids may have a promoting role in tumor
development (53).

Relationship of gut micrgflora to colon cancer. Some
attempts have been made to show a correlation between gut
bacteria and the risks of colon cancer. Reddy et a1 (54)
demonstrated that gut flora have a role in chemical carcino-
genesis by comparing germ-free and conventional rats given
DMH. Conventional rats treated with DMH developed more tumors
in the large bowel than germ-free rats treated with DMH.
They concluded that gut flora had a role in metabolizing DMH
to an active carcinogenic compound.

An association between the level of anaerobic bac-
teria, the extent of steroid degradation and the risk of
colon cancer among different populations has been suggested
(23). However, not all investigators agree. Hill (23) ex-
amined the fecal flora of individuals living in areas with a

high and low risk for colon cancer in an effort to identify

20
species related to the development of cancer. The same broad
group of bacteria were found in the feces of all the popula-
tions studied. However, fecal samples from British and
American subjects yielded more anaerobes than feces of
Ugandans, Indians, and Japanese. The latter low risk popula-
tions had many more facultative and aerobic bacteria in their
feces than did British and Americans. Thus, the ratio of
anaerobes to aerobes was much higher in people eating a
"Western-mixed" diet than in those on a vegetarian diet.

A greater proportion of colon cancer patients had NDC
present in their feces than did control patients with other
diseases (39). Control patients had gastrointestinal dis-
ease and diseases which were not related to gastrointestinal
tract. Higher numbers of bacteroides and clostridial species
were found in feces of people from high risk populations (8,
26). Eubacterium was found in greater numbers in feces from
people from low risk populations. However, Finegold et al.
(55,56) could not find any significant difference in the an~
aerobic species of bacteria between the high risk groups
(Japanese-American on mixed Western diet and polyps patients)
and low risk populations (Japanese-American on a Japanese
diet and patients with other disease). Moore and Holdeman
(57) also failed to find a specific species of bacteria which
is characteristic of high risk populations. Instead, they
reported that individuals in low risk areas tend to maintain
higher concentration of a few species, whereas the high risk

group had a more heterogenous flora in fecal specimens.

21

However, the flora of each person is so complex and the
variation among individuals within a population is so great,
that it is very difficult to point out differences between
groups, even if differences do exist.

Several studies have been conducted in an attempt
to change the composition of fecal flora by dietary modifi—
cation. The inclusion of wheat bran (58), pectin, guar gum,
bananas, plantain bananas, medium chain triglyceride, and
olive oil (59) had no effect on changing the gut flora.

Reddy et a1. (37) studied the composition of fecal
flora from volunteers who ate a high meat "mixed-Western"
diet or a non-meat diet. Both diets had the same protein
content but different fat levels. Total number of anaerobic
fecal bacteria were higher in specimens from the volunteers
while they were consuming the high meat diet than during the
period when the volunteers consumed the non-meat diet.
Henteges et a1. (61) conducted a similar experiment except
their high meat and non-meat diet had the same fat level but
different protein levels. No differences were found in the
total number of anaerobes in the feces of the subjects dur-
ing the suxfies of either diet. The discrepancies between
these two studie may indicate that the total fat content of
the diet, not protein, has an effect on changing the fecal
1microflora. Differences in methodology may also have pro-
duced the discrepancy.

I Even though number and type of microflora may not

«differ between the high and low risk populations, the

22
metabolic activities of flora are reported to differ. A
high proportion of the anaerobes isolated from stools from
people from high incidence areas were able to dehydroxylate
bile acids. The prOportions were very small when the or-
ganisms were isolated from.stools from the low incidence
population (8). Also a high proportion of the clostridia
isolated from stools of people from high risk populations
were able to dehydrogenate the steroid nucleus.

The metabolic activities of gut microflora can be
changed by diet. The activity of B-glucuronidase produced
by bacteria can be increased by increasing the level of fat
and meat in diet (35,62). Changes in the diet have caused
changes in the ability of flora to metabolize various chem-
icals in experimental animals (63).

Methods for studying carcinogenesis and mutagenesis.

 

At this point in colon cancer research, there is an urgent
need to identify compounds which are capable of causing
cancer. Only then can a cause and effect relationship be-
tween carcinogenic agents and diet be made, which in turn
will lead to a better understanding of the role of diet in
the prevention or development of cancer.

Even though epidemiological data indicate that diet-
ary differences are associated with the development of
cancer, it has not been demonstrated that tumor formation
can be caused by dietary manipulation in animals or man.
JMoreover, it is not likely that it will be feasible to pro-
‘duee colon cancer by dietary manipulation in the future.

Chae of the plausible methods to study the effects of diet on

23
carcinogenesis is to study chemical carcinogenesis in ani-
mals fed various diets. However, the interaction between
carcinogen and the gut components are hard to predict.
Even if a difference was found in carcinogenesis between
groups of animals fed different diets, it still would be
subject to argument what component of the diet or what meta-
bolite induced by diet was responsible for the difference.
The intrarectal instillation of test compounds into rats as
used by wynder and others (48-50) does permit identifica-
tion of carcinogens and cocarcinogens. However, all these
methods of animal bioassays are too time consuming (6 months
to 2 years) and expensive to use as screening methods to
test all suspected carcinogenic compounds. There is a des-
perate need for a rapid and reliable in_vi££g_method for
screening purposes.

In the past several years, many attempts have been
made to develop reliable in vitgg methods to test viral and
chemical agents responsible for tumor induction. Cell cul-
ture techniques have been used in many laboratories for
screening carcinogens and for elucidating fundamental cellu-
lar and molecular mechanisms of carcinogenesis. The major-
ity of the cell culture work has been conducted with fibro-
blasts which produce sarcomas upon inoculation into host
.animals. Most human cancers originate from epithelial
‘tissues, and it would obviously be desirable to study epithel-

:ial cells in culture. However, researchers have not yet

24
successfully grown epithelial cells in culture for subse-
quent demonstration that they can be transformed.

There are two common cell types used for the demon—
stration of transformation. One is primary and secondary
cultures from Syrian hamster embryo cells (64). The other
is permanent mouse cell lines derived from C3H mouse pro-
state. The hamster embryo cell system has the advantage
that the cells are diploid, whereas the prostate cells are
aneuploid. However, the hamster cells have a short life time
and low cloning efficiency, while the mouse prostate cells
have an unlimited life span and very high cloning efficiency.
There are other cell types used by various laboratories.
C3H/10T1/2 cells derived from C3H mouse embryos, Balb/3Ts cell
lines from mouse embryo cells, Chinese hamster lung cells,
and rat embryo fibroblasts are some of the commonly used
cell types.

The biggest limitation in using cell culture tech-
niques to demonstrate cell transformation is that no univer-
sal marker exists to differentiate transformed cells from
normal cells. The most frequently used markers for the re-
cognition of transformed cells are: 1) changes in morphol-
ogy of cells and colonies, 2) changes in growth properties,
3) chromosomal aberrations, 4) immortality in culture, and
5) ability to produce tumors when injected into syngenetic
animals. In the first method, transformed cells are charac-
terized by criss-cross random cell orientation, increased

:number of nucleoli per cell, and changes in nuclear-cytoplasmic

25
ratio. In the second method, neoplastic cells have the
ability to grow in fluid suspension, whereas most normal
cells grow only if they are attached to a surface. The
transformed cells also have the capacity to produce three
dimensional colonies in soft agar and to grow with reduced
concentrations of serum. Tumor cells lose the density-
dependent inhibition of cell division and form piled up
foci after confluency is reached. On the other hand normal
cells stop growing after confluency has been reached and
form only monolayers of cells. Scoring these piled-up foci
has been most widely used to quantitate transformation in
cell culture. In the third method, there is a correlation
between chromosomal abnormalities and development of cancer
(65). By chromosome stains, the chromosomal changes in the
cell can be detected. In the fourth method, cell longevity
is the criteria. Transformed cells demonstrate immortal-
ity in culture, while normal cells have a finite life span.
The fifth method is the ultimate criterion for demonstration
of malignant transformation. If cells suspected as being
transformed can produce tumors in a host animal, then one
can safely state that the cells were transformed.

All the above criteria, except the ability to pro-
duce tumors in animals, have been observed to some extent in
normal cells. Therefore artifacts and the bias of investi-
gators are greatly involved (66).

One of the more promising methods for screening

chemical agents for potential carcinogenic activity is to

26
determine the mutation frequencies caused by suspected
agents in microbial or in cell culture systems. It is
frequently observed that many known carcinogens also induce
mutation in cell cultures and in microorganisms (67, 68).

It is generally thought that carcinogenesis is a two
or multi-step process (69-71), consisting of initiation
and promotion stages. Many feel that the initiation step
is a mutation involving DNA damage and is an irreversible
event. The following promotion stage is a reversible pro-
cess involving gene modulation. There are two classes of
carcinogens: one class contains mutagens which will cause
the initial irreversible DNA damage, and the other class
contains promotors which will enhance tumor expression. The
initial mutation may occur in germ cells, in which case
tumors may be inherited; or it may occur in somatic cells,
in which case the tumor is not inherited. (69)

One of the most widely accepted methods to score
mutation is the method developed by Bruce Ames and co-
workers (72). They developed several series of histidine-
requiring auxotrophs of Salmonella typhimurium.(TA-1535,
TA91537, TA-1538, and TA-lOO) by specific mutation. These
.strains are easily reverted in response to a wide range of
<2hemical mutagens. By omitting histidine in culture medium,
t:he mutagen-induced revertant colonies, which do not require
Piistidine for growth, can be scored easily. Initially, only
a limited number of known carcinogens appeared to be muta-

géinic in this system. However, with the understanding that

27
a carcinogen may have to be activated by an in 2329 system
before it can act as an ultimate carcinogen, the bacterial
system was coupled with mammalian microsomal enzyme prepa-
rations to metabolize the procarcinogen to an ultimate
carcinogen. ‘With this coupled system, more and more known
carcinogens are able to cause mutation in bacteria. The
biggest limitation in this system is that the carcinogens
which are not initiators (i.e. promoters) will not be
detected. Some strains of E, 321;, yeast, fungi, and animal
cells also have been used for mutation assays using differ-
ent types of selective systems.

It seems more reasonable to use a mammalian cell
system rather than a microbial cell system to study muta-
genesis and/or carcinogenesis, if a specific and selective
method could be developed. There are three major types of
mutants isolated in mammalian cell culture, namely, tempera-
ture-sensitive conditional lethal mutants, nutritional auxo-
trophs and drug-resistant phenotypes. Temperature-sensitive
mutants have a smaller range of temperatures in which they
can grow compared to wild-type cells. In other words, the
mutants cannot grow either at the lower end of the normal
temperature range or at the higher end of the normal tempera-
ture range. This defect probably arises from a change in
the thermal kinetics of denaturation or assembly of a parti-
cular protein. Nutritional auxotrophs need certain nutrients
for growth which is not necessary for the wild-type cells.

This requirement is due to mutation which causes the

28

inactivation of an enzyme in the biosynthetic pathway of the
certain nutrients, such as amino acids or vitamins.

Drug—resistant mutants can grow in the presence of
a drug which is cytotoxic to normal cells. Drug-resistant
mutants have been isolated from the mammalian cells. Muta-
tion in the gene for hypoxanthine-guanine phosphoribosyl
transferase (HGPRTase) leads to cells which are resistant to
such purine analogues as 8-azaguanine and 6-thioguanine.
Mutation in the gene for the plasma membrane Na/K ATPase
leads to ouabain resistant cells and mutation in the gene for
protein kinase leads to dibutyryl cyclic AMP(Bt2AMP) resist-
ant cells (73). Mutants which are resistant to 5-bromodeoxy-
uridine (BrdU) and to colchicine have also been isolated (14).
BrdU, a thymidine analogue, resistant cells have decreased
thymidine kinase activity. Colchicine binds to a protein
subunit of cellular microtubles and therefore interferes with
the formation of microtubles. A wide variety of other drugs
have also been used to select mutants in culture (74).
Wild-type cells die when they are incubated with these drugs,
therefore, drug-resistant mutants can be selected by adding
the drug in the culture medium.

One of the major limitations in mutagenic studies is
that only a relatively small number of the mutant forms are
identified. The nutrition auxotrophs and drug-resistant
mutants require the hit of a mutagen on a restricted number
of target genes. If the hit occurs at another site, the

event will not be scored because of the lack of a suitable

29
detection system. Another drawback is that there are muta-
gens which do not have tumor producing effects and there
are carcinogens which do not show mutagenic activity. Many
more classes of compounds appear to be mutagenic than
carcinogenic (65).

Until recently, cancer promoters were not detected
in mutation frequency assay system. Trosko et al. (75) very
nicely demonstrated the effect of promotors in the V-79
Chinese hamster cell culture by using the ouabian and 8-
azaguanine systems.

I chose to use cells isolated from the intestine of
rat fetuses to develop a model system to screen agents that
are possibly responsible for colon cancer. Ouabain resist-
ance was the criteria chosen to determine mutation frequency.
Cells resistant to ouabain have been isolated from.mouse
and Chinese hamster cells (76), from human lymphoblasts (77,
78), Ehrlich ascites cell (79),and Hela cells (80).

Ouabain, a steroid glycoside, is a specific inhibi-
tor of the plasma membrane Na/K ATPase. Inhibition of this
enzyme decreases the active transport of potassium and cells
exposed to ouabain die due to potassium starvation. Ouabain
resistance results from a mutation in the structural gene
for ATPase due to a point mutation. The point mutation
results in a single amino acid substitution, which allows
the enzyme to be active in the presence of ouabain. Mutant
cells could: 1) have another K+-transport site with an alter-

ed response to ouabain, 2) lose all or part of the ouabain

30
’receptors in membrane, or 3) have a decreased affinity for
ouabain (74, 77).

Point mutation is due to nucleotide base-pair sub-
stitution at a speific site in the wild-type gene, such as
GC in the wild-type is replaced by AT in the mutant. Point
mutation results in a functional protein with altered
specificity due to a single amino acid substitution. By
treating cells with base-pair substitution mutagens (ethyl
methane sulphonate, MNNG, UV light), the frequency of
ouabain-resistant cells increase markedly (76, 78).

Unlike the 8-azaguanine system, selection of ouabain-
resistant mutants is not dependent upon the cell density
seeded (74) and the expression time is shorter (75). However
ouabain-resistant cells are not observed when cells are
treated with a frameshift mutagen, such as ICR-l9l. In
frameshift mutations, one or a few nucleotide bases are in-
serted into, or deleted from, a specific site within the gene.
Therefore they shift the "reading frame" of the translation
process, leading to a non-functional protein. Frame shift
mutation can be detected easily by using 8-azaguanine as a
selective marker.

The inducibility of mutation frequency by mutagens
has been used by many investigators to score mutation fre-
quency induced by environmental mutagens/carcinogens for
the purpose of screening mutagenic/carcinogenic agents and for
better understanding of the mechanisms. Most of these muta-

tion studies have been done with an established cell line

31
such as 3T3 cells andV—79 Chinese hamster cells.

I have used freshly isolated fibroblasts from the
intestine of rat fetuses since using these cells has the
advantage of utilizing cells from the target tissue. It
would be more appropriate to use intestinal epithelial
cells, but intestinal epithelial cells have not been success-
fully grown for transformation studies. Some workers have
successfully isolated intestinal epithelial cells for meta-
bolic studies, but it has not yet been possible to grow them

in vitro for a prolonged period of time.

MATERIALS AND METHODS

PART I

 

Part I was designed to study the effects of dietary
fat and fiber on steroid metabolism, intestinal transit time
and intestinal microflora in rats. In experiment 2, organic
and water soluble extracts were prepared to use for cell

culture experiments in Part II.

Experiment 1

In experiment 1, the effects of wheat bran, the most
frequently studied dietary fiber, was evaluated.

Animals and housing. Male Sprague-Dawley rats
weighing approximately 230 g were divided into 4 groups of
11 rats each. The rats were housed individually in stain-
less steel cages with raised wire floors. The animal room
was maintained at 20t1°C with controlled humidity and a 12-
hour light-dark cycle.

Qiegg. The diet composition is shown in Table 3.
Three levels of wheat bran, 2 g (Diet 2), 4 g (Diet 3), and
8 g (Diet 4), were isocalorically substituted for cornstarch
in the control diet (Diet 1). The caloric values of wheat
bran (2.13 kcal/g) and of other dietary nutrients were taken

from Composition of Foods, Agricultural Handbook No. 8, USDA,

32

33

TABLE 3. Composition of diets for experiment I
(g/lOO kcal)

Diet 1 2i2£_3 2£2£_§. 2322.1
Na-caseinate 4.7 4.2 4.2 3.3
Vitamin mixa 0.2 0.2 0.2 0.2
Mineral mixb 0.9 0.9 0.9 0.9
Lardc 3.3 3.3 3.3 3.3
Corn 0i1d 1.3 1.1 1.1 0.9
Cholesterol 0.2 0.2 0.2 0.2
L-methionine 0.07 0.07 0.07 0.07
Choline-Cl 0.04 0.04 0.04 0.04
Cornstarch ll 6 10 9 10.0 8.4
Wheat Brane - 2.0 4.0 8.0

 

aThe vitamin mix was composed of thiamin HCl 11 g; pyridox-
ine 11g; riboflavin llg; CaPantothenate 33 g; p- -aminoben-
zoic acid 55 g; menadione 25 g; inositol 50 g; ascorbic
acid 100 g; niacin 50 g; vitamin B 0.015 g; biotin 0.3 g;
folic acid 2 g; retinol acetate 1x107 IU; a-tocopherol
50,000 IU;vitamin D3 110,000 IU and cerelose to 5 kg.

bThe mineral mix (salt mix No. 4164, Teklad, Inc., Madison,

WI.) contained 6/100 g: calcium acetate H O, 6.293; calcium
diphosphate 2H2 28. 525, dipotassium phogphate, 28.443;
ferric citrate-25H20, 2.44; magnesium sulfate- 7H 0, 10. 053;
potassium iodide, 0.65; sodium diphosphate 12H2 O, 14. 630;
sodium chloride, 9.546.

cFarmer Peet's Shortenin'-Pork and beef fat-Peet Packing
Company, Chesaning, Michigan.

dMazola oil

eFeed grade wheat bran

34
1963 (82). These three levels of wheat bran were added to
provide 1.1, 2.2, and 4.4 g of neutral detergent residue
(NDR) per 100 kcal for diets 2,3 and 4, respectively. The
neutral detergent residue of wheat bran was analyzed accord-
ing to Van Soest and Wine (83). These diets were fed for
4 weeks (period I) and then all animals were switched to the
bran-free control diet for another 4 weeks (period II).

Analyses. In the beginning of the 4th week of period
I, 100 mg of chromic oxide was administered intragastric-
ally to 6 rats in each group to measure intestinal transit
time. Fecal collections were made every 8 hours for the
first day and then every 12 hours for the next 2 days. The
feces were dried to a constant weight at 60°C in a forced air
oven and stored for chromic oxide analysis. Chromic oxide
was determined according to the method of Bolin et a1. (84),
and transit time was designated as the time required for 95%
of ingested chromic oxide to be excreted.

During the 4th week of periods I and II, feces were
collected for 3 successive 24 hour periods and dried as above.
we: and dry weights were determined gravimetrically. The
samples were pooled by group for bile acid analysis.

Bile acids were extracted from aliquots of pooled
feces as described by Grundy et al. (85). Cholic acid (24-14C)
was added to each sample at the beginning of the analysis as
an internal standard to correct for incomplete recoveries

during extraction. Separation of bile acids from fatty acids

35

was as described by Makita and Wells (86). Methyl esters

of bile acids were prepared with excess of ethereal—
diazomethane. The methylated bile acids were dried,
silylated (85) and trimethylsilyl ethers (TMS ethers) were
quantitated on a GLC equipped with a flame ionization detec-
tor. The column temperature was 210°C for 20 minutes after
sample injection and it was programmed at 8°C/min to 270°C
and held at 270°C until the last bile acid was eluted.
Holding the temperature at 210°C for 20 minutes allowed
remaining fatty acid methyl esters to be separated from the
TMS ethers of bile acids. The injector temperature was 280°C;
detector temperature, 290°C; H2 flow, 30 ml/min and N2
carrier gas flow, 35 m1/min. The column was 6 ft long, 1/8
in O.D. and packed with 3% SP240l on Supelcoport 100/120.
Cholic, lithocholic, deoxycholic, chenodeoxycholic, 12-
keto-deoxycholic and 3,12-diketo-deoxycho1ic acid standards
were methylated and silylated as above with S-a-cholestane

as an internal standard. The area under the recorder trac-
ing for each standard was cut out and weighed. Standard
curves were prepared by plotting the ratio of bile acids to
S-a-cholestane vs. the quantity of bile acids injected. The
quantity of bile acids excreted in the feces was determined
from.the standard curves. An average standard curve was used
to quantitate all bile acids. Bile acids are reported as

the sum.of all bile acids. Bile acids which did not chroma-
tograph with cholic and chenodeoxycholic acids are designated

as degraded bile acids.

36

At the end of the 4th week of each experimental
periods I and II, cecums from 3 animals from.aach group were
removed for anaerobic bacterial count and in vitae bile acid
fermentation. A portion of cecal contents from each rat was
serially diluted to 10.9 concentration and incubated in an-
aerobic roll tubes at 37°C for 5 days using M98-5 medium
without hemin as described by Salanitro et a1. (87).
Colonies were counted after a 5-day incubation period. This
procedure was conducted as soon as the rats were killed and
all steps were performed aseptically and anaerobically.
Cholic acid fermentation was determined by incubating 1 m1
of the 10"1 dilution of cecal contents with 0.1 mmole of
cholic acid, 10 ml of dilution solution and l g of diet
which the animal had been eating. The fermentations were
conducted anaerobically in 50 ml serum bottles at 37°C for
24 hours. Fermentation was stopped by injecting 2‘ml of
concentrated HCl into the serum bottles. Blank fermentations
were carried out by injecting sterile dilution solution in-
stead of cecal dilutions into the serum bottles. Microbial
degradation of cholic acid was designated as the quantity
of cholic acid fermented to other bile acids minus blank

fermentations. Bile acids were extracted from.the fermenta—

tion reaction and analyzed as described for the fecal samples.

Experiment 2

 

In experiment 2, male Sprague-Dawley rats (average
260 g) were allotted into 2 groups of 10 rats each. The condi-

tions of animal care were same as experiment 1.

37

Diggg. Composition of the diets is shown in Table
4. Fat provided 45% of the energy in the high-fat, low-fiber
diet and there was no added fiber. Fat provided 7% of the
energy in the low-fat, high-fiber group and 2 g of agar per
100 kcal was added as fiber source. Agar provided 7% fiber
by weight and the caloric value of agar was excluded in the
energy calculation. The rats were fed the diets for 4 weeks.

Analyses. After the diets were fed for 3 weeks,

 

feces were collected for one week at 9 a;m, each morning.
The fecal samples were kept frozen at -20°C until all the
collections were made and then the daily collections were
pooled for each animal. The feces were dried as in experi-
ment 1. wet and dry weights were determined gravimetrically.
Bile acids and neutral steroids were extracted frmm
aliquots of pooled feces after saponification with 20% KOH
in ethylene glycol as described by Evrad and Janssen (88).

3H-cholesterol and cholic acid (24-14

C) were added to each
sample at the beginning of the analysis as internal stand-
ards to correct for incomplete recoveries during extraction.
5-a-cholestane was added to the neutral steroids as
an internal standard for GLC analysis and the dried steroids
were silylated as described for the bile acids in experiment
1. Cholesterol, coprostanol and coprostanone standards were
silylated and chromatographed with 5-a-cholestane as an

internal standard. Neutral steroid separation was on 6 ft.x

1/8 in.column packed with 3% 0V 17 on 100/120 Gas Chrome Q.

38

 

 

TABLE 4. Composition of diets in.experiment 2
(8/100 kcal)
Diets

High-fat, Low-fat,

low-fiber high—fiber
Na-caseinate 4.0 4.0
Vitamin Mixa 0. 2 0. 2
Mineral Mixb 0. 8 0. 8
Tallowc 4.2 -
Corn 0i1d 0.8 0.8
Cholesterol 0.1 -
L-methionine 0.06 0.06
Choline-C1 0.06 0.06
Cornstarch 9.8 19
Agare - 2

 

8See footnote a, Table 3

b

See footnote b, Table 3

cBeef fat was melted and filtered through cheese cloth.

dMazola oil

eDifco Bacto-Agar, Difco Laboratories, Detroit, Michigan.

39

The injector temperature, detector temperature, and column
temperature were 270°C, 290°C, and 250°C, respectively.

All other conditions were the same as described for bile
acid separation in experiment 1. A Varian CDS 111 electri-
cal integrator was used to quantitate bile acid and neutral
steroid content of the samples. Neutral steroid excretion
is expressed as the sum of cholesterol, coprostanol and
coprostanone excretion.

The separation of bile acids from fatty acids,
methylation and silylation of bile acids were done as in
experiment 1. The bile acids were separated on a 6 ft.long
x 1/8 in. O.D. column packed with.3%.SP2100 on Supelcoport
100/120.

The column temperature was 210°C for 10 minutes
after sample injection and then the temperature was programm-
ed 4°C/min to 270°C and held at 270°C until all bile acids
were eluted. Other GLC conditions were same as described
for experiment 1.

After steroid analyses, all feces were pooled by
group and the feces were extracted with chloroform and water
for use in the second part of experiment. The feces were
acidified with HCl in an aqueous slurry and excess chloro-
form was added. The mixture was mixed well in a separatory
funnel, and the chloroform layer was collected. This proce-
dure was repeated until no apparent color was extracted with
chloroform. The chloroform extract was dried in a flash

evaporator. The extract was transferred to a screw cap tube

40

‘with chloroform and dried under a stream.of nitrogen. Ali-
quots of the chloroform extracts were analyzed for bile
acid and neutral steroid contents as above.

To prepare the water extract of the feces excess
water was added to the feces and heated at 60°C for 1 hour.
Heating appeared to extract more compounds. The resultant
slurry was filtered through Whatman #1 filter paper and the
collected filtrate was dried in a flash evaporator. Both
dried fecal and water extracts were kept at 4°C until used.

At the end of the 4 week feeding period, the bile
duct was cannulated with polyethylene tubing (PE 10, Intra-
mediégz Parsippany, N.J., I.D. 0.011", O.D. 0.024"). The
rats were anaesthetized by ether. One end of the tubing was
tied ill place in the bile duct and the other end of the
tubing was routed under the skin to the midpoint of the
back and pulled out through a small puncture in the skin.
The tubing was taped in a small test tube and the test tube
was placed on back of the rat with tape in such a way that
the collected bile in tube would not be spilled by the move-
ment of rat. Bile was emptied at necessary intervals from
the tube. Bile was collected for 24 hours and food and water
were available to the rats during the bile collection. Bile
was kept frozen at -20°C for steroid analysis. Bile acids
and neutral steroids were extracted from aliquots of bile
from each rat and analyzed by GLC as described for fecal
samples. After steroid analysis, the bile was pooled by
group and dried in a vacuum oven at 60°C with a nitrogen at-

‘mosphere and stored at 4°C for the experiment in part II.

41
PART II

Part II was designed to develop a rapid, in vitro
method to screen agents which could be responsible for the
cause and development of colon cancer. Several cell cul-
ture techniques were evaluated to screen potential intesti-
nal mutagens.

1. Culture Media and Growth Conditions

Unless stated otherwise, the growth medium was made
with Eagle's Minimum.Essential Medium (EMEM) supplemented
with 10% fetal calf serum, antibiotics and fungicide (125
units/ml of penicillin, 125 Lg/ml of streptomycin, 3.44 ug/ml
of fungizone). All of the above were purchased from Grand
Island Biological Co., Grand Island, N.Y.

Buffered Saline Solution (BSS) contained 137 mM
NaCl,2.7 mM KCl, 1.0 mM CaClz, 1.0 mM MgC12-6H20, 0.15 mM
NaH2P04-H20, 1.36 mM NazHP04-7H20, 6mM NaHCO3, 5.5 mM
glucose and phenol red as a pH indicator. The pH was ad-
justed to 7.4 with 002 gas. CaC12 and MgC12'6H20 were
omitted from the B88 to prepare Ca++, Mg++ -free BSS which
was used to prepare a solution for trypsinizing cells.

Ouabain (2mM, Sigma Chemical Co., St. Louis, Mo.)
containing EMEM was prepared in batch quantities, sterilized
by filtering through a 0.22 micron filter and stored at 4°C.
This was diluted and reconstituted with 10% fetal calf
serum and antibiotics as needed. Ouabain has been reported

to be stable in solution for periods of many months (76).

42

Cultures were incubated in growth medium at 370C2h1a humi-
dified atmosphere of 5% CO2 and 95% air.

To stain and count colonies, cultures were rinsed
3 times with 388, fixed in absolute methanol for 5 minutes,
and stained with Giemsa stain for 20 minutes at room temp-
erature. After staining, the colonies were counted with a
stereo microsc0pe (lO-30X).

2. Preparation of Reagents Used.

 

Namethyl-N'-nitro-N-nitrosoguanidine (MNNG) was dis-
solved in sterile distilled water (20 ug MNNG/50 ml water).
The MNNG solution was added to culture dishes containing
growth medium to provide indicated concentrations. No
further attempt was made to sterilize the MNNG solution.

30, lZa-dihydroxy-SB-cholanic acid (deoxycholic
acid), 3a-hydroxy-58-cholanic acid (lithocholic acid), 30,
6a-dihydroxy-58-cholanic acid (hyodeoxycholic acid) and 3,
12-dione-5B-cholanic acid were dissolved in dimethyl sulf-
oxide (DMSO). The solutions were filter sterilized with a
0.2 micron FGLP solvent resistant filter (Millipore Corp-
oration, Bedford, Ma.). These sterile solutions were added
to culture dishes to provide indicated concentrations. The
final DMSO concentration was less than 0.5%.

3a, 7a, lZa-trihydroxy-SB cholanic acid (cholic
acid) and 3a, 7a-dihydroxy-58-cholanic acid (chenodeoxy-
cholic acid) were neutralized with l N NaOH, made SmM by
dissolving in growth medium and filter sterilized. These
solutions were diluted with growth medium to the indicated

concentrations before use. If cholic and chenodeoxycholic

43

acids were dissolved in DMSO and added to the culture dishes
containing growth medium, the DMSO solutions would not go
into solution or suspensions. Instead, DMSO solutions of
cholic and chenodeoxycholic acids caused protein coagulation
and pH alterations of the growth medium.

Deoxycholic and 3,12-dione-58-cholanic acids were pur-
chased from Steraloids, Inc., Wilton, N.H. Cholic and hyo-
deoxycholic acids were purchased from Sigma Chemical Co.,

St. Louis, Mo., and lithocholic and chenodeoxycholic acids
were purchased from Aldrich Chemical Co., Milwaukee, Wis.

The chloroform extract of feces from experiment 2 was
weighed, dissolved in DMSO and sterilized by filtration.
Also, the fecal water extract was weighed, dissolved in
water and filter sterilized.

Dried bile was reconstituted to the original volume
with growth medium.and filter sterilized. This bile contain-
ing medium was diluted with growth medium.to provide various
concentrations before use.

3. Isolation of Intestinal Epithelial Cells
First Approach: Sprague-Dawley rats were killed and the
large intestine was removed. The large intestine was cut
longitudinally and the inside was rinsed with BSS supple-
mented with antibiotics and fungicide (125 units/ml of peni-
cillin, 125 ug/ml of streptomycin, 50 ug/ml of gentamycin,
3.44 ug/ml of fungizone).

The mucosal surface of the intestine was removed

from the muscular layers by scraping with a scalpel. The

44

resultant tissue was trypsinized with 0.1% trypsin for 3
‘minutes to dissociate cells. This was centrifuged 1500 x
g for 3 minutes and the precipitate was suspended in cul-
ture medium (EMEM supplemented with 10% horse serum and
same concentrations of antibiotics as in B88) and plated in
a culture dish (Falcon and Corning). After 24 hours, and
then daily, the culture medium was discarded, the cells
were rinsed with B88 and fresh culture medium was added.
Second Approach: Tissue culture/cell culture technique.
Small pieces (approximately 2x2 mm) were cut from
the large intestine of a rat. These pieces were laid on a
plastic culture dish mucosal side down without medium” The
dish was rinsed with growth medium (EMEM supplemented with
10% fetal calf serum and same concentrations of antibiotics
as above) before use. The dish which contained several
pieces of intestinal explants was placed in the incubator
for 1 hour during which time the explants were allowed to
attach to the plastic surface. After 1 hour of incubation,
the growth medium was very carefully added to the dish so
that the attached pieces were not disturbed. The dish was
incubated for 4 days without medium change or disturbance.
After 4 days of incubation, the culture dish was rinsed and
fresh culture medium was added daily thereafter. Mbst of
the tissue was washed away with the daily rinsing. Epi-
thelial cells grew out from where the explants had attached

to the dish.

45

Third Approach: Epithelial cells were isolated from the

 

large intestine of Sprague-Dawley rats according to the
procedure presented by Weiser (89). Citrate, not enzyme,
was used to dissociate the epithelial cells. A piece of
sterile tubing (Bardic Feeding Tube, size 8, C.R. Bard Ind.,
Murray Hill, N.J.) was inserted in each end of the large
intestine and tied in place to fill and remove solutions
with a syringe. All the solutions were made as described
(89), filter sterilized and antibiotics and fungicide were
added (150 units/ml of penicillin, 150 ug/ml of strepto-
mycin, 50 ug/ml of gentamycin, 3.51 ug/ml of fungizone).
The crypt and villus cells were isolated together by incubat-
ing the intestine for 52 minutes with solution B.

4. Isolation of Intestinal Fibroblasts.

Pregnant Sprague-Dawley rats in the 18-19 day of
gestation were purchased. Both the small and large intes-
tine was taken from the rat fetus. The intestine was chopped
into small pieces and put into 0.2% trypsin in Ca++3Mg++ -free
B88 in a sterile culture tube for 5 minutes. The tube was
mixed vigorously during this time with a vortex mixer to
dissociate the cells. After centrifugation for 3 minutes,
the precipitate was resuspended in growth medium.and plated.
.After 6 hours, the medium was changed since there was a
yellow secretion in the intestine which greatly decreased
cell viability. Thereafter, the cells were washed and cul-

ture medium was changed daily. The attached cells were tryp-

sinized with 0.2% trypsin whenever necessary to prevent

46

confluency. The trypsinized cells were collected by centri-
fugation and replated until there were a sufficient number
of fibroblasts to conduct an experiment.

5. Protocols of Experiments.

 

Fibroblasts from.the intestine of a rat fetus were
subcultured 2-5 times before the experimental treatment.
The fibroblasts were plated in 15 cm plates and grown to
about 20% confluency.

On the day of treatment, the compounds at designated
concentrations were added to the plates and incubated for
3 hours. After the 3 hour incubation, the cultures were
washed with BSS twice and growth medium was added to the
plate. The medium was changed daily for 3 days (expression
time). After 3 days cells were trypsinized and counted with
a hemocytometer. Plating efficiency was determined by seed-
ing 200 cells in 6 cm plates with growth medium. Triplicate
dishes were used for measuring plating efficiency. Culture
medium was changed every 3 days. After 8 days, cells were
fixed, stained and counted. Plating efficiency was calcu-

lated as follows:

Plating efficiency g Average_number agocolonies in plate (A)

 

Ouabain resistant mutants (Oua-R.mutants) were deter-
mined by plating 3 x 104 cells in 6 cm plates with 1.0 mM
ouabain containing medium. Four to 8 replicate dishes were
used. Medium was changed every 3 days. After 14 days the

cells were fixed and stained, and colonies were counted.

47

Oua-R mutants (mutation frequency) was calculated per 105

survivors, taking into account the plating efficiency of

each plate as follows:

5
5 . _ 10 ,
Oua-R mutants/10 surv1vors — (pIating efficiency)x(3x10‘*)x

average number of ouabain
resistant colonies in muta-
genesis plates (B)

To determine optimal ouabain concentration and ex-
pression time to select Oua-R mutants, different concentra-
tions of ouabain in medium and different times for the ex-
pression period were used as indicated. Protocols for the
determination of plating efficiency (A) and mutagenesis (B)
are summarized in Figure 2.

Most mutagenic compounds are cytotoxic so that per-
cent cell survival needed to be determined. Percent survi-
val for the test compounds was determined by plating 200
cells in 6 cm plates. 16-18 hours later, the cells were in-
cubated with various concentrations of the test compounds
for 3 hours. The culture dishes were then washed twice with
388 and fresh growth medium was added to the dish. Medium
Was changed every 3 days and the cells were fixed and stained

after 8 days. Percent survival was calculated as follows:

73. survival ___ Average number ogogolonies in dishes (C)x 100%

Figure 2 also shows the protocol for determination of cyto-

1lioxicity for the various test compounds.

48

Figure 2. Protocols for (A) plating efficiency, (B) muta-

genesis, and (C) cytotoxicity.

49

. N gUHh

 

 

women 8525:. women woman So 0

ooafiwum mm ~00me aw oonmmfi ogomaoo \mHHoo com.
m llmhmv mlv—Tmudon mlllwrmufion ma: :oHIL

on

 

 

 

 

 

 

 

modicum 85.308 awmomao
a_naxam zs_H ea annea\aaana aoexm Ame
Tl name an
poops 5258 moved mound” :8 n." 5”
oouwafimarfifi aw owned” mpaaomaoo cousin 9300
Alllumtnww mlwfmuaon m "330qu
oocfimumm w modem. whom will

 

 

fl

_asanes eusonw ta eunaa\nnaee com A<v

50

The tumor promoting effect of bile acids were tested
as described by Trosko et a1. (75) with some modifications.
Cultures were treated with 1.5 ug MNNG/m1 of medium for 3
hours, washed with BSS and growth medium.was added. After
the expression period, cells were trypsinized and plated for
the determination of plating efficiency and Oua-R.mutant
selection as described above. At this time various bile
acids, at indicated concentrations, were added to the plates
used for both plating efficiency and mutagenesis determina-
tion. Medium was changed every 3 days and each time bile
acids were added. After 2 weeks, cultures were fixed,
stained and colonies were counted.

6. Doubling Time Determination.

Fibroblasts in 15 microscopic fields per plate were
counted every 24 hours for 3 consecutive days. The average
number of cells were plotted against time to determine doubl-
ing time. Doubling times were determined on two different
occasions with a different source of cells on each occasion.

Triplicate plates were utilized on each ocassion.

RESULTS

PART I

Experiment 1

 

Energy consumption and weight gain were similar
among groups throughout the experimental periods (Table 5).
Both wet and dry fecal mass were increased as the level of
bran in the diet increased. When the rats were switched to
the bran-free diet (period II), fecal mass decreased to
levels similar to those in controls in period I. Intesti-
nal transit time was inversely proportional to the level of
bran in the diet. No significant differences in total bile
acid excretion were found among rats fed different levels of
bran containing diets. However, fecal concentration of bile
acid (mg ”of bile acids per g of wet feces) decreased as the
level of bran in the diet increased since fecal weights
were greater when bran was added in the diet. The fecal
bile acid concentration decreased by 25, 50 and 65% of con-
trol group for diets 2 (2 g of bran per 100 kcal), 3 (4g of
bran), and 4 (8 g of bran), respectively. The percent of
fecal bile acids that were degraded increased with level of
bran in the diet. When rats were fed the control diet during
period II this difference in bile acid degradation dis-

appeared. Cholic acid degradation also increased in the

51

52

 

onwan «men Amen amen oaea>.mm
Nwmm n+~m ~+~m ~+en naaem
HH ooauom 0.:owumompwoo N
mWee aWem HHWaa NHmAN onea>.mm
ao+ma «n+me a+an m+ea anode
H ooauom o.cowumomuwoo N
im.a«o.a e.Hmm.m m.a«o.m w.~-a.oe anode ens w\ws .noaenaedaadoa
o.mwa.na e.a+a.me a.HwH.mH e.~wo.aa ann\aae\wa .doaaanaae
H unease .anaan when Hanan
aa.me.H~ aa.me.an a.oma.~e o.nwo.me anaon mane aaaanna
ao.omn.o ao.o«n.o eo.omw.o mo.ome.o enn\nan\m .aana.ann
oH.o+H.H HH.o«H.H th.o+m.H NH.o+a.o una\nan\w .aana use
HH ooanom .aowuouoxo Hooch
ame.ow~.m ao.o«n.a mo.owm.e no.o«e.o eaa\ane\m Hanna nae
sum.owe.a NH.o«n.~ aeH.o«H.~ eao.owo.e unn\aae\w name not
. H oowuom .aowuouoxo amoom
a.~ a.~ m.~ a.~ HH noanea
m.m o.m a.e H.n H donned
umu\%mo\w .cwmw uswwoz
NA em an em aHH tanned
an as me an eH season
umu\%mo\amox .co«u@aamaoo hwuoam
a been A dean N amen we use:

 

a“ abooo mnu mo unmucoo Hmanouomn cam Emaaonmuoa.oaououm .ao«uoaom Ho3om

.cmun umona mo mao>oa ooomuw pom mush

.m udm<fi

53

.umou m.uuoaaan he ooumou
mums mmoaoquMHo uaoaumouh . 306 v 3 Houuaoo Baum uaouommwo mauamoHMHawwms

. aoa x mucouaoo Hmooo mo w Hoe mwuouomn oHan> mo Hohasz

0H m

.moHom mafia humoaooom mo No
.m.m M Gonzo
.muoHo ooHMquun ouo3 HH oowuom nH muowv HH<o

.Amam>wuoommou . .m.N.H muowo How
Hoax OOH you some umo£3 mo w m.q.~.ov .cwun umo£3 mo mHo>oH pooch ooaamucoo muownn

.GOHueHHomoo uoHo How m oHan comm

 

 

H omm.a a.omm m ~.owe.m ~.om~.a HH eoHuee
e.o+H.m ~.o+~.e e.o-H.a ~.o+~.m H weaned
+ w .uswwos Amoco
©.m N.n N.© o.n HH UOHHmm
N.¢ m.¢ 0.5 0.5 H flOHHOm
wucaoo Hmauouomm
a emHe m ndHa N uaHn nH ueHa

 

A. . . emscHuaoov m Haney

54

in 31252 fermentations from cecal contents of bran-fed rats,
however, the linear increase in cholic acid degradation with
increased levels of bran in the diet was not observed in the
32:21EEQ fermentation. When rats were fed a bran-free diet,
the percent bile acid degradation in feces and in viggg for-
mentation was comparable to that of control rats in period I.

When a low level of bran (2g of bran) was added in
the diet, the number of anaerobic bacteria per g of cecal
contents stayed the same as in the control group, but the
concentrations of anaerobic bacteria in cecal contents
decreased with higher levels of bran (4 g or 8 g of bran) in
the diet. Since there was no difference in cecal weights be-
tween groups, it is likely that the number of anaerobic
bacteria per cecum also decreased with increased levels of
bran in the diet.

Experiment 2

There was no difference in energy intake and weight
gain between the two groups of rats for the 4 week period
(Table 6). Both wet and dry fecal weights were significantly
higher for the group fed the low-fat, high-fiber diet com-
pared to the group fed the high-fat, low-fiber diet.

Total bile acid excretion was similar between the
two groups, however, the concentration of bile acids was
lower in the group fed the low-fat, high-fiber diet. Total
neutral steroid excretion and neutral steroid concentrations
were significantly greater in rats fed the high-fat, low-

fiber diet compared to rats fed the low-fat, high-fiber diet.

.umou u m.ucmo5um ab.oouwou ouo3
mooaouommfio ucoEummHH .Amo.o v av masouw N ozu Cowsuon unmonchHm ma mucouommans

.ocoamumoumoo cam Hoamumoumoo mo Na

 

55

 

.m.mw cmoZm
~o.o«eo.o ~o.ow~H.o HEHme
.cowumuucoocoo owououm Hmuuao:
H.H-H.m e.H-N.H He\we.eoHenaetaatoa nHan eHHe
+ + eHHm
iuo.o-m.o ~.owm.~ oceanomoueoo\aocmumoumoo
am.ew Hm n.~w mm wtoHanentman N
*m.o ¢.~ m.mum.m~ mooom um3 w\wE GOHumuucoocoo
am.Haa.mH e.~HH.eH nea\ann\me .coHeeaaxa
+ + moHououm Hmuuaoa Hmowh
sm.0um.~ m.o:~.o moomm uo3 w\wE .COHumuucoocoo
e.oHH.~H N.HHH.~H ena\ane\ma .toHeanexa
+ + anHen eHHe Hanan
ano.o-o.~ Ho.oHH.H aaa\ann\m .aans Han
sm.o4m.m mm cum N umu\%wo\w .mde use
+ + COHuouoxm Hmoom
m.q w.¢ umu\hmo\w.:«mw umwfio3
on mm umu\>mo\amox .cowunESmcoo h umcm
uoHo uoHo

HeeHH-ewHe .eaH-3oH naeHH-3oH.ane ewH:

 

.aaHn naeHH-ewHe .nnH-3oH n no uaHe naeHe-3oH enH-ewHa n HaneHa

pom much Scum mama :H mam mooom :H EmHHonmuoE vacuouw cam mews Hooch .m mqm<H

56

However, more extensive degradation of cholesterol to
coprostanol and coprostanone was seen in the feces of rats
fed the low-fat, high-fiber diet. In the high-fat, low-
fiber group, more cholesterol was degraded to coprostanol
than to coprostanone; however, rats fed the low-fat, high-
fiber diet excreted more coprostanone than coprostanol in
the feces.

There was no difference in bile acid or neutral
steroid concentration in bile collected from the bile duct

of rats fed the two diets.

PART II

1. Intestinal Epithelial Cell Isolation

It was not possible to obtain cells using the scrap-
ing method. Cells were dissociated with the use of citrate
but they did not attach to the bottom of the plate and did
not grow.

Figure 3 shows the epithelial cells isolated by the
tissue culture/cell culture technique. The pieces of tissue
degenerated with time (approximately 5 days) and the cells
grew out from the vicinity of tissue explant attached to
the surface of the plates. Many of these cells survived
for up to 5 weeks, but there was no sign of increasing cell
numbers.

In all of the above trials the biggest problem en-

countered was contamination. Since the large intestine has

Top
Figure 3.

Bottom

Figure 4.

57

Phase micrographs of epithelial cells isolated
from rat intestine by tissue culture/cell cul-

ture method (x 655) (see material and method).

Phase micrographs of fibroblasts isolated from

intestine of rat fetus. (x 464).

58

 

59

a large number and many varieties of microorganisms, it was
almost impossible to eliminate the growth of microflora.

2. Isolation of Fibroblasts and Measurement of Doubling
Time.

Figure 4 shows the fibroblasts isolated from the in-
testine of rat fetus. The problem of microbial contamina-
tion was greatly reduced when cells were isolated from fetal
tissue. Doubling time of these cells were 20 hours.

3. Mutagenesis Study

In order to measure the dose response of fibroblasts
to MNNG, cells were treated with different levels of MNNG for
3 hours and then the Oua-R colonies were selected at 1.0 mM
ouabain in growth mediumr The number of Oua-R colonies in-
creased linearly, whereas cell survival was decreased as MNNG
concentration increased (Fig. 5). 1.5 ug MNNG/ml of medium
was used thereafter to optimize conditions for selection of
Oua-R mutants. This concentration of MNNG gives high muta-
tion frequencies with reasonably high percent survival.

To determine the optimum.ouabain concentration to
select Oua-R.mutants, cells treated with 1.5 pg MNNG/m1 of
medium and non-treated control cells were grown in medium
with different concentrations of ouabain. The MNNG treated
cells had more Oua-R colonies than non-treated cells at all
concentrations of ouabain (Fig. 6). When cells were grown
in medium with ouabain concentrations greater than 1.0 mM,
the majority of the cells did not survive in either the con-

trol or the MNNG treated groups. Cells forming colonies at

Figure 5.

60

Cytotoxicity and ouabain resistant mutant

response to MNNG dose.

5

. . ouabain resistant mutants per 10

survivors.

o -------- 0 percent survival

61

- Io2

panama 9220.

3.0

2.5

2.0

L?)
MNNG DOSE (pg/ml)

LO

0.5

 

 

IO

20325 no. .3 £532 .5381 5326

FIGURE 5 .

62

FIGURE 6. Effect of ouabain concentration on control and

MNNG treated cells.

. lcontrol cells~

o -------- o MNNG treated cells (1.5 ug/ml)

Ouoboin Resislont Mutonls per IOS Survivors

63

I05

TTTTTT

:—

   
 

lllrll

T

T

 

IO 1 4

L

0.5 |.O 1.5 2.0 2.5
Ouabain Concentrations (m M 1

FIGURE 6 .

64

ouabain concentrations greater than 1.0 mM in the medium
are considered ouabain resistant or mutant cells. Also,

as the concentration of ouabain in the medium increased,
the colony size decreased. A concentration of 1.0 mM
ouabain in the medium.was used for all the subsequent ex-
periments because this concentration of ouabain optimized
the difference between MNNG-induced mutation and spontaneous
mutation and because 1.0 mM ouabain allowed more satisfac-
tory growth of mutant colonies than higher concentrations.
Figure 7 shows colonies grown in either growth medium or in
1.0 mM ouabain-containing medium.

To determine the optimum-expression time, plates
were treated with 1.5 ug MNNG/ml for 3 hours. MNNG-treated
cells were then trypsinized and replated, and 1.0 mM ouabain
was then added on subsequent days to optimize for selection
of Oua-R cells. Plates with more than 3 days of expression
time were trypsinized and subcultured into 15 cm plates to
prevent confluency. Figure 8 show the results. Mutation
frequency increased linearly until 3 days of expression time,
but decreased with time thereafter.

Bile acids were found to be cytotoxic but the degree
of cycotoxicity of the various bile acids was different.
Figure 9 shows the cytotoxicity of six different bile acids
when cells were treated with graded concentrations of these
acids. There appear to be two groups of bile acids in terms
of cytotoxicity. Lithocholic, deoxycholic and chenodeoxy-
cholic acids were much more cytotoxic than cholic, hyodeoxy-

cholic and 3,12-dione-58-cholanic acids.

65

FIGURE 7. Bright field micrograph of Giemsa stained cells
(x 688) (top) colony grown in growth medium for
8 days, (bottom) colony grown in 1.0 mM ouabain-

containing medium for 14 days.

66

 

67

FIGURE 8. Expression time of MNNG (1.5 ug/ml)-induced
ouabain resistant mutations. Expression times
shown are intervals between MNNG treatment and

exposure to ouabain-containing medium (1.0 mM).

68

 

 

 

IS

17.

16

15

J4

13

.12

Jl.

WZHPMUWH. . WHHHHWWHEwVIVMIvO
m 5 I

mmo>_>m3m no. mun. mkz<kaz ._.z<._.m_mmm 2.4930

EXPRESSION TIME( DAYS)

FIGURE 8 .

FIGURE 9.

69

Cytotoxicity of selected bile acids

.—————I

Cholic acid

hyodeoxycholic acid
3,12-dione-58-cholanic acid
deoxycholic acid
chenodeoxycholic acid

lithocholic acid

Percent Survival

IOO

 

70

 

"A;\ ~~“~~.~ _ -
A 8‘ ~7‘-H~ I
\ .k- ~:““
I\ \.k . ~~“‘~!
IO . i \‘k.\

.. \ .
.. \ . \
.- . \\\
- \ \o
.. ' \

\ \
.. . x X

\ \
\
- . \
\
\ x .
.. . \\
\. o \\
\
\ \
- . \
\‘ \\
\. \\
I 1 _A \ l T. ' 1 1
0.5 LO '5 2.0 2.5 3.0

Bile acid concentrations (mM )

FIGURE 9 .

71

Since a preliminary experiment showed that deoxy-
cholic acid induced a greater mutation frequency than the
spontaneous mutation frequency, cells were treated with
graded levels of deoxycholic acid (0.1 mMFl.0 mM) to see if
there was a dose response. As shown in Table 7, mutation
frequency increased until deoxycholic acid concentration was
0.5 mM. Concentrations greater than 0.5 mM, however,
decreased the mutation frequency. This decrease is probably
due to the high cytotoxicity of deoxycholic acid at these
concentrations so that most cells died rather than undergoing
mutation. At 1.0 mM there were so few cells living after
treatment that a meaningful mutation frequency could not
be obtained.

The variations in mutation frequency within each ex-
periment and within a treatment from 3 experiments are shown
in Table 8. Standard error between experiments were approx-
imately 20% of the mean values.

Table 9 shows the results of several experiments in
which the mutagenic and comutagenic effect of MNNG and
various bile acids were examined. Since the spontaneous
mutation frequency (control value) was different from.ane
experiment to another, the values are presented as a percent
of the control values. Since induced mutation frequencies
by deoxycholic acid was highest at 0.5 mM (Table 7), other
bile acids were tested for their mutagenicity at this con-

centration. Chenodeoxycholic and lithocholic acids at 0.5 mM

72

TABLE 7. Cytotoxicity and ouabain resistant mutants
response to graded levels of deoxycholic acid.

 

 

 

Concentration of Oua-R mutants
deoxycholic acid, mM % survival 165 survivors
0 (control) 21 1128

0.1 23 166

0.3 19 155

0.5 19 256

0.7 11 171

1.0 5 -

 

3Mean values were from 8 plates per each concentration.
Results were from 2 experiments.

73

TABLE 8. Variations in mutation frequency within each
' “ experiment and within a treatment from three

 

 

 

 

 

experiments
Experiment Treatment
1.0 ug MNNG Deoxycholic
'ml of medium acid, 0.5 mM
1 2038186 2221128
2 140 1'12 230112
3 280 1“108 3931332
2 = 2081’“: 40 2 =282’5 56

 

8Mean 4 standard deviation for 4-8 plates per value. Value
is expressed as % of control.

bAverage mean value in 3 experiments t standard error.

74

.oHom dawn Sumo Ham moumam mum om: uaoaflummxo 50mm .wusmaauamxo mnH Ham maam> nmozn

.Amoonuaa.vcm Hmwuouma oomv

.omZQ GH oa>H0mmHo ohms soH£3 mowam mafia Honuo Ham omza auw3 common» ouoz mHHoo Houu
icoo .maomwumnaoo mvHom oHHonohxoooocono pom uwaoso Ham ooumouuna oHoB mHHma Houuaoum

.lfll

 

 

 

e3 - mm :84 .38 032858on

ONH .. mm 28 or“ .Euom oggimmuofioavuuaé

- NR man an m5 .33 8:29?er

ASH and sea za_m.o .pwom OHHonoosuHA

- RH SH as no .38 8391383888

mwa mom amm SE o.~ .oHom oaaono

mm u moa :8 o.H .vHom afiaosu

mow «NH ooa «Houucoo
o.H m.o o

Haxw: .cowumuuaoocoo wzzz
Amaonuaoo mo uaooHoav hoaoavoum nowumuaa o>HumHom ucoaumoya

 

.wowom mafia oouooaam mo uoommo owcowmunaoo paw venomous: .m mqm<9

75

produced a higher mutation frequency (Oua-R mutants) than
spontaneous mutation frequency. However, cholic, hyodeoxy-
cholic and 3,12-dione-SB-cholanic acids did not show in-
creased mutation frequency at 0.5 mM. Therefore, these bile
acids were tested at 1.0 mM concentration. None of these
bile acids produced a greater mutation frequency than the
spontaneous mutation frequency at this concentration. How-
ever, when the concentration of cholic acid was increased to
2.0 mM, a greater mutation frequency than spontaneous muta-
tion frequency was produced. Hyodeoxycholic and 3,12-diane-
SB-cholanic acids were not tested at this level.

Mutagenic effects of MNNG and bile acids were not
additive at two MNNG dose levels, 0.5 and 1.0 ug MNNG/ml.
Instead the mutation frequency was lower when cells were
treated with bile acids and MNNG than when treated with bile
acids alone.

The chloroform extract from feces of rats fed the
low-fat, high-fiber diet was more cytotoxic to cells than
the chloroform extract of feces from rats fed the high-fat,
low-fiber diet (Table 10). This is probably due to the
higher concentration of bile acids in the fecal extract of
low-fat, high-fiber fed rats. When cells were treated with
a low dose of extract (0.5 mg extract/m1) from rats fed the
high-fat, low-fiber diet, the mutation frequency was similar
to the mutation frequency of cells treated with DMSO only.

However, when cells were treated with 0.75 mg of extract/m1,

76

.mucoaeuoaxo N Baum auo3.muaamom

sapwooa mo Ha\uoauuxo anamouoano mo we

a

.aowumuucoocoo some Ham moumam use Baum waaam> amozm

 

 

 

 

 

 

- o H.me H.HH Haxwa o.H
com HH e.me a.mn Ha\math.o neeHH-emHe
.umm:3oa
eeH aH H.Nm H.mm Hexwa n.o
enH HH H.He ~.en HEHwa o.H
HaH HH w.om H.~a Ha\wa HH.o neeHH-soH
.enH emHm
NOH Hm e.o~ H.n~ eHa\ma_ m.o
nNHH NN - - Aomznv Houuaoo
mmwom oHHm mwfiououm Hmuuaoz
muo>fi>u5m moa Asflwooa.aa\w:v ucaa
mucmu:B,Mumso Hm>a>uam N numouu Sumo aw nowumuuaoocoo ucoaumoua uoan

 

.umwo Hosannnwws .umMIBOH a no uofio HonHMI30H .umm
nswfin a cam much Baum macaw mo muomuuxa Ehomouoano mo uommmo owcmwmuaz .oH mqm<H

77

there was a 76% increase in Oua-R.mutants compared to the
control. A greater extract concentration (1.0 mg of
extract/m1) did not cause a further increase in mutation.

Unlike the cells treated with the chloroform ex-
tract of feces from the rats fed the high-fat, low-fiber
diet, there was 30% increase in Oua-R.mutants when cells
were exposed to 0.5 mg of extract/m1 of medium from rats
fed the low-fat, high-fiber diet. When the cells were
treated with 0.75 mg of extract/m1, the mutation frequency
was comparable for both groups treated with the same con-
centration of extracts.

When cells were exposed to water extracts of feces
from rats fed both diets, cell survival was as high as for
control cells at all extract concentrations tested. The
water extracts from feces of rats fed both diets produced
higher mutation frequencies than the spontaneous mutation
frequency at all the levels studied (Table 11). However,
there was no dose response and there was no difference in
mutation frequency between the extract from rats fed a high-
fat, low-fiber diet and a low-fat, high-fiber diet.

Bile was very cytotoxic. When cells were treated
with bile that was reconstituted with medium to provide 20%
of the original bile concentration,no cells survived. With
concentrations of 0.01 to 0.1 of original concentrations of
bile, cell survival decreased and Oua-R mutants increased
linearly over control level. There was no difference in

mutation whether the bile was from rats fed the high-fat,

78

TABLE 11. Mutagenic effect of water extracts of feces
from.rats fed a high-fat, low-fiber diet or
a low-fat, high-fiber diet.

 

Diet Treatment Oua-R mutants
105 survivors

 

 

 

 

Control 78
High-fat, low-fiber 4 mg/mlb 14
10 mg/ml l4
Low-fat, high-fiber 4 mg/ml 13
10 mg/ml 9
25 mg/ml 15

 

aValues were from one experiment with 8 plates per concen-
tration.

bmg of water extract/m1 of medium.

79
low-fiber diet or from.rats fed the low-fat, high-fiber
diet (Table 12).

Comutagenic effects of MNNG with fecal extracts
and bile from.rats fed the high-fat, low-fiber diet were
tested. Each compound was tested at one dose level (0.75
‘mg chloroform extract/m1, 4 mg water extract/m1 and 10% of
original concentration of bile) which showed a greater
mutation frequency than spontaneous mutation frequency
(Tables 10-12). Like various bile acids, the mutagenic
effects of MNNG and either chloroform extracts of feces or
bile were not additive (Table 13). There was a slight in-
crease in comutagenicity with water extract of feces and
MNNG, however, the difference should not be considered sig-
nificant without further verification.

Since bile acids are often proposed as cancer pro-
motors, the promoting effect of bile acids were tested in
this system as in the scheme presented by Trosko (75). The
results are shown in Table 14. Since deoxycholic and litho-
cholic acids are very cytotoxic, most cells did not survive
and form colonies after 2 weeks of incubation with these
bile acids in the medium, Also, colony formation was very
poor. Therefore, the decreased mutation frequency when
cells were incubated with these bile acids probably is not
meaningful. When cells were grown in 0.2 mM deoxycholic
acid medium for 2 weeks, no cells survived.

When cells were grown in medium with cholic acid at

0.1 mM and 0.2 mM and chenodeoxycholic acid at 0.1 mM,larger

80

.AmoOSHma was mamwuouaa oomv Edwpoa mo HE\oHHn mo Han

.SOHumuuaooaoo Ham moumam o auwz ucoaauomxo ono Baum mums muadmomm

 

 

 

 

 

- o NNOH N.eH N.o

NH mH HHH e.e H.o
neeHH-emHe
eN NH NNH H.N mNo.o .uaH-seH

NH NN Hm ma.o HHo.o

- o mNOH aN N.o

NH m sen NH H.o
aaeHH-3oH
aN mH emH o.m mNo.o .eaH-emHm

aH HN e.oe N.H eHHo.o

aN NN - - HHenedooe o

uuuuuu Aabwooa Ha\w:vnunuu

muo>w>uammoa moHom oawm Houauwoaoso nowumhucooaoo

.mufimufia MINHHO HN>H>HUm N mHHm “THO

 

.adHn HeeHH-ean .unH-soH no HeeHH-soH .unH
-ewHH a dam nuns He needs aHHe seam eauoeHHoe aHHe Ho oedema aHdomneaz .NH mHm<H

81

TABLE 13. Mutagenic and comutagenic effects of fecal
extracts and bile from rats fed a high-fat,
low-fiber diet.

 

Treatment Relative mutation fre-
quency (percent of
control)

 

Mutagenicity comutagenicity with
MNNG (0.5 ug/ml)

 

Controla 100b 147
Chloroform 175 147
extract (0.75 mg/ml)

Water extract 175 196
(4 Ins/m1)

Bile (0.1 original 257 142
conc.)

 

aControl cells were untreated for water extract of feces
and bile comparisons. Control cells were treated with
DMSO for fecal chloroform extract test.

bResults were from one experiment with 5-7 plates per

tr ea tment .

82

 

 

 

TABLE 14. Promotor effect of selected bile acids.a
Treatment Oua-R mutants
105Tsurvivors
Control 297b
DMSO, 0.05% 363
Deoxycholic acid, 0.1 mM 174
Lithocholic acid, 0.1 mM 103
Cholic acid, 0.1 mM 684
Cholic acid, 0.2 mM 548
Chenodeoxycholic acid, 0.1 mM 618

 

aAll cells were treated with 1.5 ug MNNG/m1 and grown in
growth medium for 3 days. Cells were replated for plating
efficiency and for mutagenesis by adding bile acids in
growth medium at the indicated concentrations. Both groups
of cells were cultured for 2 weeks.

bResults were from one experiment with 7-8 plates per treat-
ment.

83

colonies were formed compared to the cells incubated with
deoxycholic or lithocholic acids. Mutation frequencies
were increased in this particular experiment demonstrating
that some bile acids can act as mutagenic promotors. It
is surprising that chenodeoxycholic acid was less cyto-
toxic than deoxycholic and lithocholic acids at 0.1 mM

since they showed similar cytotoxicities at 0.5 mM (Fig. 9).

DISCUSSION

Many biological metabolites in the colon are thought
to be toxic to colonic cells (76). These cell toxins could
be important agents which cause colon cancer. A larger
stool mass and shorter intestinal transit time might be very
significant factors in preventing colon cancer. Heavier
stools have a diluting effect on toxic substances, and a
shorter transit time reduces the duration of exposure of the
tissues to toxic substances.

The greater fecal mass and shorter transit time in
bran-fed rats agree with the results of other investigators
(27,29). Inclusion of agar, a purified fiber source, in the
diet also increased stool weights. Increase in fecal mass
is often correlated with bowel regularity (30). Daily bile
acid excretion was increased in bran-fed animals, even
though the increase in excretion was not statistically signi-
ficant. However, fecal bile acid concentration was lower
and decreased linearly with increasing levels of bran in the
diet. Other workers have also reported a greater bile acid
excretion in feces when fiber was added to diet. Rats fed
Purina chow or cellulose-containing diets had greater fecal
excretion of cholic acid and its degradation products than

rats fed a purified semi-synthetic diet (90). Jenkins et a1.

84

85

(27) also reported increased bile acid excretion and reduced
fecal bile acid concentration in human subjects fed a diet
containing wheat fiber.

The second experiment was designed to determine how
dietary fat and fiber content affected steroid metabolism.
Diets high in beef fat (9, 91) and low in dietary fiber (10)
have been implicated as a cause of the colon cancer through
an alteration in steroid metabolism and gut function. For
these reasons, high beef-fat (tallowL low-fiber and low-fat,
high-fiber (agar) diets were compared. Agar was used as the
source of fiber since it has been shown in our laboratory
that agar has a great capacity to increase stool weight and
decrease transit time (28). High cholesterol ingestion has
been associated with an increased excretion of fecal bile
acids by humans (92), and high fat diets (either lard or corn
oil) also increased fecal bile acid excretion in rats (25).
Since the high-fat, low-fiber diet was also a high cholesterol
diet, it was expected that the daily fecal bile acid excre-
tion would be greater for the rats fed the high-fat, low-fiber
diet than for rats fed the low-fat, high-fiber diet. wa-
ever, there was no difference in daily fecal bile acid ex-
cretion between these two dietary groups, and as shown in the
experiment 1, fiber in the diet actually increased the daily
fecal bile acid excretion. Therefore, it seems that the high
fat content of one diet and the high fiber content of the
other diet offset each other as far as daily bile acid ex-

cretion was concerned. Perhaps the most important fact was

86

that the bile acid concentration of feces from rats fed the
low-fat, high-fiber diet was several-fold lower due to the
dilution effect of the larger fecal mass.

As expected, neutral steroid excretion and concen-
tration was higher for rats fed the high-fat, low-fiber
diet than for rats fed the low-fat, high-fiber diet. It is
well established that diets rich in fat and cholesterol
lead to high excretion of neutral steroids in feces (25).

If bile acids and neutral steroids are important
carcinogens or cocarcinogens, then the reduced steroid con-
centration noted for the high fiber diets in experiment 1
and 2 would decrease the quantity of these steroids to which
the colonic cell surface is exposed.

It has been proposed that the inclusion of fiber in
the diet may influence the type and number of microflora in
the large bowel. A modification in bowel microflora could
lead to an alteration in the rate and extent to which cholic
or chenodeoxycholic acids are metabolized to secondary bile
acids such as deoxycholic, lithocholic, and corresponding
ketone derivatives. This would also influence the metabolism
of cholesterol to coprostanol and coprostanone.

Results in experiment 1 showed a decrease in concen-
tration of cecal anaerobic bacteria with the inclusion of bran
in the diet. Even though the absolute number was fewer, it
is doubtful that this decrease in anaerobic bacteria concen-

tration has physiological importance. In this type of work,

87

a difference of more than one log is considered physiologic-
ally important (93).

Inclusion of bran in the diet increased bile acid
degradation as bran concentration in the diet increased, and
the inclusion of agar in the diet also increased the degrada-
tion of neutral steroids. The trend of bile acid degrada-
tion was similar for feces and in_vi££g fermentation studies.
Therefore, it appears that the total number of anaerobic
bacteria may not be directly related to degradation of bile
acids.

A change in bacterial species present in the cecum
and colon could change the extent of bile acid and neutral
steroid degradation. Since bacterial concentrations and cecal
weights were similar for rats fed all levels of bran and
there was a decrease in transit time for bran-fed rats, the
increased bile acid degradation is likely due to a change in
types of intestinal microflora. The difference in microbial
activity due to diet was demonstrated in experiment 2 when
neutral steroid metabolism is examined. It is of interest
to note that the ratio of fecal coprostanol to coprostanone
is very different for rats fed the two different diets.

These data suggest that microorganisms which metabolize
cholesterol to coprostanol are predominant in rats fed the
high-fat, low-fiber diet, whereas bacteria which convert
coprostanol to coprostanone are more prevalent in the large
bowel of rats fed the low-fat, high-fiber diet. The prob-

able difference in types of intestinal microflora in this

88

experiment is likely due to both the fat and the fiber con-
tent of the diet. The possibility that gut ecology caused
the change in neutral steroid metabolism rather than
changes in bacterial types, however, cannot be excluded.

It is generally believed that secondary bile acids
and cholesterol metabolites are more active carcinogens or
cocarcinogens than primary bile acids and cholesterol (94).
Many investigators have postulated that high fiber (95)
and low fat (96) diets result in decreased steroid degrada-
tion in the colon, and therefore, decreased incidence of
colon cancer. The data in this study are contradictory to
this hypothesis, at least with respect to the effects of bran
fiber, agar, and tallow in rats. If fiber is beneficial in
reducing colon cancer, its major effects will be through the
dilution of carcinogens or cocarcinogens in colonic contents
as Hill suggested (96), and/or through the reduction in time
in which the colonic cells are exposed to these agents.

It was unexpected to find similar bile acid and
cholesterol concentrations in bile of animals fed either a
high-fat, low-fiber or a low-fat, high-fiber diet. With a
diet high in fat, it is expected that more bile acids would
be secreted into the gut to aid in fat digestion and absorp-
tion. However, one needs to know the concentration of bile
acids in bile and the bile flow rate before total bile acid
secretion can be ascertained. The flow rate of bile was not
determined since bile was collected under stressful condi-

tions immediately after the bile duct was cannulated.

89

Therefore, bile flow probably did not reflect physiological
conditions since the rats secreted widely varying amounts
of bile during the same 24 hour period. The difference in
bile fIOW‘WaS not related to diet, but rather seemed to be
due to a difference in response of each rat to stress and
in the success of creating a free-flowing cannula.

Reddy et a1. (25) reported a greater biliary concen-
tration of bile acids when rats were fed a high fat diet
rather than when the rats were fed a low fat diet. But
their rats were fed high and low fat diets for 10 weeks and
the bile was collected only for 4 hours. In this study, the
rats were fed their respective diets for only 4 weeks and
bile was collected for 24 hours. Rats in this study may have
secreted more dilute bile with time, since they did not con-
sume much of the available food for the 24 hour period follow-
ing surgery. If early differences in bile acid concentration
existed between groups, this difference may have been obscured
with prolonged collection time.

The experiments which were designed to develop an
intestinal cell culture method to screen mutagens were con-
ducted on the assumption that mutagenesis plays a role in
carcinogenesis and that ouabain resistant cells are true
mutants .

Ouabain is a specific inhibitor of the plasma mem-
brane Na/K ATPase. Structure of ouabain is shown below.

ATPase is responsible for the active transport of sodium and

90
potassium, In the presence of ouabain, cells can not sur-
vive due to the inhibition of this transport system.
Ouabain resistant mutation is due to a genetic alteration
in the structural gene for the ATPase system which allows
the survival of cells by permitting transport activity in
the presence of ouabain (78). A frameshift mutation can
not contribute to ouabain resistance, since a frameshift
mutagen would cause a mutation which results in a complete
loss of enzyme activity. Thus, the potassium transport
function would be completely blocked in the presence of
ouabain (78, 97). For the same reason, when ouabain re-
sistance is used as a marker for mutagen screening, deletion-

type mutagens will go undetected.

 

C6H1105 0H

OUABAIN (cngqqolz)

91

The results of this study indicate that mutants can
be selected by using ouabain as a selective agent and MNNG
'is an effective mutagen for inducing ouabain resistant muta-
tion in intestinal fibroblasts isolated from rat fetuses.
MNNG induced Oua-R.mutation in a dose dependent manner.
Structure of MNNG is shown below. ‘

Factors which must be determined in establishing
protocols for mutagenesis assays include 1) dosage of the
potentially mutagenic agent, 2) expression time for induced
mutations, 3) conditions for optimal detection of mutant
phenotypes.

Oua-R mutants could be selected when ouabain was
present at 1.0 mM, since ouabain inhibits the growth of wild-
type cells at concentrations equal to or greater than 1.0 mM.
In studies by other investigators, a ouabain concentration
of 1.0 mM was needed to select Oua-R.mutants from wild-type
cells of other rodents such as mouse and hamster cells (76,

97). However, it should be noted that a 1.0 uM concentrations

H C
3 \\'-_-
It r----NH----— N02
0201/ ii
NH

N-l‘lETHYL-N ’ -N ITRO-N-N I TROSOGUAN ID I NE

92

of ouabain was sufficient to select Oua-R mutants in human
cells (77, 78).;

As shown in Figure 7, cells grown in ouabain-
containing medium exhibited both reduced size of individual
cells as well as fewer cells per colony. Baker et a1. (76)
have also observed a decrease in colony size with increasing
ouabain concentration. Colonies surviving 0.5 mM ouabain
probably resulted from incomplete inhibition of wild-type
cell multiplication by ouabain rather than by the presence of
ouabain resistant cells.

The frequency of ouabain resistant cells reached its
maximum when they were allowed to grow in growth medium.for
3 days after mutagen treatment. The 3 day expression time
before the cells were exposed to ouabain allowed approximat-
ely 3 doubling times, provided that mutagen-treated cells
grow at the same rate as normal cells. If more than 3 days
were allowed for expression, the mutation frequency decreased.
Chang et al. reported that UV-induced Oua-R mutation fre-
quency was maximum with 2 days of expression time in Chinese
hamster cells (97). If more than 2 days for expression was
allowed, they also found a decline in mutation frequency. A
MNNG concentration of 1.0-1.5 ug/ml was effective in inducing
mutation in this study as in other studies (98).

In many mutation frequency studies, cells are not
subcultured after treatment with a mutagen. Huberman et a1.
(99) used 8-azaguanine as marker and Chang et a1. (97) used

ouabain as a selective agent to determine mutation frequency,

 

93

but neither investigator subcultured the cells after muta-
gen treatment. Instead they seeded a specific number of
cells into plates, exposed the cells to a mutagen, allowed
growth in medium for the appropriate expression period, and
then put drug-containing medium in the plates to select
'mutants. When this protocol was followed for intestinal
fibroblasts from rat fetus, Oua-R cells did not form dis-
tinctive colonies and scoring was very difficult. But,
when the cells were subcultured after the expression period,
the Oua-R cells formed distinctive colonies and were easier
to score. The protocol used in this study is more widely
practiced with cells grown in suspension (73) than cells
attached to surfaces.

The spontaneous mutation frequency in the cell sys-
tem used in this study was higher than reported for other
cell systems, and the spontaneous mutation frequency varied
from one experiment to another. The reported frequency of
Oua-R colonies in wild-type cell populations (spontaneous
mutation frequency) was 2.0 x 10'6 in Chinese hamster (V-79)
cells (97), 1.0 x 10‘6 in human lymphoblastoid lines (77),
and 1.0 x 10-5 in Chinese hamster ovary (CHO) cells (76).
Spontaneous mutation frequency could vary 2 or more orders
of magnitude not only between different cell lines but also
in the same line assayed at different times (74). In the
cell system used here, the spontaneous mutation frequency

5 5

ranged from.7.0 x 10- to 112 x 10- , In one report by

Trosko et al. the spontaneous mutation frequency of Oua-R

94

cells also varied from 2.1 x 10"5 in one experiment to 3.6
x 10'6 in another experiment (75).

The higher spontaneous mutation frequency found in
this study as compared to other cell systems is likely due
to the difference in cell types. The intestinal fibroblasts
of the rat fetus have a finite life span and must be fre-
quently isolated from intact animals. Nbst frequently used
cell types for cell culture work are established cell lines
which have an unlimited life span. It is likely that many
of the cell characteristics have changed during the constant
subculturing.

It has been reported that the fibroblasts isolated
from different organs have different characteristics. Skin
and lung fibroblasts derived from the same human fetuses
showed several differences (100). Fetal lung cultures
had faster cell replication rate, greater 3H-thymidine in-
corporation into DNA, higher cell numbers at confluency,
smaller cell volumes, decreased cellular RNA and protein con-
tents and longer life span in culture compared to fetal skin
fibroblasts. Also, these two fibroblast cultures responded
differently to the addition of hydrocortisone to culture
medium. Examination of cornea, heart, and skin fibroblasts
from the same embryonic chicks also revealed differences in
fibroblasts isolated from different organs (101). They were
different in saturation densities, in sensitivity to treat-
ment with trypsin and EDTA, in sensitivity to EDTA alone and

in the mode of cell arrangement at confluency. For these

95

same reasons fibroblasts from different animals and sources
could exhibit differences in ouabain resistance.

There could be many reasons why the spontaneous
mutations noted in this study varied from experiment to
experiment. One possible explanation for variation in spon-
taneous mutation frequency is that cells used for each
experiment were isolated from different animals. Addition-
ally, the number of subcultures before treatment was differ-
ent. Consequently trypsinization could conceivably change
the characteristic of cells. However, in the experiments
where the spontaneous mutation frequency was high, the muta-
tion frequency of treated cells was also high and vice versa.
Therefore, for the purpose of screening mutagens, the vari-
ation in spontaneous mutation would not affect the comparison
of results as long as a control group is included in each
experiment.

Cloning efficiency (percent survival) of non-treated
fibroblasts was relatively low (approximately 20%). It
seems that the cloning efficiency of freshly isolated cells
is lower than the cloning efficiency of cells from estab-
lished cell lines. Maher et al. reported that the cloning
efficiency of fibroblasts from.human foreskins was 20-35%
(102), whereas the cloning efficiency of 90-100% for aneu-
ploid V-79 Chinese hamster cells (75, 99) was reported.

Even though the cloning efficiency of cells used in this
study was relatively low, the lower cloning efficiency would

not affect the results of experiments.

96

Since bile acids have been postulated as the agents
which could cause colon cancer, selected bile acids have
been tested in this system. The bile acids tested here are
the ones reported to be found in the feces (103) and in
bile of rat. 3,12-dione-58-cholanic acid was included
since a bile acid with a GLC relative retention time similar
to 3,12-dione-SB-cholanic acid was often noted in rat feces
when the bile acids were quantitated in experiments 1 and 2
of part I. Also, the diketo-acid is even more degraded than
deoxycholic acid which is reportedly a potential carcinogen.

Interestingly, the various bile acids had different
cytotoxicities and different solubilities. Since bile acids
are lipophilic rather than hydrophilic, difficulty was en-
countered in incorporating bile acids into culture medium.
When deoxycholic, hyodeoxycholic, and 3,12-dione-58-cholanic
acids were dissolved in DMSO, they went into solution easily
in medium.without changing the medium pH. However, the
lithocholic acid in DMSO did not go into solution in medium,
therefore suspensions of lithocholic acid were used. When
cholic and chenodeoxycholic acids were dissolved in DMSO,
they changed the pH of the medium.upon addition and caused
protein coagulation. Therefore, these acids were dissolved
directly into medium after neutralization by NaOH. Thus
two methods were utilized to incorporate bile acids into
medium.

There appear to be two groups of bile acids among

the bile acids tested as far as cytotoxicity is concerned.

 

97

Deoxycholic, lithocholic and chenodeoxycholic acids were
more cytotoxic than cholic, hyodeoxycholic and 3,12-dione-
58-cholanic acids. When cells were incubated with deoxy-
cholic, lithocholic and chenodeoxycholic acids at 1.0 mM
for 3 hours, mast cells did not survive. However, cholic,
hyodeoxycholic, and 3,12-dione-58-cholanic acids had rela-
tively high percent survival when they were incubated at
2.0 mM concentrations for 3 hours.

Even though the bile acids were cytotoxic to cells
in a mammalian cell system, Silverman and Andrews did not
find any noticeable cytotoxic effect of bile acids in
Salmonella tester strains used for the Ames test (104).

It appears that there is a direct relationship between bile
acid cytotoxicity and mutagenicity. It has been generally
thought that mutation frequency increases with decreasing
cell survival (74). More cytotoxic bile acids (deoxycholic,
lithocholic and chenodeoxycholic acids) produced a greater
mutation frequency at a lower concentration (0.5 mM), while
less cytotoxic bile acids (cholic, hyodeoxycholic and 3,12-
dione-SB-cholanic acids) did not show an increased mutation
frequency even at 1.0 mM. However cholic acid at 2.0 mM
was mutagenic. Even though hyodeoxycholic and 3,12-dione-
SB-cholanic acids did not show mutagenic activity at 1.0 mM,
it is possible that these steroids would induce mutation at
higher concentrations, since a dose dependent response was

noted for deoxycholic acid and cholic acid.

 

98

The dosage difference among various bile acids that
was required to produce mutagenesis agrees with the animal
studies done by Wynder and coworkers (49,50). They found
tumor promoting effects of deoxycholic, lithocholic, cholic,
and chenodeoxycholic acids after intrarectal instillation
of MNNG in rats. 1.0 mg sodium taurodeoxycholic and litho-
cholic acids were given at each time whereas 20 mg of sodium
cholate and sodium chenodeoxycholate were used to produce
approximately the same tumor promoting effect.

Deoxycholic acid has been found to be mutagenic in
bacteria (105), but not in the Salmonel1a-mamma1ian-micro-
some test (104). The different results with different assay
systems emphasizes the importance of using a system which is
more closely related to the human.

Even though a comutagenic effect of bile acids and
MNNG was not observed in this study with rat intestinal
fibroblasts, bile acids could be comutagenic in this system
with other known carcinogens, such as 2-aminoanthracene
(2-AA). Lithocholic acid and its conjugates showed co-
mutagenic activity with suboptimal amounts of 2-AA in TA-
1538, even though bile acids themselves were not mutagenic
(104). However, comutagenesis was not observed when litho-
cholic acid was tested with MNNG, 2-acety1-aminofluorene
(AAF), 3-methy1cholanthrene (MCA), and benzo(a)pyrene (BP).
The decrease in mutation frequency when cells were treated
with bile acids and MNNG agrees with the results of

Silverman and Andrews (104). When Salmonella tester strains

 

99

were treated with lithocholic acid and MNNG, the number of
revertants decreased by 25% compared to the cells treated
‘with solvent and MNNG only. The reason for the decreased
mutation frequency when cells were treated with bile acids
and MNNG, as compared to bile acids or MNNG only, is not
known. Since bile acids are surface-active agents, they
may change membrane permeability and inhibit MNNG transport
across the membrane. Untransported MNNG could then react
with bile acids in such a way as to inhibit bile acid trans-
port or activity. This is purely speculation, but appro-
priate experiments could be designed to prove or disprove
the hypothesis. Also, the levels of bile acids tested here
for comutagenicity was the optimal concentration needed to
induce maximum mutation frequencies. Suboptimal concentra-
tions of bile acids may show synergistic effects with chemi-
cal mutagens which would demonstrate comutagenicity.

The extracts of rat feces fed with either high-fat,
low-fiber or low-fat, high-fiber diets, were tested for
mutagenic and comutagenic activity in this system. Since
the extracts were added to the medium on a weight basis,
extract concentration in the medium does not represent the
physiological concentration of these extracts in feces of
either groups.

Both chloroform and water extracts of feces from
rats fed either of the two diets produced higher mutation
frequencies than the spontaneous mutation frequency.

Since extracts from both diets increased the mutation

100

frequency at different concentrations, it seems that the
concentration of fecal metabolites in the colon could play
an important role in colon carcinogenesis. The finding
that a water extract did not exhibit cytotoxicity but did
exhibit mutagenicity is contrary to the general belief
that there is a direct relationship between cytotoxicity
and mutagenicity.

It was also interesting to find the mutagenic
effects of bile. It has been postulated that conjugated
bile acids are less carcinogenic or non-carcinogenic com—
pared to unconjugated bile acids. Since most bile acids in
bile are present in the conjugated form, conjugated bile
acids may also enhance mutation. It is possible that other
components of bile, but not the conjugated bile acids, in-
duced the mutation.

In the study of Silverman and Andrews (104), not
only lithocholic acid but also the conjugates of lithocholic
acid were found to be comutagenic with 2-AA. One can still
argue that the conjugated bile acid may have a lower muta-
genic potency than free bile acids. This possibility is
suggested by the work of Silverman and Andrews (104). The
number of revertants decreased by 40% and 20% when bacteria
were treated with taurolithocholic acid + 2-AA and with
glycolithocholic acid + 2-AA compared to when bacteria were
treated with lithocholic acid + 2-AA, respectively.

It was not surprising to find that there was no

difference in.mutagenic activity of the bile from the two

 

101

dietary groups, since the steroid concentrations of the

bile were not different. The mutation inducibility of

bile is consistent with work of Chomchai et a1. (47).

They found more tumors in the colon of DMH treated rats when
the bile duct was diverted from.the proximal half to the
distal half of the small intestine. Chloroform and water
extracts of feces and bile did not show comutagenic effects
with MNNG as in the case of bile acids.

Since the chloroform and water extracts of feces and
bile were shown to elicit mutagenic activity in the rat in-
testinal fibroblast assay system, the next step would be to
separate these extracts into more definitive groups of com-
pounds. These groups could then be tested for the mutagenic
effect until the compounds responsible for the mutagenesis
are identified.

Trosko et a1. (75) nicely demonstrated the tumor
promoting effect of 12-0-Tetradecanoyl-phorbol-13-acetate
(TPA) in UV-irradiated Chinese hamster cells, using ouabain
and 6-thioguanine resistance as selective markers. Whereas
TPA did not show any cytotoxic effect, bile acids were cyto-
toxic and incubation of cells with bile acids for 2 weeks
killed or markedly inhibited the growth of cells, especially
in ouabain medium. The promotor study was carried out only
once. Results of the present experiment only suggest the
possibility of developing an experimental protocol to test
promotors in this cell system. However, to adapt this system

for tumor promotor studies, all the conditions should be

102
optimized with a known promotor, like TPA. The dose of
test compounds would have to be tested, especially for
cytotoxic compounds like bile acids. It should not be
concluded that some bile acids have tumor promoting
effects from the result of this experiment only.

There still are many limitations in using this sys—
tem as a screening technique. One problem is that colonies
are small and are difficult to detect. Another problem is
that calculation of mutation frequency depends entirely on
the number of cells seeded. Therefore, the accuracy of
calculating mutation frequency is dependent upon the accuracy
of counting cells by a hemocytometer. Variation of the cell
counts by a hemocytometer was large and it is likely that a
bias was involved. Most experiments presented here were
repeated at least twice. One way to reduce variation and
bias is to repeat each experiment two or three times. Also,
a large number of replicate dishes perhaps 20-30 plates or
more should be used for each datum point.

Even though many of the compounds tested showed
mutagenic activity in intestinal fibroblasts cultures, it
would be presumptious to speculate that these compounds
would also cause mutation or tumorigenesis in an intact
animal. Even if any of the compounds are proven to be muta-
genic or carcinogenic in animals, the extrapolation of such
results to man should be exercised with great caution.

I pursued this study purely to develop a method for

screening purposes. If a compound is identified as a

103

mutagen by an in_vi££g system, an in_vizg_assay system should
follow; At the same time, efforts should be made to relate
the presence of the compound as a function of diet, micro-
flora metabolism and gut function. Only then can meaningful
recommendations be made for the general public in an attempt

to reduce the risk of developing colon cancer.

CONCLUSIONS AND RECOMMENDATIONS

The most notable change associated with dietary
fiber in the diet was an increased fecal mass and reduced
intestinal transit time. With the experimental designs
employed here, the variations in dietary fat and fiber
content did not affect the total daily bile acid excretion
in rats. However, the fecal bile acid concentration was
lower when dietary fiber is incorporated in the diet due to
the larger fecal mass. Since bile acids are suspected
carcinogens or cocarcinogens, dietary fiber in the diet may
play a role in decreasing the incidence of colon cancer by
diluting bile acids in the large bowel. Also, dietary fiber
might have a protective role in colon carcinogenesis by
reducing the time which colonic cells are exposed to carcino-
genic or cocarcinogenic compounds.

The difference in bile acid degradation in rats fed
graded levels of bran and the difference in neutral steroid
degradation in rats fed diets with different levels of fat
and agar indicate that the metabolic activity of gut micro-
flora can be modified by diet.

Fibroblasts were isolated from the intestine of rat
fetuses to screen potential mutagens and/or carcinogens

which can be found in the gastrointestinal tract. Increased

104

105

ouabain resistant mutation frequencies were produced when
cell cultures were treated with MNNG, bile acids, organic
and water extracts of rat feces and rat bile. Since a dose
response was seen in most of the experiments, the concen-
tration of bile acids and/or intraluminal components is
likely to be an important factor in development of colon
cancer.

Results of this study indicate that intestinal cell
culture techniques are feasible for screening mutagens and/
or carcinogens which can be found in the gut. However,
additional work should be done in an attempt to improve this
cell culture technique before it is recommended for routine
use. Such factors as the optimal cell age, optimum condi-
tion to increase colony size and the number of replicate
plates required for statistical validity should be evaluated.
Also, many known carcinogens and mutagens should be tested
to ascertain mutagenic activity in this system,

Since frameshift mutation is not detected by employ-
ing ouabain as a selective agent, a selective agent which
can detect frameshift mutation, such as 8-azaguanine, should
be developed. The basic parameters to select mutants in
this cell culture system.have been determined and adaptation
of the present protocol for 8-azaguanine use should not be
too difficult. Unless a compound shows negative results in
several systems, it should not be concluded that the com-

pound in question does not have mutagenic activity.

106

Some of the compounds which show positive results
in this mutation assay should be tested in 12,2133 systems.
The ultimate criterion will be to inject mutated cells in-
to an immunologically incompetent animal to test for

tumorigenesis.

BIBLIOGRAPHY

BIBLIOGRAPHY

1. Kassira, E.,L. Parent and G. Vahouny. Colon cancer.
An epidemiological survey. Digestive Disease 21:
205, 1976.

2. wynder, E.L., I.J. Brass and T. Hirayama. A Study of
the epidemiology of cancer of the breast. Cancer
13:559. 1960.

3. Buell, P. and J. E. Dunn Jr. Cancer mortality among
Japanese Issei and Nisei of California. Cancer 18:
656, 1965.

4. Stemmermann, G.N. Patterns of disease among Japanese
living in Hawaii. Arch. Environ. Health 20:266, 1970.

5. Staszewski, J. and W; Haenszel. Cancer mortality among
the Polish-born in the United States. J. Nat.
Cancer Inst. 35:291, 1965.

6. Mass, N. and B. Modan. Epidemiological aspects of neo-
plastic disorders in Israeli migrant population.
IV. Cancer of the colon and rectum. J. Nat. Cancer
Inst. 42:529. 1969.

7. Wynder, E.L. and B.S. Reddy. Metabolic epidemiology of
colorectal cancer. Cancer 34:801, 1974.

8. Hill, M.J. Metabolic epidemiology of dietary factors
in large bowel cancer. Cancer Res. 35:3398, 1975.

9. Wynder, E.L. and B. Reddy. Studies of large-bowel cancer:
Human leads to experimental application. J. Nat.
Cancer Inst. 50:1099, 1973.

10.Burkitt, D.P. Epidemiology of cancer of the colon and
rectum. Cancer 28:3, 1971.

ll.Lowenfels, A.B. and M.E. Anderson. Diet and cancer.
Cancer 39:1809, 1977.

12.Maugh, T.H. Vitamin A: Potential protection from car-
cinogens. Science 186:1198, 1974.

107

13.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

108

Rogers, A.B., B.J. Herndon and P.M. Newberne. Induc-
tion by dimethylhydrazine of intestinal carcinoma
in normal rats and rats fed high or low levels of
vitamin A. Cancer Res. 33:1003, 1973.

Rogers, A.B., O. Sanchez, F.M. Feinsod and P.M. Newberne.
Dietary enhancement of nitrosamine carcinogenesis.

Cancer Res. 34:96, 1974.

Schrauzer, G.N. Selenium.and cancer: A review. Bia-
inorganic Chem. 5:275,l976.

Schwartz, M.R. Role of trace elements in cancer. Cancer
Res. 35:3481, 1975.

Clayson, D.B. Nutrition and experimental carcinogenesis:
A review. Cancer Res. 35:3292; 1975.

McLean, A.E.M. and P.N. Magee. Increased renal carcino-
genesis by dimethyl nitrosamine in protein deficient
rats. Br. J. Exp. Path. 51:587, 1970.

Dodd, G.D. Genetics and cancer of the gastrointestinal
system. Radiology 123:263, 1977.

Stewart, H.L. Geographic patholo y of cancer of the
colon and rectum. Cancer 28: 5, 1971.

Haenszel, W., J.W. Berg, M. Segi, M. Kurihara and F.B.
Locke. Large-bowe cancer in Hawaiian Japanese. J.
Nat. Cancer Inst. 51:1765, 1973.

Lemon, F.R., R.T. Walden and R.W. WOods. Cancer of the
lung and mouth in Seventh-Day Adventists. Prelim-
inary report on a population study. Cancer 17:486,

1964.

Hill, M.J., J.S. Crowther, B.S. Drasar, G. Hawksworth,
V.Aries and R.E.O. Williams. Bacteria and etiology
of cancer of large bowel. Lancet 1:95, 1971.

Coombs, MaM., T.S. Bhatt and C.J. Croft. Correlation
between carcinogenicity and chemical structure in

igggopenta (a)phenathrenes. Cancer Res. 33:832,

Reddy, B.S., S. Mangat, A. Sheinfil, J.H. Weisbur er
and E.L. wynder. Effect of type and amount a diet-
ary fat and 1,2 - dimethylhydrazine on biliary bile
acids, fecal bile acids, and neutral sterols in rats.
Cancer Res. 37:2132, 1977.

Bill, M.J. Bacteria and the etiology of colonic cancer.
Cancer 34:815, 1974.

27.

28.

29.

30.
31.

32.

33.

34.

35.

36.

37.

38.

39.

109

Jenkins, D.J.A., M.S. Hill and R.H. Cummings. Effect
of wheat fiber on blood lipids, fecal steroid ex-
cretion and serum iron. Amer. J. Clin. Nutr. 28:

1408, 1975.

Forsythe, W.A., W;L. Chenoweth and M.R. Bennink.
Laxation and serum cholesterol in rats fed plant
fibers. Submitted to J. Food Sci.

Payler, D.K., E;W. Pomare, K;w. Heaton and H.E. Harvey.
The effect of wheat bran an intestinal transit.
Gut 16:1209, 1975.

Cummings, J.H. Dietary fibre. Gut 14:69, 1973.

Hill, M.J. Steroid nuclear dehydrogenation and colon
cancer. Amer. J. Clin. Nutr. 27:1475, 1974.

Hill, M.J. and V. Aries. Faecal steroid composition
and its relationship to cancer of the large bowel.
J. Pathol. 104:129, 1971.

Reddy, B.S. and E.L. Wynder. Large-bowel carcino-
genesis: Fecal constituents af populations with
diverse incidence rates of colon cancer. J. Nat.
Cancer Inst. 50: 1437, 1973.

Reddy, B.S., J.H. Weisburger and E.L. wynder. Fecal
bacterial B-glucuronidase: Control by diet. Science
183:416, 1974.

Reddy, B.S., A. Mastromarina and E.L. wynder. Further
leads on metabolic epidemiology of large bowel
cancer. Cancer Res. 35:3403, 1975.

Reddy, B.S. and E.L. wynder. Metabolic epidemiology of
colon cancer. Cancer 39:2533, 1977.

Reddy, B.S., J.H. Weisburger and E.L. wynder. Effects
of high risk and low risk diets for colon carcino-

genesis on fecal microflora and steroids in man. J.
Nutr. 105:878, 1975.

Aries, V.C., J.S. Crowther, B.S. Drasar, MrJ. Hill and
F.R. Ellis. The effect of a strict vegetarian diet
on the faecal flora and faecal steroid concentration.

J. Pathol. 103:54, 1971.

Hill, M.J., B.S. Drasar, RtE.O. Williams, T.W; Meade,
A.G. Cox, J.E.P. Simpson and B.C. Morson. Faecal
bile-acids and clostridia in patients with cancer
of the large bowel. Lancet 1:535, 1975.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

110

Reddy, B.S., T. Narisawa, D. Vukusich, J.H. Weisburger
and E.L. Wynder. Effect of quality and quantity
of dietary fat and dimethylhydrazine in colon car-
cina enesis in rats. Prac. Soc. Exptl. Biol. Med.

151: 37, 1976.

Nigra, N.D., D.V. Singh, R.L. Campbell and M.S. Pak.
Effect of dietary beef fat on intestinal tumor
formation by azoxymethane in rats. J. Nat. Cancer

Inst. 54:439, 1975.

Cook, J.W., E.L. Kennaway and N.M. Kennaway. Produc-
tion of tumors in mice by deoxycholic acid. Nature
145:627, 1940.

Shear, M.J., J. Leiter and A. Perrault. Studies in
carcinogenesis, XV. compounds related to 20-methyl-
cholanthrene. J. Nat. Cancer Inst. 2:99, 1941.

Lacassagne, A., N.P. Buu-Hoi and F. Zaldela. Carcino-
genic activity of apocholic acid. Nature 190:1007,
1961.

Lacassagne, A., N.P. Buu-Hoi and F. Zajdela. Carcino-
genic activity in situ of further steroid compounds.
Nature 209:1026, 1966.

Nigra, N.D., N. Bhadrachari and C. Chomchai. A rat
model for studying colonic cancer: Effect of choles-
tyramine an induced tumors. Dis. Colon Rectum 16:

438, 1973.

Chomchai, C., N. Bhadrachari and N.D. Nigra. The effect
of bile on the induction of experimental intestinal
tumor in rats. Dis. Colon Rectum 17:310, 1974.

Reddy, B.S., T. Narasawa, J.H. Weisburger and E.L.
Wynder. Promoting effect of sodium deoxycholate on
colon adenocarcinomas in germfree rats. J. Nat.
Cancer Inst. 56:441. 1976.

Narisawa, T., N.E. Magadia, J.H. Weisburger and E.L.
wynder. Promoting effect of bile acids on colon
carcinogenesis after intrarectal instillation of N-
methyl-N'-nitro-N-nitrosoguanidine in rats. J. Nat.
Cancer Inst. 53:1093, 1974.

Reddy, B.S., K. Watanabe, J.H. Weisburger and E.L. wynder.
Promoting effect of bile acids in colon carcinogenesis

in germ-free and conventional F344 rats. Cancer Res.
37:3238, 1977.

51.

52.

53.

54.

55.

56.

57.

58.

59.

601

61.

62.

63.

111

werner, B.K. deHeer and H. Mitschke. Cholecystectomy
and carcinoma of colon. Z. Krebsforsch. 88:223, 1977.

Pomare, E.W. and K;W. Heaton. The effect of chole-
cystectomy on bile salt metabolism. Gut. 14:753, 1973.

Lowenfels, A.B. and A. Sonni. Distribution of small
bowel tumors. Cancer Letters 3:83, 1977.

Reddy, B.S., J.H. Weisburger, T. Narisawa and E.L. wynder.
Colon carcinogenesis in germ-free rats with 1,2-
dimethylhydrazine and N-methyl-N'-nitro-N-nitroso-
guanidine. Cancer Res. 34:2368, 1974.

Finegold, S.M., H.R. Attebery and V.L. Sutter. Effect
of diet on human fecal flora: comparison of Japanese
andaAmerican diets. Amer. J. Clin. Nutr. 27:1456,
197 .

Finegold, S.M., D.J. Flora, H.Rt Attebery and V.L. Sutter.
Fecal bacteriology of colonic polyp patients and
control patients. Cancer Res. 35:3407, 1975.

Moore, W.E.C. and L.V. Holdeman. Discussion of current
bacteriological investigations of the relationships
between intestinal flora, diet, and colon cancer.
Cancer Res. 35:3418, 1975.

Drasar, B.S., D.J.A. Jenkins and J.H. Cummings. The in-
fluence of a diet rich in wheat fibre on the human
faecal flora. J. Med. Microbiol. 9:423, 1976.

Drasar, B.S. and D.J.A. Jenkins. Bacteria, diet, and
large bowel cancer. Amer. J. Clin. Nutr. 29:1410,
976.

Maier, B.R., MHA. Flynn, G.C. Burton, R.R. Tsutakawa
and D.J. Hentges. Effect of a high-beef diet on

bowel flora: a preliminary report. Amer. J. Clin.
Nutr. 27:1470, 1974.

Hentges, D.J., B.Rt Maier, G.C. Burton, MHA. Flynn, R.R.
Tsutakawa. Effect of a high-beef diet on the fecal
bacterial flora of humans. Cancer Res. 37:568, 1977.

Reddy, B.S., S. Mangat, J.H. weisburger and E.L. Wynder.
Effect of high-risk diets for colon carcinogenesis
on intestinal mucosal and bacterial B-glucuronidase
activity in F344 rats. Cancer Res. 37:3533, 1977.

Scheline, R.R. Metabolism.of foreign compounds by
astrointestinal microorganisms. Pharmacol. Rev. 25:
51, 1973.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

112

Heidelberger. C. Chemical oncogenesis in culture. Adv.
Cancer Res. 18:317, 1973.

Stoltz, D.R., L.A. Poirier. C.C. Irving. H.F. Stich,
J.H. Weisburger and H.C. Grice. Evaluation of short-

term tests for carcinogenicity. Toxicol. Appl.
Pharmacol. 29:157. 197 .

Freeman, A.E. and R.H. Huebner. Problems in interpre-
tation of experimental evidence of cell transforma-
tion. J. Nat..Cancer Inst. 50:303. 1973.

McCann, J.. E. Choi. E. Yamasaki and B.N. Ames. Detec-
tion of carcinogens as muta ens in the Salmonella/

microsome test: Assay of 3 0 chemicals. Proc. Nat.
Acad. Sci. U.S.A. 72:5135. 1975.

Brusick. D.J. In vitro mutagenesis assays as predictors
of chemical carcinogenesis in mammals. Clin. Toxicol.
10:79. 1977.

Knudson. A.G. Jr. Mutation and human cancer. Adv.
Cancer Res. 17:317. 1973.

Mondal. S.. D.W. Brankow and C. Heidelberger. Twa-
stage chemical oncogenesis in cultures of C3H/10T1/2
cells. Cancer Res. 36:2254. 1976.

Boutwell. R.R. The function and mechanism of promoters
of zarcinogenesis.CRC Critic. Rev. Toxicol. p. 419.
197 .

Ames, B.N.. W.E. Durston, E. Yamasaki and F.D. Lee.
Carcinogens are mutagens: A simple test system com-
bining liver homogenates for activation and bacteria
{3;3detection. Proc. Nat. Acad. Sci. U.S.A. 70:2281,

Friedrich, U. and P. Coffino. Mutagenesis in S49 mouse
lymphoma cells: Induction of resistance to ouabain.
6-thioguanine. and dibutyryl cyclic AMP. Proc. Nat.
Acad. Sci. U.S.A. 74:679. 1977.

 

Thompson. L.H. and L.M. Baker.Isolation of mutants of
cultured mammalian cells. Method in Cell Biology,
Vol. V1. p. 209.

Trosko. J.E.. C. Chang. L.P. Yotti and E.H.Y. Chu.
Effect of phorbol myristate acetate on the recovery
of spontaneous and ultraviolet light-induced 6-thio-
guanine and ouabain-resistant Chinese hamster cells.
Cancer Res. 37:188. 1977.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

113

Baker. R.M.. D.M; Brunette. R. Mankovitz. L.H. Thompson,
G.F. Whitmore. L. Siminovitch and J.E. Till.
Ouabain - resistant mutants of mouse and hamster
cells in culture. Cell 1:9. 1974.

Lever. J. E. and J. E. Seegmiller. Ouabain - resistant
human lymphoblastoid lines altered in the (Na+ + KT)-
dependent ATPase membrane transport system. J. Cell
Physiol. 88:343. 1976.

Lever, J. E. Genetic alternation in the (Na+ + K+ ) ATPase
transport system.expressed in human lymphoblasts and
their isolated plasma membrane. J. Cell. Physiol.
89:811. 1976.

Mayhew. E. Ion transport by ouabain resistant and
sensitive Ehrlich ascites carcinoma cells. J. Cell.
Physiol. 79:441. 1972.

Rosenberg, H.M. Variant Hela cells selected for their
resistance to ouabain. J. Cell Physiol. 85:135.
1975.

Heidelberger. . Chemical carcinogenesis. Ann. Rev.
Biochem. 44: 79, 1975.

United States Department of Agriculture. Composition of
foods. Agricultural Handbook No. 8. 1963.

Van Soest, P.J. and R.H. Wine. Use of detergents in
the analysis of fibrous feeds. IV. Determination of
plant cell wall constituents. J. Assoc. Office.
Anal. Chem. 50:50. 1967.

Bolin, D.Wl. R.P. King and E.W. Klosterman. A simpli-
fied method for the determination of chromic oxide
(Crzeg) when used as an index substance. Science

116 4.1952.

Grundy, S.M., E.H. Ahrens and T.A. Miettinen. Quanti-
tative isolation and gas-liquid chromatographic
2ngéysis 2f total fecal bile acids. J. Lipid Res.

: 7. 19 5.

Makita. M.. and W;W. we11s. Quantitative analysis of
fecal bile acids by gas liquid chromatography. Anal.
Biochem. 5:523. 1963.

Salantitro. J.P.. I.G. Fairchilds and Y.D. Zgornicki.
Isolation. culture characteristics. and identifica-
tion of anaerobic bacteria from the chicken cecum.

Appl. Microbiol. 27:678. 1974.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

114

Evrard, E. and G. Janssen. Gas-liquid chromatographic
determination of human fecal bile acids. J. Lipid
Res. 9:226, 1968.

Weiser. M.M. Intestinal epithelial cell surface membrane
glycoprotein synthesis. J. Biol. Chem. 248:2536. 1973.

Portman, O.W. and P. Murphy. Excretion of bile acids
and B-hydroxysterols by rats. Arch. Biochema Biaphys.
76:367. 1958.

Reddy. B.S.. T. Narisawa. R. Maronpot. J.H. Weisburger
and E.L. wynder: Animal models for the study of
dietary factors and cancer of the large bowel.Cancer
Res. 35:3421. 1975.

Shurpalekar. K.S.. T.R. Doraiswamy. O.E. Sundaravalli
and M.N. Rao. Effect of inclusion of cellulose in
an "atherogenic" diet on blood lipids of children.
Nature 232:554. 1971.

Chung. K;W.. G.E. Fulk and S.J. Silverman. Dietary
effects on the composition of fecal flora of rats.
Appl. Environ. Microbiol. 33:654. 1977.

Reddy. B.S. Role of bile metabolites in colon carcino-
genesis. Cancer 36:2401. 1975.

Pomare, E.W. and K.W. Heaton. Alteration of bile salt

metabolism by dietary fibre (bran). Brit. Med. J.
4:262. 1973.

Hill. M.J. Colon cancer: A disease of fibre depletion
or of dietary excess? Digestion 11:289. 1974.

Chang. C., J.E. Trosko and T. Akera. Characterization
of ultraviolet light-induced ouabain resistant
mutations in Chinese hamster cells. Mut. Res. in
press.

Shimada, H.. H. Shibuta and M. Yoshikawa. Transformation
of tissue-cultured xeroderma pigmentosum fibroblasts
by treatment with N-methyl-N'-nitro-N-nitrosoguanidine.

Nature 264:547. 1976.

Huberman, E.. L. Aspiras. C. Heidelberger, P.L. Grover
and P. Sims. Mutagenicity to mammalian cells of
epoxides and other derivatives of polycyclic hydro-
carbons. Proc. Nat. Aca. Sci. U.S.A. 68:3195. 1971.

100.Schneider. E.L.. Y. Mitsui. K.U. Au and 3.5. Shorr.

Tissue-specific differences in cultured human diploid
fibroblasts. Exp. Cell Res. 108:1. 1977.

101.

102.

103.

104.

105.

115

Conrad. G.W.. G.W. Hart and Y. Chen. Differences
in vitro between fibroblast-like cells from cornea.
HEart. and skin of embryonic chicks. J. Cell Sci.
26:119. 1977.

Maher, V.M., J.J. McCormick and P.L. Grover. Effect
of DNA repair an the cytotoxicity and mutagenicity
of polycyclic hydrocarbon derivatives in normal and

xeroderma pigmentosum human fibroblasts. Muta. Res.
43:117. 1977.

Madsen. D.. M. Beaver. L. Chang. E. Bruckner-Kardoss
and B. WOstmann. Analysis of bile acids in con-
ventional and germfree rats. J. Lipid Res. 17:
107. 1976.

Silverman, S.J. and A.W. Andrews. Bile acids: Co-
mutagenic activity in the Salmonella-mammalian-

microsome mutagenicity test. J. Nat. Cancer. Inst.
59:1557. 1977.

Jensen. KHA.. I. Kirk, G. Kplmark and M. Westergaard.
Chemically induced mutations in neurospora. Cold
Spr. Harb. Symp. Quant. Biol. 16:245. 1951.

"‘llllllllllllllllllf