JOurn.31 or in.lnerr.3adua.3-e :--earch
.Oluiie -, issue 2 - r, o. eiier.':eceni'i er -i..i:
Colon Cancer Cell Proliferation in Estrogen Receptor Beta (ERB)
Knockout Mice: A step towards Elucidating Alternative Estrogen
Replacement Therapy based on Phytoestrogens with High ERB Binding Activity
Despite decades of controversy concerning the role of estrogens in the prevention of chronic diseases, it is clear
that estrogens exert a protective effect on the risk of colon cancer.1 Due to the elevated rates of breast and
ovarian cancer associated with estrogen replacement therapy (ERT), the incorporation of phytoestrogens in the diet
is becoming an increasingly attractive alternative to traditional estrogen replacement therapy.2 The primary goal
of this project is to further elucidate alternatives to ERT based on phytoestrogens with high estrogen receptor p
(ERp) binding activity that will effectively reduce colon cancer.
ERp is one of two known estrogen receptor subtypes through which estrogenic effects are mediated. Both ERa
and ERp are class I members of the nuclear hormone receptor family and act as ligand-activated nuclear
transcription factors, yet they act on different gene promoters and have different tissue distribution. While
ERa dominates in some specific tissues and is mainly involved in reproductive events, ERp is the more
generally expressed estrogen receptor.3 The lab of Martha Campbell-Thompson was the first to determine that ERp
is expressed in rat gastrointestinal epithelial tissues.4 It has since been found that ERp is selectively lost in
malignant colon tissue.5,6 Such findings indicate that ERp is the dominant receptor in the colon and could
potentially play a role in cell proliferation.
The respective roles of ERa and ERp have been investigated through the use of knockout mice, in which deletion
of one or both of the receptors allows correlation with changes in the effects of estrogen on different tissues
and responses. ERp knockout (BERKO) mice were first generated by inserting a neomycin resistance gene into exon
3 of the coding gene, using homologous recombination in embryonic stem cells.7 RNA analysis
and immunocytochemistry revealed that these mutants lack any means of synthesizing normal ERp. BERKO
mice exhibit phenotypes distinctly different than those of aERKO mice.7 They develop normally and, as young
adults, they are indistinguishable grossly and histologically from their littermates. Unlike males lacking ERa,
male BERKO mice are fully fertile and reproduce normally. Likewise, females are fertile and exhibit normal
sexual behavior, though they produce fewer and smaller litters than wild-type mice due to reduced
ovarian efficiencies.7 Clearly, these models prove useful for further demystifying the roles of the individual
Because mice and other rodents very rarely have spontaneous colon cancer, they must be given a carcinogen
to induce colon tumors. In this study, azoxymethane (AOM) was used to introduce colon cancer to both wild type
and BERKO mice to observe the relative proliferation of colon cancer cells. AOM is one of the primary
carcinogens used to study the effect of dietary agents on colorectal cancer in animal models.8 The induced
tumors share several histopathologic characteristics with human tumors. Thus, this mouse model of
colon carcinogenesis provides a useful method of studying the pathological and molecular changes associated
with sporadic colon cancer in humans.
For examining the potential of dietary agents in chemoprevention trials, bromodeoxyuridine (BrdU) staining
has proven to be a reliable measure of gastrointestinal cell proliferation, which can be used as
intermediate biomarkers of colorectal cancer. BrdU staining allows visualization and comparison of cell
proliferation among different genotypes and genders. Incorporation of BrdU into the DNA of cells allows
identification of the cells in the S phase of the cycle, and thus a representation of actively proliferating cells.9
In light of previous findings on predisposing factors and colon cancer risk, it is hypothesized that the
homozygous male BERKO mice will show the greatest proliferative activity following AOM-induced
colon carcinogenesis. It has been found that, at all ages, incidence and mortality rates for colorectal cancers
are lower in women than in men.10 Further, considering the dominant role of ERp in the colon and colon cancer
cell lines, it follows that the absence of the receptor in BERKO mice will result in increased cancer
susceptibility.11 Following the same logic, it is anticipated that the heterozygous female BERKO mice will
show greater proliferative activity than wild type females.
MATERIALS AND METHODS
Eight mice from Taconic (Germantown, NY) were used: 2 heterozygous female and 2 homozygous male
ERp Knockout mice, 2 female and 2 male wild type C57 strain. A week after arrival, they were all placed on a
soy-free AIN 763 (Tekland, WI) diet to eliminate dietary phytoestrogens. In the second week (7/7/04), all
mice received the first of two azoxymethane (AOM) injections (15 mg/kg) given intraperitoneally (IP). AOM
was obtained from the National Cancer Institute (NCI) and dissolved in ethanol followed by a further dilution in
sterile PBS. The mice were anesthetized by isoflurane inhalation (0.6 1/min oxygen flow, dose 1-4%, to effect)
during injections. The second AOM injection was given a week later (7/14/04). The mice were held for thirteen
weeks after the first injection (total 14 weeks) and weights were taken weekly. For each week that
weight measurements were taken, mean weights of each animal group were calculated and one-way
ANOVAs indicated significant weight changes among groups. Post hoc t-tests followed and revealed which
animal groups had greater weight changes relative to other groups (p< 0.05).
After 14 weeks the mice were sacrificed. Bromodeoxyuridine (BrdU, dissolved in PBS, 10 mg/ml, 200 mg/kg
body weight) injections were given IP an hour prior to death, with 10 minute intervals between animals. Mice
were anesthetized by isoflurane inhalation, and necropsies were completed by cervical dislocation. Blood
was immediately drawn from the heart for the extraction of serum, and organs were perfused with PBS to
remove remaining blood. Liver, spleen, and uterine weights were taken. After flushing the gastrointestinal
organs with phosphate-buffered saline (PBS) and removing the mesentary, the small and large intestines
were examined for tumors. Sections of the liver, small intestine, and colon were added to RNAse Later. The
following organs were harvested for subsequent histopathological analysis: colon, small intestine, liver, gall
bladder, spleen, pancreas, salivary glands, and part of the sternum. Tissues were fixed in PLP, processed to
paraffin, and examined by hematoxylin and eosin (HE) stains. Tail snips were taken for confirmation of the
genotypes provided by the supplier.
Slides were deparaffinized in xylene and rehydrated in a graded alcohol series. Antigens were unmasked by
citrate (10 mM, pH 6.0) retrieval in a microwave oven. After quenching endogenous peroxidase with 3%
H202, blocking buffer (10pl goat serum diluted in 90pl PBS) was applied to each slide for 30 minutes.
Primary antibody solution (100 pl) containing 2 pl mouse anti-BrdU monoclonal antibody (BD Biosciences) diluted
in 1:50 blocking buffer was then applied to each slide and incubated overnight. This was followed by 100 pl
of secondary antibody (anti-mouse IgG), applied for 30 minutes. Sections were stained using a Vectashield Elite
ABC kit (Vector Laboratories) used in conjunction with a Streptavidin/Bioton kit (Vector Laboratories).
Antibody binding was visualized with the chromagen 3,3'-diaminobenzidine (DAB). Finally, sections
were counterstained with hematoxylin for 10 seconds. As a negative control, sections were incubated with a
non-specific mouse IgG monoclonal antibody. As a positive control, mouse ovary and uterus were included in the run.
Using the Zeiss Axioskop Plus with Axiocam HR color camera, 5 images were taken of the 4 colon sections for
each animal using the 20x objective. Incorporation of BrdU, recognized as dark brown nuclear stains, were
evaluated in vertically oriented colonic crypts. The Metamorph software program was used to count and
calculate areas of positive BrdU nuclei in colon crypts. Averages were found for each animal and ultimately for
Genomic DNA was isolated from 1.2 cm tail snips taken from a wild type male and the four BERKO mice (mice 3,
5, 6, 7, 8) using a DNAeasy Tissue Kit (Qiagen, Hilden, Germany) and following the manufacturer's
instructions. Subsequent PCR analysis was carried out on 5 pl of the DNA. The following primers received
from Qiagen (Alameda, CA) were used:
intron 2 (5'-TGGACTCACCACGTAGGCTC-3')
neo (5'-GCAGCCTCTGTTCCACATACAC-3'), and
exon 3: (5'-CATCCTTCACAGGACCAGACAC-3').
Intron 2 and exon 3 were used to confirm wild type genotypes, while intron 2 and neo were used to confirm
mutant genotypes. Samples were held in the thermal cycler for 30 cycles: denaturation at 940C for 30
seconds, annealing at 550C, extension at 720C for 30 seconds, and an ending hold time at 750C for 5
minutes. Samples were electrophoreses in a 2% agarose gel with molecular weight markers and a positive control
of ERP plasmid DNA.
Weights measured over the 13-week period revealed the greatest weight gain in the homozygous male BERKO
mice (Figure 1). The homozygous male that showed the most significant weight change (mouse 7) also had a
notably fatty liver (2.36 g). No gross lesions were observed on the small and large intestines except in
BERKO homozygous males (mice 7 and 8) and one wild type female (mouse 2). However, smooth raised lesions
were likely lymphoid tissues. An ANOVA and post hoc t-test revealed the most significant difference in weight
gain among groups in the 4 to 10 week measurement period that followed the second AOM injection.
The homozygous male BERKO mice showed significantly greater weight change than both the heterozygous
BERKO females and the wild type females (p < 0.05). Likewise, wild type males showed significantly greater
weight gain than both BERKO and wild type females (p < 0.05).
SWid Type Female
42 fild Tipe Male
38* Hterozygous Female
16 a "oWmozyouS Mawe
Figure 1. The relative weights of wild type mice (C57 strain) and ER-beta knockout mice taken over
a thirteen week period, starting on the date of the first azoxymethane (AOM) injection and ending on
the date all mice were sacrificed. Average weights for each group (n = 2) are shown.
BrdU imaging data show similar trends (Figure 2). Male BERKO mouse colon tissues show significantly greater
BrdU-stained cell counts and cell area averages than all other groups. Though it appears that wild type males
showed a greater average cell count than BERKO and wild type females, the ANOVA and post hoc test revealed
no significant difference due to this small sample size. Likewise, there was no significant difference between
averages found for BERKO and wild type females. Images of BrdU-stained nuclei in colonic crypts contrast
the proliferative activity of a wild type female with a wild type male (Figure 3). These images can be compared
with ERp expression throughout the colonic crypt and uterine epithelium (Figure 4). Immunohistochemistry
also allows comparison of aberrant crypt foci (ACF) with normal epithelium (Figure 5).
* Wild Type Females
a 8 Heerozygous BERKO Females
150 a 0 O Wild Type Males
10D0 L EHomozygous BERKO Males
Count Area (pm^2)
Figure 2. Average BrdU-stained cell counts and areas for colon tissue samples of each animal group (n
= 2). Scoring was accomplished using Metamorph software program. Means and standard
deviations found through a one-way ANOVA and post hoc t-test indicated significant differences
among groups. Standard deviations are denoted by error bars; same symbols indicate no
significant difference among samples.
Figure 3. BrdU-stained nuclei in colonic crypts contrasts the proliferate activity of Fl, an Erb wild
type female, with M1, an Erb wild type male.
A ERB3 in Colon
A*r d'. ' ~-.Wt*
* Goblet ceb
] Daughter ces
Figure 4. ERP expression was observed in both epithelium and throughout the crypt (A, B). Mouse
IgG control showed no staining (C). No differences were detected in ERp intensity levels or distribution
in the colon of wild-type males and females (D). Uterine epithelial expression of ERp in
an ovariectomized rat with control replacement compared to lowered ERp expression in
uterine epithelium of an ovariectomized rat with estrogen replacement (E).
AC F -
Figure 5. Immunohistology of an aberrant crypt foci (ACF) in a BERKO male mouse. ACF are early
lesions that progress to colon tumors. Immunohistochemistry for p-catenin (A), a member marker,
and BrdU (B), cell proliferation, shows increased nuclear and cytoplasmic levels in the ACF and
numerous S phase cells, respectively, compared to the normal epithelium (NE).
The PCR results of genotyping confirmed genotypes provided by the supplier.
For determining the relative effects of chemical carcinogen-induced colon cancer on wild type and BERKO mice,
the observed BrdU incorporation into cells served as the primary colonic biomarker. BrdU-stained cells are
indicators of DNA synthesis and thus allow comparison of proliferative activity among animal genotypes and
genders. It was hypothesized that both male and female BERKO mice would show greater proliferative activity
than their wild type counterparts. Moreover, it was predicted that the male BERKO mice would show the
greatest proliferative rate due to greater cancer susceptibility of the male gender.10
As anticipated, the homozygous male BERKO mice showed significantly higher weight gain, BrdU-stained cell
area, and BrdU-stained cell count averages. The higher proliferative rate indicated by these values implies
increased susceptibility to cancer. These findings suggest that without the presence of ERp to somehow facilitate
cell proliferation, males are at greater risk of developing colon cancer. Estrogen replacement therapy (ERT) has
been reported to have protective effects on the incidence and size of colonic polyps.12 It follows that ERp,
the dominant estrogen subtype in the colon, may significantly contribute to observed protective effects. This
may explain why wild type males have fewer and smaller BrdU-stained cells than BERKO males.
Both wild type females and BERKO females showed less weight gain than all males. Although it appeared that
all females also showed less proliferative activity than wild type males, BrdU-stained cell count averages did
not reveal significant differences between groups. However, clinical studies give strong indications that
female hormones decrease colon cancer risk even in women with genetic dispositions.10 This protective effect
of female hormones has been observed in similar studies using animal models for chemical carcinogen-induced
ERp expression in normal colonic epithelium, particularly at the bottom portion of colonic crypts, suggests that
it plays a significant role in the growth and regeneration of normal colonic mucosa.11 Therefore, it is interesting
that the highest levels of BrdU-labeled cells are observed in the same area, the lower proliferative zone of the
crypt. Indeed, malignant tissues show a marked loss of ERp , suggesting that its absence may allow increased
cell proliferation. In contrast, overexpression of ERp results in the apoptosis of cancerous colonocytes.6
Moreover, previous studies on human colon cancer cells have shown that overexpression of ERp resulted in
reduced anchorage independent colony formation, fewer epidermal growth factor (EGF) receptors, and
decreased proliferation in response to EGF.14 Certainly, the varying levels of ERp expression in a large number
of colorectal cancers (CRC) have significant implications for the treatment and prevention of CRC.15
Rodent models provide guidance in the selection of prevention approaches to human colon cancer. Use of
the appropriate genetically engineered mouse (GEM) models allows relatively fast and inexpensive testing
of compounds selected as potential preventative agents.16 The BERKO mouse model provides a valuable means
of studying the specific functions of ERp. Such models have begun to give insight into the metabolism of estrogens
in the colon. For instance, knockout models have been used to describe instances that the estrogen receptors work
in opposite ways. In colon carcinogenic cells, for example, ERp activates apoptosis whereas ERa acts as a
Although analysis of these knockout mice has been key to expanding our understanding of the physiological roles
of estrogens, this approach is not without drawbacks. For example, the developmental effects of estrogens cannot
be separated from those of adulthood because the BERKO mice produced to date are not conditional
knockouts.17 Currently, the goal is to establish improved models that can be used for the predictive testing
of preventive responses in humans.16
The AOM model is frequently used to test putative chemopreventative agents for colon cancer. The organo-
specific carcinogen produces colon tumors in susceptible mouse strains that exhibit many of the pathological
features found in sporadic forms of the human disease. Like human tumors, AOM-induced tumors show
microsatellite instability and are often mutated on K-ras and B-catenin genes.8 While such similarities in
tumor pathogenesis exist, important distinctions must be made between rodent carcinogen models and human
CRC. AOM-induced tumors in rodents are seldom mutated at the Apc gene (15%), are never mutated at the
p53 gene, and have a low tendency to metastasize.8 Animal models have shown that male rats have a higher risk
of developing colon cancer compared with their female counterparts when exposed to dimethylhydrazine,
the carcinogen from which AOM is derived.15 Though males are generally more susceptible to colon cancer,
the specific effects of the carcinogen must be taken into account when considering the differences observed in gender.
While the results of this study are telling, the small group sample sizes must be taken into account. At the time
this study was conducted, the lab could only manage eight animals. However, similar studies involving AOM-
induced carcinogenesis in BERKO mice are currently being conducted with increased sample sizes. Additionally,
the PCR primers used for genotyping continue to be investigated and have been proven successful in other
trials.7 This project will be incorporated into a more extensive publication detailing the role of
phytoestrogen supplementation in colon cancer prevention. Once it has been conclusively demonstrated that
BERKO mice show greater proliferative activity in the colonic epithelium, subsequent studies will test
whether phytoestrogens with high ERp activity will reduce experimentally induced colon cancer by
decreasing proliferative activity. Ultimately, such studies may provide a mechanistic explanation for the effects
of estrogens on the chemoprevention of colon cancer.
1. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ, et al.
2. Guo JY, Li X, Browning JD Jr, et al. Dietary soy isoflavones and estrone protect ovariectomized ERalphaKO and
wild-type mice from carcinogen-induced colon cancer. Nutr. 2004;134:179-182.
3. Gustafsson J. Estrogen receptor beta- a new dimension in estrogen mechanism of action. J Endocrinol.
4. Campbell-Thompson M. Estrogen receptor alpha and beta expression in upper gastrointestinal tract with regulation
of trefoil factor family 2 mRNA levels in ovariectomized rats. Biochem Biophys Res Commun. 1997;240:478-483.
5. Campbell-Thompson ML, Lynch IJ, Bhardwaj B. Expression of estrogen receptor subtypes and ERb isoforms in
colon cancer. Cancer Research. 2001:61:632-640.
6. Qiu Y, Waters CE, Lewis AE, Langman MJ, and Eggo MC. Oestrogen-induced apoptosis in colonocytes
expressing oestrogen receptor beta. Journal of Endocrinology. 2002;174:369-377.
7. Krege JH, Hodgin JB, Couse JF, et al. Generation and reproductive phenotypes of mice lacking estrogen
receptor beta. Proc Nat/ Acad Sci USA. 1998;95:15677.
8. Corpet DE, Pierre F. Point: From animal models to prevention of colon cancer. Systematic review of
chemoprevention in min mice and choice of the model system. Cancer Epidemiol Biomarkers Prev. 2003;12:391-400.
9. Campbell-Thompson M, Lauwers GY, Reyher KK, Cromwell J, Shiverick KY. 17Beta-estradiol
modulates gastroduodenal preneoplastic alterations in rats exposed to the carcinogen N-methyl-N'-
nitro-nitrosoguanidine. Endocrinology. 1999;140: 4886-4894.
10. Ries LA, Wingo PA, Miller DS, et al. The annual report to the nation on the status of cancer, 1973-1997, with a
special section on colorectal cancer. Cancer. 2000;88:2398-2424.
11. Foley EF, Jazaeri AA, Shupnik MA, Jazaeri 0, Rice LW. Selective loss of estrogen receptor beta in malignant
human colon. Cancer Res. 2000;60:245-248.
12. Chen MJ, Longnecker MP, Morgenstern H, Lee ER, Frankl HD, Haile RW. Recent use of hormone replacement
therapy and the prevalence of colorectal adenomas. Cancer Epidemiol Biomarkers Prev. 1998;7:227-230.
13. Smirnoff P, Liel Y, Gnainsky J, Shany S, Schwartz B. The protective effect of estrogen against chemically
induced marine colon carcinogenesis is associated with decreased CpG island methylation and increased mRNA
and protein expression of the colonic vitamin D receptor. Oncol Res. 1999; 11:255-264.
14. Campbell-Thompson M, Bhardwaj B. Expression of estrogen receptor (ER) subtypes and ERbeta isoforms in
colon cancer. Cancer Res. 2001;61:632-640.
15. Xie L, Yu J, Luo H. Expression of estrogen receptor b in human colorectal cancer. World J Gastroenterol.
16. Green JE, Hudson T. The promise of genetically engineered mice for cancer prevention studies. Nat Rev
17. Harris HA, Katzenellenbogen JA, Katzenellenbogen BS. Characterization of the Biological Roles of the
Estrogen Receptors, ERp and ERB, in Estrogen Target Tissues in Vivo through the Use of an ERP-Selective
Ligand. Endocrinology. 2002;143:4172-177.
Back to the Journal of Undergraduate Research
College of Liberal Arts and Sciences I University Scholars Program I University of Florida |
ï¿½ University of Florida, Gainesville, FL 32611; (352) 846-2032.