| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Departments of Nutritional Sciences [J. S., Y-I. K.], Medicine [K-J. S., Y-I. K.], Pathology [A. M.], and Surgery [S. G.], University of Toronto, Toronto, Ontario, M5S 1A8 Canada; Division of Gastroenterology [Y-I. K.], Department of Medicine, St. Michael’s Hospital, Toronto, Ontario, Canada M5B 1W8; and Samuel Lunenfeld Research Institute Centre for Cancer Genetics [C. A., S. G.], Mount Sinai Hospital, Toronto, Ontario, Canada M5G 1X5
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
Two animal studies, performed in the dimethylhydrazine colorectal cancer rat model, have supported a causal relationship between folate deficiency and colorectal cancer (22 , 23) . These studies have also shown a dose-dependent protective effect of modest levels of dietary folate supplementation up to four times the basal dietary requirement (22 , 23) . Levels of dietary folate greater than four times the dietary requirement did not convey additional benefits; in fact, there was a nonsignificant trend toward increased colorectal tumorigenesis in rats fed a supraphysiological dose of folate (20 times the daily requirement; Ref. 23 ). This study suggests that supplemental folate may have two distinct actions in this model. At modest levels of supplementation beyond the dietary requirement, folate seems to possess an inhibitory effect on the genesis and progression of colorectal neoplasms (22 , 23) . However, exceptionally high supplemental folate levels may promote the progression of chemically induced colorectal neoplastic foci (23) . In support of this latter finding, dietary folate supplementation exceeding the basal requirement by 1000 times increased the development of ACF,3 the probable earliest precursor of colorectal cancer (24) , compared with a control diet in another rat study in which azoxymethane (a metabolite of dimethylhydrazine) was used (25) .
Although some similarities do exist, tumor development in chemical rodent models of colon cancer differs in several important respects from that observed in humans (26 , 27) . The chemically induced carcinomas often arise from flat foci of dysplasia rather than from adenomatous polyps. The relatively high doses of genotoxic chemical carcinogens differ from the natural etiological causes involved in most cases of sporadic colorectal cancer in humans. Most importantly, molecular genetics of chemical rodent models are significantly different from those observed in human colorectal carcinogenesis. For example, although K-ras mutations are as frequently mutated in these rodent models as in human colorectal cancer (28 , 29) , the Apc and p53 genes, two commonly mutated genes in human colorectal cancer (30) , are either mutated to a much lesser extent or not mutated at all in these rodent models (31, 32, 33, 34, 35) . Also, recent evidence suggests that chemically induced colorectal carcinogenesis in rodents involves alterations in unique colon cancer susceptibility genes that are not implicated in human colorectal cancer (36 , 37) .
Recently developed genetic murine models characterized by the spontaneous development of small intestinal and colonic tumors have provided an excellent opportunity to investigate the effects of environmental and genetic factors in both familial and sporadic colorectal carcinogenesis. One such model, Apc+/- Msh2-/-, was generated by crossing Min (multiple intestinal neoplasia; Apc+/-) mice with Msh2-/- mice (38) . Min mice carry a heterozygous germ-line mutation at codon 850 of the mouse Apc gene and develop approximately 25–75 small intestinal adenomas and 1–5 colorectal adenomas by 160–180 days, at which time they become moribund and die from anemia and intestinal obstruction (39 , 40) . The wild-type Apc allele is lost in polyps from Min mice (41) , and, therefore, this model resembles the syndrome of familial adenomatous polyposis coli in humans (42) . The MSH2 gene is one of several mismatch repair genes that ensure accurate replication of the genome during cell division (43) . Germ-line mutations in the mismatch repair genes have been implicated in HNPCC and somatic mutations in about 15% of sporadic colorectal cancer (43) . Mutations in the mismatch repair genes result in a "mutator phenotype," in which loss of postreplicative DNA repair increases the mutation rate, and results in a replication error phenotype or microsatellite instability (43) .
Apc+/- Msh2-/-mice display an accelerated intestinal adenoma phenotype and develop numerous dysplastic colonic ACF (38) . They develop approximately 350 small intestinal tumors, 8 colon tumors, and 55 ACF by 80 days of age, at which time they become moribund and die of anemia or bowel obstruction (38) . The average time required for a nascent tumor to develop into a macroscopically visible adenoma in these mice was estimated to be 42 and 27 days in the small intestine and colon, respectively (38) . The Apc+/- Msh2-/- mouse model provides many advantages compared with previous animal models of colorectal neoplasia. These mice develop intestinal neoplasms spontaneously without the need for carcinogens. In particular, in contrast to other genetic models of colorectal cancer, Apc+/- Msh2-/- mice develop colonic ACF, the probable earliest precursor of colorectal cancer observed in humans (24) . This allows this model to be used in studies examining early events of colorectal neoplasia. The Apc+/- Msh2-/- mouse model is phenotypically and genotypically more similar to human colorectal cancer, thereby making this model more clinically relevant than other available animal models. Because of the accelerated nature of tumorigenesis in this model, it is possible to test the effects of chemopreventive agents on initiation and progression of small intestinal and colorectal polyps and ACF in a relatively short time.
This study investigated the effects of dietary folate on the initiation and promotion of intestinal tumorigenesis in Apc+/-Msh2-/- mice. Furthermore, possible mechanisms by which dietary folate can modulate small intestinal and colorectal tumorigenesis in this model were investigated.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
Mice.
Min (Apc +/-Msh2+/+) mice were bred at the Samuel Lunenfeld
Research Institute on the C57BL/6J strain (original breeding pair from
The Jackson Laboratory, Bar Harbor, ME). The generation of
Msh2-/- mice has been described previously (44
, 45)
. Apc +/-Msh2+/- mice were generated by crossing
male Min mice with female Msh2+/- mice. Male Apc+/-
Msh2+/- mice were then crossed with female Msh2+/-
mice to generate Apc+/- Msh2-/- mice as described
previously (38)
.
Genotyping.
Ear punch tissue was processed for genotyping using PCR-based assays to
determine Apc and Msh2 status as described
previously (38
, 44)
.
Diet.
Mice were fed an amino-acid-defined diet containing either 0 or 8 mg
folate/kg diet (Dyets, Bethlehem, PA; Ref. 46
). These
diets constitute standard means of inducing folate deficiency or
providing supplemental dietary folate in rodents (46)
. The
diet containing 0 mg folate/kg produces progressive folate deficiency
of a moderate degree without anemia, growth retardation, or premature
death through week 5, after which systemic folate indicators stabilize
(22)
. Although this diet is completely devoid of folate,
intestinal microflora are capable of de novo synthesis of
folate, some of which is incorporated into the tissue folate of the
host; this prevents severe folate deficiency (47)
. The
degree of folate deficiency produced by this diet is comparable with
that associated with the increased risk of colorectal cancer in humans
(3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17
, 19, 20, 21)
. The diet containing 8 mg folate/kg
represents folate supplement 4 times the basal dietary requirement for
rodents (i.e., 2 mg folate/kg; Ref. 48
). This
level of folate was chosen because the 8 mg/kg level has consistently
provided a degree of chemoprevention against colorectal cancer in
previous rodent studies (22
, 23)
. These diets contained
50 g cellulose/kg, 60% calories as carbohydrates, 23% fat, and
17% L-amino acid. The amount of methyl donors,
methionine, choline and vitamin B12, were
8.2 g, 2.0 g and 50 µg per kg diet, respectively. Diet and
water were provided ad libitum. Detailed composition of the
diets is shown in Tables 1
2
.
|
|
Enumeration of Small Intestinal Adenomas, Colonic ACF, and
Tumors.
Mice were killed by cervical dislocation. Intestines were immediately
removed and flushed with Krebs buffer solution to remove fecal debris.
The entire length of the small intestine and colon was opened
longitudinally, laid flat on Whatman filter paper, and fixed for at
least 3 h in 10% neutral buffered formalin. The mucosa of the
fixed small intestine and colon was stained with methylene blue and
examined in a blinded fashion for tumors and ACF by gross inspection
and light microscopy as described previously (49)
. All of
the small intestine and colon tumors and colonic ACF were counted in a
blinded fashion. Previous studies have shown that all small intestinal
tumors are adenomas and the majority of colonic ACF are dysplastic in
this model (38)
. Representative small bowel tumors
(adenomas) and all of the colonic tumors were processed in a standard
manner for H&E staining and histologically analyzed by a
gastrointestinal pathologist [A. M.] blinded to the study
groups.
Folate Concentration Determination.
At the time of death, blood was withdrawn from the heart into
vacutainer tubes containing EDTA using a preheparinized 18-gauge needle
and was centrifuged at 800 x g for 10 min at
4°C. Serum was stored at -70°C in 0.5% ascorbic acid for the
serum folate assay. Serum folate concentrations were measured by a
microtiter plate assay using Lactobacillus casei as
described previously (50)
.
DNA Extraction.
Areas corresponding to microscopically confirmed small intestinal
and colonic adenomas on H&E staining were marked on matched paraffin
blocks. DNA from adenomas was extracted as crude preparations from
paraffin blocks using proteinase K lysis mix [10 mM
Tris-HCl (pH 8.0), 100 mM KCl, 2.5 mM
MgCl2, 0.45% Tween 20, and 1 mg/ml proteinase
K] as previously described (45)
. Care was taken to avoid
contamination from adjacent nonneoplastic tissues. The sections were
homogenized in the lysis mix and digested for 1 h at 65°C
followed by 10 min at 95°C. Extracted DNA was stored at -20°C
until DNA methylation and microsatellite assays. Small intestinal and
colonic DNA from normal mucosa was extracted from areas corresponding
to normal histology from paraffin blocks in a similar fashion
(45)
. DNA from the liver, snap-frozen at the time of
sacrifice and stored at -70°C, was extracted by a standard technique
using a lysis buffer containing proteinase K followed by phenol,
chloroform, and isoamyl alcohol organic extraction (51)
.
Genomic DNA Methylation Determination.
The methylation status of CpG sites in genomic DNA from the liver was
determined by the in vitro methyl acceptance capacity of DNA
using [3H-methyl]SAM as a methyl donor and a
prokaryotic CpG DNA methyltransferase, Sss1, as previously described
(52)
. The manner in which this assay is performed produces
a reciprocal relationship between the endogenous DNA methylation status
and the exogenous [3H]methyl incorporation.
Briefly, 500 ng of liver DNA were incubated with 2 µCi
[3H-methyl]SAM (New England Nuclear, Boston,
MA), 3 units Sss1 methylase (New England Biolabs, Beverly, MA), 1x
Sss1 buffer [120 mM NaCl, 10
mM Tris-HCl (pH 7.9), 10 mM
EDTA, and 1 mM DTT] in a total volume of 30 µl
at 30°C for 1 h. The in vitro methylated DNA was
isolated from a 15-µl aliquot of the reaction mixture by filtration
on a Whatman DE-81 ion exchange paper (Fisher Scientific, Springfield,
NJ). The DNA was washed three times with 20 ml of 0.5
M sodium phosphate buffer (pH 7.0), followed by 2
ml of 70% ethanol and 2 ml of absolute ethanol. The filters were
air-dried and the resulting radioactivity of the DNA retained in the
filters was measured by scintillation counting using a nonaqueous
scintillation fluor. The amount of radiolabel bound to a filter
from an incubation mixture lacking Sss1 was used as background and was
subtracted from the values obtained with mixtures containing DNA. The
background value was always <1% of the uptake observed with DNA
samples. All of the analyses were performed in duplicate.
Genomic DNA methylation status assessed by the in vitro methyl acceptance capacity of DNA was confirmed by a methyl-sensitive restriction digestion method. Three samples from folate-depleted and -supplemented groups at 3 and 6 week diet starts were selected as representatives of their respective groups because their serum folate concentrations were closest to the respective means of their groups. Genomic liver DNA was digested with HpaII (New England Biolabs) overnight at a final concentration of 20 units/µg DNA at 37°C in a buffer provided by the supplier. One additional DNA sample from each group was selected and digested with MspI (New England Biolabs). HpaII and MspI are isoschizomers that cleave the sequence 5'-CCGG-3' between the two cytosine residues. HpaII is unable to cleave CCGG if the internal cytosine is methylated whereas MspI can cleave CCGG regardless of the methylation of the internal cytosine residue (53) . A HpaII specific oligonucleotide (5'-TAT AGC CGG CTA TA-3') was added in 10-fold molar excess of genomic DNA to increase the cutting efficiency of HpaII (54) . Digested DNA (150 ng/lane) was separated on a 0.8% agarose gel and transferred to a nylon membrane using a Southern blot technique under standard conditions (55) . A mouse centromeric minor satellite repeat sequence derived from plasmid MR150 (generously provided by Dr. Janet Rossant, Samuel Lunenfeld Research Institute, Toronto, Ontario, Canada) was used to hybridize restricted DNA fragments as previously described (56) . pMR150 was labeled with digoxigenin-11-dUTP using the PCR DIG Probe Synthesis kit (Boehringer Mannheim, Montreal, Quebec, Canada) and hybridized with the membrane that was subsequently exposed according to the manufacturer’s protocol.
Microsatellite Instability Assay.
Microsatellite instability was detected by comparison of
electrophoretic mobility of amplified normal and neoplastic DNA using
primers from five loci on mouse chromosomes 1 (D1 Mit4), 2
(D2 Mit16), 5 (D5 Mit10), 6 (D6 Mit8),
and 10 (D10 Mit2) as described previously (Research
Genetics, Huntsville, AL; Refs. 38
, 45
, 46
). Matched
normal and neoplastic DNA from the small intestine and colon were
selected from each dietary group at two time points. Primer labeling
and PCR amplification were performed according to the protocol supplied
by Research Genetics for Mouse Map Pairs with the following
modification. A single primer (10–15 pmol) from each primer pair was
end-labeled using T4 polynucleotide kinase (New England Biolabs) and
[
-32P]dATP (New England Nuclear) according
to Sambrook (55)
. A 4-µl aliquot of the PCR products was
mixed with formaldehyde dye mix (2 µl), denatured at 85°C for 2 min
and electrophoresed on 7% polyacrylamide gels under denaturing
conditions for 2–3 h. Gels were dried and exposed to X-ray film for
12–72 h. A positive case was defined as one showing instability at one
or more loci, confirmed in two independently performed PCR reactions.
Statistics.
The distribution of each variable was assessed graphically to determine
whether it was normally distributed. Those variables that were not
normally distributed were subjected to logarithmic transformation
before performing a significance test. Comparisons of means between the
folate-depleted and -supplemented groups were assessed by Student’s
t test. Statistical analyses were performed using SYSTAT 5
for Macintosh (Systat, Evanston, IL). All of the significance tests
were two-sided, and were considered statistically significant if the
observed significance level was less than 0.05. Results are expressed
as mean ± SE of the untransformed data.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
). No premature deaths occurred. This finding indicates that
folate deficiency in the mice that were fed 0 mg folate/kg was
moderate, otherwise severe growth retardation or premature death would
have occurred (57)
. The average daily food consumption,
which was determined on a preassigned day of each week, was also
similar between the two groups (data not shown).
|
). The mean serum folate concentrations of the folate-depleted and
-supplemented mice were comparable with those observed in rats placed
on similar diets (22
, 23)
.
|
). This protective
effect of dietary folate supplementation on the development of small
intestinal adenomas was consistently observed in the duodenum, jejunum,
and ileum (P < 0.02, Table 3
).
In the colon, the folate-supplemented mice developed significantly
fewer ACF (by 2.8-fold) compared with the folate-deficient mice
(P = 0.028; Table 3
). Although
folate-supplemented mice had a 2.8-fold lower incidence of
colonic adenomas compared with the folate-depleted mice, this fell
short of statistical significance (P = 0.1;
Table 3
).
Genomic DNA Methylation Status.
The extent of genomic DNA methylation was determined in the liver.
There was no difference in the extent of genomic DNA methylation as
assessed by the in vitro methyl acceptance capacity of DNA
and a methyl-sensitive restriction digestion method between the
folate-depleted and -supplemented groups (Fig. 2, A and C
). These data suggest that a moderate
degree of folate deficiency of 8 weeks duration did not result in a
significant degree of genomic DNA hypomethylation.
|
). These data suggest that widespread microsatellite
instability is probably not a major factor leading to accelerated small
intestinal and colonic adenomas in these mice. These data also suggest
that superimposed folate deficiency and supplementation had no
significant observable effect on microsatellite instability.
|
|
Effects of Dietary Folate Deficiency and Supplementation Beginning at 6 Weeks of Age (i.e., after the Establishment of Neoplastic Foci) on Intestinal TumorigenesiS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
). The mean
weight of the folate-deficient mice increased in parallel to that of
the folate-supplemented mice and remained 16–19% lower than that of
the folate-supplemented mice at each time point throughout the
experiment (P < 0.03, Fig. 1B
).
No premature deaths were observed. The average daily food consumption
was also similar between the two groups (data not shown).
Serum Folate Concentrations.
The mean serum folate concentration was significantly lower, by
4.8-fold, in mice placed on the folate-depleted diet than those on the
supplemented diet (P = 0.004; Table 3
). The
mean serum folate concentrations of the folate-depleted and
-supplemented mice were comparable with those observed in rats placed
on corresponding diets (22
, 23)
.
Development of Small Intestinal Adenomas and Colonic ACF and
Adenomas.
Mice fed the folate-deficient diet had a 4.2-fold lower number of small
intestinal adenomas compared with those fed the folate-sufficient diet
(P = 0.001; Table 3
). When each of the
segments of the small intestine was examined individually, this pattern
was observed throughout the length of the small intestine
(P < 0.02; Table 3
).
In the colon, no significant difference was observed in the number of
ACF and adenomas between the two diet groups (Table 3)
.
Genomic DNA Methylation Status.
The mean incorporation of [3H-methyl]SAM into
hepatic DNA, which is inversely related to the extent of genomic DNA
methylation, was 56% lower in the folate-deficient group than in the
folate-supplemented group, which indicated a significantly higher
degree of genomic DNA methylation in the folate-deficient group
compared with the folate-supplemented group (P
< 0.05; Fig. 2B
). This finding was confirmed by the
methyl-sensitive restriction digestion method, in which lower molecular
weight HpaII-digested fragments were observed in the
folate-supplemented group compared with the folate-deficient group
(Fig. 2D
). These data suggest that a moderate degree of
folate deficiency of 5 weeks duration paradoxically increased the
extent of genomic DNA methylation.
Microsatellite Instability.
A total of 20 small intestinal and colonic adenomas from 12 mice were
analyzed for microsatellite instability. No difference in
electrophoretic mobility was observed between normal and tumor tissue
from either of the diet groups at the five loci tested (Fig. 3B
).
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
. However, the number of small intestinal adenomas in the moderately
folate-deficient mice was not significantly different from that
observed in Apc+/-Msh2-/- mice on a control diet
containing the basal dietary requirement of folate (i.e., 2
mg folate/kg diet/day) for the same duration of time (38)
.
This is likely related to the fact that by 11 weeks of age, the small
intestine of Apc+/-Msh2-/- mice is maximally saturated
with adenomas and, hence, no new adenomas can develop
(38)
. Thus, the possibility that folate deficiency might
have had an additional promoting effect on the development of small
intestinal adenomas beyond the expected maximum number of adenomas
observed in Apc+/-Msh2-/- mice on control diets, cannot
be ruled out because the effect of this genotype is already maximal for
any additional effects of folate deficiency to be seen.
In contrast to its effect given before the establishment of neoplastic
foci, dietary folate had an opposite effect on the development of small
intestinal adenomas when given after the establishment of neoplastic
foci. In this situation, a moderate degree of folate deficiency induced
by dietary depletion of folate significantly decreased the number of
small intestinal adenomas by 4.2-fold compared with folate
supplementation at four times the basal dietary requirement for rodents
(Table 3)
. Dietary folate supplementation did not significantly
increase the number of small intestinal adenomas beyond the maximum
number of adenomas observed in Apc+/-Msh2-/- mice on a
control diet containing the basal dietary requirement of folate for the
same duration of time (38)
. Again, this is probably
related to the saturation of the small intestine with adenomas by 11
weeks of age in Apc+/- Msh2-/- mice (38)
.
Therefore, the possibility that folate supplementation might have had
an additional promoting effect on the development of small intestinal
adenomas beyond the expected maximum number of adenomas observed in
Apc+/- Msh2-/- mice on control diets cannot be ruled out.
The effect of dietary folate on the development of colonic ACF and
adenomas was different from that observed for small intestinal
adenomas. Dietary folate supplementation, started before the
establishment of neoplastic foci, significantly decreased the number of
colonic ACF by 2.8-fold compared with a moderate degree of folate
deficiency (Table 3)
. Although the folate-supplemented mice had a 67%
reduction in the number of colonic adenomas compared with the
folate-deficient mice, this fell short of statistical significance
(Table 3)
. In contrast, dietary folate intervention started after the
establishment of neoplastic foci had no significant effect on the
development of colonic ACF and adenomas (Table 3)
.
The protective effect of dietary folate supplementation of a moderate
degree (4 times the basal requirement) on small intestinal adenomas and
colonic ACF, started before the establishment of neoplastic foci, in
Apc+/- Msh2-/- mice corroborates previous studies in
dimethylhydrazine-treated rats (22
, 23)
. Our data also
support prior epidemiological observations that suggest a 40%
reduction in the risk of colorectal adenoma and cancer in individuals
with the highest dietary folate intake compared with those with the
lowest intake (3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
. The present study also confirms
findings from a recent prospective study involving 88,756 female
nurses, which showed a 75% reduction in colorectal cancer risk in
women using multivitamin supplements containing 
400 µg folic acid
for 15 years after controlling for known confounding factors
(17)
. Except for one study (8)
, however, it
is not known whether dietary or supplemental folate taken in these
epidemiological studies was started before neoplastic foci were present
in the colon (3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
. Randomized folate intervention
studies in human subjects free of colon adenomas or cancer are
necessary to confirm our present findings in humans. Taken together,
these observations suggest that a modest degree of folate
supplementation may prevent the development of colorectal neoplasia.
However, our study suggests that folate intervention should be started
before the establishment of neoplastic foci.
Although the folate-supplemented mice had a 67% reduction in the
number of colonic adenomas compared with the folate-deficient mice when
intervention with dietary folate started before the establishment of
neoplastic foci, this fell short of statistical significance (Table 3)
.
This may be related to the fact that the number of histologically
verified colonic adenomas per mouse was too few for an adequate
statistical comparison. Another possibility is that the number of mice
in each group was too small to detect a statistically significant
difference in the number of colonic adenomas between these two groups
(i.e., type II error). This finding concerning colonic
adenomas may seem to be inconsistent compared with the protective
effect observed for colonic ACF. However, although Apc+/-
Msh2-/- mice develop numerous ACF, these ACF do not seem to
contribute significantly to the colonic adenoma population
(38)
. What factors limit the progression of ACF to
adenomas in this model is unclear. It may simply be a lack of time;
there may not be enough time for adenomas to develop in these mice
inasmuch as they have a life span of only 11–13 weeks
(38)
. It also seems that ACF develop postnatally whereas
colonic adenomas arise perinatally in these mice (38)
.
The inhibitory effect of dietary folate deficiency started after the establishment of neoplastic foci, on small intestinal tumorigenesis in the present study, is not entirely surprising. Folate is an important factor for a number of metabolic pathways that involve the transfer of one-carbon groups (58) . Among such pathways are the biosyntheses of purines and thymidylates, and as such, folate plays a key role in DNA replication and cell division (58) . Consequently, folate deficiency in tissues with rapidly replicating cells results in ineffective DNA synthesis. In neoplastic cells, in which DNA replication and cell division are occurring at an accelerated rate, interruption of folate metabolism has been observed to cause ineffective DNA synthesis, which results in the inhibition of tumor growth (59) . Indeed, this has been the basis for antitumor therapy using a number of antifolate agents, including methotrexate and 5-fluorouracil (59) . It has been shown in experimental models that growth of a transplanted cancer is inhibited in folate-deficient rats (60) , that folate deprivation reduces the growth of virally induced cancers (61) , and that the time required for developing nerve sheath tumors in transgenic mice is significantly delayed by restricting the level of folate in diet (62) . Furthermore, the addition of folate to established tumors has been shown to result in an "acceleration phenomenon." For instance, children with acute leukemia treated with folate supplementation experienced an accelerated progression of leukemia (63) . Taken together, these observations suggest that folate deficiency, started after the establishment of neoplastic foci, has an inhibitory effect of tumor progression or may even cause tumor regression.
In contrast to its inhibitory effect on small intestinal adenomas, dietary folate deficiency, started after the establishment of neoplastic foci, had no significant effect on the number of ACF and adenomas in the colon. It may be that rapidly developing small intestinal adenomas with accelerated tumorigenesis in these mice are most susceptible to the effect of folate depletion, whereas colonic ACF and adenomas with a normal or slower growth rate may not. Another possibility is a lack of time for folate deficiency to exert its effect on established colonic neoplastic foci because most mice die by 11 weeks of age (38) .
The mechanisms by which dietary folate can modulate intestinal tumorigenesis have not been clearly elucidated. In the present study, two of the proposed mechanisms, genomic DNA methylation and mismatch repair, were investigated. Folate is an important factor in DNA methylation (64) , which is an important epigenetic determinant in gene expression (an inverse relationship), in the maintenance of DNA integrity and stability, and in the development of mutations (65 , 66) . Genomic and proto-oncogene-specific DNA hypomethylation seems to be an early, and consistent, event in carcinogenesis (65 , 66) including that of colorectal cancer (30 , 53 , 67) . In addition, site-specific hypermethylation at specific CpG islands located near or at the promoter region of tumor suppressor and mismatch repair genes seems to be an important mechanism in gene silencing in colorectal carcinogenesis (68 , 69) .
In the present study, a moderately folate-deficient diet, provided for
8 weeks (i.e., a 3-week start), did not produce a
significant degree of genomic DNA hypomethylation in the liver compared
with a folate-supplemented diet (Fig. 2, A and C
). This finding is consistent with previous observations
made in rats subjected to the similar diets (23
, 70)
. This
suggests that alterations in genomic DNA methylation are not a major
mechanism by which folate supplementation that started before the
establishment of neoplastic foci significantly reduced the number of
small intestinal adenomas and colonic ACF in Apc+/-
Msh2-/- mice. In contrast, a moderately folate-deficient diet,
provided for 5 weeks (i.e., a 6-week start), increased the
extent of genomic DNA methylation in the liver compared with a
folate-supplemented diet in the present study (Fig. 2, B and D
). This finding may appear paradoxical but is consistent
with a prior observation that demonstrated that severely
folate-depleted rats had a 59% increase in the extent of genomic DNA
methylation in the liver compared with control rats at 6 weeks of
folate depletion (71)
. This may be related to the
observations that states associated with diminished availability of SAM
result in an enhancement of DNA methyltransferase activity, the enzyme
responsible for DNA methylation (72
, 73)
. Therefore,
moderate folate deficiency of 5 weeks is associated with an increased
degree of genomic DNA methylation attributable to a compensatory
up-regulation of DNA methyltransferase, but, by 8 weeks of folate
deficiency, this effect is no longer observed. The increased extent of
genomic DNA methylation associated with moderate folate-deficiency of 5
weeks’ duration was associated with a significant reduction in the
number of small intestinal adenomas compared with
folate-supplementation.
Because the small intestines and colons of the Apc+/- Msh2-/- mice were covered with numerous adenomas and ACF, it was technically impossible to scrape only the normal mucosa excluding adenomas and ACF to obtain normal small intestinal and colonic DNA. Determining the extent of genomic DNA methylation from intestinal mucosal scrapings containing varying proportions of normal cells, adenomas, and ACF would not have been informative. In the present study, therefore, the effect of dietary folate on genomic DNA methylation in normal cells was determined only in the liver. Because both site-specific hypo- and hypermethylation play a role in carcinogenesis (65 , 66) and because folate may modulate DNA methylation in a site-specific manner (71) , it would be of interest to study the effect of folate on sites within proto-oncogenes and tumor suppressor and mismatch repair genes that are implicated in intestinal carcinogenesis from microdissected normal cells, adenomas, and ACF.
Another mechanism that was investigated in the present study pertains to the effect of folate on DNA mismatch repair. Folate is an essential factor for the de novo biosynthesis of purines and thymidylate and, hence, plays an important role in the maintenance of DNA integrity and stability (58) . Accumulating in vivo and in vitro evidence suggests that folate deficiency is associated with DNA damage (71 , 74) , deoxynucleotide pool imbalances leading to misincorporation of uracil into newly synthesized DNA (74 , 75) , and impaired DNA excision (76) and mismatch (77) repair.
The present study, however, suggests that widespread microsatellite instability, which is observed in the majority of HNPCC and 15–20% of sporadic colorectal cancers in humans (43) , is likely not a major factor leading to accelerated small intestinal and colonic tumorigenesis in Apc+/- Msh2-/- mice. Even folate deficiency and supplementation had no significant effect on microsatellite instability, which suggests that modulation of microsatellite instability is not likely a mechanism by which dietary folate modulates small intestinal and colonic tumorigenesis in this model. The microsatellite instability phenotype is observed in some mismatch repair-deficient murine models including Msh2-/- (44 , 45) , Mlh1-/- (78) , and Pms2-/- (79) mice, whereas it is not seen in other models such as Msh6-/- (80) . Although lymphoid and skin cancers and a small fraction of small intestinal and colonic cancers from Msh2-/- mice (44 , 45) display microsatellite instability, it is rarely observed in small intestinal and colonic polyps from Apc+/- Msh2-/- mice (38) . Despite the lack of widespread microsatellite instability in Apc+/- Msh2-/- mice, it is possible that Msh2 deficiency may cause replication errors in one or a few critical genes resulting in accelerated intestinal tumorigenesis.
Several limitations associated with the Apc+/- Msh2-/- murine model need to be acknowledged: (a) the predominant phenotype in this model is the development of small intestinal adenomas in contrast to colon polyps in humans; (b) Apc+/- Msh2-/- mice do not develop small intestinal or colonic adenocarcinoma because they become moribund as a consequence of florid polyposis; (c) as previously described, the contribution of ACF to the adenoma population in this model is not clearly established; (d) this model may reflect only inherited types of accelerated tumorigenesis such as familial adenomatous polyposis and HNPCC and not sporadic colorectal carcinogenesis. Lastly, despite Msh2 deficiency, widespread microsatellite instability is absent in tumors. Nonetheless, the Apc+/- Msh2-/- mouse seem to be an excellent model to study chemopreventive effects of dietary factors and drugs on colorectal carcinogenesis because of the spontaneous development of small intestinal and colonic adenomas and ACF, genetic similarities to human colorectal cancer, and the accelerated nature of tumorigenesis, which provides an opportunity to determine effects of chemopreventive agents on both initiation and progression of tumorigenesis in a relatively short time.
The mice that were randomized to receive 0 mg folate/kg for the 6-week
start study had an unexpected 16% lower mean weight at the start of
the diet and remained 16–19% lower than the folate-supplemented mice
during the 5 weeks of dietary intervention despite similar average
daily food consumption (Fig. 1B
). However, the
folate-deficient mice demonstrated no growth retardation and their mean
weight increased in parallel to that of the folate-supplemented mice
during the study period (Fig. 1B
). Because calorie
restriction may inhibit tumor development in chemical and genetic
knockout models (81
, 82)
, the possibility that the
significantly lower weights of the folate-deficient mice might have
contributed to lower numbers of small intestinal adenomas compared with
the folate-supplemented diet when folate intervention was started at 6
weeks of age (i.e., after the establishment of neoplastic
foci) cannot be ruled out. However, one observation against this
possibility is that the folate-deficient mice had significantly higher
numbers of small intestinal adenomas and colonic ACF compared with the
folate-supplemented mice when dietary intervention was started at 3
weeks of age (i.e., before the establishment of neoplastic
foci) despite the fact that their mean weight was 8–18% lower (albeit
statically nonsignificant) compared with the folate-supplemented
mice during the study period (Fig. 1A
). The present study
did not include a group receiving a diet containing the basal dietary
requirement of folate (i.e., 2 mg folate/kg diet) and
comparisons were made to a historic control receiving this diet from a
previous study (38)
. Future studies examining the
dose-response effect of dietary folate in this model are warranted to
confirm our present findings.
In summary, the present study suggests that dietary folate supplementation, at 4 times the basal requirement, significantly protects against the development of small intestinal adenomas and colon ACF if started before the establishment of neoplastic foci, in this genetically predisposed mouse intestinal tumorigenesis model. However, if started after the establishment of neoplastic foci, dietary folate seems to have an opposite effect in this model. These data suggest that the timing of folate intervention is critical in providing an effective and safe chemopreventive effect on colorectal carcinogenesis. Notwithstanding the limitations associated with this model, our data suggest that the optimal timing of folate intervention should be established before folate supplementation can be used as a safe chemopreventive agent against colorectal cancer. Furthermore, the mechanisms by which folate modulates colorectal carcinogenesis need to be elucidated.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported in part by grants from the American
Institute for Cancer Research (to Y-I. K. and S. G.), the Medical
Research Council of Canada (to Y-I. K.), the Medical Research Council
of Canada/Canadian Association of Gastroenterology/BYK Canada
(to Y-I. K.), and the St. Michael’s Hospital Health Sciences Research
Program (to Y-I. K.). Y-I. Kim was supported in part by a scholarship
from the Medical Research Council of Canada. Presented in part at the
1999 American Association for Cancer Research meeting, April 10–14,
1999, Philadelphia, PA, and published in abstract form in Proceedings
of the American Association for Cancer Research 1999, 40: 3490. 
2 To whom requests for reprints should be
addressed, at Department of Medicine, Room 7258, Medical Sciences
Building, University of Toronto, 1 King’s College Circle, Toronto,
Ontario, M5S 1A8 Canada. Phone: (416) 978-1183; Fax: (416) 978-8765;
E-mail: youngin.kim{at}utoronto.ca
3 The abbreviations used are: ACF, aberrant crypt
foci; CpG, cytosine-guanine (dinucleotides); HNPCC, hereditary
nonpolyposis colorectal cancer; SAM,
S-adenosylmethionine.
Received 11/ 8/99. Accepted 5/ 3/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSEffects of Dietary Folate...DISCUSSIONREFERENCES |
|---|
