Notes
The following principal investigators and institutions, listed according to accrual of patients, participated in this EORTC Head and Neck Cancer Cooperative Group study: J.-L. Lefebvre (Centre Oscar Lambret, Lille, France): D.
Chevalier (Hdpital Claude Huriez. Lille); B. Luboinski (Institut Gustave-Roussy, Villejuif, France): L. Traissac (H6pital Pellegrin, Bordeaux, France): G.
Andry (Institut Jules Bordet, Brussels, Belgium); J.-P. Rame (Centre Francois
Baclesse. Caen. France): S. Crispino (Ospedale San Gerardo. Monza. Italy); C.
Chiamenti (Ospedale Verona, Italy): J. M. H. van den Beek (University Hospital, Maastricht, The Netherlands); J.-M. David (Centre Claudius Regaud,
Toulouse. France): and J. Bemier (Ospedale San Giovanni. Bellinzona, Switzerland). Medical supervisor for the study: D. Lacombe (Brussels): scientific coordinator: R. Sylvester (Brussels): and scientific expert for the interim analysis
(December 1993): M. K. B. Parmar (statistician, Medical Research Council Cancer Trial Office, Cambridge, U. K.).
Supported by Public Health Service grants CA11488-20 through CA11488-25
from the National Cancer Institute. National Institutes of Health, Department of
Health and Human Services.
Manuscript received September 22, 1995: revised February 29, 1996; accepted
April 22. 1996.
Wheat Bran and PsyIlium Diets: Effects on
Af-Methylnitrosourea-Induced Mammary
Tumorigenesis in F344 Rats
Leonard A. Cohen, Zhonglin Zhao, Edith A. Zang, Tin T. Wynn,
Barbara Simi, Abraham Rivenson*
Background: Experimental and epidemiologic evidence suggests that increased dietary fiber is associated with
decreased breast cancer risk. Little is known about the role
played by different types of Tiber and, particularly, mixtures
of soluble and insoluble fibers similar to those consumed by
human populations in reducing breast cancer risk. High intake of fiber may suppress bacterial hydrolysis of biliary
estrogen conjugates to free (absorbable) estrogens in the
colon and thus may decrease the availability of circulating
estrogens necessary for the development and growth of
breast cancers. Purpose: The purpose of this study was to
evaluate the effect of wheat bran (an insoluble fiber) and
psyllium (a soluble fiber) alone and in combination on overall estrogen status, on fecal bacterial fJ-D-glucuronidase (a
key diet-responsive estrogen-deconjugating enzyme) activity,
and on the induction of mammary tumors in rats treated
with yV-methylnitrosourea (MNU). Methods: One hundred
fifty virgin female F344 rats were fed the NIH-07 diet from
28 days of age until 50 days of age; they were then given a
single dose (40 mg/kg of body weight) of MNU by tail vein
injection. Three days later, they were randomly assigned to
one of five experimental dietary groups (30 animals per
group). Soft, white wheat bran (45% dietary fiber content)
and psyllium (80% dietary fiber content) were added to a
modified (high-fat) American Institute of Nutrition (AIN)76A diet at the following percents, respectively: 12% + 0%
(group 1), 8% + 2% (group 2), 6% + 3% (group 3), 4% +
4% (group 4), and 0% + 6% (group 5). Blood, urine, and
feces were collected and analyzed by radioimmunoassay
techniques for estrogens. Cecal contents were analyzed for
bacterial [3-D-glucuronidase activity. After 19 weeks on the
experimental diets, the rats were killed, and mammary
tumors were counted and classified by histologic type.
Cumulative tumor incidence was evaluated by the KaplanMeier life-table method and the logrank test. Tumor numJournal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
ber was evaluated by the chi-squared test of association, and
tumor multiplicity was evaluated by the Mantel-Haenszel
chi-squared test. All statistical tests were two-tailed. Results:
As the level of psyllium relative to that of wheat bran increased, the total tumor number and multiplicity of mammary adenocarcinomas in rats decreased as a statistically
significant linear trend across groups 1-5 (f <.O5). Compared
with the group given wheat bran alone, the group given the
1:1 (wheat bran:psyllium) combination had maximum
protection against mammary tumorigenesis, while the
groups given the 4:1 or 2:1 (wheat bran:psyllium) combination or psyllium alone had intermediate protection. No
statistically significant differences in circulating estrogens or
urinary estrogen excretion patterns were observed among
the five experimental groups. Fecal estrogen excretion, however, decreased with increasing levels of psyllium (P<.01),
and cecal fi-D-glucuronidase activity exhibited a decreasing
trend with respect to the increasing psyllium content of the
diet across groups 1-5 (P<.0l). Conclusions: The addition of
a 4%:4% mixture of an insoluble (wheat bran) fiber and a
soluble (psyllium) fiber to a high-fat diet provided the maximum tumor-inhibiting effects in this mammary tumor
model. Although increasing levels of dietary psyllium were
associated with decreased cecal bacterial p-D-glucuronidase
activity, these changes were not reflected in decreased circulating levels of tumor-promoting estrogens. Therefore, the
mechanism(s) by which mixtures of soluble and insoluble
dietary fibers protect against mammary tumorigenesis
remains to be clarified. [J Natl Cancer Inst 1996;88:899-907]
*Affiliations of authors: L. A. Cohen, Z. Zhao, T. T. Wynn, B. Simi (Division
of Nutrition and Endocrinology), E. A. Zang (Division of Epidemiology), A.
Rivenson (Research Animal Facility), American Health Foundation, Valhalla,
NY.
Correspondence to: Leonard A. Cohen, Ph.D., Division of Nutrition and Endocrinology, American Health Foundation, 1 Dana Rd., Valhalla, NY 10595.
See "Notes" section following "'References."
ARTICLES
899
Epidemiologic evidence provides support, with some exceptions (1,2), for the hypothesis that dietary fiber or compounds of
plant origin associated with dietary fiber may be protective
against breast cancer (3-7). Moreover, studies by Cohen et al.
(8) and Arts et al. (9) have directly demonstrated that supplemental wheat bran suppressed the development of chemically
induced rat mammary tumors. Several mechanisms have been
proposed by which dietary fiber may exert its protective effects
(10). Primary among these mechanisms is the concept that fiber
exerts its preventive effects on breast cancer by altering the
enterohepatic recirculation of absorbable estrogens (//). Studies
(12-14) have shown that women consuming high levels of fiber
exhibit (a) decreased urinary excretion of conjugated (nonabsorbable) estrogens, (b) decreased serum estrogens including
estrone sulfate (E|S), the principal circulating form of estrogen,
and (c) enhanced fecal elimination of estrogens. On the basis of
these and other findings, it has been proposed (//) that the observed alterations in estrogen disposition were a consequence of
the growth-suppressing effect of fiber on deconjugating ((3-Dglucuronidase-producing and sulfatase-producing) colonic bacteria. By suppressing bacterial hydrolysis of biliary estrogen
conjugates to free (absorbable) estrogens in the colon, high fiber
intake was hypothesized to decrease the availability of circulating estrogens necessary for breast cancer growth and development (13).
Based largely on earlier epidemiologic studies linking low
fiber intake to increased colon cancer risk, Lanza et al. (15) and
several government agencies (16-20) have proposed that total
dietary fiber intake (currently, 10-12 g/2000 calories) should be
increased to 25 g or more and should come from a broad variety
of food sources for the promotion of normaL bowel health and
for optimal protection against colon cancer. However, with the
exception of the study by Alabaster et al. (21) in a rat colon
tumor model system, no studies have been reported concerning
the influence on experimental cancer development of mixtures
of soluble and insoluble fibers, similar to those found in the U.S.
diet.
The purpose of our study, therefore, was to assess the effects
of mixtures of two fiber-rich isolates—soft, white wheat bran
(which is largely insoluble) and psyllium (a common source of
soluble fiber derived from the husk of the plant Plantago
ovala)—on /V-methylnitrosourea (MNU)-induced mammary
tumor development in rats. In addition, by comparing serum,
urinary, and fecal estrogen levels with cecal bacterial (i-Dglucuronidase activity, we tested the hypothesis that mixtures of
wheat bran and psyllium selectively alter the enterohepatic recirculation of estrogens.
Materials and Methods
Animal Care and Adherence to Guidelines
Animals received proper care and maintenance in accord with institutional
guidelines. Three rats were housed together in a polyethylene cage that contained hardwood shavings and was covered with a filter top. The animal room
was controlled for temperature (24 "C ± 2 *C). light (12-hour cycle), and
humidity (50%). Diets were provided in powdered form, and tap water was
provided ad libitum. Stainless-steel "J"-type powder feeders were used to
prevent scattering of food.
900
ARTICLES
The experimental protocol used (see below) was approved by the American
Health Foundation Institutional Animal Care and Use Committee. Animal care
was conducted with strict adherence to institutional guidelines and to guidelines
specified in the "Guide for Care and Use of Laboratory Animals" (U.S. Department of Health and Human Services publication No. 85-23, 1985).
Protocol for Experimental Tumor Induction
One hundred fifty virgin female inbred F344 rats (starting age, 28 days old;
Charles River Breeding Laboratories. North Wilmington, MA) were maintained
on the standard Open Formula Rat and Mouse Ration (NIH-07) diet (4.5% fat.
23.5% protein. 50% carbohydrate, and 4.5% fiber) (Zeigler Bros., Gardners. PA)
(22) until 50 days of age. All rats were then assigned to one of five groups of 30
animals each by recognized randomization procedures (23) to equalize initial
weight. At 50 days of age, all rats received a single dose (40 mg/kg of body
weight) of MNU (Ash Stevens Inc., Detroit. MI) by tail vein injection. MNU
was dissolved in a few drops of 3% acetic acid and diluted with distilled water to
give a stock solution of 10 mg of MNU/mL, which was administered within 2
hours of preparation (24).
Dietary Fiber Supplements and Administration of Specified
Diets
A certified lot of the soft, white wheat bran was supplied by the American
Association of Cereal Chemists, St. Paul, MN. Before being packed and stored,
the bran was processed through an enzyme-deactivation steamer system. The
major components of soft, white wheat bran are as follows: total carbohydrate,
63.8%: fat, 6.8%; protein, 16.0%; ash, 6.5%: and moisture, 6.0%. Total dietary
fiber was 45.0%, neutral detergent fiber was 30.0%, starch was 16.0%, and pectin was 2.0%; 100% of the particles passed through an American Association of
Cereal Chemists' #70 mesh screen (0.212 mm ± 10 |im). The above analyses
were conducted for the American Association of Cereal Chemists by Medallion
Laboratories, St. Paul, MN. Psyllium husk fiber (U.S. Pharmacopeial Convention, 22nd ed.) was obtained from J. B. Laboratories. Inc., Holland, MI. The
major components of psyllium are as follows: total carbohydrate, 80%; fat, 2%3%; protein, 4%-5%; moisture, 8%-10%: and total dietary fiber, 80%.
Diets were prepared in 15-kg lots in our kitchen and were stored at 4 "C until
use. The diets were formulated to account for the carbohydrate, protein, and fat
contributed by the psyllium and soft, white wheat bran. The diets provided approximately 18 calories per day as fat (40% of total calories). Rats were fed
three times per week, and feeders were removed and washed after each feeding.
Three days after MNU administration, the rats were transferred to one of five
different experimental diets (Table 1). The diets were formulated in our kitchen
on the basis of a modification of the American Institute of Nutrition (AIN)-76A
diet (25) (Table I). The standard AIN-76A diet has 5% fat. 20% protein. 65%
carbohydrate, and 5% fiber. All ingredients were obtained from Dyets Inc.,
Bethlehem. PA. The rats remained on these diets for the duration of the experiment (19 weeks).
Observation Schedule
At weekly intervals, beginning 4 weeks after MNU injection, each rat was
weighed, and the location of palpable tumor(s) and the date they were found
were recorded.
Collection of Urine and Feces
Approximately 18 weeks after the administration of MNU. six (20%) of 30
rats from each group were placed in metabolism cages, and their urine and feces
were collected over a 24-hour period. Urine (typically, 10-15 mL every 24
hours) was collected on dry ice in 50-mL conical centrifuge tubes and then
thawed and centrifuged at 200g for 20 minutes at 4 'C. The clear supernatant
was collected and stored at -20 "C until use. Fecal pellets were collected, placed
in tightly capped 20-mL scintillation vials, and stored at -20 "C until analysis.
Necropsy, Serum and Cecal Collection, and Histopathology
Approximately 19 weeks after MNU administration, the experiment was terminated. Blood was drawn from the rats by heart puncture under ketamine anesthesia between 11 AM and 1 PM in order to minimize variation in hormone levels
(26). The blood was collected in evacuated, sterile Autosep tubes (Terumo
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
Table 1. Experimental diets*
Group I:
12% wheat bran
Casein, g
Corn starch, g
Dextrose, g
Methionine, g
Corn oil. g
Fiber, g
Mineral mix, g
Vitamin mix, g
Choline bitartrate.
Total, g
Total kcal
% carbohydrate
% protein
%fat
21.6
31.4
10.5
0.4
18.7
12.0
4.2
Group 2:
8% wheat bran
+ 2% psyllium
22.2
32.2
10.7
0.4
19.0
Group 3:
6% wheat bran
+ 3% psyllium
Group 4:
4% wheat bran
+ 4% psyllium
Group 5:
6% psyllium
22.5
32.6
10.8
0.4
19.2
9.0
4.3
22.9
33.0
10.9
0.4
19.4
8.0
4.3
23.5
33.5
12.0
0.4
1.2
0.2
1.2
0.2
1.2
0.2
19.9
1.2
0.2
100
10.0
4.2
1.2
0.2
100
100'
100
100
422
432
437
441
456
39.6
20.0
39.8
39.8
20.6
39.6
39.8
20.7
39.4
39.8
20.7
39 4
6.0
4.3
39.9
20.7
39.4
*The composition of the diets was adjusted so that the rats in the different dietary groups consumed approximately equal amounts of protein, minerals, and
vitamins, despite differences in caloric density. F344 females consume approximately 44 kilocarlones per day.
Medical. Elkton, MD). Serum and red blood cells were then separated by
centrifugation at 200^ for 20 minutes at 4 *C. Serum was stored at -20 'C until it
was assayed. Blood was taken during metestrus, diestrus, and late estrus (not
proestrus). Rats were then killed by carbon dioxide inhalation, and mammary
tumors (classified as palpable or nonpalpable but grossly visible) were excised,
fixed in 10% buffered formalin, embedded in paraffin blocks, and stained with
hematoxylin—eosin for histologic examination. At necropsy, trie ceca were excised, tied off at the open end, and placed in vials flushed with N-, to maintain
anaerobic conditions. The vials were then sealed and stored at 4 *C until use. For
assay, the vials containing the ceca and the empty vials were weighed, and the
weight of the cecal contents was determined by difference. The cecal contents
were placed on aluminum foil under a steady stream of a mixture of N2, CO2,
and H, gases. The cecal contents were then suspended in phosphate-buffered
saline (pH 7.0), vortexed, and then centrifuged at 600 rpm for 5 minutes (4 - C).
The supernatant was then saved, and the pellet was discarded. Histologic diagnosis of mammary tumors was based on criteria outlined by Young and Hallowes (27) and Russo et al. (25).
Serum Estrogen Assay
The method for conjugated and unconjugated serum estrogen assay was a
modification of that described by Adlercreutz et al. (29). To precipitate out
protein and to remove lipid and free, unconjugated estrogens from their binding
proteins, we added 2.3 mL of 100% methanol to each 1-mL serum sample. The
mixture was then placed in a refrigerator at 4 'C for 20 minutes, after which it
was centrifuged at 2300,? for 10 minutes at 4 'C; the supernatant was saved. We
then added 1 mL of 70% methanol to the supernatant and treated the mixture as
described above. The two supematants were then combined and dried under
nitrogen. Distilled H2O (2 mL) was then added to the'residue, and the mixture
was extracted two times with 5.0 mL diethyl ether. The ether layer, containing
unconjugated estrogens (extraction efficiency, >95%), was then blown dry under
Nj, and the residue was brought up in buffer and assayed by use of a commercially available radioimmunoassay (RIA) kit (Wein Laboratories Inc., Succasunna, NJ). The aqueous layer, containing conjugated estrogen, was placed under a
stream of N2 gas to remove residual ether, and we then added 0.25 mL of arylsulfatase solution (250 U) (type H-l from the snail Helix pomalia; Sigma
Chemical Co.. St. Louis, MO) that had been pretreated with activated, neutralized charcoal (Sigma Chemical Co.) in a sodium acetate buffer (pH 4.2). The
reaction mixture was incubated in a shaker bath overnight at 37 'C. Before use,
the crude enzyme preparation was purified by treatment with activated charcoal
as follows: Charcoal (200 mg) was added to I mL of the enzyme preparation
(2500 U/mL). The suspension was then agitated for 2 hours at room temperature,
after which it was centrifuged at 18 900^ for 30 minutes at room temperature.
Before use, the supernatant was filtered through Celite (Aldrich Chemical Co.,
Milwaukee. Wl). Following enzymatic hydrolysis, the mixture was extracted
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
two times with 5.0 mL diethyl ether to obtain free estrone (E,). The ether phase
was then dried under N, gas, and the residue was assayed by RIA for E, by use
of the RIA kit (Wein Laboratories Inc.). The E, assay is highly specific (crossreactivity with 17P-estradiol [E,] and estriol [E3] <0.025) and sensitive (10
pg/mL detectable).
Radioactive aliquots were counted in a liquid scintillation counter (LKB;
Wallac Inc., Gaithersburg, MD), and RIA data were processed by use of a computer program obtained from Robert Maciel Associates Inc., Arlington, MA
Based on radiometric measurement using [6,7- H]E,S (54 Ci/mmol) (Du Pont
NEN Research Products, Boston, MA), the efficiency of E,S recovery from
rat serum was 70%. Intra-assay variation (coefficient of variation) was 3.0%7.5%, and interassay variation (coefficient of variation) was 8.8%-16.5%. The
data were expressed as picograms per milliliter of E,S (as E,) after adjustment
for extraction efficiency.
Urinary Estrogen Analysis
The method used was a modification of that described by Adlercreutz et al.
(29). Urinary estrogens are present primarily as water-soluble conjugates.
Hence, a deconjugation step is required prior to RIA. For E, and E;, 0.3 mL of
urine was added to 0.7 mL of H2O; for E3, 0.06 mL of urine was brought up to I
mL with H2O. After vortex mixing, 0.2 mL of (J-D-glucuronidase (type H-5 from
H. pomalia) was added to the unne sample. The reaction solution, containing
2000 U P-r>glucuronidase activity per sample, was then vortex-mixed and incubated overnight in a shaker bath at 37 'C. It was then extracted twice with 5.0
mL diethyl ether. The deconjugated estrogens present in the ether fraction were
dried, and the residue was used for RIA of E,, Ej, and E3 with the commercially
available kits (Wein Laboratories Inc.). The Ej and E3 kits are highly specific
(cross-reactivity with E, [or E, or E3J <0.025%) and sensitive (10 pg/mL detectable). All extractions were done in duplicate. The data were expressed as
picograms of estrogen per milligram of creatinine to account for variations in
total 24-hour urinary output.
Assay of Unconjugated Fecal Estrogens
The method used was based on that of Adlercreutz and JHrvenpSa (30). Approximately 70%-8O% of rat fecal estrogens are present in the unconjugated
form (7). Fecal pellets were weighed (wet weight) and then freeze-dried. Following lyophilization, the pellets were weighed again (dry weight) and then
were ground to a powder with a mortar and pestle, and to 0.25 g of powdered
feces was added 5 mL of 70% methanol. The mixture was sonicated for 10
minutes, vortex-mixed for I minute, and centrifuged at 2500g for 10 minutes at
4 'C. The clear supernatant was saved, and the procedure was repeated with the
remaining pellet. The two supemalants were then pooled and placed at -20 "C
overnight to precipitate lipids and protein. The mixture was centrifuged at 2500#
ARTICLES
901
for 10 minutes at 4 'C, and the methanol layer was dried under N2 gas until approximately 2 mL remained. Free estrogens were extracted with ether after 1 mL
(pH 4.2) acetate buffer was added to the mixture. The ether layer was then
decanted and dried under N2 gas.
The dried residue was brought up in 70% methanol and passed through a
DEAE-Sephadex A-25 column (acetate form). The column was eluted with 1.0,
1.0, and 6.0 mL of 70% methanol. The eluates were combined in a 16 x 100-mm
glass test tube and dried under N, gas. The resulting residue was dissolved in I
mL ethanol and vortex-mixed, and 0.05 mL was transferred to three different
tubes for E,, Ej, and E3 assay. The tubes were dried again and then assayed by
RIA. Data were expressed as nanograms per gram of feces or nanograms per 24
hours of fecal collection.
Cecal (i-D-Glucuronidase Assay
The cecal fS-D-glucuronidase assay was based on the method described by
Kulkarni and Reddy (31). The cecal contents were weighed, mixed into a slurry
in 0.1 M phosphate buffer, sonicated on ice for 30 seconds, and then centrifuged
at AOg for 15 minutes at 4 *C. Supernatant (1 mL) was added to a 15-mL tube,
sonicated on ice, and centrifuged at 26 I85g for 20 minutes at 4 "C. The supernatant was saved, and 0.1 mL was added to a 16 x 125-mm glass test tube. To
each tube was added 0.8 mL of 0.1 M phosphate buffer and 0.1 mL of
phenolphthalein-P-r>glucuronide, and the reaction mixture was then mixed and
incubated at 37 'C for 30 minutes. The reaction was stopped by the addition of
2.5 mL of alkaline glycine buffer (pH 10.25). After dilution of the sample by the
addition of 2.5 mL of distilled H2O, the reaction mixture was allowed to stand
for 10 minutes at room temperature. Free phenolphthalein was then measured by
absorption spectrophotometry at 540 nm. (J-D-Glucuronidase activity was expressed as micrograms of phenolphthalein liberated per hour per gram of cecal
content or as micrograms of phenolphthalein liberated per hour per gram of total
cecal content.
Statistical Analysis
Tumor-free survival was estimated separately for each group by the KaplanMeier product-limit estimate for censored data (32). The survival distributions
for the five groups were then compared by the logrank test (33J4). The purpose
of the analysis was to test the null hypothesis that the survival distributions of all
groups were equal. In addition to the overall test of significance, pairwise comparisons between groups were made.
Tumor incidence (expressed as the percentage of tumor-bearing animals) was
compared among the groups by the chi-squared test of association. Overall doserelated associations between tumor incidence and the psyllium content of the
diet across groups 1-5 were tested by use of Armitage's test for linear trends in
proportions (35). Tumor multiplicity among the groups was compared by use of
the Mantel-Haenszel (36) chi-squared test. In addition, overall dose-related associations between tumor multiplicity across groups 1-5 were tested by linear
regression analysis.
Estrogen levels in urine, feces, and serum were compared among the groups
by use of single classification of analysis of variance followed by Tukey's adjustment (37) for multiple comparisons.
The overall weight gains in the animals of all groups were compared by use
of single classification analysis of variance with repeated measures (38). The test
of interest was the interaction between weight and time to evaJuale the null
hypothesis of no difference in weight gain over time among the groups. Pairwise
comparisons among the groups were also conducted.
All statistical tests were two-tailed and were considered statistically significant at P<.05. Significance tests for all pairwise comparisons were adjusted for multiple comparisons by multiplying the actual P value by the
number of comparisons made (in this case, 10) for the evaluation of statistical
significance.
Results
Tumor Yields
As the ratio of wheat bran to psyllium decreased, there was an
overall decline in mammary tumor yields. Maximal protection
was seen in MNU-treated rats fed a diet that included supplemental wheat bran-psyllium at a ratio of 1:1, and moderate
protection was observed in MNU-treated rats fed diets containing supplemental wheat bran—psyllium at 4:1 and 2:1 or psyllium alone, when compared with the group fed wheat bran.
When assessed in terms of percent tumor incidence, there was a
borderline statistically significant negative trend (P<.08) (Table
2). When tumor yields were assessed in terms of tumor multiplicity or total number of tumors, a statistically significant linear
trend (f<.05) was found across groups 1-5. The effect of psyllium was most evident when adenocarcinomas only were included in the analysis (Fig. 1). Tumor latency (i.e., time to
development of first tumor) was not significantly delayed in
group 4 compared with that seen in the other four treatment
groups (Fig. 2).
Table 2. Effect of fiber mixtures on mammary tumor incidence* and number
All tumorst
Group No.
1
2
3
4
5
Diet
A1N-76A + 12% wheat bran
AIN-76A + 8% wheat bran +
2% psylhum
AIN-76A + 6% wheat bran +
3% psyllium
AIN-76A + 4% wheat bran +
4% psyllium
AIN-76A + 6% psyllium
Adenocarcinoma onlyll
No. of rats with tumor/
total No. of ratsj
%
Total No.
of tumors§
No. of rats with tumor/
total No. of ratsj
%
Total No.
of tumors§
24/30
27/30
80
90
83
58
16/30
18/30
53
60
35
33
24/30
80
62
12/30
40
19
20/30
67
54
9/30
30
15
25/30
83
59
13/30
43
18
*Tumor incidence: All pairwise comparisons were not statistically significant when adjusted for multiple comparisons. The unadjusted comparison of group 2 and
group 4 was statistically significant for both all tumors and adenocarcinoma only (P = .0286 and P = .0195, respectively). Armitage's test for linear trends across
groups, P<,08 for adenocarcinoma only.
tTumor number, for all tumors: When unadjusted for multiple comparisons, group 1 versus group 2, P<.053; group I versus group 4, P<,0132: and group 1 versus
group 5, P-c.0439. When adjusted, all pairwise comparisons were not statistically significant by chi-squared test.
^Number of tumor-bearing rats/total number of rats at risk (n = 30). Since no group had 100% incidence, the numerator is always less than the denominator.
§Total number of tumors per group was assessed by chi-squared analysis, whereas multiplicity was assessed by Armitage's test.
IITumor number, for adenocarcinoma only: When unadjusted for multiple comparisons, group 1 versus group 3. /)<.0295: group I versus group 4, /><.0O47: group
1 versus group 5, P<.0196; group 2 versus group 4. P<.047. All other pairwise comparisons were not statistically significant by chi-squared test.
902
ARTICLES
Journal of Die National Cancer Institute, Vol. 88, No. 13, July 3, 1996
25 r
Fig. 1. Histograms depicting adenocarcinoma multiplicity as a function of diet group. Group 1 (AIN-76A +
12% wheat bran); group 2 (A1N-76A + 8% wheat bran
+ 2% psyllium); group 3 (AIN-76A + 6% wheat bran +
3% psyllium); group 4 (AIN-76A + 4% wheat bran +
4% psyllium); group 5 (AIN-76A + 6% psyllium). Solid
bar = frequency of animals with no rumors; diagonal
hatched bar = frequency of animals with one tumor,
open bar = frequency of animals with two tumors;
crosshatched bar = frequency of animals with three or
more tumors. All pairwise group comparisons were not
statistically significant when adjusted for multiple comparisons. The unadjusted comparison of group 1 and
group 4 was statistically significant (P - .045) by the
Mantel-Haenszel chi-squared test. Linear regression
showed an overall dose-related decrease in tumor multiplicity (P<.05).
20
Group 1
Group 2
Group 3
Group 4
Number of tumors
The histopathologic profile of the mammary tumors induced
by MNU included intraductal and invasive papillary carcinoma
with desmoplastic features and fibroadenoma with ductal carcinoma in situ. The latter category appeared as a fibroadenoma
Group 1 Group 2 ••
Group 3 •-
•
i
Group 4 0.50
Group 5 -
ibil ity
o
I:
rSj7\
mmf'if~
i
:
!• >.< f
0- 4 0 -
'1
1
Q.
0.30
:
10
bi
21
umi
«
IS
r
o
,
0.20
25
24
|
i
:
1
Cecal (3-D-Glucuronidase
" 1 2 3
I
1
1
0.10
1 ~
1—
2B
i !
n r\n
1
i
u.uu
c
35
70
105
Estrogens
Single-point serum or urinary conjugated estrogen levels did
not vary in any consistent way with the wheat bran/psyllium
ratio in the diet or with tumor yields (Table 3).
Fecal estrogen excretion profiles exhibited a statistically significant overall linear decrease as psyllium levels increased
(Table 3) for all estrogens. Fecal estrogen excretion was highest
in the wheat bran group and lowest in the psyllium group. To
account for differences in total fecal output, we assessed fecal
estrogens in terms of both nanograms of estrogen excreted every
24 hours and nanograms excreted per gram of fecal dry weight.
Although differences among groups were more pronounced
when expressed in terms of 24-hour output, similar trends were
observed regardless of how the data were expressed. As previous studies (9,13) noted, the concentration of estrogen was
about 500 times higher in urine and feces than in serum.
0.70
0.60
with small clusters of malignant epithelial cells confined to the
intraductal space.
140
Days
Fig. 2. Kaplan—Meier life-table curves for cumulative mammary adenocarcinoma incidence by dietary group. Ordinate: proportion of tumor-bearing
animals per unit time. Abscissa: days after /V-methylnitrosourea treatment. All
pairwise group comparisons were not statistically significant when adjusted for
multiple comparisons. The unadjusted comparison of group 1 and group 4 was
statistically significant (P = .04) by the logrank test. Numbers represent number
of tumor-free animals at designated time points.
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
To account for differences in cecal mass (Fig. 3) among the
treatment groups, {J-D-glucuronidase activity was expressed per
gram of cecal mass and per total grams of cecal mass. Regardless of the mode of expression, a statistically significant linear
decrease in P-D-glucuronidase activity was found as the relative
proportion of psyllium to wheat bran increased (Fig. 3). The
highest levels were consistently found in the wheat bran group
and the lowest levels in the psyllium group. These results indicate that psyllium is a very effective suppressor of (J-Dglucuronidase-producing cecal bacteria.
Cecal and Fecal Weights
Cecal and fecal wet weights varied inversely as the psyllium
content of the diet increased (Table 3). Cecal wet weights inARTICLES 903
Table 3. Blood, urine, and fecal estrogen profiles as a function of diet*
Diet supplement
Wheat bran, %
Psyllium, %
Serum eslrogenst
Estrone sulfate
Estrone
Hp'-Estradiol
Urinary estrogens^
Estrone
Estriol
I7f5-Estradiol
Total
Fecal estrogens§
Estrone
Estriol
np-Estradiol
Total
Weight of fecaJ and
cecal comentsll
Fecal weight
Wet
DryU
Cecal weight: wet#
Group 1
(n = 5)
Group 2
(n = 6)
12
0
2
177 ± 16
164± 21
92± 13
64±9
5.2 ±0.6
1.0 ±0.2
0.8 ±0.1
7.0 ±0.8
Group 3
(n = 6)
Group 4
(n = 6)
Group 5
(n = 6)
6
3
4
4
0
6
83±9
76 ±18
171 ±34
85 ±14
61 ± 11
172 ±29
96 ± 15
61 ± 15
201 ±25
97 ± 7
61 ± 10
6.1
1.1
0.7
7.9
6.0
0.9
0.6
7.5
5.7 ± 1.4
0.9 ±0.1
0.7 ± 0.2
7.3 ±1.6
6.3 ± 2.6
0.8 ±0.1
0.7 ±0.2
7.9 ± 3.0
± 1.4
±0.3
±0.2
±1.8
±1.9
±0.1
±0.1
±2.0
14 ± 2
8± 1
14±2
36 ±5
14 ±2
7± 1
19 ± 6
40 ± 6
12 ±2
8±1
13 ± 3
33 ±5
12 ± 2
8± 1
12± 1
32 ± 3
10±2
6± 1
8± 1
23 ± 4
1.81 ±0.73
1.36 ±0.56
2.33 ±0.13
1.88 ±0.47
1.43 ±0.26
2.85 ± 0.20
1.97 ±0.57
1.52 ±0.45
3.28 ±0.50
1.66 ± 0.42
1.25 ± 0.36
3.69 ± 0.68
1.32 ± 0.85
1.00 ± 0.63
4.14 ±0.22
•Five rats were randomly selected from group 1, and six rats each were randomly selected from groups 2-5. Blood, urine, feces, and cecal contents analyzed were
obtained from the same rats.
fValues = means ± standard deviation, pg/mL. Estrone sulfate (as estrone) after adjustment for extraction efficiency (70%). Analysis of variance, not statistically
significant (NS). Test for linear trend across groups by regression analysis, NS.
tValues = means ± standard deviation, ng/mg of creatinine. Analysis of variance comparing diet groups, NS. Test for linear trend across groups by regression
analysis, NS.
§Values = means ± standard deviation, ng/g of fecal dry weight. Test for linear trend across groups by regression analysis statistically significant at P<.0\ for all
estrogens. Group 2 differs from groups 4 and 5 for 17(5-esiradiol and total estrogens (P<.0\). Groups 1 and 2 differ from groups 4 and 5 for 17P-estradiol and total
estrogens (P<.05). Groups I and 2 differ from group 5 for estrone (P<.05). Group 1 (P<.0\) and groups 3 and 4 (/"<.05) differ from group 5 for estriol. All P values
were based on analysis of variance followed by Tukey's adjustment for multiple comparisons.
HValues = means ± standard deviation, g; 24-hour collection.
flTest for linear trend by regression analysis, NS.
#Test for linear trend by regression analysis, P<.000\.
5000
a.
I
4000
«
S
3000
E
CD
r
2000
-o
S
Fig. 3. Cecal pVt>glucuronidase activity plotted as a
function of diet group. There was a statistically significant, overall dose-related negative trend across
groups 1-5 by linear regression analysis (P-c.01).
Values are means ± standard deviation (n = 6). The
standard deviation for group 5 was 60.
I
1000 ~
2
3
4
Study group
904
ARTICLES
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
200
- ' '
Mean weight, g
175
150
Group 2
125
/
/
—
Group 4
-—
Group 5
Fig. 4. Mean absolute animal weight gain plotted as a
function of time. Intervals represent time points in weeks,
beginning at /V-methylnitrosourea (NMU) administration
and ending 19 weeks after NMU administration. There
was a higher rate of weight gajn by group I compared
with that by groups 4 and 5 (adjusted /><.05) and group 3
(adjusted P<.0l). Group 2 also had a higher rate of weight
gain than group 5 (adjusted P<.05). None of the other
pairwise comparisons showed a statistically significant
difference.
100 '
I
!
a
7
I
0
11
16
19
Week
creased significantly across groups 1-5, most likely as a result of
the high water-retention properties of psyllium while residing in
the cecum. The opposite was the case with fecal wet and dry
weights, although the trend toward lower weights in the psyllium-containing groups did not achieve statistical significance.
The lower fecal weight in the psyllium-containing groups may
be attributed to increased reabsorption of water from psylliumcontaining feces in the large bowel and to the greater fermentability of psyllium compared with wheat bran. The percentage
of fecal water in all five groups was similar, it averaged 25%
and varied from a low of 16% to a high of 34%.
Animal Weight Gains
Rats fed wheat bran alone (group 1) gained weight to a statistically significantly greater extent than rats in groups 4 and 5
(Fig. 4).
Small Group Sample Sizes
In Table 3, none of the three one-way analyses of variance
were statistically significant, perhaps because of the small group
sample sizes. In an effort to determine the power of these comparisons, we noted that groups 2 and 5 differed the most across
all three serum estrogen variables. Power analyses were
simplified and based on just this two-group comparison with
Student's t test (corrected for type I error to resemble a Tukey's
post hoc comparison). For E|S, the probability of detecting a 37unit (201 - 164) difference between the two groups on the basis
of the given sample sizes and standard deviations (power) was
found to be less than 45%; 11 rats in each group are needed for
the power to be greater than 80%. Similarly, for E|, the power to
detect a 15-unit difference was less than 50%, and 10 rats are
needed in each group to achieve at least 80% power. For E2, a
15-unit difference has less than 15% power, so that more than
25 rats are needed in each group to attain 80% power based on
the giveji standard deviations. In terms of how large a difference
would have been statistically significant (i.e., to find statistical
significance 80% of the time), we would need to see a 55-unit
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
difference in E|S levels between the two groups, a 20-unit difference for E| and a 35-unit difference for E2 (based on given
sample sizes and standard deviations).
Discussion
To our knowledge, this is the first study to demonstrate the
inhibitory effects of mixtures of soluble and insoluble fibers in
an experimental mammary tumor model. Previous studies conducted by us (8) and other investigators (9) have demonstrated
that diets supplemented with wheat bran inhibit MNU-induced
mammary tumor development. In these studies, the inhibitory
effect of fiber was most evident in rats fed diets high in fat (8).
It is of interest that the effect of fiber was most pronounced
when assessed in terms of mammary adenocarcinoma. When
fibroadenomas with ductal carcinomas in situ were included in
the analysis, the effect was less dramatic. Since ductal carcinomas in situ are considered to be part of a morphologic continuum leading from an original initiating event to a fully
developed carcinoma (59), this suggests that one effect of fiber
may be to act at the clonal expansion stage of the carcinogenic
process.
Differences in body weight gain were noted among the
groups fed wheat bran and psyllium. The reason for these differences is unclear, since the diets were formulated to be
isonutrient with respect to one another. One possibility is that
availability of nutrients in wheat bran was greater than that in
psyllium. Nonetheless, the overall differences in animal weights
among the treatment groups, though small, could account in part
for the differences observed in tumor yields.
The mechanism(s) by which fiber may inhibit breast cancer
have been reviewed {10,40). They include reduction of circulating tumor-promoting estrogens via suppression of bacterial P-Dglucuronidase activity in the colon and cecum (//); direct
binding of estrogens in the colon by fiber (8); or the presence of
isoflavonoids (10), phytates (41), and protease inhibitors (42) in
fiber, all of which have demonstrated anticancer properties. In
ARTICLES
905
this study, we tested the hypothesis that tumor inhibition is the
direct consequence of the ability of dietary fiber to alter the
enterohepatic recirculation of estrogens. According to this
hypothesis, a decrease in circulating estrogens, particularly EiS,
the most abundant and presumably diet-responsive circulating
estrogen (43,44), coupled with a decrease in P-D-glucuronidase
and a concomitant increase in estrogen excretion, would be expected in the group fed 4% wheat bran-4% psyllium compared
with the group fed wheat bran alone. While increasing levels of
psyllium did indeed suppress P-D-glucuronidase activity [as was
shown previously by Roberts-Andersen et al. (45)], there was no
concomitant change in circulating or urinary estrogen concentrations.
It should be noted that, since serum estrogen measures in this
study represent one point in time, in contrast to urinary and tecal
measurements, which represent 24-hour clearance levels, the
possibility exists that changes in serum levels actually did occur
at earlier points in the experiment but were not detected. This
possibility could be readily tested by cannulation followed by
longitudinal assays during the course of a feeding study. In this
manner, transient alterations in circulating estrogens, imposed
on circadian and estrus rhythms, could possibly be detected.
Also worth noting is the fact that, because of the relatively small
sample size (n = 6), there may have been insufficient power to
detect differences between groups. Future studies with larger
sample sizes will address this problem.
Precisely how mixtures of soluble and insoluble fibers interact physiologically in the gut is poorly understood. Soluble
fibers such as psyllium are known to alter the emulsification and
lipolysis of fat, interact with sterols, readily form gels with
water, and are susceptible to bacterial fermentation to shortchain fatty acids such as butyrate (46). Also, it is interesting
that, in the case of psyllium, when compared with cellulose,
fecal aerobes have been reported to increase with respect to
anaerobes, colonic pH is decreased, and P-D-glucuronidase is
suppressed (45). On the other hand, wheat bran, an insoluble
fiber, does not readily undergo fermentation and, consequently,
does not lower the pH of the gut: however, it does bind
estrogens (8) and also suppresses P-D-glucuronidase activity
(45). Psyllium, therefore, appears to share certain properties
with insoluble fiber, i.e., it can cause a suppression of bacterial
p-D-glucuronidase and an increase in the relative numbers of
fecal aerobes. Psyllium also has properties normally associated
with soluble fibers, i.e., high water retention, fermentability, and
lowering of colonic pH. On the basis of our results, psyllium exerts a suppressing effect on P-D-glucuronidase activity in a doserelated manner. Nonetheless, although addition of psyllium to
wheat bran clearly modulates the microbial ecology of the
cecum, the relationship, if any, between bacterial p-D-glucuronidase activity and mammary tumor development remains to
be clarified.
It is of interest that our results are in close accord with those
reported by Alabaster et al. (21) in the dimethylhydrazine-induced colon tumor model. Since the incidence of and rates of
mortality from breast and colon cancers are concordant in many
populations (47-51), the possibility may be considered that a
common environmental factor is involved in the cause and
prevention of both diseases. Minor constituents present in wheat
906
ARTICLES
bran and/or psyllium, including phytates (41), protease inhibitors (42), and isoflavonoids (Adlercreutz H: unpublished
data), may, therefore, merit more scrutiny in the future.
References
(/) Willett WC, Hunter DJ, Stampfer MJ, Colditz G, Manson JE, Speigelman
D, et al. Dietary fat and fiber in relation to risk of breast cancer. An 8-year
follow-up [see comment citations in Medline]. JAMA 1992;268:2037^4.
(2) Graham S, Marshall J, Mettlin C, Rzepka T, Nemoto T, Byers T. Diet in
the epidemiology of breast cancer. Am J Epidemiol 1982; 116:68-75.
(3) Lubin F, Wax Y, Modan B. Role of fat, animal protein, and dietary fiber in
breast cancer etiology: a case-control study. J Natl Cancer Inst 1986;
77:605-12.
(4) van 't Veer P, Kolb CM, Verhoef P, Kok FJ, Schouten EG, Hermus RJ, et
al. Dietary fiber, beta-carotene and breast cancer results from a case-control study IntJ Cancer 1990,45:825-8.
(5) Baghurst PA, Rohan TE. High-fiber diets and reduced risk of breast cancer. IntJ Cancer 1994;56:173-6.
(6) Lee HP, Gourley L, Duffy SW, Esteve J, Lee J, Day NE. Dietary effects on
breast-cancer risk in Singapore [see comment citation in MedlineJ. Lancet
1991;337:1197-200.
(7) Howe GR, Hirohata T, Hislop TG, Iscovich JM, Yuan JM, Katsouyanni K,
et al. Dietary factors and risk of breast cancer, combined analysis of 12
case-control studies [see comment citation in Medline]. J Natl Cancer Inst
1990;80:561-9.
(8) Cohen LA, Kendall ME, Zang E, Meschter C, Rose DP. Modulation of Nnitrosomethylurea-induced mammary tumor promotion by dietary fiber
and fat [see comment citations in Medline]. J Natl Cancer Inst
1991;83:496-501.
(9) Arts O , de Bie AT, van den Berg H, van 't Veer P, Bunnik GS, Thijssen
JH. Influence of wheat bran on NMU-induced mammary tumor development, plasma estrogen levels and estrogen excretion in female rats. J
Steroid Biochem Mol Biol 1991 ;39:193-202.
(10) Adlercreutz H. Western diet and Western diseases: some hormonal and
biochemical mechanisms and associations. Scand J Clin Lab Invest Suppl
1990;201:3-23.
(//) Gorbach SL. Estrogens, breast cancer, and intestinal flora. Rev Infect Dis
1984;6 Suppl 1:S85-9O.
(12) GoWin BR, Adlercreutz H, Gorbach SL, Woods MN, Dwyer JT, Conlon T, et
al. The relationship between estrogen levels and diets of Caucasian American
and Oriental immigrant women. Am J Clin Nutr 1986;44:945-53.
(13) Goldin BR, Gorbach SL. Hormone studies and the diet and breast cancer
connection. Adv Exp Med Biol 1994,364:35^6.
(14) Woods MN, Gorbach SL, Longcope C, Goldin BR, Dwyer JT, Morrill-LaBrode A. Low-fat, high-fiber diet and serum estrone sulfate in premenopausal women. Am J Clin Nutr 1989:49:1179-83.
(15) Lanza E. Shankar S. Trock B. Dietary fiber. Chap 12. In: Micozzi M,
Moon T, editors. Macronutnents. Investigating their role in cancer. New
York: Marcel Dekker, 1991:293-320.
(16) US Department of Agriculture and Department of Health and Human Services. Dietary guidelines for Americans. 3rd ed. Washington (DC): US
Govt Print Off, 1990.
(IT) Public Health Service, US Department of Health and Human Services.
Healthy People 2000: National health promotion and disease prevention
objectives. Washington (DC): US Govt Print Off, 1990.
(IS) US Department of Health and Human Services. The Surgeon General's
report on nutrition and health. PHS Publ No. 88-50210. Washington (DC):
US Govt Print Off. 1988.
(19) World Health Organization. Diet, nutrition and the prevention of chronic
diseases. Geneva: World Health Organization, 1990.
(20) Health and Welfare Canada. Report of the Expert Advisory Committee on
Dietary Fiber. Ottawa: Health Protection Branch, Health and Welfare,
1985.
(21) Alabaster O, Tang ZC, Frost A, Shivapurkar N. Potential synergism between wheat bran and psyllium: enhanced inhibition of colon cancer. Cancer Lett 1993:75:53-8.
(22) Bieri JG, Stoewsand GS, Briggs GM, Phillips RW, Woodward JC. Knapka
JJ. Report of the American Institute of Nutrition Ad Hoc Committee on
Standards for Nutritional Studies. J Nutr 1977; 107:1340-8.
(23) US Department of Health, Education and Welfare. Guidelines for carcinogen bioassay in small rodents. DHEW Publ No. (NlH)76-801.
Washington (DC): US Govt Print Off. 1976:1-65.
(24) Chan PC, Head JF, Cohen LA, Wynder EL. Influence of dietary fat on the
induction of mammary tumors by iV-nitrosomethylurea: associated hor-
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
(25)
(26)
(27)
(28)
(29)
(JO)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(59)
(40)
(41)
(42)
mone changes and differences between Sprague-Dawley and F344 rats. J
Natl Cancer Inst 1977:59:1279-83.
Newbeme PM. Bien JG, Briggs GM, Nesheim MC. Control of diets in
laboratory animal experimentation. Anim Res News 1978;21:A1 -A 12.
Butcher RL, Collins WE, Fugo NW. Plasma concentration of LH. FSH.
prolactin, progesterone and estradiol-17fi throughout the 4-day estrous
cycle of the rat. Endocrinology 1974;94:1704-78.
Young S, Hallowes RC. Tumours of the mammary gland. IARC Sci Publ
1973:5:31-73.
Russo J, Russo IH, Rogers AE, van Zwietan Ml. Gusterson B. Pathology
of rumours in laboratory animals. Tumours of the rat. Tumours of the
mammary gland. IARC Sci Publ 1990,99:47-78.
Adlercreutz H, Fotsis T, Heikkenen R. Current state of the an in the
analysis of estrogens. In: Gorog S, editor. Advances in steroid analysis.
Budapest, Hungary: Akademia Kiado, 1982:3-33.
Adlercreutz H, Jarvenpaa P. Assay of estrogens in human feces. J Steroid
Biochem 1982; 17:639-45.
Kulkami N, Reddy BS. Inhibitory effect of Bifidobacienum longum cultures on the azoxymethane-induced aberrant crypt foci formation and fecal
bacterial f}-glucuronidase. Proc Soc Exp Biol Med 1994:207:278-83.
Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958:53:457-81.
Peto R, Peto J. Asymptotically efficient rank invariant procedure. J R Stat
Socl972;135A:l85-207.
Helwig JT, Council KA, editors. Statistical analysis user's guide. Statistics
version. 5th ed. Raleigh (NC): SAS Institute, 1985.
Armitage P. Tests for linear trends in proportions and frequencies.
Biometrics 1955; 11:375-85.
Mantel N, Haenszel W. Statistical aspects of the analysis of data from
retrospective studies of disease. J Natl Cancer Inst 1959;22:719-48.
Miller RG. Simultaneous statistical inference. 2nd ed. New York:
Springer-Verlag, 1981:37.
SAS Institute. SAS for linear models: a guide to the ANOVA and GLM
procedures. Raleigh (NC): SAS Institute, 1981.
Boone CW, Kelloff GJ, Steele VE. Natural history of mtraepithelial
neoplasia in humans with implications for cancer chemoprevention
strategy. Cancer Res 1992:52:1651-9.
Gerber M. Fiber and breast cancer from epidemiological studies to intervention studies. Dietary Fiber Bibliography Reviews 1994;2:5-10.
Shamsuddin AM. Inositol phosphates have novel anticancer function. J
Nutr 1995;125(3 Suppl):725S-732S.
Troll W, Kennedy AD, editors. Protease inhibitors as potential cancer
chemopreventive agents. New York: Plenum Press, 1992.
(43) Carlestrom K. Estrone sulfate (EIS)—an estrogenic pre-hormone in postmenopausal women? In: Fioretti P. Flamighy C, Jasonni VM. Melis GB.
editors. Postmenopausal hormone therapy, risks and benefits. New York:
Raven Press, 1987:29-34.
(44) Adlercreutz H. Gorbach SL. Goldin BR, Woods MN. Dwyer JT,
Hamalainen E. Estrogen metabolism and excretion in Oriental and
Caucasian women [see comment citation in Medline] [published erratum
appears in J Natl Cancer Inst 1995:87:147]. J Natl Cancer Inst 1994;
86:1076-82.
(45) Roberts-Andersen J, Mehta T. Wilson RB. Reduction of DMH-induced
colon tumors in rats fed psyllium husk or cellulose. Nulr Cancer
1987;10:129-36.
(46) Jacobs LR. Relationship between dietary' fiber and cancer: metabolic,
physiologic, and cellular mechanisms. Proc Soc Exp Biol Med 1986:
183:299-310.
(47) Wynder EL, Hyams L, Shigematsu T. Correlations of international cancer
death rates. An epidemiological exercise. Cancer I967;2O:113-26.
(48) Drasar BS, Irving D. Environmental factors and cancer of the breast and
colon. Br J Cancer 1973;27:167-72.
(49) Howell MA. The association between colorectal cancer and breast cancer.
J Chronic Dis 1976:29:243-61.
(50) Burbank F. A sequential space-time cluster analysis of cancer mortality in
the United States: etiologic implications. Am J Epidemiol 1972;95:393417
(5/) Levine RS, Witte J. Geographical correlations of colon and breast cancer
[letter]. Lancet 1984:2:222.
Notes
Supported by a grant from the Kellogg Company, Battle Creek, MI, and the
Anne Skowronski Memorial Fund.
We thank M Epstein, J. Braley, and R. Hamid for assistance in the animal
bioassay; B. Pittman for assistance in the statistical analysis; and A. Banow for
preparation of the manuscript.
Presented in part at the 86th Annual Meeting of the American Association for
Cancer Research, Toronto, Ontario, Canada, March 21, 1995, and at the
American Institute for Cancer Research's Fifth Annual Conference on Diet and
Cancer, Washington, DC, September 1-2, 1994.
Manuscript received September 5, 1995; revised January 30, 1996; accepted
April 22, 1996.
Need the latest information on
clinical trials for HIV and AIDS?
Call the AIDS Clinical Trials
Information Service:
1-800-TRIALS-A
(1-800-874-2572) Monday through Friday, 9 a.m. to 7 p.m. eastern time.
Free up-to-date information
about more than 200 federally
and privately sponsored HIV
and AIDS clinical trials and the
drugs used in the trials.
———-^" Strict
Personalized assistance from
English- and Spanish-speaking
health specialists.
TTY/TDD access: 1-800-243-7012.
A service of the U.S. Department of Health and Human Services, Public Health Service. The Centers
for Disease Control, Food and Drug Administration, National Institute of Allergy and Infectious
Diseases, and National Library of Medicine have collaborated to provide this service.
Journal of the National Cancer Institute, Vol. 88, No. 13, July 3, 1996
ARTICLES
907