Download pdf - Suppressing effect of perilla oil on azoxymethane - Carcinogenesis

Carcinogenesis vol.17 no.6 pp. 1291-1296, 1996

Suppressing effect of perilla oil on azoxymethane-induced foci ofcolonic aberrant crypts in rats

Nobuhito Onogi, Masataka Okuno, Chihito Komaki,Hisataka Moriwaki, Toshihiko Kawamori1,Takuji Tanaka1, Hideki Mori1 and Yasutoshi Muto2

First Department of Internal Medicine and 'First Department of Pathology,Gifu University School of Medicine, 40 Tsukasa-machi, Gifu 500, Japan2To whom correspondence should be addressed

We have investigated the modulatory effect of dietaryperilla oil which is rich in the n-3 polyunsaturated fattyacid, a-linolenic acid, on the development of azoxymethane(AOM)-induced colonic aberrant crypt foci (ACF) in maleF344 rats. Animals were given three weekly subcutaneousinjections of AOM (15 mg/kg body weight) to induce ACF.The rats were fed a basal diet containing either 12% oliveoil, 12% safflower oil, 12% perilla oil, 6% perilla oil plus6% olive oil, or 3% perilla oil plus 9% olive oil for 5weeks, starting 1 week before the first dosing of AOM. Allrats were sacrificed 2 weeks after the last AOM injection.The amount of food consumed and body weight gain wereidentical among every dietary group. The frequency ofACF was significantly lower in the rats fed 12% perilla oilthan in those fed 12% olive oil or 12% safflower oil (P <0.01 and P < 0.05, respectively). The suppressive effect ofperilla oil was dose-dependent, as the number of ACF was20.7, 40.7 and 47.4% of those of the 12% olive oil-fedcontrols in rats fed 12% perilla oil, 6% perilla oil plus 6%olive oil and 3% perilla oil plus 9% olive oil, respectively.Perilla oil significantly reduced ras expression as well asthe AgNORs count (cell proliferation biomarkers) in thecolonic mucosa, as compared with olive oil or saffloweroil (P < 0.01, respectively). Marked increases in n-3polyunsaturated fatty acids in membrane phospholipidfractions and decreased PGE2 levels were observed incolonic mucosa of perilla oil-fed rats. These results suggestthat perilla oil, even in small amounts, suppresses thedevelopment of aberrant crypt foci, and is therefore apossible preventive agent in the early stage of colon carcino-genesis.

Introduction

Epidemiologic studies have shown that dietary factors play akey role in the etiology of colorectal cancers (1-3). Western-style diets, rich in animal fat and poor in fiber, have beensuggested to associate with an increased risk of colon cancer(4). On the other hand, studies in Alaskan and GreenlandEskimos have suggested that their low incidence of coloncancer is linked to their high dietary consumptions of fish oil

•Abbreviations: ACF, aberrant crypt foci; AgNORs, silver-stained nucleolarorganizer region protein; AOM, azoxymethane; DHA, docosahexaenoic acid;EGFR, epidermal growth factor receptor; EPA, eicosapentaenoic acid; GAPDH,glyceraldehyde-3-phosphate dehydrogenase; MUFA, monounsaturated fattyacid; PGE2, prostaglandin E2; PUFA, polyunsaturated fatty acid; SFA,saturated fatty acid.

products (5,6). Thus, the types or composition of fatty acidsin oil may be significant determining factors in colon carcino-genesis.

Fish oil contains an appreciable amount of n-3 polyunsatur-ated fatty acids (PUFAs*), namely eicosapentaenoic acid (EPA,C20:5 n-3) and docosahexaenoic acid (DHA, C22:6 n-3) anda small amount of the n-6 PUFA, linoleic acid (C18:2 n-6).Perilla oil is also rich in vegetable n-3 PUFA, a-linolenic acid(C18:3 n-3). a-Linolenic acid is converted to EPA or DHAafter enzymatic desaturation and elongation in the liver. Atten-tion has been paid to n-3 PUFAs in the prevention of coloncarcinogenesis. Studies in animal models have shown thatsuch oils rich in n-3 PUFAs suppress chemically induced coloncancer in rats (7-12). n-3 PUFAs have also been reported tosuppress rectal cell proliferation in humans (13,14). Thesereports have shown that colon tumorigenesis was suppressedwhen n-3 PUFA was administrated in both the initiation andpost-initiation phase (10,11) or in the post-initiation phaseonly (12). However, limited information is currently availableon the effect of perilla oil in the initiation phase of coloncarcinogenesis.

Aberrant crypt foci (ACF) have been recently reported tobe putative preneoplastic lesions of colon cancer in bothrodents (15) and humans (16), and ACF with large numbersof crypts (four or more) per focus have been proposed asintermediate biomarkers for colon carcinogenesis (17-19).ACF have been shown to be associated with c-fos (20) andras mutations in rats (21,22) and humans (23). They havebeen demonstrated to be useful biomarkers for screeningchemopreventive agents against colon cancer (18). Recently,DHA has been reported to suppress the formation and growthof ACF, as well as colon tumors, in rats (19).

Here, we report an inhibitory effect of perilla oil on thedevelopment of ACF induced by azoxymethane (AOM) inrats. The aim of the present study was to examine whetherperilla oil is effective in the initiation phase of colon carcino-genesis and to gain further insight into its mechanism. Theeffect was dose-dependent, and was accompanied by a reduc-tion in colonic cell proliferation as revealed by the silver-stained nucleolar organizer region protein (AgNORs) count aswell as ras expression. A rapid replacement of fatty acids inthe mucosal phospholipid fraction by n-3 PUFAs and asuppression of prostaglandin E2 (PGE2) synthesis in thecolonic mucosa of perilla oil-fed rats was also found, suggestingone of the mechanisms of perilla oil's antitumor effect.

Materials and methods

Animals

Four-week-old male F344 rats (Shizuoka Laboratory Animal Center,Hamamatsu, Japan) were quarantined for I week, and randomized intoexperimental and control groups. Animals were housed, two to three ratseach, in a plastic cage in a holding room under constant conditions of 22 ±2°C, 50 ± 10% humidity, and a 12-h-light-dark cycle. They had free accessto drinking water and food.

© Oxford University Press 1291

Dow

nloaded from https://academ

ic.oup.com/carcin/article/17/6/1291/1785337 by guest on 24 N

ovember 2021

N.Onogi et al.

Chemicals and dietary fats

AOM was purchased from Sigma Chemical Co., St Louis, MO. The dietaryfats, olive oil, safflower oil and perilla oil were supplied by the AjinomotoCo., Tokyo, Japan. The fatty acid composition of the oils was as follows:olive oil—9.6% palmitic acid (C16:0), 2.7% stearic acid (C18:0), 80.6% oleicacid (C18:l), 5.8% linoleic acid (C18:2 n-6) and 0.6% ot-linolenic acid (C18:3n-3); safflower oil— 6.2% palmitic acid, 1.6% stearic acid, 11.2% oleic acid,80.5% linoleic acid and 0.4% oc-linolenic acid; and perilla oil—-6.2% palmiticacid, 2.0% stearic acid, 17.7% oleic acid, 15.2% linoleic acid and 56.0% a-linolenic acid.

Diets

The experimental diets were prepared once a week by adding olive oil,safflower oil or perilla oil to the basal laboratory diet (Oriental Yeast Co.,Tokyo, Japan). The final w/w composition of the diet was 20% casein, 59%sucrose, 4% cellulose, 0.15% choline chloride, 4% mineral mixture, 1%vitamin mixture and 12% test oil. The diet had standard levels of nutrients.The final composition of test oils was 12% olive oil in the 012 diet, 12%safflower oil in the S12 diet, 12% perilla oil in the P12 diet, 6% perilla oilplus 6% olive oil in the P6O6 diet and 3% perilla oil plus 9% olive oil in theP3O9 diet. The ratios of monounsaturated fatty acid (MUFA) : n-6 PUFAs :n-3 PUFAs were 1:0.07:0.01 in the O12 diet, 1:7.2:0.04 in the S12 diet, and1:0.74:2.7 in the PI2 diet. The food was sealed in air-tight plastic bags undernitrogen gas and stored at —20°C until use. The food in the animal cages wasshaded from light and changed every other day.

Experimental design

A total of 100 rats were divided into ten groups of ten rats each (Figure 1).The 6-week-old rats in groups 1-5 were given subcutaneous injections ofAOM (15 mg/kg body weight) once a week for 3 weeks. The rats in groups6-10 served as AOM-negative controls. The rats in groups 1 and 6 were fedthe O12 diet. Groups 2 and 7 rats were fed the S12 diet, groups 3 and 8 ratsthe PI2 diet, groups 4 and 9 rats the P6O6 diet and groups 5 and 10 rats theP3O9 diet, starting at 5 weeks of age. The daily intake of the diets wasrecorded. All rats were provided with the diet and tap water ad libitum, andwere weighed weekly. All animals were killed at 4 weeks after the firstadministration of AOM, and complete necropsies were performed. The colonsof the five rats in each group were used for examining ACF and AgNORs,and those of the remaining five rats were used for the analyses of PGE2 andfatty acid compositions.

Identification of ACF

At the termination of the study, the colons of the five animals in each groupwere flushed with saline, slit open longitudinally from cecum to anus, placedbetween two pieces of filter paper and fixed in buffered 10% formalin solutionfor 24 h. The colons were stained with 0.2% methylene blue in salineaccording to the method of Bird (24). The number of ACF/colon and thenumber of aberrant crypts in each focus were counted microscopically at amagnification of X40. The criteria used to identify an aberrant crypt focustopographically were as follows: (i) increased size; (ii) thicker epithelial celllining; and (iii) increased pericryptal zone relative to normal crypts.

PGE2 levels

For determination of PGE2 levels in the colonic mucosa, the colons of thefive rats in each group were slit longitudinally, placed between two pieces ofslide-glass, and immediately frozen in liquid nitrogen. The mucosal layer,which attached to one piece of the glass, was separated easily from themuscular and serosal layers, and was scraped off using a brazer. Homogenateswere prepared by homogenizing the mucosa in ice-cold phosphate-bufferedsaline containing 1 mM EDTA, 0.1 mM indomethacin and 100 U/ml aprotininusing a glass-Teflon homogenizer. The debris and nuclei were removed bycentrifugation at 2000 g for 10 min at 4°C. After adding [3H]PGE2 (DuPont/NEN, Boston, MA) as an internal standard, the supernatants were acidified topH 3.5 with acetic acid and centrifuged at 2000 g for 10 min to removepellets. PGs were extracted on a Bond-Elute C-18 column (AnalytichemInternational, Harber, CA), and the PGE2 level in the extract was assayedusing a PGE2 -[I25I] RIA kit (DuPont/NEN). The data were calculated asmg/g wet weight of tissue sample.

Fatty acid compositions

The fatty acids in the colonic mucosa were extracted by the method of Folchet al. (25), and phospholipids were isolated according to the method of Rouseret al. (26). Pentadecanoic acid was added as an internal standard. The sampleswere subjected to methanolysis in 5% HCl in methanol at 80°C for 2 h undernitrogen. Fatty acid methyl esters were extracted with n-hexane and analysedby gas-chromatography (Shimadzu GC-9A, Shimadzu, Kyoto, Japan) with anHR-SS-10 column (Shimadzu). The oven temperature was programed from5O-22O°C. The identification and quantitation of each fatty acid were made

1292

Group no.

1 - 5

6-10

3 i

t

2

ttest

test

3

toil

oil

4 5

t

weeks

10 rats each

10 rats each

Fig. 1. Experimental protocol, i, AOM, 15 mg/kg body weight, viasubcutaneous injection; t , sacrifice. Test oil: 12% olive oil (groups 1 and6); 12% safflower oil (groups 2 and 7); 12% perilla oil (groups 3 and 8);6% perilla oil plus 6% olive oil (groups 4 and 9); and 3% perilla oil plus9% olive oil (groups 5 and 10).

with authentic standard mixtures (Sigma) using a CR-3A Chromatopacintegrator (Shimadzu).

AgNORs countThe colon was divided into three equal portions, labeled the upper, middleand lower colon, and fixed in 10% buffered formalin. All portions of thecolon were embedded in paraffin, and two serial sections (3 (im thick) weremade. One section was used for staining AgNORs, and the other was stainedwith hematoxylin and eosin for histologic examination. AgNORs staining wascarried out as described previously (27). The AgNORs count of the mucosalepithelium of the colon was determined in all epithelial cells lining the 20well-oriented crypts, in which the base, lumen and top of the crypt couldbe observed completely. The AgNORs-stained nuclei were counted at amagnification of X400. Data are expressed as number of AgNORs/nucleus.

Northern blot analysisTotal RNA was extracted according to the method of Chomczynski and Sacchi(28). Northern hybridization was performed as described previously (29),utilizing 32P-labeled probe specific either for human c-H-ras (Japanese CancerResearch Resources Bank) or glyceraldehyde-3-phosphate dehydrogenase(GAPDH).

Statistical analysisThe significance of the differences between group means was analyzed byDunnet's Mest.

Results

The mean body weight and food intake of the rats receivingAOM were lower than those of rats without AOM treatmentin each dietary group (Table I). However, no significantdifference was observed in either the mean body weights orthe mean food intakes of the AOM-treated rats in the differentdietary groups. No neoplasm was found on macroscopic ormicroscopic examination of the organs of any rat in each group.

The rats treated with AOM (groups 1-5) showed a 100%incidence of ACF (Table II). No ACF developed in the colonsof rats without AOM treatment (groups 6-10). Both the numberof ACF/colon and the number of aberrant crypts/colon weresignificantly decreased in the S12 group as compared with the012 group (P < 0.01, respectively) (Table II). The number ofACF/colon and the number of aberrant crypts/colon were alsosignificantly reduced in the P12 group as compared with the012 (P < 0.01, respectively) and S12 group (P < 0.05,respectively) (Table II). The number of aberrant crypts/focusin the PI2 and S12 groups were significantly smaller than thatin the 012 group (P < 0.05). The suppressive effect of perillaoil on the development of ACF was obvious in both small- tolarge-sized foci, as the number of foci consisting of one toeight crypts was equally decreased by perilla oil administration,as compared with safflower (P < 0.05) or olive oil feeding(P < 0.01) (Figure 2). The suppressive effect of perilla oil onthe number of ACF/colon and on the total number of aberrantcrypts/colon was dose-dependent (Table II). Both numberswere significantly lower in the P3O9 group than in the 012

P5T

Dow



ovember 2021

Suppression of colonic aberrant crypts in rat

Table I. Mean body weight and mean diet intakes of rats in each test group

Group no.Diet

1O12

2S12

3Pl2

4P6O6

5P3O9

With AOM treatmentNo. of ratsMean body weight (g)Mean food intake (g/rat/day)

1022912.9

± 10± 0.6

10224 ± 411.0 ± 1.0

10227 ± 1212.5 ± 1.1

10233 ±13.3 ±

111.6

10232 ±13.6 ±

142.0

Group no.Diet

Without AOM treatmentMean body weight (g)Mean food intake (g/rat/day)

6Ol2

25012.2

± 10± 0.6

7Sl2

245 ±13.4 ±

121.0

8Pl2

24913.0

± 9±0.6

9P6O6

238 ±12.9 ±

5.02.1

10P3O9

243 ±12.1 ±

141.5

Rats were fed with a diet containing 12% olive oil (O12), 12% Safflower oil (S12), 12% perilla oil (P12), 3% perilla oil plus 9% olive oil (P3O9), or 6% perillaoil plus 6% olive oil (P6O6).Values represent mean ± SD, at the termination of the study (n = 10).

Table II. Effect of test oils on AOM-induced ACFs in rat colon

Group no. Treatment Incidence11 No. ofACF/colon

No. of aberrantcrypts/colon

No. of aberrantcrypts/focus

AOM HAOM 1AOM HAOM 1AOM -,

f- O12

h S12

h Pl2

h P6O6

h P3O9

5/55/55/55/55/5

155.0 ± 10.881.8 ± 8.6b

40.0 ± 14.8bc

78.5 ± 11.3b

91.4 ± 13.8b

260.4 ± 16.2119.6 ± 11.8b

50.3 ± 3.1b-c

126.5 ± 21.0b

148.6 ± 26.3b

1.68 ± 0.081.46 ± 0.04b

1.54 ± 0.09b

1.61 ± 0.081.62 ± 0.07

For definitions of dietary groups, see the note to Table I."Numbers of rat colon with ACF/total numbers of colon are scored•"Significantly different from the AOM + O12 group; P < 0.01.Significantly different from the AOM + S12 group; P < 0.05.Values are mean ± SD (n = 5).

Group no.

1. O12 group

2.Si2group

3. P12 group

1 2 3 4 5 6 7 8 9 10

No. of aberrant crypts per focus

Fig. 2. Size distribution of AOM-induced ACF in the colon of O12, S12and PI2 groups. Columns, mean; bars, SD {n = 5). (a) Significantlydifferent from the O12 group; P < 0.01. (b) Significantly different from theS12 group; P < 0.05.

group, indicating that as little as 3% perilla oil inhibited thedevelopment of ACF. The suppression was also obvious inACF of all the sizes (Figure 3).

The fatty acid compositions of the colonic mucosa in the012, S12 and P12 groups with AOM treatment are shown inTable III. The data were summarized by grouping the fattyacids into saturated fatty acids (SFAs), MUFAs, n-6 PUFAsand n-3 PUFAs. MUFAs and n-6 PUFAs were the majorcomponents in the O12 group and in the S12 group, respect-ively, corresponding to the composition of the dietary fattyacids. Only trace amounts of n-3 PUFAs were detected in the

Group no.

1 1 1.Oi2 group

tZZZl 5. P3O9group

4. P6O6group

3. P12 group

' 2 3 4 5 6 7 8 9 10No. of aberrant crypts per focus

Fig. 3. Dose-dependent effect of perilla oil on the development of AOM-induced colonic ACF in the 012, P3O9, P6O6 and P12 groups. Columns,mean; bars, SD (n = 5). (a) Significantly different from the O12 group;P < 0.01. (b) Significantly different from the P3O9 and P6O6 groups;P < 0.01.

012 and S12 groups. In the P12 group, n-3 PUFAs increasedmarkedly. EPA and DHA were the major n-3 PUFA compon-ents. The mucosal n-3 PUFAs/MUFAs ratio was 0.98 in theP12 group and 0.04 in the 012 group. The n-3 PUFAs/n-6PUFAs ratio was 1.10 in the P12 group and 0.03 in theS12 group.

A high PGE2 level in the colon was observed in the S12group (Figure 4). The PGE2 level in the P12 group wassignificantly (P < 0.05) decreased, to approximately half thatof the S12 group. The PGE2 level of the 012 group was

1293

Dow



ovember 2021

N.Onogi el at.

Table

Groupno.

HI. Fatty acid

Treatment

composition

SFAs

C16:0

of the phospholipid fractions

MUFAs

C18:0 C16:1

of the

C18

colonic

:1

mucosa of AOM-treated

n-6 PUFAs

C18:2 C20:4

rats in the O12,

n-3 PUFAs

C18:3

S12 and

C20:5

P12 groups

C22:6

Others

A O M + O12 27.7 ± 2.5 11.3 ± 2 . 5 4 . 0 + 1 . 1 32.2 ± 4.1" 6.0 + 0.7 12.3 + 1.3 0 . 1 + 0 . 1 0.4 + 0.1 1 .1+0.2 5.0 ± 1.2

39.0 ± 5.0 36.2 + 5.2' 18.3 ± 2.0 1.6 + 0.4

AOM + S12 17.2 + 2.3 6.7 ± 0.4 2.9 ± 0.6 14.4 i: 1.0" 43.5 ± 3.0* 9 .3+1.7 0.2 i 0.1 1.3 ± 1.4 0.2 + 0.1 4 . 4 + 1 . 7

23.9 + 2.7 17.3 + 1.6 52.8 + 47° 1.7 + 1.6

AOM + P12 18.2 ± 1.6 10.3 ± 2.0 2.7 + 0.6 20.5 ± 1.8 16.6 ± 1.4 4.1 ± I.I 16.0 ± 6.0" 5 .6+1 .7 1.1+0.8 4.8 + 3.5

28.5 + 3.6 23.2 + 2.4 20.7 ± 2.5 22.7 + 8.5°

For definitions of dietary groups, see the note to Table I."Significantly different from the other two groups; P < 0.05.SFA, saturated fatty acid; MFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid.Values represent per cent of total fatty acids in each oil (mean ± SD, n = 5).

UJ

Q.

• I :

• T

-a)

Group no

O12 group

2

S12 group

3

P 12 group

Fig. 4. PGE2 levels in the colonic mucosa of AOM-treated rats in the 012,S12 and P12 groups. Columns, mean; bars, SD (n = 5). (a) Significantlydifferent from the O12 and S12 groups; P < 0.05.

liable IV. Number of cells/crypt column and AgNORs counts of colonicmucosa in each group

Group no. Treatment No. of cells/crypt column

AgNORscount/nucleus

AOM + O12AOM + S12AOM + P12AOM + P6O6AOM + P3O9

42.4 + 3.932.5 + 2.0*33.4 + 2.9*34.3 + 3.2*35.3 + 2.9*

.60

.51

.35

.41

0.140.170.2 l b

0.12b

.54 ± 0.16

For definitions of dietary groups, see the note to Table I."Significantly different from the AOM + O12 group; P < 0.01.Significantly different from the AOM + O12 group; P < 0.05.Values are mean ± SD (n = 5).

approximately the same as that of the S12 group, although theoccurrence of ACF was much higher in the 012 group.

The mean number of cells/crypt column was significantlysmaller (P < 0.01) in the S12 and the P12 groups than in theO12 group (Table TV). The mean AgNORs count/nucleus inthe P12 group was significantly lower {P < 0.01) than thosein the O12 and S12 groups. Both numbers were suppressed in

ras

GAPDH

0,2 group S,2 group P)2 group

Fig. 5. ras mRNA expression in colonic mucosa of AOM-treated rats in the012, S12 and PI2 groups. Total RNA was isolated from each colonicmucosa, fractionated through 1% agarose-formaldehyde gels, transferred tonylon membranes, and hybridized with 32P-labeled probes for ras (upperbands) and GAPDH (lower bands).

a dose-dependent fashion by increasing the amount of perillaoil (Table IV).

The expression of c-H-ras mRNA was reduced in the PI2group as compared to those in the S12 and 012 groups (Figure5). The expression in the SI2 group was slightly reduced ascompared to those in the 012 groups.

Discussion

In the present study, perilla oil clearly suppressed the develop-ment of AOM-induced ACF in rat colons in a dose-dependentfashion, suggesting that dietary n-3 PUFA was effective in theinitiation phase of colon carcinogenesis. Treatment with perillaoil also reduced the AgNORs number and the height of thecolonic mucosa. After 6 weeks of perilla oil treatment, thefatty acid compositions of mucosal membrane phospholipidswere completely changed, and there were significantly reducedPGE2 levels in the colonic mucosa. It is noteworthy that a 3%perilla oil diet reduced the number of ACF to less than halfthat of 12% olive oil controls. These results are consistentwith the report by Narisawa et al. (11) in which a 3% perillaoil feeding significantly suppressed the incidence of colontumors, as compared with 12% safflower oil. In the currentstudy, there was no significant difference in food consumptionor body weight gain among the dietary groups. Therefore, thesuppressing effect of perilla oil on occurrence of ACF wasnot related to calorie intake.

Unexpectedly, the development of ACF was highest in the

1294

Dow



ovember 2021

Suppression of colonic aberrant crypts in rat

olive oil group, and the results were reproducible in threeindependent experiments (data not shown). Epidemiologicstudies have indicated a low mortality from colon cancer inthe areas where a large amount of olive oil is consumed(30,31)- However, the effect of oleic acid is not so obvious inanimal experiments, as a previous study reported a similarincidence of colon tumors in rats fed diets supplemented eitherwith palm oil, rich in oleic acid (40% of total fatty acids) orwith safflower oil (10). Moreover, the results in the presentstudy were obtained in the initiation phase, therefore the tumorpromoting effect of safflower oil might not be exerted in suchan early phase of carcinogenesis. In fact, the development ofaberrant crypt foci has also been reported to be enhancedby the supplementation with olive oil as compared to thesupplementation with corn oil which is rich in n-6 PUFAs(32). One possible mechanism is that olive oil may positivelyinfluence carcinogen activation in the liver (33). Further studieswill be needed to clarify the effect of oleic acid and linoleicacid on the development of colonic ACF.

Although the mechanism(s) of the tumor promoting effectof animal fat and of the lack of such an effect by fish orperilla oil is not fully understood, some explanations havebeen proposed. One is that high dietary fat increases thesecretion of secondary bile acids, which are known to act ascolon tumor promoters (34,35). However, a detailed experimentby Narisawa et al. (10) has shown that the amount andconcentration of total and secondary bile acids in the feces arenot different between rats fed perilla oil and those fed saffloweroil. Therefore, it seems unlikely that the inhibitory effect ofperilla oil can be attributed to a change in fecal bile acids.Another possible explanation is altered PG synthesis. PGs,particularly the type-2 series, are believed to be closelyinvolved in colon carcinogenesis, as increased PG levelshave been found in colon cancer tissues (36). PGE2 induceshyperproliferation in colonic mucosa (37), and inhibitors of PGsynthesis, such as indomethacin, inhibit colon carcinogenesis inrats (38^tO). n-3 PUFA has been reported to inhibit theproduction of the type-2 series of eicosanoids, including PGE2,from arachidonic acid (41^43). In the present study, theconcentration of PGE2 in the colonic mucosa was suppressedby perilla oil. However, the PGE2 concentration did notcorrelate with the development of ACF in the olive oil group.Since the effect of PGE2 have been reported to be exerted inthe promotion phase (44), it might not be the case in theinitiation phase. These results led us to consider anothermechanism by which perilla oil exerts its inhibitory effect onthe development of ACF. In the present study, perilla oilsignificantly reduced AgNORs count/nucleus, suggesting thatperilla oil decreased the number of cells in S-phase, and thusprotected DNA from injury by AOM. The observation thatperilla oil suppressed c-H-ras expression, another cell prolifera-tion biomarker, also supports the possibility that perilla oilexerted an antiproliferative effect on the mucosal cells andthus protected them from the carcinogenic stimuli by AOM.Marked alterations in fatty acid composition were also observedin the colonic mucosal membranes of rats fed perilla oil. Oralsupplementation with n-3 PUFAs is known to induce a selectiveincorporation of n-3 PUFAs and a competitive exclusion ofn-6 PUFAs in the membrane phospholipid fractions (45). Aschanges in the ratio of n-3 to n-6 PUFAs in the membranecould affect the function of the membrane itself and/or mem-brane bound receptors, such as epidermal growth factor receptor(EGFR) (46), it may be possible that such change in membrane

compositions altered the sensitivity of the cells to growth andcarcinogenic stimuli (47). Another possible mechanism maybe antioxidant action which is exerted by micronutrients, suchas vitamin E, in the oils. The concentrations of vitamin E were465 (AM in olive oil, 1023 |iM in safflower oil and 1493 |iMin perilla oil, which we measured using high performance liquidchromatography (Onogi et al., unpublished observations).Therefore, the development of ACF may be related inverselyto the amount of dietary intake of vitamin E, as has beensuggested very recently (48). Further studies are currently underinvestigation in our laboratory to elucidate the mechanism ofperilla oil's effect.

Our data that a vegetable n-3 PUFA, a-linolenic acid,effectively inhibited the development of colonic ACF, an earlybiomarker of colon carcinogenesis, support previous reportsin which supplementation with perilla oil inhibited colontumorigenesis in rats (10-12). Moreover, even a small amountof perilla oil (3%) could exert a significant suppressing effect.These findings support the possibility for clinical use. Inaddition, a-linolenic acid in perilla oil is much more chemicallystable than EPA and DHA in fish oil (49), suggesting that theuse of perilla oil seems advantageous for the long-term clinicalapplications. In fact, a clinical trial in which perilla oil isbeing used for the treatment of inflammatory bowel diseases,including Crohn's disease, has already started in Japan. Weare now preparing for a clinical trial for an interventionalstudy to prevent colon cancer with the use of perilla oil.

Acknowledgements

We thank Messrs Kenta Kajiwara, Tetsuo Kobayashi, Ichiro Sounaka, ToshioMaki (Ajinomoto Inc.), Tetsuya Shintani and Seiji Hiraku (Ono Pharmaceut-icals) for their excellent technical assistance. This work was supported byGrants-in-Aid from the Ministry of Education, Science and Culture (05770350,M.O., 05670463, H.M.).

References

l.Wynder.E.L., Kajitani.T., Ishikawa.S., Dodo.H. and Takano.A. (1969)Enviromental factors of cancer of the colon and rectum. Cancer, 23,1210-1220.

2. WeisburgerJ.H. and Wynder.E.L. (1987) Etiology of colorectal cancerwith emphasis on mechanism of action and prevention. In DeVita.V.T.Jr,Hellman.S. and Rosenberg,S.A. (eds), Important Advances in Oncology,J.B.Lippincott Co., Philadelphia, pp. 197-220.

3. WeisburgerJ.H. (1991) Causes, relevant mechanisms, and prevention oflarge bowel cancer. Semin. Oncol., 18, 316-336.

4.Tajima,K., Hirose.K., Nagasawa.N., Kuroishi.T. and Tominaga.S. (1985)Urban-rural difference in the trend of colorectal cancer mortality withspecial reference to the subsites of colon cancer in Japan. Jpn. J. CancerRes., 76, 717-728.

5.Blot,W.J., LanierA, Fraumeni,J.F.Jr and Bender.T.R. (1975) Cancermortality among Alaskan natives. J. Natl Cancer Inst., 55, 546-554.

6. Bang,H.O., DyerbergJ. and Hjorne,N. (1976) The composition of foodconsumed by Greenland Eskimos. Ada Med. Scand., 200, 69—73.

7.Reddy,B.S. and Sugie.S. (1988) Effect of different levels of omega-3 andomega-6 fatty acids on azoxymethane-induced colon carcinogenesis inF344 rats. Cancer Res., 48, 6642-6647.

8.Reddy,B.S., Burill.C. and RigottyJ. (1991) Effect of diets high in omega-3 and omega-6 fatty acids on initiation and postinitiation stage of coloncarcinogenesis. Cancer Res., 51, 487-491.

9.Minoura,T., Takata,T, Sakaguchi,M., Takada,H., Yamamura.M., Hioki.K.and Yamamoto.M. (1988) Effect of dietary eicosapentaenoic acid onazoxymethane-induced colon carcinogenesis in rats. Cancer Res., 48,4790-4794.

lO.Narisawa.T., Takahashi.M., Kotanagi.H. et al. (1991) Inhibitory effect ofdietary perilla oil rich in the n-3 polyunsaturated fatty acid a-linolenicacid on colon carcinogenesis in rats. Jpn. J. Cancer Res., 82, 1089-1096.

ll.NarisawaJ., Fukaura,Y., Yazawa,K., Ishikawa.C, Isoda,Y. andNishizawa.Y. (1994) Colon cancer prevention with a small amount of

1295

Dow



ovember 2021

N.Onogi et al.

dietary perilla oil high in alpha-linolenic acid in an animal model. Cancer,73, 2069-2075.

12.Hirose,M., Masuda,A., Ito,N., Kamano.K. and Okuyama,H. (1990) Effectsof dietary perilla oil, soybean oil and safflower oil on 7,12-dimethylbenz[a]anthracene (DMBA) and 1,2-dimethylhydrazine (DMH)-induced mammary gland and colon carcinogenesis in female SD rats.Carcinogenesis, 11, 731-735.

13.Bartram,H.-P., Gostner.A., Scheppach.W., Reddy.B.S., Rao.C.V., Dusel.G.,Richter.F. and Richter,A. (1993) Effects of fish oil on rectal cellproliferation, mucosal fatty acids, and prostaglandin E2 release in healthysubjects. Castroenterology, 105, 1317-1322.

14.Anti,M., Armelao.F., Marra.G. et al. (1994) Effects of different doses offish oil on rectal cell proliferation in patients with sporadic colonicadenomas. Gastroenterology, 107, 1709-1718.

15. McLellan.E.A. and Bird,R.P. (1988) Aberrant crypts: potential preneoplasticlesions in the murine colon. Cancer Res., 48, 6187-6192.

16.Pretlow,T.P., Barrow.B.J., Ashton.W.S., O'Riordan,M.A., Pretlow,T.G.,JurcisekJ.A. and Stellato,T.A. (1991) Aberrant crypts: putativepreneoplastic foci in human colonic mucosa. Cancer Res., 51, 1564-1567.

17.Pretlow,T.P., O'Riordan,M.A., Somich.G.A., Amini.S.B. and Pretlow.T.G.(1992) Aberrant crypts correlate with tumor incidence in F344 rats treatedwith azoxymethane and phytate. Carcinogenesis, 13, 1509-1512.

18.Kawamori,T., Tanaka,T., Hara.A., Yamahara,J. and Mori.H. (1995)Modifying effects of naturally occurring products on the development ofcolonic aberrant crypt foci induced by azoxymethane in F344 rats. CancerRes., 55, 1277-1282.

19Takahashi,M., Minamoto.T., Yamashita.N., Yazawa.K., Sugimura,T. andEsumi.H. (1993) Reduction in formation and growth of 1,2-dimethylhydrazine-induced aberrant crypt foci in rat colon bydocosahexanoic acid. Cancer Res., 53, 2786-2789.

20. Stopera,S.A., Davie.J.R. and Bird.R.P. (1992) Colonic aberrant crypt fociare associated with increased expression of c-fos: the possible roleof modified c-fos expression in preneoplastic lesions in colon cancer.Carcinogenesis, 13, 573-578.

21.Stopera,S.A. and Bird.R.P. (1992) Expression of ras oncogene mRNA andprotein in aberrant crypt foci. Carcinogenesis, 13, 1863-1868.

22.Vivona,A.A., Shpitz.B., Medline.A., Bruce.W.R., Hay.K., Ward.M.A.,Stern.H.S. and Gallinger.S. (1993) K-ras mutations in aberrant cryptfoci, adenomas and adenocarcinomas during azoxymethane-induced coloncarcinogenesis. Carcinogenesis, 14, 1777-1781.

23. Pretlow.T.P., Brasitus.T.A., Fulton,N.C, Cheyer,C. and Kaplan.E.L. (1993)K-ras mutation in putative preneoplastic lesions in human colon. /. NatlCancer Inst., 85, 2004-2007.

24. Bird.R.P. (1987) Observation and quantification of aberrant crypts in themurine colon treated with a colon carcinogen: preliminary findings. CancerLett., 37, 147-151.

25.Folch,J., Ascoli.I., Lees,M., Meath.J.A. and LeBaron.F.N. (1951)Preparation of lipid extracts from brain tissue. J. Biol. Chem., 191,833-841.

26. Rouser,G., Krichevsky.G. and Yamamoto.A. (1967) Columnchromatographic and associated procedures for separation anddetermination of phosphatides and glycolipids. In Marinetti.G.V. (ed.),Lipid Chromatographic Analysis. Dekker, New York, vol.1, pp. 99-120.

27.Tanaka,T, Kojima,T, Suzui.M and Mori,H. (1993) Chemoprevention ofcolon carcinogenesis by the natural product of a simple phenolic compoundprotocatechuic acid: Suppressing effects on tumor development andbiomarkers expression of colon tumorigenesis. Cancer Res., 53, 3908-3913.

28. Chomczynski,P. and Sacchi.N. (1987) Single-step method of RNA isolationby acid guanidinium thiocyanate-phenol-chloroform extraction. Anal.Biochem., 162, 156-159.

29.Okuno.M., Caraveo.V.E., Goodman.D.S. and Blaner.W.S. (1995)Regulation of adipocyte gene expression by retinoic acid and hormones:effects on the gene encoding cellular retinol-binding protein. J. Lipid Res.,36, 137-147.

30.Macqart-Moulin,G., Riboli,E., CoreeJ., Charnay,B., Berthezene.P. andDay.N. (1986) Case-control study on colorectal cancer and diet inMalseilles. Int. J. Cancer, 38, 183-191.

31.LaVecchia.C, Harris,R. and Wynder.E.L. (1988) Comparativeepidemiology of cancer between the United States and Italy. Cancer Res.,48, 7285-7293.

32. Bird.R.P. and Lafave,L.M.Z. (1995) Varying effect of dietary lipids andazoxymethane on early stages of colon carcinogenesis: enumeration ofaberrant crypt foci and proliferative indices. Cancer Detect. Prevent., 19,308-315.

33.Alldrick,A.J., Rowland,I.R., Lake.B.G. and Flynn.J. (1987) High levels ofdietary fat: alteration of hepatic promutagen activation in the rat. J. NatlCancer Inst., 79, 269-272.

34. Reddy.B.S. and Maeura.Y. (1984) Tumor promotion by dietary fat inazoxymethane-induced colon carcinogenesis in female F344 rats: influenceof amount and source of dietary fat. /. Nail Cancer Inst., 72, 745-750.

35.Cummings,J.H., Wiggins.H.S., Jenkins.D.J.A., Houston,!-!., Jivraj,T.,Drasar,B.S. and Hill,M.J. (1978) Influence of diets high or low in animalfat on bowel habit, gastrointestinal transit time, fecal microflora, bile acid,and fat excretion. J. Clin. Invest., 61, 953-963.

36. Bennett.A., Del Tacca.M., Stamford.I.F. and Zebro,T. (1977) Prostaglandinsfrom tumours of human large bowel. Br. J. Cancer, 35, 881-884.

37.Tutton,P.J.M. and Barkla.D.H. (1980) Influence of prostaglandin analogueson epithelial cell proliferation and xenograft growth. Br. J. Cancer, 41,47-51.

38. Narisawa.T., Takahashi.M., Niwa.M., Fukaura.Y. and Wakizaka,A. (1987)Involvement of prostaglandin E2 in bile acid-caused promotion of coloncarcinogenesis and antipromotion by the cyclooxygenase inhibitorindomethacin. Jpn. J. Cancer Res., 78, 791-798.

39. Narisawa,T., Sato.M., Tani.M., Kudo.T., Takahashi.T. and Goto,A. (1981)Inhibition of development of methylnitrosourea-induced rat colon tumorsby indomethacin treatment. Cancer Res., 41, 1954-1957.

40.Pollard,M. and Luckert.P.H. (1983) Prolonged antitumor effect ofindomethacin on autochthonous intestinal tumors in rats. J. Natl CancerInst., 70, 1101-1105.

41. Culp.B.R., Titus.G.R. and Lands.W.E.M. (1979) Inhibition of prostaglandinsynthesis by eicosapentaenoic acid. Prostaglandins Med., 3, 269-278.

42. Corey.E.J., Chih,C. and Cashman.J.R. (1983) Docosahexaenoic acid is astrong inhibitor of prostaglandin but not leukotriene biosynthesis. Proc.NatlAcad. Sci. USA, 80, 3581-3584.

43.Rao,C.V. and Reddy.B.S. (1993) Modulating effect of amount and typesof dietary fat on ornithine decarboxylase, tyrosine protein kinase andprostaglandins production during colon carcinogenesis in male F344 rats.Carcinogenesis, 14, 1327-1333.

44.Reddy,B.S. (1992) Inhibitors of the arachidonic acid cascade and theirchemoprevention of colon carcinogenesis. In Wattenberg,L., Lipkin.M.,Boone.C.W. and Kelloff.G.J. (eds), Cancer Chemoprevention. CRC Press,Boca Raton, pp. 153-164.

45.Awad,A.B., Chattopadhyay,T.P. and Danahy.M.E. (1990) Effect of dietaryfat composition on rat colon plasma membrane and fecal lipid. J. Nutr.,119, 1376-1382.

46. Stubbs.C.D. and Smith.A.D. (1984) The modification of mammalianmembrane polyunsaturated fatty acid composition in relation to membranefluidity and function. Biochim. Biophys. Ada, 779, 89-173.

47.Turini,M.E., Basu.T.K. and Clandini.M.T. (1990) Prostaglandins—diet—cancer: a review. Nutr. Res., 10, 819-827.

48. Shivapurkar,N., Tang,Z., Frost,A. and Alabaster,O. (1995) Inhibition ofprogression of aberrant crypt foci and colon tumor development by vitaminE and p-carotene in rats on a high-risk diet. Cancer Lett., 91, 125-132.

49.Cho,S.Y, Miyashita,K., Miyazawa.T., Fujimoto.K. and Kanada.T. (1987)Autooxidation of ethyl eicosapentaenoate and docosahexaenoate. J. Am.Oil Chem. Soc, 64, 876-879.

Received on November 28, 1995; revised on February 23, 1996; accepted onMarch 13, 1996

1296

Dow



ovember 2021