(CANCERRESEARCH52, 5906-5912, November1, 19921

ABSTRA@

The effectof increasingamountsof wheatbran (0, 5, 10, 20%)inAIN-76 semisynthetic diet on colonic luminal short chain fatty acids,epithelial cell histone acetylation, and cytokinetics, was studied for 2weeks In groups of 10 male Sprague-Dawley rats. Luminal contentswere removedfrom the colon at sacrifice, quick frozen, and analyzed forshort chain fatty acids by gas-liquid chromatography. Histone acetyladon wasassessedin cells isolatedfromthe same @nim*Is.Cellproliferation was measured after a short pulse in vivo with I3Hlthymldine.Colonicluminal butyrate levelswere lowerin the 0 and 20% fibergroups, and higher in the 5 and 10% fiber groups. In contrast, cellpruliferadon, as determined by labeling index, was higher in the 0 and20%fiber groups, and lowerin the 5 and 10%fiber groups. This resultedIn a significant inverse correlation between luminal butyrate levels andcolonic cell proliferation. In addition, there was a positive linear conelatlon between luminal butyric acid levels and colon epithelial cell histone acetylation. From these data it was concluded that colonic butyratelevels can be modulated by the addition of wheat bran to the diet andthat butyrate can modulate DNA synthesis (calculated as labeling index) in the proliferative compartments of colonic crypts. The localization of dividing cells was unchanged and no induction of terminal differentiatlon was detectable (contrary to what has been observed fortransformed cells in culture).

INTRODUCTION

Colonic epithelial cell proliferation and differentiation havebeen well characterized both in normal (1—3)and in pathological conditions (4, 5). Recently it was suggested that the measurements of these parameters in gastrointestinal cells mightassist studies of nutritional intervention attempting to inhibitneoplasia (6).

Many recent epidemiological (7—9)and experimental (10—14)studies have focused on the role of dietary fibers and theirgastrointestinal metabolites on colonic epithelial cell physiology and carcinogenesis. The widely divergent results of thesestudies could be explained, in part, by taking into account differences in fiber composition, particularly the fermentable cornponents (15). Among the main fecal byproducts of fermentablefibers are (SCFA)3 (16, 17). One ofthem, butyric acid, has beendescribed to produce quite striking effects on all the tested celllines of neoplastic or transformed origin (18, 19). These effects

Received8/10/92; accepted8/24/92.The costs of publicationof this article weredefrayedin part by the paymentof

pagecharges.This articlemust thereforebeherebymarkedadvertisementin accordance with 18 U.S.C.Section 1734solelyto indicatethis fact.

I Supported in part by Associazione Italiana per Ia Ricerca sul Cancro (AIRC

1990-92) (L. C. B.), American Cancer Society Grant SIG-7A (M. L), and AmericanInstituteforCancerResearch(J. R. L.).

2 To whom requests for reprints should be addressed, at Department of ChemicalCarcinogenesis,Istituto Nazionaleper Ia Ricercasul Cancro,1ST,VialeBenedetto XV no. 10, 16132Genoa, Italy.

3 The abbreviations used are: AIN-76, American Institute ofNutrition semisyntheticdiet SCFA,shortchainfattyacidsAACC,AmericanAssociationofCerealChemists dThd, thymidine.

are seen even at concentrations obtainable in vivo, with fecaland cellular concentrations being correlated (20, 21). In fact, inseveral instances the expression of constitutive genes was increased (22, 23) together with a few usually silent genes (23—25). The net result was a marked inhibition of both DNA andprotein synthesis (18, 19, 26), resulting in a preferential blockof the cell cycle in early G1 (26), eventually leading to terminalcellular differentiation (18, 19). Butyric acid also has an effecton several enzymatic activities (18, 19, 27—29),including histone deacetylases (30). This class of enzymes is inhibited evenby trace amounts of butyric acid with resultant histone hyperacetylation (30). Acetylation of the lysine residues in the basicamino terminal region of the histones neutralizes their positivecharges, weakening the interaction with the phosphate group ofthe DNA strand, and therefore opening the structure of thebasic unit of the chromatin, the nucleosome (31, 32). Association with hyperacetylated histones has been shown to be a necessary but insufficient condition for the expression of a genesequence (33, 34). It also makes the DNA more prone to carcinogen damage and damage repair (35, 36).

In this paper studies are presented involving the modulationof colonic epithelial cell proliferation by variable contents offiber in the diet. We thought it would be helpful in understanding molecular mechanisms involved, to use a very well-definedtype of semisynthetic diet (AIN-76) (37) with the addition of 0,5, 10, and 20% of a very well-characterized fiber: AACC hardred wheat bran. The amount and composition of fecal SCFAwere carefully monitored and the level of histone acetylationwas analyzed for each individual rat. Furthermore, the proliferative index was determined for 50 crypt columns/rat.

MATERIAlS AND METHODS

Diet. The A!N-76 semisyntheticstandard diet (59.6% glucose;23.6% casein vitamin-free; 0.4% D,L-methlonrne;8% corn oil USP;0.4% choline bitartrate; 2% AIN-76 vitamin mix; 6% AIN-76 mineralmix, Dyets, Inc., Bethlehem, PA) was the basal diet for all studies (37).Note that glucose is substituting for starch, which may escape digestionin the small intestine and act like dietary fiber, and had to be eliminatedas a variable.The diet was used in its standard formula or was uniformlydiluted by the addition ofO, 5, 10, or 20% wheat bran (AACC certifiedhard red wheat bran, St. Paul, MN). The wheat bran supplement wasanalyzed by the AACC and contained 42.7% total dietary fiber (Association of Official Analytical Chemists Method 43.Al4). Diets werefreshly prepared every other day and stored at 4C.

Animals. In each diet study a group oftwenty 100-g male rats (5pmgue Dawley Charles River, Italy) was singly housed. The distance fromthe wire grid floor to the cage base was 10 cm to prevent rats fromreaching their feces. They received the diet for 2 weeks with food andwater ad libitum. During the study rats were weighed weeklyand monitored daily for food intake. Weight gain was calculated as the difference between initial and final body weight. Energy intake was evaluatedfrom the food intake by using physiological fuel values of4.0 KcaI/g for

5906

Modulation of Colonic Epithelial Cell Proliferation, Histone Acetylation,and Luminal Short Chain Fatty Acids by Variation of Dietary Fiber(Wheat Bran) in Rats'

Lidia C. Boffa,2 Joanne R. Lupton, Maria R. Mariani, Marcello Ceppi, Harold L Newmark,Alessandra Scalmati, and Martin Lipkin

Department of Chemical Carcinogenesis, Istitulo Nazionale per Ia Ricerca su! Cancro, 1ST, 16132 Genoa, Italy fL. C. B., M. R. M., M. C.]; Irving WeinsteinLaborato?y for Gastrointestinal Cancer Prevention, Gastroenterology and Nutrition Service, Memorial Sloan-Kettering Cancer Center, New York, New York 10021(A. S., H. L N., M. LI; and FacultyofNutrition, TexasA&M University,CollegeStation, Texas77843-2471(1. R. Li

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

MODULATION OF COLONIC EPITHELIAL CELLS BY DIETARY FIBER

protein and carbohydrate and 9.0 Kcal/g for lipids, while dietary fiberwas considered to provide 0 calories. Daily energy intake was calculatedby multiplying the energy density of each diet by the respective dailyfood intake. Food intake, energy intake, and body weight were determined on 20 rats/diet group. Data were analyzed by one wayanalysis ofvariance. Where variance was significant (P < 0.05), means were saparated by using Duncan's test (38). At the end of the experiments, ratswere sacrificed by cervical dislocation between 11 n.m. and 12 noon tominimize the effects of diurnal variation.

At the time of sacrifice the colon was quickly resected, openedlengthwise, and the feces were removed, frozen, and stored at —20°Cuntil analysis for short chain fatty acids. The colon was placed immediately into ice-cold isotonic saline. Histones from 12 rats/diet groupwere analyzed and cell proliferation indexes were assessed in the memaining 8 animals.

Cell Proliferation !ndexes. One h prior to sacrifice 8 animals/groupwere given injections of [3HJthymidine (Amersham International,United Kingdom; 2 Ci/mmol) 1 @@Ci/gof body weight. Sigmoid colonswere fixed in 10% buffered formalin acetate, pH 6.9—7.1,and stabilizedwith 1.5% methanol (39). Paraffin sections were processed in a blindedfashion for autoradiography and were developed and stained with hematoxylin and eosin (39). In glands longitudinally sectioned from baseto lumen, the total number of cells per “cryptcolumn―(each side of thelength of the crypt) and the number and position of labeled cells werescored (40). For each individual rat the labeling index is defined as thetotal number of [3H]dThd-labeledepithelial cells (in all the 50 cryptsassayed) divided by the total number ofcells counted in the same cryptcolumn. This index was calculated not only for the total crypt but alsofor each of its five proliferative compartments of equal size (compartment 1 at the base ofthe crypt and compartment 5 at the crypt surface)(40). The significance of any apparent change in cell number, labeledcells, and labeling index of a group of rats, with the variation of thepercentage of wheat bran supplement to the diet, was estimated byconventional analysis (using all the individual parameters in eachgroup) with an updated version of a program based on the MannWhitney test (41).

Analysis of Level of Histone Acetylation. Histones were preparedfrom colon epithelial cells of 12 rats ofeach group. Colons were rinsedin KC1 (25 mM), MgCl2 (5 mM), phenylmethylsulfonyl fluoride/1,2-epoxy-3-(p-nitrophenoxy)propane (0.1 mM), sodium butyrate (5 mM),Tris-HCI buffer (50 mM), pH 7.2. Epithelial cells were scraped from themucosal layer and placed in chilled buffer-sucrose(0.32 M).Nuclei wereprepared as previously described (42). Histones were extracted fromwashed nuclei with 0.25 MHC1,precipitated with 9 volumes of acetoneat 4°C,and freed from nonhistone proteins and nucleic acid fragmentsby ion-exchange chromatography on Bio-Rex 70 (Bio-Rad) as alreadydescribed (43). Electrophoretic separation of histone acetylated forms,mainly H4 and H3, was obtained on gels containing 15%polyacrylamide, 5.5% acetic acid, 8 Murea, and 0.2% Triton X-100 (43). Gels werestained with 0.25% Coomassie Brilliant Blue-R and analyzed directlywith a scanning laser spectrophotometer (LKB 2202 Ultrascan) connected to a recorder-integrator (LKB).The statistical significanceof thedifference between the amount of each acetylated forms of histone H4and its total level of acetylation in the different dietary groups wasevaluated in Table 5, using Duncan's means separation (38).

Analysis of Short Chain Fatty Acids. Rat focal samples were storedat —20Cuntil use. They were then placed in liquid nitrogen and groundto a fine powder in a chilled mortar and pestle. To approximately 0.3 gof powdered sample weighed accurately in a 1.5-mI microcentrifugetube, 500 @tlof 3 mM ethylbutyric acid (internal standard) in 70%ethanol were added. Each tube was then vortexed, stored briefly at 4°C,and subsequently shaken for 20 mm (44). Samples were then centrifuged for 20 mm at 11,500 x g at 4°Cand 100 @lof the resultingsupernatant were combined with 100 gdof 3 mr@iheptanoic acid in 70%ethanol (45). Immediately before injection in the gas chromatograph,20 @dof 10% H3P04 were added to the mixture and 1 Mlwas analyzed(46, 47). All solutions and samples were kept refrigerated during use.Chromatographic determinations were performed in a gas chromatograph Hewlett-Packard 5890 Series II, with flame ionization detector

equipped with a 30-m, 0.53-mm inside diameter, HP-FFA-PTA-TPAcapillary column (Hewlett-Packard) and a I-rn, 0.53-mm inside diameter, deactivated glass capillary precolumn (Supelco) (48). A HewlettPackard Series II Integrator was used to plot and integrate data. Thechromatographic conditions used were: injector at 165°C,detector at220°C,column flow, 2.21 ml/min helium, make-up flow, 28 mL/min N2,flame, 430 mi/mm air, 30 ml/min H2. Samples were concentrated onthe column head at 80°Cfor 30 s. The temperature was increased at70°C/mmto a temperature of 145C for 5 mm. A temperature gradientof 5°C/mmwas continued until the final temperature of 185°Cwasreached. After the analysis the column was allowed to regenerate at200°Cfor 5 mm and the temperature was reduced for the next run.Short chain fatty acid standards containing 2-ethylbutyric acid, heptanoic acid, and H3P04 were made fresh, stored at —70°C,and chromatographed at the beginning and end of daily sample runs. Blankscontaining only the internal standard solutions in 70% ethanol werealso chromatographed daily. Response factors and retention times forthe individual fatty acids weredetermined from the standard runs. Fattyacid composition of samples was calculated from the appropriate standard response factor. Data were obtained from a sample of 10 rats/dietary group. Statistical evaluation of the difference in amount ofSCFAs for each group was performed by using Duncan's mean separation.

Mathematical and Statisdcal Analysis of Data. Cellular labeling in

dexes were fitted in a logistic regression model according to the equation (49):

J@=exp(fl,x1)/1 +exp(Z$,x1)

where I@is the labeling index, x1are the covariates (specific group ofeach individual; the fecal butyric acid concentration) and @,the parameters, derived from the data, that indicate how I, is modified withdifferent values ofx1. Data analysis was performed with GLIM package(50). This model is the most suitable for the statistical analysis oflabeling index (as percentage) since it is based on a binomial distribution. It has been used in order to evaluate the diet, butyric acid, and theirconcurrent effect on I@.The importance of each variable is evaluatedaccording to its deviance, defined as the difference between the vanability not accounted for by the model in the presence and absence ofthe variable studied (Table 4). The deviance has approximatively thetrend of a x2 with degrees of freedom equal to the number of independent parameter taken into consideration during model fitting. The percentage of variability accounted for by the model and covariates considered was estimated with the pseudo R2 test (51).

In the study of a possible correlation between level of histone acetylation and fecal butyric acid concentration, the mean value of eachgroup was fitted in a linear regression model by using the acetylationlevel as the dependent variable. The significance (P) of the model wasalso evaluated (52).

RESULTS

Effect of Diet on Food and Energy Intake and Weight Gain.Table 1 summarizes results of food and energy intake andweight gain in groups ofrats ingesting 0, 5, 10, and 20% dietarywheat bran. At all fiber supplementation levels, rats consumeda greater weight of food than when wheat bran was not added tothe diet. This was presumably due to the diet energy dilutionwith fiber supplementation. Energy densities of the diet were:4.05 kcal/g, 0%; 3.93 kcal/g, 5%; 3.80 Kcal/g, 10%; 3.57Kcal/g, 20%. When food energy values are calculated, allgroups of rats excluding the one with highest level of fibersupplementation, consumed equivalent amounts of energy.This was true in both the first and second week. Initial bodyweights were similar across groups. Body weights at the end ofthe first and second week and weight gains are in the same rankorder as energy intakes with 5 > 0 > 10 > 20% wheat bran after1week. After 2 weeks there was no difference between the 0 and

5907

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Table I Effectofdiet onfood and energyintake and rat bodyweightValuesare expressed as mean ±SEM on a sample size of 20rats/group.Means

within a rownot sharing the samesuperscriptare significantlydifferentatP

Table 3 Mean amount, in p.w/gwetweight,ofluminal SCFAin groupsoflO rats maintained with AIN-76 diet supplemented with differentamountsofwheatbranResults

are means ±SEM on a sample size of 10 rats/group. Values ineachrownot sharing a common superscript are significantlydifferent by onewayanalysis

of variance by using Duncan's means separation (P < 0.05)(52).Wheat

bran0%

5% 10%20%AcidAcetic

22.20 ±2.6― 30.30 ±2.9― 28.90 ±1.4― 82.70 ±7.6°Propionic5.30 ±2.6― 10.20 ±0.8° 9.60 ±1.1° 7.70 ±@Isobutyric0.07 ±0.05c 0.99 ±0.09 0.67 ±0.08k' 0.26 ±0.09cButyric1.15 ±0.lSc 7.28 ±0.86°-―9.53 ±l.18a 6.23 ±0.81―Isovalenc0.22±0.l0'@ l.34±0.ID 0.86±0.12―1.89±1.65°―Valeric0.13 ±0.07c 1.17 ±0.12° 0.76 ±0.09― 0.16 ±O.06cTotal

29.10 ±3.2c 5131 ±39b 50.36 ±2.5― 98.93 ±7.1°

Table 4 Evaluationofeffects ofeach covariateon labeling indexesas calculatedby the model that takes into account the 2 covariates and theirinteractionCovariate

flu SEPGroup

1 (O%)@ —2.04700 0.06033 20%) were hyperacetylated.

In the study ofa possible correlation between level of histoneacetylation in colon epithelial cells and fecal butyric acid concentration, it was found that the mean value of each group(Table 5) can be fitted in a regression model where acetylationlevel is the dependent variable. The slope of the linear relationwas 10.04 with SE of 4.52 and P < 0.05; the correlation therefore is significant and positive.

DISCUSSION

In the literature there are several experimental (10—17)andepidemiological (7—9)reports on the influence of fiber (amountand composition) in the diet on colon epithelial cell metabolismand carcinogenesis. In many cases the results are conflicting orcontradictory.

In this paper a clarification of some of the issues was attempted by using a well described model system: inbred rats,standardized housing conditions, well defined semisynthetic diets supplemented with different percentages of thoroughlycharacterized fiber, and the experiments were ended at the sametime of the day.

In previous experiments we have seen that when diets areuniformly diluted by the addition of the fiber supplement therats merely consumed a greater weight of diet in order to compensate. Since diets are designed on an energy basis, if rats haveequivalent energy intakes they will have equivalent intakes of allnutrients except fiber.

In this study the highest fiber group (20%) did not eat asufficiently greater weight of food to entirely compensate forthe energy dilution (Table 1). Thus this group of rats had alower energy intake than the other three groups; this was reflected also in a lower weight gain (Table 1). Although this is acomplicating issue in our study, we do not feel that it shouldaffect the interpretation of our results. In fact, in almost everydiet-large bowel cancer study (53, 54), the group ofanimals thatgains the least weight would develop the lowest number oftumors. In our case, however, the “marker―for cancer risk, anincrease in cell proliferation, was highest not lowest in thisgroup. It should also be noted that a 20% wheat bran-supplemented diet is not really a 20% fiber-supplemented diet butrather an 8.5% fiber diet (since wheat bran is only 42.7% dietary

acetate followed the same pattern as total SCFA concentrations; however, the 5 and 10% levels ofsupplementation did notproduce acetate values that were significantly higher than controls. In contrast, levels of propionate, isobutyrate, and valerateincreased over control values with 5 and 10% but returned tocontrol level with 20% wheat bran supplementation in the diet.There was no significant diet effect on levels of isovalerate.

Butyrate was the SCFA ofinterest because ofits known effecton histone deacetylase (30). Its concentration increased in thecolonic lumen with increasing amounts ofwheat bran up to the10% supplementation level, but the butyrate concentration ofthe 20% fiber dietary group was back to the 5% level (Table 3).

Mathematical Analysis of Data. On the basis of the logisticregression model we have established that all the covariatesavailable have a significant effect on I(labeling index), althoughthe percentage of variability they account for when evaluatedwith the pseudo R2 test, was found to be only 47%.

The evaluation of the effect of each covariate on the labelingindex I, is analyzed in Table 4 through the (fl@)in the equationgiven in “Materialsand Methods.― All groups of rats with adifferent diet determine a highly significant (P < 0.001) andinversely proportional effect to the labeling index (e.g., —2.04700). Butyric acid too has an inversely proportional effectover I, but with a lower significance. The concurrent effect ofthe two variables though, is in part modulated by a factor ofinteraction between the different groups of rats and butyricacid. This interaction has a directly proportional effect on thedefinition off,, with a significance that varies within the groups(from significant P < 0.05 to nonsignificant P = 0.8) (Table 4).

Effect of Diet on Histone H4 Acetylatlon. Total purified histone preparations from colon epithelial cells of rats maintamed with 0, 5, 10, and 20% wheat bran-supplemented diet,and a colon carcinoma-derived cell line HT29, were separatedinto their components by gel electrophoresis according to theirlevel of acetylation. Histone H4 acetylated forms can be easily

5909

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Table 5 Hisione H4 acetylation in rat colon epithelial cells: correlation with percentage ofwheat bran in the dietResults are means ±SEM on a sample size of 12 rats/group. Values in each column not sharing a common superscript are significantly different

of variance by using Duncan's means separation (P < 0.05) (52).byone wayanalysisAcetyl

groups―%

ofwheat bran 0 1 2 34Totale0

8.6 ±0.8° ND173.8 ±0.8°10.0±0.7c 8.0 ±0.7―137.4 ±0.8―8.5±0.7°

11.8±0.7°144.5±0.9°@6.5±0.7― 12.1±0.8°132.5 ±0.9°'6.4±0.8'@ 4.8 ±0.9'@95.1 ±0.8°

Table 6 Correlationbetweencellproliferation,histoneacetylation,andbutyricacidconcentration in colonic epithelial cells ofrats maintained on dietswithdifferenifiber

supplements%a

jb Acetylation―BAd0

0.130±0.013 73.8±0.81.15±0.1550.066 ±0.005 137.4 ±0.8 7.28 ±0.86100.057 ±0.004 144.5±0.9 9.53 ±1.18200.084 ±0.006 132.5 ±0.9 6.23 ±0.81a

Percentageofwheat bransupplementedto the AIN-76 basicdiet ±SE.b Labeling index of total colon epithelial cells in a homogeneous group ofrats(number

of[3Hjthymidine-labeledcells/totalnumberofcells scored)±SE.CGlobal level of histone H4 acetylation ± SE.

d Luminal BA (butyric acid) concentration, in @a.i/gwet weight, in groupof10rats maintained on diet supplemented with different amount of fiber(seefootnote

a).

MODULATIONOF COLONICEPITHELIALCELLSBY DIETARYFIBER

44.2 ±0.9― 46.4 ±0.9― 0.8 ±0.7cS 23.9±0.8° 40.4±1.0° 17.7±0.8°

10 24.0 ±0.8° 39.5 ±0.9° 16.2 ±0.8°20 28.2±0.9@ 4l.8±0.9c 11.4±0.9―HT29 42.1 ±1.0°' 36.7 ±0.9@' 10.0±0.8b

d Amount of non-, mono-, di-, ui-, tetraacetylated forms as percentage of total histone H4 (average of 4 determinations).e Global acetylation level of histone H4 = 1 x % mono- + 2 x % di- + 3 x % tn- + 4 x % tetraacetylated form ±SE.1Not detectable or within SE.

fiber). The 20% wheat bran-supplemented diet would correspond to a human diet of approximately 43 g/day. This isdefinitely a high fiber diet but not “nonphysiological―nor “inordinately high.―

Before describing an overall evaluation of the data we wouldlike to point out two peculiar effects of the fiber-free diet. Theabsence of fiber seems to cause a marked hyperplastic effect oncolonic crypt epithelial cells (Table 2). This type of effect hasbeen previously achieved, in this model system, only with anutritional stress diet (55). Furthermore, even in the absence ofknown precursors of fermentable material in the diet, luminalbutyric acid concentration was relatively high (1. 15 @LM/gwetweight; Table 3), leaving its origin so far undetermined.

When we analyzed the cellular parameters of the groups ofrats in our studies, we noted that variation in results did notcorrelate linearly with the percentage of wheat bran in the diet.In particular, an increase in the percentage of wheat bran from0 to 10% led to a progressive decrease of [3H]dThd-labeledcrypt column cells and labeling index, but a 20% supplementcauses an inversion of the trend (Table 2).

Butyric acid is the only one of the SCFAs produced by fiberfermentation, whose concentration has an effect that could account, at least in part, for the above noted labeling indexes(Tables 3 and 4). Since a mathematical evaluation of an appropriately chosen logistic regression model showed that the interpolation of the covariate analyzed (Table 4) can justify only47% of the experimental distribution of labeling indexes, thismeans that other covariates are involved in the definition of theevent. Although these covariates are necessary for the accuratedefinition of the model, it is unlikely that a single one of themhas a determinant role as a modulator of cellular proliferation.

Butyric acid is known as an inducer of terminal differentiation of cultured cells of neoplastic origin (18, 19). Among theeffects produced it causes a partial block of cells in early G1,with consequent inhibition of DNA synthesis. An inhibition ofthis type could be involved in modulating the labeling index ofcolonic crypt cells of different groups of rats (Table 6). The

correlation coefficient between labeling index (1) and luminalbutyrate concentration is strongly negative (—0.98; P < 0.01). Itshould be noted though that an opposite effect has been previously observed at very low butyric acid concentrations (in therange of 0.1 mM) in some normal primary cells in culture (56).

Cell accumulation (hyperplasia) in the colon of the fiber-freediet group of rats could not be accounted for simply by theabsence ofthe brush-offeffect ofundigested fibrous material. Infact, hyperplasia appears, even if to a lesser extent, also in thegroup with the highest wheat bran supplementation (20%). Thiseffect could be accounted for to some extent by the difference ofcrypt cell proliferation in the different groups (Table 2).

The state of morphological differentiation of colonic epithehal cells was not studied extensively, but no change was observed: this type of evaluation might show variations in differentiated cell distribution throughout the 5 crypt compartmentswith different proliferative capacity (Table 2).

Butyric acid is also a modulator of chromatin primary (nucleosome) structure because of its ability to inhibit histonedeacetylation, therefore loosening the interaction between theDNA and these structural proteins. In colonic epithelial cells ofdifferent groups ofrats the level ofhistone acetylation appearedto be modulated by the cellular concentration of butyric acidthat is in equilibrium with its fecal concentration (Table 6) (44).The correlation coefficient between the level of histone acetylation and butyric acid concentration in each group is positive(0.96; P < 0.05), while the labeling index (1) is negative (—0.97;P < 0.05) (Table 6). This means that butyric acid concentrationcontrols to some extent not only colonic epithelial cell primarystructure (and therefore its accessibility to carcinogen damageand damage repair), but also cell proliferation.

In the study of a possible correlation between luminal butyricacid concentration and histone acetylation, the mean value ofthis last parameter was conventionally evaluated for eachgroup. In fact a fairly large number of rats is needed to obtainand analyze an electrophoretic separation ofthe different acetylated forms. Consequently there is an averaging effect of thedata and the level ofacetylation cannot be utilized as a covariatein Table 4.

A complicating factor in the interpretation of the results arethe data for the 20% wheat bran diet group. There is no satisfactory explanation for the increase in cell number per crypt,labeled cells, and labeling index (corresponding to a lower concentration ofluminal butyric acid and consequent lower degreeof histone acetylation), with an increase of the wheat bran supplement from 10 to 20%. However, several parameters shouldbe taken into account: the variation in concentration of luminalbutyric acid itself and the decreased transit time through theintestine, with consequent shorter exposure of fecal material tointestinal flora and epithelial cells to the fermentation products.

5910

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

MODULATION OF COLONIC EPITHELIAL CELLS BY DIETARY FIBER

These findings have further implications for both rodent andhuman studies. They can also explain some data in the literature, correlating amount of sodium butyrate to 1,2-dimethylhydrazine-induced rat intestinal neoplasia (57).

Thus the 5% wheat bran diet decreased both colonic epithehal cell hyperproliferation and hyperplasia, compared with thefiber-free diet. However, when wheat bran content was furtherenriched to 10 and 20%, increased hyperplasia and hyperproliferation gradually recurred. These findings may help to explain previous similar data in rodent models (10, 58—60).Theyalso substantiate the hypothesis that complexities may arise inclinical human fiber trials (61) in which subjects are expected tobegin clinical trials consuming different amounts of fiber,thereby contributing to lower, higher, or unchanged proliferative indices which are expected to be more uniform. Unfortunately, differences of this type have sometimes tended to belumped together in the past.

From the results obtained in this study we postulate that,while diets containing moderate amounts of fiber may haveprotective effects on cell proliferation, differentiation, and carcinogenesis, fiber-free diets and diets supplemented with toomuch fiber would have the potential to promote colon carcinogenesis.

REFERENCES

1. Leblond, C. P., and Stevens, C. E. The constant renewal of the intestinalepithelium in the albino rat. AnaL Rec., 100: 357—377,1948.

2. Lipkin, M., Bell, B., and Sherlock, P. Cell proliferation kinetics in the gastrointestinal tract ofman. I. Cell renewal in colon and rectum. J. Clin. Invest.,42: 767—776,1963.

3. Lipkin, M., and Quastler, H. Cell population kinetics in the colon of themouse. J. Clin. Invest., 41: 141—146,1962.

4. Lipkin, M., Blattner, W. A., Fraumeni, J. F., Jr., Lynch, H. T., Deschner, E.E., and Winawar, S. Tntiated thymidine labelingdistribution in the identification ofhereditary predisposition to colon cancer. Cancer Res., 43: 1899-1904, 1983.

5. Lipkin, M., Blattner, W. A., Gardner, E. J., Burt, R. W., Lynch, H. T.,Deschner, E. E., Winawar, S., and Fraumeni, J. F., Jr. Classification and riskassessment of individuals with familial polyposis, Gardner syndrome andfamilial non-polyposis colon cancer from (3H@dThd-labeling patterns in coIonic epithelial cells. Cancer Res., 44: 4201—4207,1984.

6. Lipkin,M. Biomarkersof increasedsusceptibilityto gastrointestinalcancer:new application to studies of cancer prevention in human subjects. CancerRes.,48:235—245,1988.

7. Heilbrun, L. K., Nomura, A., Hankin, J. H., and Stemmermann, G. N. Dietand colorectalcancerwith specialreferenceto fiberintake. Int. J. Cancer,44:1—6,1989.

8. Freudenheim, J. L., Graham, S., Horvath, P. J., Marshall, J. R., Haughey, B.P., and Wilkinson, G. Risks associated with source of fiber and fiber components in cancer of the colon and rectum. Cancer Res., 50: 3295—3300,1990.

9. Trock, B., Lanza, E., and Greenwald,P. Dietary fiber,vegetables,and coloncancer: critical review and meta-analyses of the epidemiologic evidence. J.Natl. Cancer Inst., 82: 650—661,1990.

10. Jacobs, L. R. Enhancement of rat colon carcinogenesis by wheat bran consumption during the stage of 1,2-dimethylhydrazine administration. CancerRes., 43: 4057—4061,1983.

11. Robblee,N. M., McLellan,E. A., and Bird, R. P. Measurementof theproliferativestatusofcolonicepitheliumas a riskmarkerforcoloncarcinogenesis: effects of bile acid and dietary fiber. Nutr. Cancer, 12: 301—310,1989.

12. Reddy, B., Eagle, A., Katsifis, S., Simi, B., Bartmam, H. P., Pemno, P., andMahan, C. Biochemicalepidemiology of colon cancer effect of types ofdietary fiber on fecal mutagens, acid, and neutral sterols in healthy subjects.Cancer Res., 49: 4629—4635,1989.

13. Heitman, W. D., Ord, A. V., Hunter, K. E., and Cameron, I. L Effect ofdietary celluloseon cell proliferation and progressionof l,2-dimethylhydrazinc-induced colon carcinogenesis in rats. Cancer Rca., 49: 5581—5585,1989.

14. Alberts, D. S., Einspahr, J., Rees-McGee, S., Ramanujam, P., Buller, M. K.,Clark, L., Ritenbaugh, C., AtWOOd,J., Pethigal, P., Earnest, D., Villar, H.,Phelps, J., Lipkin, M., Wargovich,M., and Meyskens,F. L., Jr. Effectsofdietary wheat bran fiber on rectal epithelial cell proliferation in patients withresection for colorectalcancers.J. Natl. Cancer Inst., 82: 1280—1285,1990.

IS. Nyman, M., and Asp, N. G. Fermentation of dietaryfibrecomponents in therat intestinal tract. Br. J. Nutr., 47: 357—366,1982.

16. Cummings, J. H. Short chain fatty acids in the human colon. Gut, 22:763—779,1981.

17. Sakata, T. Effects of indigestible dietary bulk and short chain fatty acids onthe tissue weight and epithelial cell proliferation rate of the digestive tract inrats. J. Nutr. Sci. Vitaminol., 32: 355—362,1986.

18. Prasad, K.N. Butyricacid:a small fattyacidwithdiversebiologicalfunctions.LifeSd.,27:1351—1358,1980.

19. Kruh,J. Effectsofsodium butyrate,a newpharmacologicalagent, on cells inculture. Mol. Cell. Biochem., 42: 65—82,1982.

20. Demigne, C., Remesy, C., and Rayssiguier, Y. Effects of fermentable carbohydrates on volatile fatty acids, ammonia and mineral absorption in the ratcaecum. Reprod. Nutr. Dcv., 20: 1351—1359,1980.

21. Demigne,C., and Remesy,C. Stimulationofabsorption ofvolatile fatty acidsand minerals in the cecum of rats adapted to a very high fiber diet. J. Nutr.,115: 53—60,1985.

22. Jacobsson, K. G., Riesenlfeld, J., and Lindahl, U. Biosynthesis of heparin:effects of n-butyrate on cultured mast cells. J. Biol. Chem., 260: 12154—12159,1985.

23. Philippe, J., Drucker, D. J., Chick, W. L., and Habener, J. F. Transcriptionalregulation of genes encoding insulin, glucagon,and angiotensinogenby sodium butyrate in a rat islet cell line. Mol. Cell. Biol., 7: 560—563,1987.

24. Griffin, M. J., Price, G. H., Bazzel, K. L., Cox, R. P., and Ghosh, N. K. Astudy ofadenosine 3'-5'-cyclic monophosphate, sodium butyrate and cortisolas inducers of HeLa alkaline phosphatase. Arch. Biochem. Biophys., 164:619—623,1974.

25. Leder, A., and Leder, P. Butyric acid, a potent inducer oferythroid differentiation in cultured erythroleukemic cells. Cell, 5: 319—322,1975.

26. Xue, S., and Rao, P. N. Sodium butyrate blocks HeLa cells preferentially inearly G1 phase ofthe cell cycle. J. Cell. Sci., 51: 163—171,1981.

27. Dexter, D. L, Crabtree, G. W., Stoeckler, J. D., Savarese, T. M., Ghoda, L.Y., Rogler-Brown, T. L., Parks, R. E., Jr., and Calabresi, I. P. N,N-Dimethylformamide and sodium butyrate modulation of the activities of purinemetabolizing enzymes in cultured human colon carcinoma cells. Cancer Res.,41:808—812,1981.

28. Tsao, D., Monta, A., Bella, A., Jr., Luu, P., and Kim, Y. S. Differentialeffects of sodium butyrate, dimethyl sulfoxide, and retinoic acid on membrane-associated antigen, enzymes, and glycoproteins of human rectal adenocarcinoma cells. Cancer Res., 42: 1052—1058,1982.

29. Kelly, R. E. Sodium butyrate induced alterations in lysosomal enzyme activity. In Vitro Cell. Dcv. Biol., 21: 373—381,1985.

30. Boffa, L. C., Vidali, G., Mann, R. S., and Allfrey, V. G. Suppression ofhistone deacetylation in vivo and in vitro by sodium butyrate. J. Biol. Chem.,253: 3364—3366,1978.

31. Bode, J., Gomez-Lira, M., and Schroter, H. Nucleosomal particles open asthe histone core becomes hyperacetylated. Eur. J. Biochem., 130: 437—445,1983.

32. Ausio, J., and van Holde, K. E. Histone hyperacetylation: its effects onnucleosome conformation and stability. Biochemistry, 25: 1421—1428,1986.

33. Hebbes, T. R., Thorne, A. W., and Crane-Robinson, C. A direct link betweencore histone acetylation and transcriptionally active chromatin. EMBO J., 7:1395—1402,1988.

34. Boffa, L. C., Walker, J., Chen, T. A., Sterner, R., Mariani, M. R., andAllfrey, V. G. Factors affecting nucleosome structure in transcriptionallyactive chromatin: histone acetylation, nascent RNA and inhibitors of RNAsynthesis. Eur. J. Biochem., 194: 811—823,1990.

35. Bohr, V. A., Phillips, D. H., and Hanawalt, P. C. Heterogeneous DNAdamage and repair in the mammalian genome. Cancer Res., 47: 6426—6436,1987.

36. Ramanathan, B., and Smerdon, M. J. Enhanced DNA repair synthesis inhyperacetylated nucleosomes. J. Biol. Chem., 264: 11026—11034,1989.

37. Report of the American Institute of Nutrition ad Hoc Committee on Standards for Nutritional Studies. J. Nutr., 107: 1340—1348,1977.

38. Duncan, D. B. Multiple range and multiple-F. Biometrics, 11: 1—42,1955.39. Manual of Histological and Special Staining Technics. Ed. 2, pp. 1—2.New

York: McGrawHill Book Co., 1960.40. Lipkin, M., Enker, W. E., and Winauer, S. J. Tritiated-thymidine labeling of

rectal epithelial cells in “non-prep―biopsies of individuals at increased riskfor colonic neoplasia. Cancer Lett., 37: 153—161,1987.

41. Norusis, M. J. Statistical Package for Social Sciences, Chicago, IL: SSPSInc., 1990.

42. Boffa, L. C., Vidali, G., and Allfrey, V. G. Changes in nuclear non-histoneprotein composition during normal differentiation and carcinogenesis of intestinal epithelial cells. Exp. Cell Res., 98: 396—410,1976.

43. Allegra,P., Sterner, R., Clayton, D. F., and AlIfrey,V. G. Affinitychromatographic purification of nucleosomes containing transcriptionally active DNAsequences. J. Mol. Biol., 196: 379—388,1987.

44. Remesy, C., and Demigne, C. Partition and absorption of volatile fatty acidsin the alimentarycanal of the rat. Ann. Rech. Vet., 7: 39—43,1976.

45. Brighenti, F., Testolin, G., Canzi, E., Ferrari, A., Wolever, T., Ciappellano,S., Porrini, M., and Simonetti, P. Influence oflong term feeding of differentpurified dietary fibers on the volatile fatty acid profile, pH and fiber.degrading activity of the cecal contents of rats. Nutr. Res., 9: 761—765,1989.

46. Traitler, H., Fleming, S. E., and Koellireuter, B. Analysis of volatile fattyacids in biological specimens by capillary gas chromatography. J. High Res@olut. Chromatogr. Chromatogr. Commun., 11: 124—129,1988.

47. Remesy,C., and Demigne,C. Determinationof volatilefatty acids in plasma

5911

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

MODULATIONOF COLONICEPITHELIALCELLS BY DIETARY FIBER

after ethanolic extraction. Biochem.J., 141: 185—190,1974.48. Quigley, J. D., III, Smith, Z. P., and Heitmann, R. N. Changes in plasma

volatile fatty acids in response to weaning and feed intake in young calves. J.Dairy Sci., 74: 258—263,1991.

49. Breslow, N. E., and Day, N. E. Statistical methods in cancer research. In.IARC Scientific Publications No. 32, Vol. I. Lyon, France: InternationalAgency for Research on Cancer, 1980.

50. Baker, R. J., and Nelder, J. A. in. Royal Statistical Society GeneralizedLinear Interactive Modeling (GLIM), Release 3.7. Boston, MA: NumericalAlgorithmus Group, 1978.

51. Nelder, J. A., and Wedderborn, R. W. M. Generalized linear models. J. R.Stat. Soc. (A), 135: 370—384,1972.

52. Kleinbaum, D. G., Kupper, L. L., and Muller, K. E. In: The University ofNorth Carolina, Chapel Hill, Applied Regression Analysis and Other Multivariable Methods. Chapel Hill, NC: PWS-Kent Publishing Company, 1988.

53. Klurfeld, D. M., Weber, M. M., and Kritchevsky, D. Inhibition of chemicallyinduced mammary and colon tumor promotion by caloric restriction in ratsfed increased dietary fat. Cancer Res., 47: 2759—2762,1987.

54. Albanese, D. Caloric intake, body weight, and cancer: a review. Nutr. Cancer,9:199—217,1987.

55. Newmark,H. L., Lipkin, M., and Maheshwari,N. Colonic hyperplasiaand

hyperproliferation induced by a nutritional stress diet with four componentsof western-style diet. J. Nati. Cancer Inst., 82: 491—496,1990.

56. Carpaneto, E., Mariani, M. R., Femni, S., Boffa, L. C., Lipkin, M., andNewmark, H. Stimulation of DNA synthesis in primary culture of humanlymphocytes by sodium butyrate. Proc. Am. Assoc. Cancer Rca., 32: 331,1991.

57. Freeman, H. J. Effects of differing concentrations of sodium butyrate onl,2-dimethylhydrazine-induced rat intestinal neoplasia. Gastroenterology,91: 596—602,1986.

58. Jacobs, L. R., and Lupton, J. R. Effect of dietary fibers on rat large bowelmucosal growth and cell proliferation. Am. J. Physiol., 346: G378—G385,1984.

59. Jacobs, L. R., and Lupton, J. R. Relationship between colonic luminal pH,cell proliferation,and colon carcinogenesisin l,2-dimethylhydrazinetreatedrats fed high fiber diets. Cancer Res., 46: 1727—1734,1986.

60. Lupton, J. R., Coder, D. M., and Jacobs, L. R. Long term effectsof fermentablefiberson rat colonicpH and epithelialcellcycle.J. Nutr., 118:840—845,1988.

61. Schatzkin, A., Freedman, L. S., Schiffman,M. H., and Dawsey,S. M. Validation of intermediate end points in cancer research. J. NatI. Cancer Inst.,82:1746—1752,1990.

5912

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

1992;52:5906-5912. Cancer Res Lidia C. Boffa, Joanne R. Lupton, Maria R. Mariani, et al. of Dietary Fiber (Wheat Bran) in RatsAcetylation, and Luminal Short Chain Fatty Acids by Variation Modulation of Colonic Epithelial Cell Proliferation, Histone

Updated version

http://cancerres.aacrjournals.org/content/52/21/5906

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

.pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/52/21/5906To request permission to re-use all or part of this article, use this link

on June 8, 2021. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/content/52/21/5906http://cancerres.aacrjournals.org/cgi/alertsmailto:pubs@aacr.orghttp://cancerres.aacrjournals.org/content/52/21/5906http://cancerres.aacrjournals.org/