DifferentialEffects of SodiumButyrate,DimethylSulfoxide,andRetinoicAcid on Membrane-associatedAntigen, Enzymes,andGlycoproteinsof HumanRectalAdenocarcinomaCells1
D. Tsao, A. Morita, A. Bella, Jr., P. Luu, and V. S. Kim2
Gastrointestinal Research Unit, Veterans Administration Medical Center, and Department of Medicine, University of California School of Medicine,San Francisco, California 94121
and purine analogs (37), antimetabolites (42) and vitamin A(27, 38). Different inducers of differentiation may not act bythe same mechanisms and may induce different phenotypicexpression (14, 35, 17). For example, variant clones of murineerythroleukemic cells which are resistant to one inducer maydifferentiate in the presence of another inducer. The ratio of @6minor globin chains in these cells may vary depending oninducer (35).
We have reported previously that sodium butyrate andDM503 had considerable effects on the biochemical propertiesof poarly differentiated human colon cancer cells and thatthese effects were reversible (20). The studies of the effects of3 putative differentiating agents (sodium butyrate, DMSO, andretinoic acid) on human colonic cancer cells have been extended to several other tumor cell lines. In the present study,the effects of the 3 differentiating agents on cellular morpholagy, growth rate, cell surface glycoproteins, and several membrane-associated enzymes are compared using a human rectalcarcinoma cell line, HAT-i 8, which is poorly differentiated.
MATERIALS AND METHODS
Treatmentof Cell Cultureswith Inducers.A humanrectalcancerline,HRT-i8, developedfroma patientwith primaryadenocarcinomaof rectum was kindly given to us by Dr. W. A. F. Tompkins (44). Thecell line was maintained as monolayer cultures in Dulbecco's modifiedEagle'smediumsupplementedwith 5% fetal bovineserum,penicillin,(100 units/mI), and streptomycin (100 @tg/ml).The treatment of cellswith sodium butyrate, DMSO, or retinoic acid was carried out byculturingthemin the aforementionedmediumsupplementedwith sodium butyrate (2 mM or 5 mM), DMSO (2% v/v), or retinoic acid (3.3x 10-@M), and the growth mediumwas changed every fourth dayunless otherwise indicated. The concentration of each inducer usedhad no toxic effects on these cells as tested by trypan blue dyeexclusionstudies.
Cell GrowthStudies.Forcell growthstudies,35-mmplastictissueculture dishes (Falcon No. 3001) were used. Cells were plated inquadruplicateat 1.2 x 1O@cells/dish and were harvestedat 1-dayintervals using trypsin. The number of cells was determined using aCoulterCounter.Resultswereplottedon semilogarithmicpaper,anddoubling times were calculated from the curves. Saturation densitieswereobtainedfromthe sameexperimentsby determiningthe numberof cells in the plateau regions of the growth curves which represent themaximumcell numberper 35-mmtissueculturedish.Growthin semisolidagar(DifcoLaboratories,Detroit,Mich.)mediumwasperformedby a slightlymodifiedmethodof DeLarcoandTodaro(8) asdescribedpreviously(20). Colony-formingefficienciesin soft agar were determined by counting colonies containing 20 or more cells 2 weeks afterinoculation.
3 The abbreviations used are: DMSO, dimethylsulfoxide; CEA, carcinoembry
onic antigen; PBS, phosphate-buffered saline, 0.01 M NaH2PO4-Na2HPO4/0.15M NaCI, pH 7.4; SDS, sodium dodecyl sulfate; Con A, concanavalin A: RCA,Ricinus communis agglutinin; WGA, wheat germ agglutinin.
1052 CANCERRESEARCHVOL. 42
ABSTRACT
The effects of sodium butyrate, dimethyl sulfoxide (DMSO),and retinoic acid on the growth, morphology, carcinoembryonicantigen content, cell surface membrane-associated enzymeactivities, and glycoprotein profiles of a human rectal adenocarcinoma cell line (HRT-18) in culture were compared. Allthree agents reversibly caused a marked increase in doublingtimes, a decrease in saturation densities, and a markedlyreduced colony-forming efficiency in soft agar. Only butyratecaused gross morphological changes including cell enlargement, flattening, and increased membranous process formation. Carcinoembryonic antigen content was increased duringculture in butyrate, while it was reduced by DMSO and unchanged by retinoic acid.
The activities of membrane-associated enzymes were alteredsignificantly in the butyrate-treated cells. For example, anincrease in the activities of alkaline phosphatase (10-fold), yglutamyl transpeptidase activity (3-fold) and sucrase activity(2-fold) was observed, while those of aminooligopeptidase andK@-stimulatedphosphatase actually showed slight decreases.DMSO- or retinoic acid-treated cells showed a marked decrease in alkaline phosphatase activity, but other enzyme activities remained unchanged. Surface protein-labeling patternsof lactoperoxidase-catalyzed iodinated HAT-i 8 cells showedno significant change from the control cells following treatmentwith DMSO or retinoic acid. The most prominent changecaused by butyrate treatment was the appearance of a majorglycoprotein band with an apparent molecular weight of60,000.
These data indicate that the use of butyrate, DMSO, andretinoic acid may provide useful information concerning theidentification of differentiation-associated markers of humanrectal cancer cells. Furthermore, these agents, although havingsimilar effects on the growth properties, have different effectson the morphology and on the biochemical properties of humanrectal cancer cells.
INTRODUCTION
Biochemical and morphological differentiation has been induced in some normal and cancer cell lines by a variety ofagents, including polar organic compounds (10, 11, 14, 15,21, 43), bisacetamides, short-chain fatty acids (23, 24), purine
I Supported by USPHS Grant CA-i 4905 from the National Cancer Institute
through the National Large Bowel Cancer Project and the VeteransAdministrationMedical Research Service.
2 To whom requests for reprints should be addressed at: Gastrointestinal
Research Unit, Veterans Administration Medical Center, 4150 Clement Street,San Francisco, Calif. 94121.
Received May 22, 1981 ; accepted December 2, 1981.
Growth propertiesof a human rectal adenocarcinoma ceII line(HRT-18)Conditions8Doubling
time(hr)Saturation
density(x 1o-@cells)Colony-forming
efficiency in softagar(%)Control
SodiumbutyrateDMSORetinoic acid22
776739.50.0
i2.48.8
42.089
00
38
Differentiating Agents on Colorecta! Cancer Ce!!s
Preparation of Membrane and Cytoplasmic Fractions. For thedeterminationof CEA and membrane-associatedenzymeactivities,cells cultured in 60-mm plastic tissue culture dishes were used. Thecells were scraped after being washed 4 times and suspended in PBS.Thecellsuspensionwassonicatedandcentrifugedat 100,000 x g for1 hr to obtain the membrane pellet and the cytoplasmic fraction.
CEA Assay. CEA activity was determinedby the direct zirconylphosphate gel assay method as suggested by Hoffman-La Roche Inc.
Enzymeend ProteinAssays.All enzymeassayswerecarriedoutat37° and under conditions in which the reactions were linear withrespect to enzymeconcentrationsand time. Sucrase activity wasmeasuredby themethodof Dahlqvist(7).TheL-aminopeptidaseassaywas carried out as described previously (32) using phenylalanylglycineas a substrate. K@-stimuIatedphosphatasewas estimated according tothe methodof Murer et a!. (31). Alkalinephosphataseactivity wasmeasuredusingp-nitrophenylphosphateassubstrate.Theincubationmixture consisted of 0.5 M NaHCO3buffer, pH 10.0, and 28 mMsubstratein a final volumeof 0.5 ml.The reactionwasterminatedbythe addition of 1 ml 1 N NaOH, and the extent of hydrolysis wasdeterminedspectrophotometricallyat 410 nm.‘y-Glutamyltranspeptidaseactivitywas measuredusing‘y-glutamylnitroanilideas substrate.The reaction volume was 0.8 ml containing 50 mM Tris-HCI, pH 8.4,and3 mMsubstratewithor without40 mMdiglycine.Thereactionwasstopped by the addition of 0.8 ml 3 M acetic acid, and the productswere measuredspectrophotometrically at 470 nm. Enzymeactivity wasestimated by calculating the difference between the duplicate samplesincubated in the presence or absence of diglycine. Protein was estimatedbythemethodof Lowryeta!. (28)usingcrystallinebovineserumalbuminas standard.
Surface Labeling of Cells. Labeling of the surface proteins of HRTi 8 was carried out by lactoperoxidase (L 8257; Sigma Chemical Co.,St. Louis, Mo.)- catalyzed iodinationmodifiedfrom a previouslydescribed method (i 9, 45). Cells cultured in 75-sq cm tissue cultureflasks were labeled in situ after being washed4 times with PBS.Lactoperoxidase-catalyzed iodination of cell surface proteins was carnedoutbyadding4 mlPBScontaining3 mt@iglucoseat pH7.4, 0.1 mlof Na'261(1 mCi/mi), and 0.03 ml of lactoperoxidase(5.5 units)to aflask of washedcells.Thereactionwasinitiatedby the additionof 10@Ll(12.5 units) glucose oxidase (type V; Sigma) and was allowed tocontinue for 15 mm at 25°.The labeled cells were washed 4 times withPBS and were scraped from the flask in 1 ml PBS. The cells weresonicated twice for 15 sec, and aliquots of the cell homogenate wereusedfor SDS-gelelectrophoresis.
SDS-GelElctrophoresls. SDS-polyacrylamide slab gel electrophoresisin a discontinuousbuffersystemwascarriedoutaccordingto themethod of Laemmli (22) with a vertical slab gel apparatus (Bio-RadLaboratories,Richmond,Calif.).Sampleswere dissolvedin 0.021 MTris buffer, pH 6.8, containing 2% SOS, 5% 2-mercaptoethanol,0.001% bromophenolblue,and 10% glycerol.Justbeforeapplicationto the gel, sampleswereheatedat 100°for 3 mm.Gelthicknesswas1.5 mm, and the current appliedwas 50 ma/gel. Gels were stainedwith 000massiebrilliantblue for proteinsaccordingto the methodofFairbankset a!. (i 2). High-molecular-weightproteinstandards(BioRadLaboratories)for SDS-gelelectrophoresisare myosin,/1-galacto
@ sidase,phosphorylaseB, bovineserumalbumin,andovalbumin.Autoradiography. Gels containing ‘251-labeledproteins were dried
and exposed to X-ray fIlm (Kodak X-Omat A film).Lectin Affinity Chromatography.ConA-Sepharosewaspurchased
from Pharmacia(Uppsala,Sweden).RCA was purified from castorbeans according to a modification of a method of Nicolson and Blaustein (33), but no further separation of the lectin with a molecularweight of 120,000 (RCA1) from that with a molecular weight of 60,000(RCA2)was attempted.WGA was purifiedas describedby the methodof Bassett(2). Both purified lectinswere coupledto CNBr-activatedSepharose as described by March et a!. (29). 1251-labeledcells weresolubilizedin I % NonidetP-40(ParticleData,Inc., Elmhurst,Ill.) andweresonicatedfor 15 sec.Thesolutionwasthenincubatedat 25°for30 mm,diluted10-foldwithPBS,andspunat 100,000 x g for 60 mm.
The supernatant was applied to a lectin affinity column (volume, 0.5ml) equilibrated with PBS-O.1% Nonidet P-40. The column was washedfree of radioactivity with the same buffer, and the bound glycoproteinswere eluted with the appropriate sugar dissolved in PBS containing0.1 % Nonidet P-40. The eluted materials were analyzed by SOS
polyacrylamide gel electrophoresis.
RESULTS
Effect of Inducers on Growth Properties. Chart i shows thegrowth curves of HAT-i 8 cells in the presence and absence ofsodium butyrate, DMSO, and retinoic acid. Sodium butyratecaused a marked reduction in the growth rates of HAT-i 8 cellsat 2 mp,i and complete inhibition of growth at 5 mM. At 2 mMsodium butyrate had little effect on the viabilities of the cells asassessed by trypan blue exclusion, but at 5 mM dye uptakebegan to occur on the fourth day. Normal growth rate wasrestored upon removal of butyrate. DMSO caused reduction inthe growth rates similar to that observed with 2% butyrate.Aetinoic acid at 3.3 X iO2 M caused less inhibition of thegrowth rate than was observed with the other 2 agents. Thegrowth rate was restored to normal upon removal of DMSO orretinoic acid. Cells remained viable with both DMSO or retinoicacid as shown by trypan blue dye exclusion. The in vitro growthproperties of treated cells are summarized in Table i . Thedoubling times of cells in the inducer-containing media were77 hr (2 m@butyrate), 67 hr (DMSO), and 39 hr (retinoic acid),all considerably longer than that of untreated cells. The saturation densities of inducer-treated cells were reduced markedlyto 25% (butyrate) and i 8% (OMSO) of the control value.Aetinaic acid-treated cells showed only a slight reduction in
IU)
a
‘U
@0
az
‘UU
DAYS
Chart i . Growth curves of HRT-i 8 cells in the presence ( ) and absence(—) of sodium butyrate (S. . . .5, 2 mM; S——S. 5 mM), DMSO (2%), and
retinoic acid (3.3 x 1o' M).Arrow, time at which each agent was removed fromcultures.
Table 1
8 FInal concentrations in the media were: sodium butyrate, 2 mM; DMSO, 2%;
0.Tsaoeta!.saturation density. Both sodium butyrate-treated cells andDMSO-treated cells did not form colonies in soft agar. Aetinoicacid treatment resulted in the formation of about one-half asmany colonies as the control cells. Chart 2 shows the photomicrographs of HAT-i 8 cells grown in the presence and absence of inducers. After 4 days, butyrate-treated cells are
200K@
116K94K
68K
43K -@
Chart 2. Autoradiograms of ‘251-labeledsurface membrane proteins of HRT18 cells. Cells were cultured in the presence of the following agents for 8 days:Lane 1, control; Lane 2, control (alcohol): Lane 3, sodium butyrate (2 mM);Lane4, DMSO(2%);Lane5, retinoicacid(3.3 x 1o' M).Thesameamountof proteinwas applied to each lane. K, molecular weight in thousands.
Table 2Effects on CEA content of HRT-18 cells
spread out with a polygonal epithelioid configuration. Thetreated cells also have a marked increase in the number andthe length of processes projecting from the expanded cytoplasm. DMSO- or retinoic acid-treated cells displayed a lesspronounced increase in the length of membranous processesthan did butyrate and showed no change in the cell size.
The CEA content of butyrate-treated cells was increased 17-fold in the membrane fraction and 40-fold in the cytoplasmicfraction compared to control cells (Table 2). By contrast, CEAcontent was reduced in the DM50-treated cells. Retinoic acidtreated cells showed no change.
As shown in Table 3, butyrate, DMSO, and retinoic acid hadvarying effects on the specific activities of 5 plasma membraneassociated enzymes. Butyrate caused an increase in the activities of alkaline phosphatase (i 0-fold), ‘y-glutamyltranspeptidase (3-fold), and sucrase (2-fold) but had no effect on aminopeptidase. K@-stimuIatedphosphatase activity was reducedto about one-half of that in the untreated cells. DMSO causedno change in the enzyme activities except for alkaline phosphatase which showed a decrease (one-third). Aetinaic acidcaused a decrease in the activities of alkaline phosphatase(one-tenth) and ‘y-glutamyltranspeptidase (one-half), while theactivities of the other 3 enzymes remained unchanged.
Chart 2 shows a typical autoradiogram of SDS-polyacrylamide gels of 125l-labeledsurface membrane proteins of HATi 8 cells cultured in the presence and absence of inducingagents. The only consistently noticeable change in the labelingpatternwasobservedwith butyratetreatmentwhichcausedanincrease in the labeling of a protein with an apparent molecularweight of 60,000. The increase in the labeling of this proteinwas more marked when cells were treated with 5 mM butyratethan with 2 mM butyrate (Chart 3). To further characterize thenature of the surface membrane proteins with molecularweights of 60,000, the 1251-labeledcells were salubilized innon-ionic detergent, Nonidet P-40, and run on 4 different lectinaffinity columns in 0.1% Nonidet P-40. The columns werewashed, and the bound glycoproteins were eluted with specificsugars (Chart 4). In these experiments, radioactively labeledmembrane proteins at amounts below the binding capacity ofthe lectin column were applied to ensure sufficient binding ofglycoproteins to the columns. The fractions separated by lectinaffinity chromatography were further resolved on SDS-polyacrylamide gel electrophoresis followed by radioautography.The labeled peak that passes through the column containsmostly dialyzable low-molecular-weight radioactive materials(70 to 90%). The results are shown in Chart 5 where it is notedthat the protein with molecular weight of 60,000 binds campletely to ACA-Sepharose and also binds well to WGA- andCon A-Sepharose. It binds very poorly to peanut agglutininSepharose.
Table3
V
0
12345
@Finalconcentrations in the media were: sodium butyrate, 2 mM;DMSO, 2%;retinoic acid, 3.3 x i0' M.
b Mean ± S.D. of 4 experiments.
a Final concentrations in the media were: sodium butyrate, 2 mM; DMSO, 2%; retinoic acid, 3.3 x 1 0@ M.b Mean ±S.D. of 4 experiments.
Chart 5. Autoradiograms of iodinated cells surface proteins bound to lectinSepharose column. P. fraction not bound to column: S. bound fraction elutedwith a specific sugar.
reversibly caused these tumor cells to acquire in vitro properties that are considered to be more consistent with normal cellgrowth (i 0, 40). The concentration of each inducer used in thisstudy was approximately one-half of that which caused camplete inhibition of cell growth or that affected cell viability. Incontrast to the growth studies, only butyrate caused markedchanges in cellular morphology. The significance of butyrateinduced cellular enlargement and flattening and of the increased number and length of membranous processes is notclear.
The tissue CEA contents have been reported to vary depending on the degree of differentiation of human colorectaltumors (5, 9); poorly differentiated or anaplastic tumors generally have a lower CEA content than do well-differentiatedcolorectal tumors. Our data showed that butyrate and DMSOcaused reversible changes in the CEA content of HAT-i 8 cellswhile retinoic acid had no effect. Butyrate caused a marked(20- to 40-fold) increase in the CEA content of HAT-i 8 cells.Since higher CEA content also seems to correlate with betterdifferentiated human colorectal tumor cell lines (25), theseresults may indicate the induction of a better-differentiatedphenotype in the human rectal cancer cells by butyrate. Usingimmunofluorescent studies, Hager et a!. (i 8) reported an increase in membrane-associated CEA of 2 human colon cancercell lines following a treatment with another polar solvent (N,Ndimethylformamide).
In the small intestine, alkaline phosphatase, ‘y-glutamyltranspeptidase, sucrase, and aminooligopeptidases are associatedwith the microvillous membrane, while K4-stimulated phosphatase is localized in the basolateral membrane of mucosal cells(3i , i 6). Furthermore, the activities of these enzymes arehigher in the more-differentiated villous cells than in the lessdifferentiated crypt cells (34). Although comparable data arenot available for the colon, the alkaline phosphatase activityhas been reported to be lower in colon cancer tissues than in
456Chart 3. Autoradiograms of ‘25l-labeledsurface membrane proteins of butyr
ate-treated and control HAT-i 8 cells. Cells were cultured for 8 days. Lanes 1, 2,and3, the sameamountof proteinwasapplied:Lanes4, 5, and 6, the sameamount of radioactivity applied. Lanes 1 and 4, control; Lanes 1 and 5, sodiumbutyrate (2 mM);and Lanes 3 and 6, sodium butyrate (5 mM).
‘S$5
‘S
Chart 4. Lectln affinity chromatography of iodinated surface membrane protelnsof sodiumbutyrate-treatedHAT-i8 cells.Arrow,pointat whichtheelutingsugar was applied. The columns were eluted at 10 mI/hr. and 0.2-mI fractionswere collected. The following sugars were used to elute the glycoproteins boundto column:ConA-Sepharose(5EPH.),0.2Ma-rnethylmannoside;ACA-Sepharose, 0.2 N lactose; peanut agglutinln-Sepharose (PNA), 0.2 M lactose: WGASepharose, 0.5 M N-acetylglucosamine.
DISCUSSION
Inthe presentstudy,sodium butyrate,DMSO, and retinoicacid all had marked effects on various growth properties ofHRT-i 8 cells including prolongation of doubling time, decreasein saturation density, and in vitro tumorigenicity as tested bygrowth in soft agar. These results suggest that all 3 inducers
normal colonic tissues (3). Therefore, the increase in the activities of alkaline phosphatase, y-glutamyl transpeptidase, andsucrase in the butyrate-treated HAT-i 8 cells may reflect amore differentiated phenotype than do the untreated cancercells. The concomitant decrease in the activities of 2 otherplasma membrane-associated enzymes, K@-stimulatedphosphatase and aminooligopeptidases in the butyrate-treatedcells, indicates that there is not a generalized increase in geneexpression. Although not shown, the induction of alkaline phosphatase appears to require new protein synthesis, since itcould be completely inhibited by cyclohexamide.
In the case of butyrate-treated colon cancer cells, it appearsthat the changes in CEA content and in the activities of membrane-associated enzymes are what would be expected of amore differentiated phenotype (3, 25, 34). However, when thesame cell line is treated with DMSO or retinoic acid, the CEAcontent is reduced by DMSO, alkaline phosphatase activity isdecreased in cells grown in DMSO or retinoic acid, and yglutamyl transpeptidase activity is decreased in cells grown inretinoic acid. If the changes in growth parameters and colonyformation in soft agar are taken as indicators of a more normalcell line, then it may be concluded that the changes in CEAcontent and in enzyme levels are not goad markers of a moredifferentiated phenotype or that the progression from a malignant to a normal cell line is not closely related to the processof differentiation. Alternatively, each inducer may causeexpression of different phenotypic markers of differentiation incolonic cells. In the absence of well-characterized differentiation markers for colonic cells, it is not yet possible to determinewhich of these conclusions are valid. Although they may not berelated to the more normal properties of the treated cells, thechanges in CEA and enzyme levels observed in this study mayprovide useful information about the mechanisms involved inthe effects of butyrate, DMSO, and retinoic acid on coloncancer cells.
The mechanisms by which various groups of substancesinduce cellular differentiation are not well understood. Butyratehas been noted to affect nuclear histone composition (46, 39)and DNA structure (43). Although DMSO and other polarcompounds alter membrane permeability and microviscosity(1 ) and also cause the single-stranded breaks in the DNA (43)to occur, the relationship between these 2 events remainsunclear. It is known that retinol or retinoic acid is required forthe normal differentiation of epithelial tissues (27). Aetinoidsalso cause differentiation of embryonal carcinoma cells (4i ). Itis thought that retinoids may act like steroid hormones byforming a complex with a binding protein which is translocatedinto the cell nucleus causing alterations in gene expression(27), but firm supportive evidence is lacking.
Carbohydrate moieties of glycoconjugates through cell surface interactions have been suggested to play an importantrole in cellular differentiation (30, 48). Significant qualitative orquantitative alterations in cell surface glycoproteins have beenreported to occur in association with normal processes ofcellular differentiation (4, 47) or during in vitro differentiation ofmouse embryonal carcinoma cells (30) or mouse neuroblastoma cells (26). A significant alteration in the cell surfaceprotein-labeling pattern was observed only with butyrate treatment. With butyrate treatment a considerable increase in thelabeling of cell surface proteins occurred. In addition, a newband having an apparent molecular weight of 60,000 ap
peared. Since this protein bound specifically to Con A-, RCA-,or WGA-Sepharose columns, it would appear that it contains aglycoprotein. Although further studies are necessary beforedefinitive conclusions can be drawn, it is possible that the cellsurfaceglycoproteinwith a molecularweight of 60,000 maybe closely related to the mechanism by which butyrate causesthe colon cancer cells to assume a more normal phenotype.Butyrate and other short-chain fatty acids are natural fermentation products of colonic bacterial flora. Moreover, the concentration of butyrate in the colonic lumen is likely to be ashigh, if not higher, than the concentrations used in our study(6, 36). Whether or not intraluminal concentrations of shortchain fatty acids play a role in the development of colorectalcancer remains to be determined. Further studies using ‘‘differentiating agents' â€m̃ay not only provide useful information onthe identification of differentiation-associated markers of human colorectal cells but may also contribute to elucidation ofthe molecular events involved in their tumorigenicity.
REFERENCES
i . Arndt-Jovin, D. J., Ostertag, W., Eisen, H., Limek, F., and Jovin, T. M.Studies of cellular differentiation by automated cell separation. Two modelsystems: Friend virus-transformed cells and Hydra attenuata. J. Histochem.Cytochem., 24: 332—347,1976.
2. Bassett, E. Large scale preparation of wheat germ agglutinin. Prep. Biochem.,5: 461—477,i 975.
3. Bell, L., and Williams, L. Histochemical demonstration of alkaline phosphatase in human large intestine, normal and diseased. Histochemistry, 60: 84—89, i979.
4. Bouhours, J. F., and Glickman, A. M. Rat intestine glycolipids. II. Distributionand biosynthesis of glycolipids and ceramide in villus and crypt cells.Biochim. Biophys. Acta, 441: 123—i33,1976.
5. Breborowicz, J., Easty, G. C., Birbeck, M., Robertson, D., Nery, A., andNeville, A. M. The monolayer and organ culture of human colorectal carcinomata and the associated “normal―colonic mucosa and their productionof carcinoembronic antigens. Br. J. Cancer, 3 1: 559—569,1975.
6. Chernov, A. J., Doe, W. F., and Gompertz, D. Intrajejunal volatile fatty acidsin the stagnant loop syndrome. Gut, 13: 103—i06,i972.
7. Dahlqvist, A. Assay of intestinal disaccharidase. Anal. Biochem., 22: 99—107,i968.
8. DeLarco, J. E., and Todaro, G. J. Growth factors from murine sarcoma virustransformed cells. Proc. NatI. Acad. Sd. U. S. A., 75: 400i —4005,i 978.
9. Denk, H., Tappeiner, G., Eckerstorfer, A., and Holzner, J. H. Carcinoembryonic antigen (CEA) in gastrointestinal and extragastrointestinal tumors andits relationship to tumor-cell differentiation. Int. J. Cancer, 10: 262—272,i972.
i 0. Dexter, D. L. N,N-Dimethylformamide-induced morphological differentiationand reduction of tumorigenicity in cultured mouse rhabdomyosarcoma cells.Cancer Res., 37: 3136—3140,i977.
1i . Dexter, D. L., and Hager, J. C. Maturation-induction of tumor cells using ahuman colon carcinoma model. Cancer (Phila.), 45: 1i 78—i184, 1980.
i 2. Fairbanks, G., Steck, T. L., and Wallach, D. F. H. Electrophoretic analysisof the major polypeptides of the human erythrocyte membrane. Biochemistry, 10: 2606—2617,i971.
13. Freeman, H. J., Ether, M. E., Garrido, A. B., and Kim, Y. S. Alterations incell surface membrane components of adapting rat small intestinal epithehum. Gastroenterology, 75: 1066—1072, i 978.
I 4. Friend, C. The regulation of differentiation in murine virus-induced erythroleukemia cells. In: A. G. McKinell, M. A. Diberadino, M. Blumenfeld, and A.D.Bergad(eds.),DifferentiationandNeoplasia,pp.202—212. Berlin:Springer-Verlag,i 980.
15. Friend, C., Scher, W., Holland, J. G., and Sato, T. Hemoglobin synthesis inmurine virus-induced leukemia cells in vitro: stimulation of erythroid differentiation by dimethyl sulfoxide. Proc. Natl. Acad. Sci. U. S. A. 68: 378—382,i97i.
i 6. Fujita, M., Ohta, H., Kawai, K., Matsui, H., and Nakao, M. Differentialisolation of microvillous and basolateral plasma membranes from intestinalmucosa: mutually exclusive distribution of digestive enzymes and ouabainsensitive ATPase. Biochim. Biophys. Acta, 274: 336—347,1972.
i 7. Gazitt, Y. Comparative study of two groups of inducers of Friend erythroleukemia cell differentiation in a chemically defined medium. Cancer Ass., 41:i184-i186, 1981.
18. Hager, J. C., Gold, D. V., Barbosa, J. A., Fligiel, Z., Miller, F., and Dexter,D. L N,N-Dimethylformamide-induced modulation of organ- and tumor-as
sedated markers in cultured human colon carcinoma cells. J. NatI. Cancer.Inst., 64: 439—445,1979.
19. Kim, V. S., Tsao, D., Hicks, J., and McIntyre, L. J. Surface membraneglycoproteins of cultured human pancreatic cancer cells. Cancer (Phila.),47: 140—146,1981.
20. KIm, V. S., Ysao, D., Siddiqui, B., Whitehead, J. S., Arnstein, P., Bennett, J.,and Hicks, J. Effects of sodium butyrate and dimethylsulfoxide on biochemical properties of human colon cancer cells. Cancer (Phila.), 45: 1185—1192, 1980.
21. Kimhi,V., Palfrey,C.,Spector,I., Barak,Y.,andLittauer,U.Z. Maturationof neuroblastoma cells in the presence of dimethysulfoxide. Proc. NatI.Aced. Sd. U. S. A., 73: 462-466, 1976.
22. Laemmli, U. K. Cleavage of structural proteins during the assembly of theheadofbacteriophageT4.Nature(Lond.),227:680—685,198i.
23. Leavltt, J., Barrett, J. C., Crawford, B. D., and Tsao, P. 0. P. Butyric acidsuppressionof the in vitroneoplasticstateof Syrianhamstercells.Nature(Lond.), 271: 262—265,i978.
24. Leder, A., and Leder, P. Butyric acid, a potent inducer of erythroid differentiation in cultured erythroleukemia cells. Cell, 5: 319—322,1975.
25. Lelbovitz,A., Stlnson,J. C., McCombs,W. B., Ill, McCoy, C. E., Mazur, K.C.,andMabry,N.D.Classificationofhumancolorectaladenocarcinomacelllines. Cancer Ass., 36: 4562—4569,i 976.
26. Uttauer, U. Z., Giovanni, M. Y., and Glick, M. C. A glycoprotein from neuritesof differentiatedneuroblastomacells. J. Biol. Chem.,255: 5448—5453,1980.
27. Lotan, R. Effects of vitaminA and its analogs(retinoids)on normal andneoplastic cells. Biochim. Biophys. Acta, 605: 33-91 , 1980.
28. Lowry, 0. H., Aosebrough, N. J., Farr, A. L., and Randall, A. J. Proteinmeasurement with the Folin phenol reagent. J. Biol. Chem., 193: 265—275,1951.
29. March, 8. C., Parikh, I., and Cuatrecasas, P. A simplified method forcyanogen bromide activation of agarose for affinity chromatography. Anal.Blochem., 60: 149—152, 1974.
30. Muramatsu, T., Gachelin, G., Nicolas, J. F., Condamine, H., Jakob, H., andJacob, F. Carbohydrates structure and cell differentiation: unique propertiesof fucosyl-glycopeptidesisolatedfrom embryonalcarcinomacells. Proc.NatI.Acad.Sd.U.S.A.,75:2315-2319,1978.
31. Murer, H., Ammann, E., Biber, J., and Hopfer, U. The surface membrane ofthesmallintestinalepithelialcell. I. Localizationof adenylcyclase.Biochim.Biophys.Acta,433:509—519, 1976.
32. Nicholson, J. A., and Kim, Y. S. One step L-aminooxidase assay for intestinalpeptlde hydrolase activity. Anal. Biochem., 63: 110-1 17, i 975.
33. Nicolson, G. L., and Blaustein, J. The interaction of Ricinus communisagglutinln with normal and tumor cell surfaces. Biochim. Biophys. Acta, 266:
543-547, 1972.34. Nordstrom, C., Dahlqvist, A., and Josefson, L. Quantitative determination of
enzymes in different parts of the villi and crypts of rat small intestine. J.Histochem. Cytochem., 15: 7i 3—721, 1967.
35. Orkin, S. H., Swan, D., Leder, P. Differential expression of a and fi globingenes during differentiation of cultured erythroleukemic cells. J. Biol. Chem.,250: 8753—8760,1975.
36. Prizont, A., Whitehead, J. S., and Kim, Y. S. Short chain fatty acids in ratswith jejunal blind loops. Gastroenterology, 69: 1254—1264, 1975.
37. Reuben, A. C., Aifkind, A. A., and Marks, P. Chemically induced murineerythroleukemic differentiation. Biochim. Biophys. Acta, 605: 325—346,1980.
38. Sasak, W., DeLuca, L. M., Dion, L. D., and Silverman-Jones, C. S. Effect ofretinoic acid on cell surface glycopeptides of cultured spontaneously transformed mouse fibroblasts (Balb/c 3Ti 2-3 cells). Cancer Res., 40: 1944—1949, i980.
39. Sealy, L., and Chalkley, A. Effect of sodium butyrate on histone modification.Cell,14:115—121,1978.
40. Stoker, M. Aegulation of growth and orientation in hamster cells transformedby polyoma virus. Virology, 23: 165—174, 1969.
4i . Strickland, S., and Mahdavi, V. The induction of differentiation in teratocarcinomous stem cells by retinoic acid. Cell, 15: 393—403,1978.
42. Terada, M., Epner, E., Nudel, U., Salmon, J., Fishbach, E., Rifkind, A. A.,and Marks, P. A. Induction of murine erythroleukemia differentiation byactinomycin D. Proc. NatI. Acad. Sci. Ii. S. A., 75: 2795—2799,1978.
43. Terada, M., Nudel, U., Fibach, E., Aifkind, A. A., and Marks, P. A. Changesin DNA associated with induction of erythroid differentiation by dimethylsulfoxide in murine erythroleukemia cells. Cancer Aes., 38: 835—840,1978.
44. Tompkins, W. A. F., Watrach, A. M., Schmale, J. D., Schultz, A. M., andHarris, J. A. Cultural and antigenic properties of newly established cellsstrains derived from adenocarcinomas of the human colon and rectum. J.NatI. Cancer Inst., 52: 1101—ii 06, i 974.
45. Tsao, D., Colton, 0. G., Chang, J. S., Buck, R. L., Hudson, B. G., andCarraway, K. L. Alterations of red cell membranes from phenylhydrazinetreated rabbits. Biochim. Biophys. Acta, 469: 61—73,i 977.
46. Vidali, G., Boffa, L C., Bradbury, E. M., and Allfrey, V. G. Butyrate suppression of histone deacylation leads to accumulations of multiacetylated formsof histones H3 and H4 and increased DNase 1 sensitivity of the associatedDNA sequences. Proc. NatI. Acad. Sci. U. S. A., 75: 2239—2243, 1978.
47. Weiser, M. M. Intestinal epithelial cell membrane glycoprotein synthesis. I.An indicator of cellular differentiation. J. Biol. Chem., 248: 2536—2@41,1973.
48. Yusuf, H. K. M., Merat, A., and Dickerson, J. W. T. Effect of development onthe gangliosides of human brain. J. Neurochem., 28: 1299—1312, 1977.
Figs. 1 to 8. Photomicrographs of HAT-i 8 cells in logarithmic phase of growth in the presence and absence of sodIum butyrate, DMSO, and retinoic acid: Figs. 1and 5, control cells; Figs. 2 and 6, butyrate (2 mM)-treated cells: Figs. 3 and 7, DMSO (2%) -treated cells; Figs. 4 and 8, retinolc acid (3.3 x 10' M)-treated cells.Figs 1 to 4, x 50; Figs. 5 to 8, x 200.
1982;42:1052-1058. Cancer Res D. Tsao, A. Morita, A. Bella, Jr., et al. Glycoproteins of Human Rectal Adenocarcinoma CellsRetinoic Acid on Membrane-associated Antigen, Enzymes, and Differential Effects of Sodium Butyrate, Dimethyl Sulfoxide, and
