Glycoprotein Synthesis and Secretion by Mucosal

Biopsies of Rabbit Colon and Human Rectum

RICHARDP. MACDERMOrr,ROBERTM. DONALDSON,JR., andJERRYS. TuMi

From the Departments of Medicine and Anatomy, Boston University School ofMedicine, Boston, Massachusetts 02118

A B S T R A CT Elucidation of mechanisms involved inthe control of colonic production of mucus requires directexamination of glycoprotein synthesis and secretion bycolonic mucosa. In the past, the limited viability of in-testinal mucosa in vitro has hampered such investiga-tions. When maintained in an organ culture system, mu-cosal biopsies of rabbit colon and human rectum remainedviable for 24 h as documented by morphologic appear-ance and a steady rate of protein synthesis and secretion.These biopsies also incorporated 'C-labeled glucosamineinto tissue. glycoproteins and secreted labeled glyco-proteins at a steady rate for 24 h. Glucosamine was pre-dominantly incorporated into macromolecules that wereultimately secreted, in contrast to leucine, which waspredominantly incorporated into tissue macromolecules.When studied by autoradiography, cultured rabbit co-lonic biopsies synthesized and secreted glycoproteinsin vitro at cellular sites and over a time-course similarto those previously described for the intestine of intactanimals. Acetylcholine consistently stimulated secretionof labeled glycoproteins but did not alter glycoproteinsynthesis. In contrast, cycloheximide inhibited glyco-protein synthesis but had no effect on the secretion ofnewly synthesized glycoproteins. Rectal biopsies frompatients with active ulcerative colitis incorporated in-creased amounts of [1'C]glucosamine into glycoproteins

This work was presented in part at the Annual Meetingof the American Federation for Clinical Research, AtlanticCity, N. J., April 1973 (1973. Clin. Res. 21: 518) andat the Annual Meeting of the American GastroenterologicalAssociation, May 1973 (1973. Gastroenterology. 64: 765).

Dr. MacDermott's present address is: Laboratory ofApplied Immunology, Walter Reed Army Institute for Re-search, Washington, D. C. Dr. Donaldson's present addressis: Chief of Medical Service, Veterans Administration Hos-pital, West Haven, Conn. 06516. Dr. Trier's present ad-dress is: Chief, Gastroenterology Division, Peter BentBrigham Hospital, Boston, Mass. 02115.

Received for publication 22 January 1974 and in revisedform 15 May 1974.

during organ culture and secreted labeled glycoproteinsmore rapidly into the incubation medium when comparedto biopsies obtained from healthy volunteers These find-ings indicate that organ culture provides a useful meansof directly examining the synthesis and secretion ofglycoproteins by healthy and diseased colonic mucosa.

INTRODUCTIONThe regulation of mucus production by colonic mucosaremains poorly understood although previous work (1-4)has clarified the sequence of events involved in intestinalsynthesis of glycoproteins. Neutra and Leblond (2) in-jected radioactive precursors of glycoproteins into intactrats and at intervals thereafter examined colonic mu-cosa by autoradiography. These workers showed thatpackaging of glycoproteins into mucus granules occursin the Golgi region of goblet cells. These mucus granulesthen coalesced to form globules which migrated to theapex of the cell where mucus is discharged into the lu-men. Using cell fractionation techniques, Forstner (3)demonstrated the progression of radioactively labeledglycoproteins from a microsomal fraction to a brushborder fraction of small intestinal cells and then intothe intestinal lumen. Similarly, Kim, Perdomo, and Nord-berg (4) showed that after synthesis of the peptideportion of the molecule in the rough endoplasmic reticu-lum, the sugar moieties are added in the smooth endo-plasmic reticulum by stepwise glycosylation. Thus,within the channels of endoplasmic reticulum membrane-bound glycosyl transferases attach one sugar at a timeto the growing glycoprotein.

In spite of this work, however, little is known aboutthe rates of glycoprotein synthesis and secretion by thecolon or about factors which influence these rates. Inorder to examine these processes directly, it is necessaryto maintain isolated colonic mucosa under steady-stateconditions for reasonably prolonged periods of time and

The Journal of Clinical Investigation Volume 54 September 1974@545 554 545

to assess secretion independent of synthesis. The limitedviability of intestinal mucosa in vitro has previouslyprevented such an approach. Recently, however, an organculture system has been developed which allows humanand rabbit small intestinal biopsies to be maintained invitro for 24 h (5, 6). Cultured biopsies of rabbit smallbowel mucosa synthesized and secreted proteins at asteady rate for 24 h with good histological preservation(6). In addition, human rectal biopsies remained viablefor 24 h when maintained in organ culture (7). Wehavetherefore used this technique to examine the biosynthesisand secretion of glycoproteins by biopsies of rabbit andhuman colonic mucosa.

METHODSCulture technique. A 10-15-cm segment of distal colon

was excised from anesthetized 2-5-kg white male NewZealand rabbits. After opening the segment lengthwise,multiple mucosal biopsies were taken with a multipurposebiopsy tube (8) (Quinton Instruments, Seattle, Wash.).For human studies, biopsies were obtained from rectums ofnormal volunteers or from patients with active ulcerativecolitis. Up to four biopsies were taken from each volunteer,of which one was fixed in Bouin's solution (7) to assessmorphology. Normal colon was obtained at the time ofsurgery from two subjects undergoing colostomy revisionafter previous colectomy for diverticulitis. The blood supplyof the colonic segment to be resected was maintained up tothe moment of removal. The 4-5-cm segment was thenopened lengthwise, and multiple mucosal biopsies weretaken with a multipurpose biopsy tube. Biopsies were cul-tured as described in detail previously (5). In brief, bi-opsies were oriented mucosal side up on a stainless steelwire screen in plastic organ culture plates (Falcon Plastics,Division of B-D Laboratories, Inc., Los Angeles, Calif.).The screens were floated on 1 ml of organ culture mediumcomposed of 90% Trowells T8 medium (Grand IslandBiological Co., Grand Island, N. Y.) and 10% fetal calfserum to which 100 U per ml penicillin and 100 ,ug perml streptomycin had been added. The biopsies were culturedfor periods of up to 24 h at 370C in a sealed McIntoshjar (Torsion Balance Co., Clifton, N. J.) in 95% 02-5%C02.

Glycoprotein and protein synthesis and secretion. To as-sess radioactive glucosamine incorporation into glycopro-teins, 1.25 1uCi D-[1-14C]glucosamine (sp act 56.5 mCi permmol, New England Nuclear, Boston, Mass.) was added toeach milliliter of medium. To measure incorporation intotissue glycoprotein, cultured biopsies were homogenized inground glass homogenizers in 1.5 ml of chilled 0.04%CaCl. Total biopsy protein was determined from an ali-quot of this homogenate by the method of Lowry, Rose-brough, Farr, and Randall (9). Another aliquot was precipi-tated with 10% trichloroacetic acid (TCA) 1 and 1%ophosphotungstic acid (PTA) and centrifuged at 15,000 rpmfor 10 min. The pellet was washed twice with 10% TCA-1% PTA and then washed twice with chloroform-methanol(1: 1 by volume). After drying in room air, the pellet wasdissolved in 1 cm' of NCS (Amersham-Searle, ArlingtonHeights, Ill.) and counted in Liquiflour-Toluene (New En-gland Nuclear) in a Beckman LS-250 liquid scintillation coun-

lAbbreviations used in this paper: PTA, phosphotungsticacid; TCA, trichloroacetic acid.

ter (Beckman Instruments, Inc., Fullerton, Calif.). Countingefficiency was determined by the external standard-channelsratio method. To study protein synthesis, 1.25 gCi of L-[U-4C] leucine (sp act 305 mCi per mmol, New England

Nuclear) or 2 /ciCi of L-[4,5-2H]leucine (sp act 55.2 mCiper mmol, New England Nuclear) were added to each milli-liter of medium. After culture, biopsies were homogenizedand treated as above.

To measure secretion of radiolabeled glycoproteins orproteins, the incubation medium was collected from eachbiopsy. Before collection of the medium, however, the bi-opsy was gently rinsed with a few drops of the mediumto obtain secreted macromolecules adherent to the top ofthe biopsy. The medium was then centrifuged at 15,000rpm for 10 min to remove cellular debris. The supernatewas dialyzed in Visking tubing for 72 h against three ex-changes of 0.15 M NaCl (200: vol: vol). After dialysis,an aliquot of the medium was counted in Bray's solution(10), to determine the amount of radioactive precursorincorporated into nondialyzable macromolecules. Controldialysis bags contained radioactive incubation medium whichhad not been incubated with a biopsy. To correct for varia-tions in biopsy size, synthesis and secretion of radiolabelwere expressed as disintegrations per minute per milli-gram of total biopsy protein as determined by the Lowrymethod (9). The amounts of RNA and DNA per biopsywere determined by previously described methods (11, 12)so that synthesis and secretion could also be expressed interms of micrograms of RNAand DNA.

Pulse-chase studies. To measure macromolecular secre-tion independent of synthesis, a "pulse-chase" system wasutilized. During the pulse period, biopsies were cultured for1 h with 5.0 ACi ['4C]glucosamine in the medium. Biopsieswere then cultured during the chase period over new me-dium containing 20 mM unlabeled glucosamine and noradioactive glucosamine. By dilution, the nonradioactiveglucosamine minimizes further labeling of tissue glycopro-teins. Therefore, during the chase period labeled glycopro-teins are secreted into the medium, and when the biopsiesand medium were processed as described above, secretioncould be distinguished from synthesis.

Autoradiography. To localize the site of glucosamineincorporation into macromolecules and to determine itssubsequent pattern of movement within epithelial cells, bi-opsies were cultured over medium to which 10 ,uCi of D-[6-'H]glucosamine (sp act 3.6 Ci per mmol, New EnglandNuclear) per ml had been added. The biopsies were cul-tured over isotope-containing medium for 40 min, afterwhich the radioactive medium was removed and replaced bymedium containing 20 mMunlabeled glucosamine. Biopsieswere then harvested after culture over this medium for 1, 1,2, 3, 4, 6, 8, or 24 h. After fixation in buffered osmiumtetroxide and embedment in epoxy resin, autoradiographswere prepared as described previously (5, 7). To determinethe distribution of radioactivity within goblet cells, thelocation of radioactivity was determined in 100 well ori-ented goblet cells for each time interval in coded autoradio-graphs. An ocular micrometer was used to divide eachgoblet cell into basal, middle, and upper thirds. If thepulse of radioactivity could not be localized to a specificthird of the cell, the cells were scored as diffusely labeled.Cells without appreciable radioactivity were tabulated asunlabeled cells.

Drug studies. The effect of cycloheximide was deter-mined by adding 1 mg of cycloheximide (Sigma ChemicalCo., St. Louis, Mo.) to each milliliter of medium. Acetyl-

546 R. P. MacDermott, R. M. Donaldson, Jr., and J. S. Trier

choline chloride (Sigma Chemical Co.) in concentrationsvarying from 10V to 107" M was added to the mediumalong with 3 X 10V M eserine (Calbiochem, La Jolla,Calif.). Atropine sulfate 1OV M (Sigma Chemical Co.) wasused to block the effects of acetylcholine.

Column chromatography. Dialyzed medium was sub-jected to gel filtration on Sepharose 2B (Pharmacia FineChemicals, Piscataway, N. J.) with pH 6.8, 0.1 M phos-phate buffer.

RESULTSTo investigate the stability of the biopsies over the 24-hincubation period, we cultured rabbit biopsies over 'C-labeled leucine for 6 h ("0-6 hour" biopsies) and com-pared the incorporation of ["C]glucosamine into tissueand secreted glycoproteins with the incorporation whichoccurred when biopsies were cultured over nonradio-active medium for 18 h and then over medium containinglabeled leucine for 6 h ("18-24 hour" biopsies). Asshown in Fig. 1, biopsies incorporated similar amountsof radioactivity into glycoproteins and secreted similaramounts of labeled glycoprotein into the medium whetherstudied immediately or after 18 h of culture. In addition,the biopsies showed maintenance of near normal histo-logical architecture at the end of the 24-h incubationperiod.

When rabbit biopsies were continuously incubated for24 h, the amount of radioactive ['H]leucine incorporatedinto tissue proteins plus secreted proteins was linear(Fig. 2B). In addition, glucosamine incorporation intoboth tissue and secreted glycoproteins (total glucosamineincorporation) was also essentially linear over the 24 h(Fig. 2A). During the first 12 h most of the "C-labeledglucosamine was incorporated into glycoproteins ex-tracted from the tissue. Subsequently, however, increas-ing amounts of incorporated radioactivity were foundin the medium, so that at 24 h the amount of macromo-lecular radioactivity recovered from the medium, whichconstituted radioactivity incorporated into secreted gly-coproteiis, was greater than that present in the tissue.In contrast to the results obtained with glucosamine,the rate of incorporation of radioactive leucine into tis-sue proteins and also into proteins secreted into the me-dium remained reasonably constant throughout the 24-hperiod. Furthermore, the amount of radioactivity presentin tissue protein greatly exceeded that found in secretedproteins throughout the period of culture.

Cycloheximide, a known inhibitor of protein synthesis(13), affected leucine and glucosamine incorporation intotissue macromolecules differently. When biopsies werenot preincubated with cycloheximide but were cultureddirectly for 18 h over medium containing cycloheximidewith either ["C]glucosamine or ["4Cileucine, the drugalmost completely abolished leucine incorporation intotissue proteins but only moderately reduced glucosamineincorporation into glycoproteins (Fig. 3). To show that

0%oZI

I-0x.zI0

2

a

ma

zI0

3001

'2r TIZ

:im

a .

S

4

2

0- 6 18 24

HOURSFIGURE 1 Incorporation of ["C]glucosamine into tissue andsecreted glycoproteins by biopsies at the beginning (0-6hours) and the end (18-24 hours) of a 24-h culture period.The 0-6 hour biopsies were cultured over medium contain-ing "C-labeled glucosamine for 6 h and then processed,while the 18-24 hour biopsies were cultured over unlabeledmedium without added glucosamine for 18 h and thentransferred to medium containing radioactivity for the re-maining 6 h. The results are expressed as mean dpm/mgbiopsy protein ±1 SEM. Each bar represents the meanresult for six biopsies.

radiolabeled glucosamine was not simply binding non-specifically to the tissue but was incorporated into macro-molecules, biopsies were first preincubated with cyclo-heximide for 1, 6, or 12 h and then cultured for an ad-ditional 18 h with either ["C]glucosamine or ["C]leu-cine. With progressively lengthening periods of prein-cubation (Fig. 3), glucosamine incorporation was in-creasingly inhibited.

The pattern of movement of labeled glycoproteinsthrough colonic cells was examined by autoradiography.Fig. 4 shows autoradiographs of tissue fixed 1 and 6h after a 40-min exposure to radioactive glucosamine.In the top panel, at 1 h, label was seen primarily in theGolgi regions of goblet cells, in mucus granules justabove the Golgi region, and throughout the cytoplasmof absorptive cells. In the bottom panel, by 6 h, labeledglycoprotein was concentrated in secreted mucus in thecrypt lumen, in the apical mucus granules of goblet cells,and in the glycocalyx of the absorptive cells. Determina-tion of the distribution of radioactivity in goblet cells incoded slides showed the sequential progression of glu-

Glycoprotein Synthesis and Secretion by Colonic Biopsies 547

lo [

A.['4C] GLUCOSAMINE

12 18 24

HOURS

l1 1,200xz

S- I,000

IL0.

a

a 600

w 400

8zU 200La

go

45H] LEUCINE

HOURS

FIGURE 2 Incorporation of radioactivity by rabbit colonic biopsies continuously exposed to1.25 gCi of ["Ciglucosamine or [4C] leucine over a 24-h period. The amount of radioactivityincorporated into macromolecules in the medium and tissue was determined as described inMethods. The results are expressed as mean dpm/mg protein ±1 SEM. Each point representsthe mean value for six to fourteen biopsies.

cosamine-labeled macromolecules from the basal cyto-plasm at i-2 h to the apical cytoplasm by 6-8 h and out ofthe cells by 24 h (Table I).

50r-

0

I-z0

IL0

z

0.

25F

T14 C] GLUCOSAMINE

i ~ ~ ~ ~ I14C1 LEUC

_z i 6NONE I HOUR 6HOURS 12HOURS

HOURSPREINCUBATION WITH CYCLOHEXIMIDE

CINE

FIGURE 3 Effect of cycloheximide on ["C]glucosamine and["C] leucine incorporation into tissue macromolecules byrabbit colonic biopsies. The biopsies were preincubated for1, 6, or 12 h with cycloheximide. Cycloheximide was keptin the incubation medium for an 18-h further incubationwith either ["C] glucosamine, as shown by the open bars,or ["C]leucine, as shown by the hatched bars. One group ofbiopsies was not preincubated with cycloheximide. Thevertical axis shows the radioactivity incorporated as per-cent of control values. The control biopsies, not shown,were incubated for the same period of time, but were notall exposed to cycloheximide. Each bar represents the mean±1 SEMfor six to eight biopsies.

This pulse-chase system was also used to quantitatethe uptake, incorporation, and secretion of radioactiveglucosamine and leucine. The rabbit biopsies. were firstpulsed with '4C-labeled glucosamine or leucine for 1 hand then transferred to chase medium containing largeamounts of unlabeled precursor. Wecould then observethe distribution of radioactivity and correlate the find-ings with the results of autoradiography. Since the ex-cess of unlabeled precursor prevents further incorpora-tion of labeled precursor molecules into the biopsy,once labeling of the tissue macromolecules had occurred,secretion could be followed independent of tissue syn-thesis.

Fig. 5 shows the changing distribution of radioactiveleucine and glucosamine for 24 h after the initial 1-hpulse period. We determined: (a) the percent of re-covered radioactivity which was found in the TCA su-pernate of homogenized biopsies (TCA-soluble glyco-proteins, proteins, and precursors), (b) the percent ofrecovered radioactivity in the TCA precipitate of ho-mogenized biopsies (TCA-precipitable glycoproteins andproteins), and (c) the percent of recovered radioactivityin nondialyzable macromolecules recovered from themedium (secreted glycoproteins and proteins). For both[14C]glucosamine and [14C]leucine, most of the radioac-tivity was initially present in the TCA supernate (acid-

548 R. P. MacDermott, R. M. Donaldson, Jr., and J. S. Trier

go

0xz

ii

I-

0

0.68

02N

aaa

0.0L0

01'

soluble form). Radioactivity subsequently disappearedgradually from the supernate and was incorporated intotissue macromolecules. The radioactivity in the TCAsupernate was completely removed by dialysis, indicatingthat the TCA-soluble radioactivity represented smallmolecules. After the pulse period there was progressiveformation of "C-labeled TCA-precipitable glycoproteinsand proteins. Peak incorporation of ["C]glucosamineinto tissue glycoproteins occurred at 3-4 h after the pulseperiod, while peak incorporation for ["C]leucine intotissue proteins occurred at 2-3 h after the pulse period.Secreted radiolabeled proteins and glycoproteins ap-peared in the medium approximately 2-3 h after the

BFIGURE 4 Autoradiographs of epithelium from biopsies cul-tured for 40 min over D-[6-'H]glucosamine containing pulsemedium and then transferred to unlabeled glucosamine con-taining chase medium for 1 h (A) and for 6 h (B). X 1,400.(A) After 1 h over chase medium the radioactivity ingoblet cells is concentrated in the Golgi zone (arrows) inthe lower third of the cells. Absorptive cell cytoplasm islightly but diffusely labeled and the lumen (L) containsno radioactive material (B) After 6 h over chase mediumradioactivity in many goblet cells is found near their luminalsurface (arrow) and the lumen (L) contains much radio-active material. Most of the radioactivity in absorptive cellsis at their apical surface in the brush border region.

TABLE ILocalization of Pulse of Radioactivity in Goblet Cells

of Cultured Rabbit Colon*

Number of goblet cells

Basal Middle ApicalTime third third third Diffuse Unlabeled

h

88 4 0 8 01 72 9 0 19 02 55 22 0 23 04 42 28 5 25 06 21 35 20 23 18 10 14 47 20 9

24 4 6 1 1 27 52

* Biopsies were cultured over medium containing D-[6-3H]-glucosamine for 40 min and then transferred to isotope-freemedium containing an excess of unlabeled glucosamine for anadditional i-24 h of culture. 100 goblet cells were scored foreach time period.

pulse period and continued in more or less linear fashionover the remaining 24 h. After 24 h approximately 40%of the incorporated ["C]leucine and ["C]glucosaminewas recovered as macromolecules which had been se-creted into the medium. These findings allowed selectionof a period of time suitable for measuring secretion oflabeled glycoproteins independent of tissue synthesis.Thus measurement of secretion into the medium wasbegun 3 h after completion of pulse labeling. At thispoint in time peak incorporation of radioactivity into tis-sue glycoproteins had already occurred and labeled mac-romolecules were beginning to be secreted into the incu-bation medium. Medium was then collected for an addi-tional 3 h.

To determine whether acetylcholine stimulated thesecretion of labeled glycoproteins, we incubated rabbitcolonic biopsies in chase medium for 3 h after pulse la-beling with ["C]glucosamine and then exposed the biop-sies to fresh chase medium containing 106 M acetylcho-line (Fig. 6). As seen in the top panel, the amount ofmacromolecular radioactivity remaining in control biop-sies decreased over 3 h (a period extending from 3 to6 h after the pulse labeling). However, exposure of biop-sies to acetylcholine resulted in a much greater fall intissue radioactivity. Concomitantly, as seen in the bottompanel of Fig. 6, the amount of radioactive macromolecu-lar material recovered from the medium was markedlyincreased when biopsies were exposed to acetylcholine.As shown in Fig. 7, maximum stimulation of glycopro-tein secretion by acetylcholine occurred with concentra-tions of 1V-106 M and acetylcholine-stimulated secre-tion was completely blocked by atropine. Although acetyl-choline stimulated secretion of labeled glycoproteins byrabbit colonic biopsies, acetylcholine over the same dose

Glycoprotein Synthesis and Secretion by Colonic Biopsies 549

range had no effect on overall incorporation of glucosa-mine into glycoproteins when examined under continuousculture conditions for 24 Ii (Table IT). In contrast tothe effects of acetylcholine, cycloheximiiide failed to in-hibit secretion of previously synthesized glycoproteinsas demonstrated by pulse-chase experiments (Table III)even though this agent clearly inhibited tissue synthesisof glycoproteins (Fig. 3).

To observe the influence of inflammation on glycopro-tein synthesis and secretion, we obtained rectal biopsiesfrom patients with active ulcerative colitis and comparedthem to rectal biopsies obtained from healthy subjects

100.

o 50

V-

0

o100

IL

zI&I 50

I

a-

1[4C] GLUCOSAMINE

2 4 6 8 24

242 4 6 8

HOURSAFTERPULSEFIGURE 5 Recovery of [14C]oglucosamine and ["4C]leucinefrom rabbit colonic biopsies cultured by the pulse-chasetechnique. Biopsies were pulse labeled with radioactivityfor 1 h and then cultured over chase medium for varyingperiods of time. The amount of radioactivity found insecreted macromolecules (0- -O), tissue macromolecules(TCA-precipitable radioactivity, *-0), or in the TCAsupernate (X- - -X) was then determined. The results areexpressed as the percent of recovered radioactivity presentin the TCA precipitate, the TCA supernate, or in macro-molecules secreted into the medium. The bars represent±1 SEM. Each point represents the mean of 10 to 14hiopsies.

xz

i

0a-C,

a-

I-C.)0a

cr

-J

w-J0

0

7-)4

5

4

3

2

4

3

2

TISSUE

CONTROL- o--o ACETYLCHOLINE A

MEDIUM.0-11

0-10,00,..O 0-11

00,11,

0-01000,

0.00,

0 3

HOURS

FIGURE 6 Effect of acetyicholine on glycoprotein secretionby rabbit colonic biopsies. Biopsies were pulse labeled for 1h with ['4C]glucosamine, then cultured for 3 h over chasemedium, and finally cultured over fresh chase medium (seeMethods) with or without acetylcholine 10' M in thepresence of eserine 3 X 10' M for an additional 3 h. Theamount of macromolecular radioactivity remaining in thetissue and the amount secreted into the medium were thendetermined. The results are expressed as dpm/mg biopsyprotein ± 1 SEM. Each point represents the mean of sixto nine biopsies.

and biopsies of undiseased bowel from a resected seg-ment of sigmoid colon. Studies of incorporation of ["4C1-glucosamine and [14C] leucine revealed essentially thesame patterns of incorporation and secretion of radio-activity for normal human colonic and rectal biopsies asdescribed in Figs. 1 and 2 for rabbit colonic biopsies. Asshown in Fig. 8, biopsies from patients with active ul-cerative colitis incorporated a significantly (P < 0.01)greater amount of [14C]glucosamine into tissue glyco-proteins than did biopsies from normal volunteers andhistologically normal colon obtained at surgery. Whenthe results were expressed in terms of dpm/rg DNAordpm/Iog RNA, the findings were not significantly altered(data not shown). The amount of incorporated ['4C]glu-cosamine recovered from the medium was also signifi-cantly (P

pendently of tissue synthesis. Biopsies wvere pulse la-beled for 1 h in medium containing [14C]glucosamine,and after 2 h in chase medium the biopsies were trans-ferred to fresh chase medium for a 4-h collection period.After the chase period, there was no significant differ-ence (0.2 < P < 0.5) in the macromolecular radioac-tivity remaining in the tissue of eight rectal biopsiesfrom four active ulcerative colitis patients in compari-son to 12 rectal biopsies from four normal volunteers(Fig. 9). However, as also seen in Fig. 9, there was afourfold increase in the amount of radiolabeled glyco-proteins found in the medium of the biopsies from ul-cerative colitis patients compared to that of control biop-sies (P < 0.01). Furthermore, when the results wereexpressed as the percent of total macromolecular radio-activity which was recovered from the medium (Fig. 9),this percentage was significantly (P < 0.01) increasedin the case of ulcerative colitis biopsies. The observedvariation in results occurred not only from patient to

to0xzw0I:a.

_

CD

iL

an3

w

wc:0wco

2

H

0

00a:a:

-j

w-J0

0o A(4

ACETYLCHOLINE

_

PLUS ATROPINE

E lo 12 lN-10(l-A CoN6CET4 ATO2CETYLCHOLINE(MOLAR CONCENTRATION )

FIGURE 7 Effect of increasing concentrations of acetyl-choline on glycoprotein secretion by rabbit colonic biopsies.Biopsies were cultured utilizing the pulse-chase system asdescribed in Methods. Conditions of culture were as de-scribed in the previous figure. The solid line shows theeffect of acetylcholine in doses from 102 to 10' M. Thebroken line shows the effect of acetylcholine, when atropine10' M was also added. Eserine, 3 X 10' M, was includedat all times. The macromolecular radioactivity secreted isexpressed as dpm/mg biopsy protein ±1 SEM. Each pointrepresents the mean of six to nine biopsies.

TABLE IILack of Effect of A cetylcholine on Glycoprotein Synthesis

by Colonic Biopsies*

Total incorporation of[14C]glucosamine

dpm/mg biopsy proteinX 10-3

Control 166±i1910-2 AI Acetylcholine 147 +2310-4 1\1 Acetylcholine 1564910-6 M'I Acetvlcholine 133 i 1410-8 M Acetylcholine 150+1610-1s NI Acetylcholine 166+21

* The medium contained eserine 3 X 10-4 i1\ alone or to-gether with varying concentrations of acetylcholine. Biopsieswere cultured for 24 h with 5.0 ACi of [14C]glucosamine perml after which the amount of radiractivity incorporated intotissue glycoproteins and into glycoproteins recovered fromthe medium was measured. The results show total incorpora-tion into tissue and secreted glycoproteins and are expressedas the mean+4SE for 6-12 biopsies.

patient but also from biopsy to biopsy in those caseswhere two biopsies were obtained from a single patient.There was no apparent correlation between glycoproteinproduction and severity of the lesion as observed histo-logically.

Dialyzed media from two normal and two ulcerativecolitis rectal biopsies were subjected to gel filtration onSepharose 2B. As seen in Fig. 10, two peaks of radio-activity were observed. The first peak, within the voidvolume, most likely represents aggregates of large mo-lecular weight glycoproteins, presumably mucus. Mediafrom two control biopsies had 37% of the secreted radio-activity recovered from the column present in this firstpeak, while media from ulcerative colitis biopsies had64% of the recovered counts in the first peak.

TABLE IIIEffect of Cyclohexirnide on Glycoprotein Secretion by

Rabbit Colonic Biopsies*

Percent of incorporatedBiopsies glucosamine secreted

Control 37%+08.2%1,000 ,ug/ml Cycloheximide 34%+5.6%

* The biopsies were pulsed for 1 h with ['4C]glucosamine.They were then placed in chase medium (see Methods) withor without cycloheximide for 18 h. The percent of incorpo-rated radioactivity secreted into the medium was then de-termined by dividing the amount of radioactivity found inthe medium by the amount found in the tissue plus themedium (total incorporation). The results are expressed asthe mean+SE for six biopsies.

Glycoprotein Synthesis and Secretion by Colonic Biopsies 551

go0

xzhi

-0

0.

0

2

2

0.11

C

hI

I-4

00.

002

za49~000

-i0

U

1201-

100-

so F

60 F

401-

20 I

ULC. COLITIS

m2 NORMAL

000

00

I

f

TISSUE

0

0

0

0

0

0

MEDIUM

2

0

xz3

22

9-

£.

C

9I

w

32 2

IL

09

FIGURE 8 Incorporation of ['4C]glucosamine into tissueand secreted glycoproteins by 13 biopsies from eight activeulcerative colitis patients (hatched bars), in comparison to18 biopsies from five normal volunteers and two surgicallyobtained normal colon specimens (open bars). Bars showmean values in dpm/mg biopsy protein and each circlerepresents one biopsy. All biopsies were cultured over ["4C]-glucosamine continuously for 24 h.

DISCUSSION

Biopsies of rabbit colon cultured for 24 h maintainedan intact and morphologically near normal mucosa asdetermined by light microscopy. Comparable resultswere previously obtained with biopsies of human rectalmucosa (7). Sustained viability was confirmed by thefact that cultured mucosal biopsies from rabbit colonand human rectum incorporated leucine into proteinsand glucosamine into glycoproteins at a steady rate for24 h. This incorporation probably represents macro-molecular synthesis rather than nonspecific attachmentof radioactive precursors since incorporation was almostcompletely prevented by the presence of cycloheximide,which prevents the transfer of amino acids from amino-acyl tRNA to a growing protein chain (13). However,cycloheximide had only a modest inhibitory effect onglycoprotein synthesis except when biopsies were pre-incubated with cycloheximide for at least 6 h beforeculturing with [14C]glucosamine. During glycoproteinsynthesis, the peptide portion of the molecule is synthe-sized first, after which the sugar moieties are incorpo-rated (1). Thus, it is only after prolonged exposure tocycloheximide that the incorporation of glucosamine intoglycoproteins is halted (14, 15), presumably because of

0 0

0

0

. 01

0

s 0*TS

TIMUE

100 _

o 0soz

I ~ ~~~~~00 9

12002&g2 _

BU

NORMAL

8COLIE

8.

to

FIGURE 9 Secretion of prelabeled glycoproteins by nine bi-opsies from four active ulcerative colitis patients (hatchedbars) in comparison to 12 biopsies from four normal volun-teers (open bars). Each point represents one biopsy. Bi-opsies were pulse labeled for 1 h with ['4C] glucosamine.After a 2-h chase period (see Methods), biopsies weretransferred to fresh chase medium and the medium col-lected for 4 h. The results in the first two sets of barsare expressed as the mean dpm/mg protein of ["C]glu-cosamine incorporated into glycoproteins remaining in thetissue or secreted into the medium. The third set of barsrepresents the percent of incorporated [14C] glucosaminerecovered from the medium: 100 X (dpm/ml protein in me-dium) /(dpm/mg protein in medium + tissue).

60-- VOID VOLUME-Aml MEDIUMFROM:

0 l ULC. COLITIS °-°-°< 40 N M°z20 t 8I0

Z 20hi

0.ibW6 12 18 24 30 36

ML OF EFFLUENTFIGURE 10 Elution pattern obtained from dialyzed mediasubjected to gel filtration on Sepharose 2B. Data areplotted as percent of total number of counts recoveredfrom the column. The broken line indicates mean resultsfor media of two biopsies from two ulcerative colitis pa-tients and the solid line indicates mean results for mediaof two biopsies from two control subjects.

552 R. P. MacDermott, R. M. Donaldson, Jr., and I. S. Trier

the depletion of either peptide precursors or the re-quired glycosyl transferases.

Total incorporation of radioactivity into macromole-cules by cultured colonic biopsies was linear with timewhen biopsies were continuously incubated in the pres-ence of either "C-labeled leucine or glucosamine. Never-theless, the pattern of distribution of incorporated radio-activity was quite different depending upon which la-beled precursor was used. When biopsies were incubatedwith "C-labeled leucine, both the amount of incorporatedradioactivity extracted from biopsy tissue and the amountof macromolecular radioactivity recovered from the me-dium increased steadily over the 24-h period. Through-out the incubation, the amount of incorporated radio-activity found in the tissue greatly exceeded that presentin the medium. Thus, most of the newly synthesizedprotein containing labeled leucine remained within thebiopsy, while only a small but relatively constant propor-tion of this newly synthesized protein was released intothe incubation medium. In contrast, colonic biopsies in-cubated with "C-labeled glucosamine incorporated thisradioactivity into glycoproteins which for the most partwere ultimately secreted into incubation medium. Indeed,by the end of 24 h more macromolecular radioactivitywas recovered from the incubation medium than wasextracted from the tissue. This preferential secretion ofnewly labeled glycoproteins is consistent with previousobservations by Forstner (3) who found that intestinalmucosa incorporated "C-labeled glucosamine into threeglycoprotein pools: a large, poorly labeled pool of struc-tural glycoproteins with a relatively slow turnover rateand two smaller, more intensely labeled pools whichturned over more rapidly, one present in the glycocalyxand the other secreted into the intestinal lumen.

When examined by autoradiography after pulse la-beling, cultured biopsies were found to synthesize andsecrete glycoproteins at cellular sites similar to thosepreviously described in the intestine of intact animals(2, 16). In addition, both the biochemical data and theautoradiographic studies were consistent with the time-course observed in previous in vivo studies (2, 3). Theinterval between exposure of intestinal mucosa to la-beled glucosamine and initial secretion of labeled gly-coproteins was 3-4 h (2, 3).

During pulse-chase experiments the amount of la-beled glucosamine incorporated into acid-precipitableglycoproteins continued to increase during the chase pe-riod and reached a maximum 3-4 h after the initial pulse.This gradual increase in macromolecular radioactivitywas accompanied by a decrease in dialyzable, acid-solubleradioactivity extracted from biopsy tissue. Presumablythis labeled material was in the process of being incor-porated into glycoproteins. Other investigators haveshown this acid-soluble form of radioactivity to be pre-

dominantly UDP-N-acetylhexosamine (15). However,the ["C]glucosamine might also be converted into galac-tose, fucose, galactosamine, sialic acid (17), and to aslight extent into leucine (14) before it is incorporatedinto glycoprotein molecules.

Acetylcholine markedly increased the rate of secretionof newly synthesized glycoproteins by biopsies of rabbitmucosa, an effect which was readily blocked by atropine.Thus, isolated colonic biopsies respond to cholinergicstimulation in a manner analogous to that of intact colonwhich is known to secrete goblet cell mucus in responseto vagal stimulation or parasympathomimetic drugs( 18) .

The responses of colonic biopsies to acetylcholine andcycloheximide suggest that in organ culture glycopro-tein synthesis and secretion can be clearly differentiated.Acetvlcholine markedly stimulated secretion of newlysynthesized glycoproteins, but had no effect on glyco-protein synthesis. In contrast, cycloheximide impairedglycoprotein synthesis, but had no effect on secretion.

Increased colonic production of mucus has been fre-quently observed in patients with ulcerative colitis andother inflammatory disorders of the colon (18). Wetherefore cultured rectal biopsies from patients withactive ulcerative colitis to determine whether the inflam-matory state influenced the biosynthesis and secretionof glycoproteins. The results indicated that both bio-synthesis and secretion were substantially increased inbiopsies from patients with colitis when compared tothose from healthy volunteers. The increased incorpora-tion of ["Ciglucosamine into tissue glycoproteins isconsistent with the finding (19) that cultured biopsiesfrom patients with ulcerative colitis have increased ratesof epithelial cell turnover and migration, which wouldrequire increased rates of glycoprotein synthesis forincorporation into new cells. In addition, the increasedsynthesis and secretion of glycoproteins observed in vitroin cultured biopsies reflects the increased mucus pro-duction observed in vivo in patients with ulcerativecolitis. Not only was there an increased secretion of["Ciglucosamine-labeled glycoproteins, but the elutionpattern of the dialyzed media fractionated by Sepharose2B gel filtration suggested an increase in the proportionof large molecular weight glycoproteins. The elutionpatterns obtained showed only a change in the relativeamounts present in the two peaks. No new peaks wereobserved to suggest an obvious change in glycoproteintype, a finding consistent with previous studies indicatingthat the composition of rectal mucus is similar in normalsubjects and patients with ulcerative colitis (20).

ACKNOWLEDGMENTSThe authors are grateful to Ms. Carol Schaier, Mr. DavidSmiley, Ms. Pamela Colony, and Ms. Elaine Markezin fortechnical assistance, Dr. Elihu Schimmel of the Boston

Glycoprotein Synthesis and Secretion by Colonic Biopsies 553

Veterans Administration Hospital for his help in locatingsuitable study patients, and Dr.. Lester Williams of BostonUniversity Hospital for providing the segments of normalcolon.

This research was supported by U. S. Public HealthService grants AM05025, AM11867, and AM17537.

REFERENCES1. Kent, P. W. 1967. Structure and function of glycopro-

teins. Essays Biochem. 3: 105-151.2. Neutra, M., and C. P. Leblond. 1966. Synthesis of the

carbohydrate of mucus in the Golgi complex as shownby electron microscope radioautography of goblet cellsfrom rats injected with glucose-H3. J. Cell Biol. 30:119-136.

3. Forstner, G. G. 1970. (1-14C) glucosamine incorporationby subcellular fractions of small intestinal mucosa.Identification by precursor labeling of three functionallydistinct glycoprotein classes. J. Biol. Cheem. 245: 3584-3592.

4. Kim, Y. S., J. Perdomo, and J. Nordberg. 1971. Glyco-protein biosynthesis in small intestinal mucosa. I. Astudy of glycosyltransferases in microsomal subfrac-tions. J. Biol. Chem. 246: 5466-5476.

5. Browning, T. H., and J. S. Trier. 1969. Organ cultureof mucosal biopsies of human small intestine. J. Clin.Invest. 48: 1423-1432.

6. Kagnoff, M. F., R. M. Donaldson, and J. S. Trier.1972. Organ culture of rabbit small intestine: prolongedin vitro steady state protein synthesis and secretion andsecretory IgA secretion. Gastroenterology. 64: 541.

7. Eastwood, G. L., and J. S. Trier. 1973. Organ cultureof human rectal mucosa. Gastroenterology. 64: 375-382.

8. Brandborg, L. L., C. E. Rubin, and W. E. Quinton.1959. A multipurpose instrument for suction biopsy ofthe esophagus, stomach, small bowel and colon. Gastro-enterology. 37: 1-16.

9. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chew. 193: 265-275.

10. Bray, G. A. 1960. A simple efficient liquid scintillatorfor counting aqueous solutions in a liquid scintillationcounter. Anal. Biochent. 1: 279-285.

11. Schneider, WV. C. 1957. Determination of nucleic acidsin tissues by pentose analysis. Methods Enzymol. 3: 680-684.

12. Schmid, P., C. Schmid, and D. C. Brodie. 1963. Thedetermination of the total deoxyribose of deoxyribonu-cleic acid. J. Biol. Chewn. 238: 1068-1072.

13. Goldberg, I. H. 1965. Mode of action of antibiotics. II.Drugs affecting nucleic acid and protein synthesis.Amer. J. Med. 39: 722-752.

14. Spiro, R. G., and M. J. Spiro. 1966. Glycoprotein bio-synthesis: studies on thyroglobulin. J. Biol. Chem. 241:1271-1282.

15. Ikehara, Y., J. Molnar, and H. Chao. 1971. Inhibitionof glycoprotein synthesis by cycloheximnide in liver andEhrlich tumor cells. Biochim. Biophys. Acta. 247: 486-495.

16. Ito, S., and J. P. Revel. 1964. Incorporation of radio-active sulfate and glucose on the surface coat of entericmicrovilli. J. Cell Biol. 23: 44A-45A.

17. Draper, P., and P. W. Kent. 1963. Biosynthesis of in-testinal mucins. 4. Utilization of (1-'4C)glucose by sheepcolonic mucosa in vitro. Biochem. J. 86: 248-254.

18. Florey, H. W. 1970. General Pathology. W. B. SaundersCompany, Philadelphia. 4th edition. 198-214.

19. Eastwood, G. L., and J. S. Trier. 1973. Epithelial cellrenewal in cultured rectal biopsies in ulcerative colitis.Gastroenterology. 64: 383.

20. Soergel, K. H., and F. J. Ingelfinger. 1964. Compositionof rectal mucus in normal subjects and patients withulcerative colitis. Gastroenterology. 47: 610-616.

554 R. P. MacDermott, R. M. Donaldson, Jr., and J. S. Trier