Gut, 1992, 33, 39-43

Production of epithelial cell growth factors by laminapropria mononuclear cells

J R Lowes, J D Priddle, D P Jewell

AbstractThe effects oflamina propria mononuclear cellculture supernatant on epithelial celi DNAsynthesis were studied using cells isolatedfrom patients with inflammatory bowel diseaseand normal controls. Supernatants fromresting and phytohaemagglutinin stimulatedcells were studied and supernatants thatstrongly promoted DNA synthesis werepooled, and growth factor activity partiallycharacterised. The effects of recombinantinterleukins-l, 2, 3, interferon-y, and granulo-cyte macrophage colony stimulating factorwere tested in the same system. Resting laminapropria mononuclear cells produce factorsthat increase DNA synthesis. Production ofthese factors is increased by phytohaemag-glutinin stimulation. No significant differenceswere found in production of these factorsbetween patients with inflammatory boweldisease and normal controls. The molecularweight of the active factor(s) lies in the region31-48 kD. Chromatofocusing produced twopeaks of activity, one in the region pk 5 5 andone around pk 6*4. The activity was heat andacid pH labile. Activity was not destroyed,however, by 005% trypsin. Recombinantgranulocyte macrophage colony stimulatingfactor was a weak stimulus to epithelial DNAsynthesis, interleukin-l was weakly inhibitorybut other cytokines tested did not have anyeffect. Granulocyte macrophage colony stimu-lating factor is probably important in control-ling epithelial cell growth.

Gastroenterology Unit,Radcliffe Infirmary,OxfordJ R LowesD P Jewell

Department ofPharmacology, SouthParks Rd, OxfordJ D PriddleCorrespondence to:Dr J R Lowes, The GeneralHospital, Steelhouse Lane,Birmingham B4 6NH.

Accepted for publication4 March 1991

Gastrointestinal epithelium is in a constant stateof orderly renewal. In inflammatory conditionsaffecting large and small bowel, the crypt cellproduction rate (CCPR) is increased. In ulcera-tive colitis CCPR is increased when the disease isin relapse and to a lesser extent it is also increasedwhen the disease is in histological remission. 2This is the net result of an increased numbers ofcells synthesising DNA, expansion of the pro-

liferative zone, and increased cell turnover rate.There is an increased risk of malignant change inthese diseases and the risk, particularly in ulcer-ative colitis, appears to be related to the extentand the duration of the disease.3The control of cell growth is complex and

multifactorial. The roles of hormones, peptideregulatory factors such as epidermal growthfactor and insulin like growth factors have beenextensively investigated. The relationshipbetween epithelial cells and immune cells of thelamina propria is attracting increased attention.4It has been suggested that cytokines releasedfrom intestinal mononuclear cells are responsiblefor the increased expression of class II major

histocompatibility proteins on epithelial cellsand soluble factors released from T cells areresponsible for this increase in cell turnover.5Continued exposure to such growth factors maybe responsible for the increased risk ofmalignantchange.

This study measures the effects on epithelialcell turnover by measuring incorporation ofbromodeoxyuridine into the epithelial cell lineHT29 after exposure to supernatant from cultureof lamina propria mononuclear cell, or knownas cytokines. Attempts have been made tocharacterise growth promoting factors producedafter culture of intestinal lamina propria mono-nuclear cells from mucosal tissue affected byulcerative colitis, Crohn's disease, or histologic-ally normal.

Methods

PATIENTSFresh surgical specimens were obtained frompatients undergoing intestinal resection forulcerative colitis (11), Crohn's disease (three),colonic neoplasia or chronic constipation (9).Where possible mucosa from inflamed and non-inflamed sites were compared in specimens frompatients with distal ulcerative colitis. In patientswith colon cancer, mucosa was taken >5 cmfrom a tumour. Patient details are listed in theTable.

LAMINA PROPRIA MONONUCLEAR CELLSLamina propria mononuclear cells wereobtained by the method of Bull and Bookman.6Briefly, surgical resected intestine was scrapedfree of excess mucus and the mucosa dissectedfrom the muscularis propria. Approximately30 mmx 5 mm mucosal strips were washed oncein 1 mM dithiothreitol (Sigma, Poole, UK) for 20minutes at room temperature and then epithelialcells were removed by shaking at 37°C in 5 mMethylene diamine tetra-acetic acid for 30minutes. This was repeated three times withintermediate washes in calcium and magnesiumfree Hank's balanced salt solution. The resultingfragments were then minced and digested inRPMI 1640 containing 10% fetal calf serum(Gibco, Paisley, Scotland) and collagenase from

Patient details

Disease group Nornal UC Crohn's

n 9 11 3Age (median) 65 42 37Range 50-77 5-65 30-46Drug therapy nil iv steroids iv steroids

sulphasalazine (5)

39

on Septem

ber 4, 2021 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.1.39 on 1 January 1992. Dow

nloaded from

Lowes, Priddle,Jewell

clostridium histolyticum (BCL, BoehringerManheim GmbH, W Germany). Digestion con-tinued for three hours (1 mg/ml collagenase (15))or 16 hours (0 3 mg/ml collagenase (eight)). Theresults from long and short digestion times werecomparable and have therefore been combined.The digest was then passed through a nylonmesh (200 gum) and the mononuclear cellsseparated by density centrifugation on Ficoll-paque (Pharmacia, Milton Keynes, UK). Theresulting preparation of mononuclear cells wereresuspended at 106 viable cells/ml in RPMI 1640,with 10% fetal calf serum, supplemented withgentamicin 1 [ig/ml. Viability was always>95%).Mononuclear cells were cultured for 72 hours

in 24 well plates (Linbro, Flow Laboratories,Rickmansworth, UK), in 1 ml aliquots (10[tg)/m, phytohaemagglutinin (WellcomeLaboratories, Beckenham, UK). Pilot experi-ments had shown that growth factor productionwas maximal after 72 hours growth in thepresence of lectin. Supernatants were subse-quently harvested, spun and filtered through a22 [im filter, frozen and stored at -20°C forfuture testing. Freezing and storing at -200C hadno detectable effect on stimulation index for up tosix months.

EPITHELIAL CELL LINEThe human colonic adenocarcinoma cell lineHT29 was obtained from Dr P Brandtzaeg,Oslo, Norway. The cells were grown in 25 cm2T-Flasks (Linbro, UK), the cells were seeded at106/flask and passaged weekly in Leibowitz-15medium (L-15) (Flow Laboratories, UK),supplemented by 10% fetal calf serum, 2 mMglutamine, and antibiotics.

EPITHELIAL CELL PROLIFERATION ASSAYThis was adapted from an ELISA techniquedeveloped for the measurement of leucocyteproliferation.7HT29 cells were seeded 5 x 104/ml, 200 ,ul/

well in 96 well plates in L-15 medium. After 48hours the medium was removed and 50 itimononuclear cell culture supernatant togetherwith 150 1t medium was added to the wells.Control wells received 50 >t RPMI 1640 and10% fetal calf serum from the same batch used inthe culture of the mononuclear cells. After afurther 24 hours, 50 Id 5 x 10-5 M bromodeoxy-uridine was added to the wells. After another 24hours the plates were washed four times withHank's balanced salt solution and air dried at370C for three to six hours. The cells were thenfixed by the addition of 200 jtl 95% ethanol(BDH, Poole, UK) for 10 minutes, emptying thewells and air drying at room temperature for 20minutes. The DNA was denatured by heating to70°C for 45 minutes in 95% formamide in 0- 15Msodium salt citrate buffer (Sigma, Poole, UK).The wells were then washed five times in pH 7 4,0-13 M phosphate buffered saline (PBS) 0-1%Tween 20 (Sigma, UK). The monoclonal anti-body Bu2Oa, directed towards the bromodeoxy-uridine incorporated into single stranded DNA,was a generous gift from Dr D Mason, John

Radcliffe Hospital, Oxford. This was used as aculture supernatant diluted 1:50 in PBS Tween.Two hundred microlitres of antibody was addedto each well and incubated at room temperaturefor 30 minutes. The plate was then washed fourtimes with PBS Tween and a second layerantibody of peroxidase conjugated rabbit anti-mouse serum (Dako, High Wycombe, UK) wasused. After washing four times with PBS Tweenthe peroxidase substrate orthophenyline diamine(OPD, Dako UK) was added at a concentrationof 0-2 mg/ml. The plate was then incubated for20 minutes in the dark at room temperature andthe reaction terminated by the addition of 100 RI2 M sulphuric acid. The coloured reactionproduct was read in a Multiscan plate reader at492 nm. The amount of cellular protein/well wasquantified by a modification of the Bradfordtechnique.8 The plates were washed four timeswith methanol, air dried for four hours at 37°C,30 >t Bradford reagent (Biorad Laboratories,Watford, UK) added and the plates shaken for20 minutes, 120 p1 water added and the platesshaken for a further 20 minutes. The opticaldensity at 595 nm was then measured in theMultiscan plate reader.

GROWTH FACTOR ACTIVITYGrowth' factor activity was expressed as incre-ment in optical density at 492 nm against control.

CYTOKINESHuman interferon-y was a generous gift of Dr GScott, Wellcome Biotechnology, Kent, and was aculture supernatant from a Chinese hamsterovary cell line transfected with recombinanthuman interferon-y gene. Recombinantinterleukin-1l (IL-1) and interleukin-2 (IL-2)were purchased from Koch-Light UK. Recom-binant interleukin-3 (IL-3) and granulocytemacrophage colony stimulating factor (GM-CSF) were the generous gift of Glaxo Research,Geneva, Switzerland.

GEL CHROMATOGRAPHYSupernatants that produced an increment of 0-2OD492 units were considered to be 'active.'Active supernatants were pooled, desalted on aSephadex G25 column and lyophilised. Theresulting powder was redissolved in a minimumvolume of PBS and fractionated on a TSKG2000SW column (Pharmacia-LKB, UK) usingan LKB HPLC pump 2150 and fractioncollector.

Active fractions in the DNA synthesis assaywere pooled, desalted, lyophilised and passedonto a chromatofocusing column (Pharmacia-LKB) and eluted with polybuffer 74 (Pharmacia-LKB).

PHYSICAL TREATMENTSActive supernatants were placed in boiling waterfor 30 minutes. Five millilitres of active super-natant was passed onto a sephadex G-25 columnequilibrated with 0-1 M glycine buffer pH 2 andeluted at five minutes, six hours, and 24 hours.

40

on Septem

ber 4, 2021 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.1.39 on 1 January 1992. Dow

nloaded from

Production ofepithelial cellgrowthfactors by lamina propria mononuclear cells

A

Inflamed Non-Inflamed llealUlcerative colitis Ulcerative colitis Crohn's disease

*p < 0 05 (Mann-Whitney U test)Figure 1: Production ofepithelial growth factor by resting and activated lamina prmononuclear cell culture supernatants. OD=optical density.

0o1 1 10 100 1000

Phytohaemagglutinin (,ug/ml)Figure 2:Phytohaemagglutinin doesnot have any detectable effect Three millilitres of active suj

alone on HT29 growth. incubated with 0-05% Trypsin forOD=optical density. 370C.

pernr30n

STATISTICAL ANALYSISUnless stated otherwise experimecarried out in quintuplicate and pointsmeans (SD). Comparisons between,non-parametric data were made using'Mann-Witney U test and paired t-testsfor parametric data.

ResultsThe supernatants of resting lamin;mononuclear cells (LPMNC) produce(cant increase in DNA synthesis in thecell line. Stimulation of the LPMphytohaemagglutinin produced a furtbcant rise in DNA synthesis. Thercdifference, however, between sui

produced from culture of cells obtaidiseased or control mucosa either in t

state or after lectin stimulation (Fig 1). Phyto-haemagglutinin alone had no effect on the rate ofDNA synthesis (Fig 2). After concentration of212 ml of active supernatant by desalting andlyophilisation, the lyophilate was redissolved in1 ml PBS and fractionated on a TSK G2000SWcolumn, 35 x045 ml fraction were collected.The column was calibrated using proteins ofknown molecular weight. Active fractions werefound in the 31-48kD range and fractions ofmolecular weight less than 20 kD were inhibitoryto DNA synthesis (Fig 3).

Active fractions from the high performanceliquid chromatography gel column were concen-

trated as previously and applied to a chromato-focusing column. The active fractions werefound to elute around pH 6-4 and 5 5 as seen in

* Figure 4. Biological activity was destroyed byD~ exposure to heat (p<0 05), pH 2 (p<0 05) but

Control not 0-05% trypsin as shown in Figure 5.

RECOMBINANT CYTOKINESropria Interferon-y, interleukin-2, interleukin-3 pro-

duced no significant effect on DNA synthesissuing awide range ofconcentrations. Interleukin-113 produced a slight decrease in the rate ofDNAsynthesis that only achieved statistical signifi-cance at a dose of 50 IU/ml (p<0 05). Granu-locyte-macrophage colony stimulating factor wasa stimulator of DNA synthesis producing amaximal increase in DNA synthesis at a dose of100 U/ml. Higher doses did not produce anygreater effect suggesting that the stimulus wassaturated at about this level (Fig 6).

DiscussionThese results show that human lamina proprialymphocytes produce factor(s) that increaseDNA synthesis in a colonic epithelial cell line.We have not shown any difference in the produc-tion of these factors by cells obtained from eitherinflammatory bowel diseases or control tissuefrom normal mucosa (albeit from colonic mucosathat had neoplastic change at a distant site).

atant was Production of active supernatants markedlyminutes at increased after lectin stimulation of the mono-

nuclear cell preparation, although in two cases ofactively inflamed ulcerative colitis, net produc-tion decreased after lectin stimulation. Themolecular weight of the growth factor activity

nts were lies in the range 31-48 kD and activity was found,represent to lie in two peaks with pK of 5-5 and 6-4groups of approximately. The biological activity of thetwo-tailed supernatant could be destroyed by exposure towere used acid pH and to high temperature, but not 005%

trypsin. This is consistent with activity lying ina native polypeptide. Using recombinantcytokines this effect was partially reproduced byrecombinant granulocyte-macrophage colony

a propria stimulating factor.d a signifi- Increased quantities of colony stimulatingepithelial factors including granulocyte-macrophageiNC with colony stimulating factor and interleukin-3 haveier signifi- been found in lamina propria mononuclear celle was no isolated from patients with ulcerative colitis and)ernatants Crohn's disease.9 Synthesis by lamina propriained from mononuclear cell culture supernatant. Gelhe resting chromatography showed that the supernatants

10

08

C,

a)

00a)

00

-5C

00-cM0)0

c--0)

00

41

on Septem

ber 4, 2021 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.1.39 on 1 January 1992. Dow

nloaded from

Lowes, Priddle,Jewell

0-4

0-3-

0.2

70

C40

0N

0

0.1ilI

*0 0,

-0*1

-0-2-

Figure 3: Stimulationindices ofhigh performanceliquid chromatographyfractions generatedfrompooled and concentratedsupernatant.

-0 3

III

T Ti.1.

*.T T

IT

IITT

.-

E m~~~'a

(00 EDBS s A.0inCae -aV

T'

t- X4mt en X Ew C;>m - N X st m w - eo 0 0 - N m et L -tD mX IXO mX1-- _-- _-- _- 'r-'NNe N NN N eNN C en X e Cn)

Fraction

ir

20 40Fraction

60 80

contain both growth inhibitory and growth pro-moting activities. No significant differences weredetected in the activity of supernatants fromcontrol subjects and either Crohn's disease orulcerative colitis subjects.

Although differences in epithelial growthfactor production between the different popula-tions were not detectable using this in vitrotechnique, small in vivo differences may beimportant in controlling crypt cell productionrate. In the experiments carried out here we arenot comparing like with like as the laminapropria lymphocyte subsets are altered ininflammatory bowel disease. We know that invitro isolation selectively enriches certain popu-lations and alters the function of isolated cells. "'

Therefore conclusions about differencesbetween control and inflammatory bowel diseasesubjects must be made with caution.

The observation that phytohaemagglutininstimulation of the cells increases growth factorproduction suggests that activation of the laminapropria mononuclear cell population leads toepithelial cell growth factor release, we have notfurther characterised the cell type producinggrowth factor, but the observation that phyto-haemagglutinin leads to an increase in produc-tion suggests that T cells may be important.

Culture supernatants will contain a widevariety of biologically active substances. Precisecharacterisation of the stimulatory factor has notbeen achieved but, of the recombinant growthfactors tested, only granulocyte-macrophagecolony stimulating factor showed a proliferativeeffect. The cytokines used, however, were pro-duced by recombinant DNA technology inbacterial plasmids. Thus they had not undergoneeukaryotic glycosylation and have potentiallydifferent biological activities. Using the same cellline in an assay of class II HLA antigen induc-tion, we have shown that it is approximately 100-fold more sensitive to interferon-y produced in aeukaryotic system than other workers havefound using an E coli derived product. ' 12

Interferon-y, interleukin-2, and interleukin-3did not have any effect on DNA synthesis. Thesecytokines were used across a wide range ofconcentrations and in the case of interferon-y atleast, HT29 cells have been shown to be sensitiveto induction of class II HLA molecules andis therefore likely to express interferon-yreceptors.'2 It has been suggested thatinterferon-y may be important in increasingepithelial cell growth rate in the murine model ofintestinal damage, and that monoclonal anti-bodies to interferon-y may abrogate the pro-liferative phase of an intestinal graft versus hostreaction.'3 Interferon-y is usually considered tohave growth inhibitory properties.'4 From the invitro data presented here it would seem unlikelythat interferon-y has a primary role in controlling

8-

7 -

M 6-

5-

0

Figure 4: Stimulationindices offractions generatedby a chromatofocusingcolumn.

\1ii \s1

4 I I I I

42

a0a?r000E-=M.+U N

on Septem

ber 4, 2021 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.1.39 on 1 January 1992. Dow

nloaded from

Production ofepithelial cellgrowthfactors by lamina propria mononuclear cells 43

06-

0-5

c 0-4-......'0

03_~~~~~~~~~~~~~~~~~~~~~~~~~.............CN .......................

a.)

<. 02 .T3.03 a 1 T ; ~~~~~~~~~~~~~~~~~~~~~~~. . . . . . . . . . . . . . . ....0 o ......... ,., , , , ...... ,,, .. ................~................

a) 0-2

o.l ..........~~~~~~~~. .. . . . . .. T

20

w -0-1

202-

~~.110 10 0000o

Granulocyte-macrophage0colony stimulating factor.(U/ml)

08.....

06~~~~~~~~~~~~~~~~..O.0. .4.

02......

0.0

Grarfec nulct-arpaeclnstimulanidxotrpn,H2 ndhatinfractore(/mI

Figure 6: Stimulation indexafter incubation withgranulocyte-macrophagecolony stimulatingfactor.

epithelial cell DNA synthesis although it mayhave an effect on crypt cell production rate byother mechanisms or be important in synergisingwith or controlling the release of other potentcytokines in vivo. This has parallels in the studyof keratinocyte growth regulation. Direct intra-dermal injection of interferon-y results in anincrease in keratinocyte proliferation,'5 butused in isolation in vitro it inhibits growth.'6Granulocyte-macrophage colony stimulatingfactor and interleukin-3, however, have beenfound to stimulate keratinocyte growth in vitro. 16An attempt was made to abrogate the increase inDNA synthesis observed by lamina propriamononuclear cell supernatant by titrated incuba-tion with anti-interferon-y antiserum and anti-granulocyte-macrophage colony stimulatingfactor antibody, but it was found that sheepantiserum or murine monoclonal antibodyalone was a potent stimulus to epithelial DNAsynthesis (data not shown).

Granulocyte-macrophage colony stimulatingfactor has been identified initially as important inthe control of haematopoietic stem clonal expan-sion.'7 It has not previously been shown to havean effect on the growth of epithelial cells. Thephysical characteristics of the factor produced in

vitro by the lamina propria cells are consistentwith those of granulocyte-macrophage colonystimulating factor, which has a great degree ofvariability in its degree of glycosylation produc-ing quite a wide range in its observed molecularweight.'8 The observation that lower molecularweight fractions of the supernatant inhibitedDNA. synthesis suggest that the increaseobserved with crude supernatants may be the neteffect of opposing actions of several cytokines.The other cytokines that we have tested,interferon-y, interleukin-2, and interleukin-3have no demonstrable effect on DNA synthesis,but interleukin- 1 was found to be weakinhibitor of DNA synthesis. The apparent lowmolecular weight of the inhibitory factors wouldbe in keeping with other known growthinhibitory factors such as arachadonic acid meta-bolites or tumour necrosis factor produced bymonocytes.These observations have implications for the

control growth of the epithelium in physiologicaland pathological states. It may relate epithelialgrowth control to the amount of antigenic stimu-lation of the immune cells in the underlyinglamina propria.This work was presented to the Autumn meeting of the BritishSociety of Gastroenterology, 1989 in Dublin and published inabstract form in Gut. This work was generouslv supported bySandoz UK.

1 Allan A, Bristol JB, Williamson RCN. Crypt cell productionrate in ulcerative proctocolitis: differential increments inremission and relapse. Gut 1985; 26: 999-1003.

2 Serafini EP, Kirk AP, Chambers TJ. Rate and pattern ofepithelial cell proliferation in ulcerative colitis. Gut 1981; 22:648-52.

3 Lennard-Jones JE, Morson BC, Ritchie JK, Williams CB.Cancer surveillance in ulcerative colitis: experience over 15years. Lancet 1983; ii: 149-52.

4 Brandtzaeg P, Sollid L, Thrane PS, Kvale D, Bjerke K, ScottH, et al. Lymphoepithelial interactions in the mucosalimmune system. Gut 1988; 29: 1116-30.

5 Macdonald TT, Spencer J. Evidence that activated mucosal Tcells play a role in the pathogenesis of enteropathy in humansmall intestine. J Exp Med 1988; 167: 1341-9.

6 Bull DM, Bookman MA. Isolation and functional characteri-zation of human intestinal mucosal lvmphoid cells. J ClinInvest 1977; 59: 966-74.

7 Magaud J-P, Sargent I, Mason DY. Detection of human whitecell proliferative responses by immunoenzymatic measure-ment of bromodeoxyuridine uptake. J Immunol Methods1988; 106: 95-100.

8 Baumgarten H. A simple microplate assay for the determina-tion of cellular protein. J Immunol Methods 1985; 82: 25-37.

9 Hapel A, Pullman WJ, Doe WF. Colony stimulating factorproduction by intestinal lamina propria cells is increased ininflammatory bowel disease. Gastroenterologv 1988; 94:A361.

10 Gibson PR, Hermanowicz A, Verhaar HJJ, Ferguson DJP,Bernal AL, Jewell DP. Isolation of intestinal mononuclearcells: factors released which affect lymphocyte viability andfunction. Gut 1985; 26: 60-8.

11 Sollid LM, Gaudernack G, Markussen G, Kvale D,Brandtzaeg P, Thorsby E. Induction of various HLA class IImolecules in a human colonic adenocarcinoma cell line.Scandj Immunol 1987; 2: 175-80.

12 Lowes JR, Radwan P, Priddle JD, Jewell DP. Induction ofclass II MHC proteins on a colonic epithelial cell line. ClinSci 1988; 74 (suppl 18): 31.

13 Mowat A McI, Felstein MV, Parrott DMV. Role of interferonsin immunologically mediated enteropathy. Gut 1988; 29:A1436.

14 Kronenberg LH. Cellular effects of interferon pp 80-2. In:Stiehm ER, moderator. Interferon: immunobiology andclinical significance. Ann Intern Med 1982; %: 80-93.

15 Kaplan G, Nusrat A, Sarno EN, Job CK, McElraith J, PortoJA, et al. Local and systemic effects of intradermal injec-tions of recombinant human interferon-y in lepromatousleprosy patients. AmJ Pathol 1987; 128: 345.

16 Hancock GE, Kaplan G, Cohn ZA. Keratinocyte growthregulation by the products ofimmune cells.JT Exp Med 1988;168: 1395-402.

17 Metcalf D. The molecular biology and functions of thegranulocyte-macrophage colony stimulating factors. Blood1986; 67: 257-67.

18 Nicolas NA, Burgess AW, Metcalf D. Similar molecularproperties of granulocyte-macrophage colony stimulatingfactors produced by different mouse organs in vitro and invivo.JTBiolChem 1979; 254: 5290-9.

on Septem

ber 4, 2021 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.1.39 on 1 January 1992. Dow

nloaded from