Gut, 1992, 33, 324-330

Antigen induced suppression in peripheral blood andlamina propria mononuclear cells in inflammatorybowel disease

H R Dalton, P Hoang, D P Jewell

AbstractUsing an autologous system, suppressor cellfunction to a range of mycobacterial antigensand Kunin antigen of peripheral blood mono-

nuclear cells and lamina propria lymphocyteshas been investigated in normal subjects andpatients with inflammatory bowel disease. Inthe peripheral blood there was reduced antigeninduced suppression in patients with Crohn'sdisease in remission to Mycobacterium para-tuberculosis, purified protein derivative (PPD),Mfortuitum, and Kunin antigens (p<005). Inpatients with ulcerative colitis in remissionthere was reduced antigen induced suppres-sion in the peripheral blood to Kunin antigen(p<0.001), M avium (p<0.01),M nonchromo-genecin, and M fortuitum (p<005). Thephenomenon of antigen induced suppressionwas largelyCD8 dependent, as depletingCD8+cells reduced the effect and the concentrationof soluble CD8 in the culture supernatant wasdirectly related to the suppressor index(r= 0.25, p<0 05). These results are likely to bea true reflection of the cell mediated responseto antigen as patients with a positive Mantouxskin test have a significantly higher suppressorindex to PPD than Mantoux negative subjects(p<OO5). These findings may have signifi-cance in the aetiopathogenesis of inflamma-tory bowel disease. However, a similar effectcould not be shown in the lamina proprialymphocytes of patients having colectomy foractive disease.

Gastroenterology Unit,Radcliffe Infirmary,OxfordH R DaltonP HoangD P JewelCorrespondence to:Dr D P Jewell,Gastroenterology Unit,Radcliffe Infirmary, OxfordOX2 6HEAccepted for publication22 July 1991

The aetiology of inflammatory bowel disease isunknown. There is increased immune activationin both Crohn's disease and ulcerative colitis,particularly at mucosal level. One hypothesis toexplain these immunological changes is thatthere is reduced suppressor cell function ininflammatory bowel disease.

After Hodgson's original description of re-

duced suppressor cell function in the peripheralblood mononuclear cells (PBMC) of patientswith inflammatory bowel disease,' several othersfollowed. Most of these PBMC studies foundsimilar results to Hodgson's.2-' However, some

workers found no difference between normaland control subjects.8 '0 Several studies of sup-pressor cell function of lamina propria lympho-cytes have been reported"-'" with discordantresults.The reasons for these discrepancies could be

several. Differing methods were used in eachstudy, the earlier studies largely using allogeneicsystems. Furthermore, in most of these studieslectin was used to induce the suppressor cell

population. This makes it even more difficult tounderstand the pathophysiological importanceof this in vitro work as, in vivo, antigen(s) will beresponsible for suppressor cell activation. Therehas been only one reported study of antigenspecific suppressor cell function in inflammatorybowel disease in which Ebert et al describedincreased suppressor cell function in inflamma-tory bowel disease to antigen derived fromMycobacterium paratuberculosis. 6

There has been renewed interest in the patho-physiological role of mycobacteria in Crohn'sdisease. 17 This has centred onM paratuberculosis,which causes a granulomatous ileocolitis inruminants. Much effort has been invested inculturing tissue specimens from patients withinflammatory bowel disease for mycobacterialorganisms, and to date M paratuberculosis hasbeen isolated from nine patients with Crohn'sdisease.'8 However, the importance of this is notclear as atypical mycobacteria have also beenisolated from normal subjects and patients withulcerative colitis.'9 Little is known about theeffect of mycobacteria on cellular mechanisms ininflammatory bowel disease, but if mycobacteriaare important in the pathophysiology, it wouldseem logical to expect differences in cellularresponses to mycobacteria in patients with thesediseases compared with normal subjects.A bacterial antigen which has created interest

in ulcerative colitis is the common enterobacterialantigen of Kunin. Kunin antigen shares epitopeswith a colonic mucopolysaccharide antigen towhich haemagglutinating antibodies are foundin patients with this disorder. In addition, anti-bodies to germ free rat colonic epitheliumdetected in the serum of patients with ulcerativecolitis cross react with Kunin antigen.20 Using animmune complex model of colitis in rabbits,Mee' et/al/showed that a chronic colitis developedin animals who had been immunised with Kuninantigen.2'The aim of this study was to investigate

cellular responses to mycobacterial and Kuninantigens in inflammatory bowel disease by study-ing antigen induced suppressor cell function inisolated PBMC and lamina propria lymphocytesof patients with Crohn's disease, ulcerativecolitis, and normal control subjects.

Patients and methods

SUBJECTSTen patients with Crohn's disease (medianage=46 years, range 32-63), nine patients withulcerative colitis (median age 49 years, range26-61), and 10 normal volunteeers (median age

324

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from

Antigen induced suppression in peripheral blood and lamina propria mononuclear cells in inflammatory bowel disease

31 years, range 28-49) were studied. Thepatients with Crohn's disease had their diagnosesconfirmed histologically or radiologically. Allwere in remission at the time of the study: nonewere taking corticosteroid medication but threewere taking sulphasalazine; five patients hadileocolonic disease and five had ileal disease only.The patients with ulcerative colitis had their

diagnoses confirmed by standard clinical, radio-logical and histological criteria. All nine patientshad disease limited to the left side; six weretaking sulphasalazine, two olsalazine, and onemesalazine. None were taking steroids orazathioprine, and all were clinically in remissionat the time of the study. The normal volunteerswere laboratory staff. None had any importantchronic or acute illness and none were taking anyregular medication.

PBMC ISOLATIONPBMC were obtained by Ficoll-isopaque densitycentrifugation. The cells were then resuspendedto a concentration of 1 x 106 cells/ml in culturemedium consisting of RPMI 1640 with HEPESbuffer (25 mmol) and 2 mmol/l glutamine(GIBCO, Paisley Scotland) plus penicillin (100U/ml), gentamicin (50 [tg/ml), and 10% fetalcalf serum.

INDUCTION OF SUPPRESSOR CELLSThreex 106 PBMC were incubated in 25 cmculture flasks for seven days at 37°C in 5% C02,with one of a series of antigens. The antigensused were Johnin purified protein derivative(PPD) (extract fromM paratuberculosis, Ministryof Agriculture, Central Veterinary Laboratory,Weybridge, Surrey, UK), tuberculin PPD(Evans Medical Ltd, Horsham, England, UK),Kunin antigen (kind gift ofProfessorW Bartnik,Warsaw, Poland), M avium 1479, M kansasii 8,M nonchromogenicin R812R, M vaccin R859R,M fortuitum II 1387 (all kindly donated by DrJ L Standford, London, UK). The latter fiveantigens were suspensions of ultrasonicatedmycobacteria made up in M/15 borate bufferedsaline. The Kunin antigen used was a lyophyl-ised extract of ultrasonicated enterobacteriacae.All these antigens were used at a dose of 10 ,ug/ml, except the Kunin antigen, which was used ata dose of 50 [tgIml. These concentrations wereconsidered optimal from dose-response experi-ments performed before the main study.

For each subject, one aliquot of isolatedPBMC was incubated for seven days withoutthe addition of antigen.

COCULTURE EXPERIMENTSAfter seven days the induced cells were ultra-centrifuged at 500 g for 10 minutes. The super-natants were removed and stored at -20°C forlater use and the cells washed three times inculture medium. The viability of the cells wasassessed by the Trypan blue exclusion methodand they were then irradiated with 2500 radsand resuspended to a concentration of 1 x 106cells/ml. Fresh autologous PBMC respondercells were isolated at the same time using Ficoll-

isopaque density centrifugation. A cocultureexperiment was then performed, using thefollowing cell mixtures:

(1) 5x 104 PBMC induced with antigen andsubsequently irradiated were mixed with 5 x 104fresh autologous PBMC plus 2 5 [ig/ml con-canavalin A (suboptimal dose).

(2) 5x 104 PBMC cultured without antigenand subsequently irradiated were mixed with5 x 104 fresh autologous PBMC plus 2-5 [ig/mlconcanavalin A.

(3) 5 x 104 PBMC induced with antigen andirradiated were mixed 5 x 104 fresh autologousPBMC alone.

(4) 5 x 104 PBMC cultured without antigenand irradiated were mixed with 5x 104 freshautologous PBMC alone.The coculture was performed at 370 C in 5%

CO2 in a 96 t-well plate (Cell-Cult, Hounslow,UK) for 64 hours. The total reaction volume ineach well was made up to 150 >t with culturemedium and all experiments were performed intriplicate. Sixteen hours before the end of thecoculture, each well was pulsed with 1iCiof [Methyl-3H] thymidine (Amersham Labora-tories, Amersham, UK). The cells were thenharvested using an automated Titertek cellHarvester (Skatron, Lier, Norway) on filterpaper discs. These were then suspended andcentrifuged in 10 ml of scintillation fluid (Opti-Phase 'Safe', LKB) and counted on a scintil-lation counter.

SUPPRESSOR INDEXThe suppressor index (SI) for each antigen wascalculated using the following formula:

1- (1)-(3) x 100%(2)=(4)

Positive SIs indicated active suppression, andnegative SIs indicated lack of suppression orhelp.

CELL DEPLETION STUDIESSeveral experiments were performed using aCD8+ cell depleted population. The cell deple-tion was performed using CD8+Dynabeads(Dynabeads M-450, Dynal, Oslo, Norway). Thecoculture experiments and SI calculation wereperformed using the same method as for theundepleted cell populations. Adequate depletionof these cell populations was confirmed byimmunofluorescent staining with the appro-priate monoclonal antibody followed by analysisby flow cytometry (FACScan, BectonDickinson).

SOLUBLE CD8 STUDIESSupernatants taken after seven days of culturewith each antigen were assayed for soluble CD8using an enzyme immunoassay test kit (cell freeT8 test kit, T Cell Sciences, Cambridge, MA,USA).

SKIN TESTSSkin tests were performed using tuberculin PPD

325

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from

Dalton, Hoang, Jewell

cells/mi. The lamina propria lymphocytes whichhad been cultured with antigen were thenirradiated (2500 rad). Coculture experimentswere then performed, using the unstimulatedlamina propria lymphocytes as responder cells,and the suppressor index calculated as describedabove. The object of these experiments was toincrease the viability of the lamina proprialymphocytes and to use autologous laminapropria lymphocytes in the responder phase ofthe assay (instead ofPBMC).

Day 3 Day 5

Time

Figure 1: Typical proliferative response ofperipheral bloodmononuclear cells (PBMC) (dpm) to optimal dose ofKunin,Mycobacterium and M vaccin at days 3, 5, and 7.

ETHICSApproval for the study was given by the CentralOxford Research and Ethics Committee.

and Johnin antigens. 0 1 ml of each antigen at a

concentration of 20 ,ug/mI was injected intra-dermally into the velor aspect of each forearm.The test was read after 72 hours and was

regarded as positive if greater than 5 mm ofinduration was present.

LAMINA PROPRIA LYMPHOCYTESAntigen induced suppressor cell function was

also assessed in lamina propria cells. Twenty twosubjects were studied in total, using colonicresection specimens.The control group consisted of macroscopic-

ally normal colonic mucosa from nine subjects(carcinoma n=4, adenoma n=2, idiopathicconstipation n= 2, and diverticular disease n= 1).Seven patients with ulcerative colitis werestudied (severe acute ulcerative colitis n=6,dysplasia n= 1) and six patients with Crohn'sdisease (ileocolonic n=2 ileocaecal stricturen= 4). Six of seven patients with ulcerative colitisand five of six patients with Crohn's disease werereceiving corticosteroids at the time of theresection.The lamina propria lymphocytes were isolated

as previously decribed22 using a modification ofthe method of Bull and Bookman.23 The medianyield ofcells from control subjects was 4 Ox 106/gtissue (range 2 5-9-2 x 106 cells/g), from patientswith Crohn's disease 11 4x 106 cells/g (range5-0-15-Ox 106), and from ulcerative colitis10-1x106 cells/g (range 4-0-13-3x106). Thelamina propria lymphocytes (5x 104) were thenincubated with each antigen for seven days,washed, irradiated, and cocultured with freshautologous PBMC and 2-5 [ig/ml concanavalinA as described above for the PBMC studies.Suppressor indices were then calculated as

described.In some experiments lOx 106 lamina propria

lymphocytes in 3 ml culture medium werecultured with the same concentration of antigensas above. In addition, lOX 106 lamina proprialymphocytes in 3 ml were cultured alone. Afterseven days the cells were centrifuged using a

Ficoll-paque gradient and washed three times inculture medium. The cell viability was assessedusing the Trypan blue exclusion method, and thelamina propria lymphocytes were resuspendedin culture medium to a concentration of 1 x 106

STATISTICSThe data was analysed by the Mann-Witney Uand Spearman correlation tests where appro-priate. A value of p<005 was taken as

significant.

Results

PBMCThe antigens were used to induce the PBMC forseven days. This time was chosen on the basis ofpreliminary experiments which showed an ade-quate proliferative response at day 7 (Fig 1). Thedose of antigen used was chosen on the basisof preliminary dose-response experiments (datanot shown). After seven days the viability of thePBMC was assessed by the Trypan blue exclu-sion method. The viability was greater than 65%in all experiments, and greater than 80% in over

90% of the experiments.Initial experiments were performed using

antigen in both the inducer and responder phase.However, in order to obtain reasonable prolifera-tion, the culture period in the responder phaseneeded to be at least seven days. This resulted inhigh background counts, making calculation of asuppressor index difficult, and also gave prob-lems in cell viability since the total assay time was14 days. Furthermore, using different antigensin the responder phase made comparisons diffi-cult. Finally, a few patients showed little, if any,proliferative response to antigen so that there

E

a)0o

a)U)

m

60 000 -

50 000 -

40 000 -

30 000 -

20 000 -

1 o0000-

-e- Normal

* Crohn's disease

ItXssoU', w- a T

0 5 10 15 20 25

Con A (itg/ml)Figure 2: Proliferative response ofperipheral bloodmononuclear cells (PBMC) (dpm) to varying doses ofconcanavalin A (Con A) after 64 hours.

12 000 -

o10000 -0L

a) 8000-

o 6000-0.(na)

4000-u

m 2000-

0Day 7

326

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from

Antigen induced suppression in peripheral blood and lamina propria mononuclearcells in inflammatory bowel disease

was no result from the responder cells makingthe calculation of a suppressor index impossible.

For these reasons a suboptimal dose of con-canavalin A (2-5 1ig/ml) was used (Fig 2). Thisgave a reproducible proliferative response after64 hours in the fresh autologous PBMC used inthe responder phase.

PBMC SUPPRESSOR ASSAYS (Table I)Patients with ulcerative colitis had reducedsuppression to Kunin antigen (medianSI=-44-3%) compared with normal subjects(+19-4%). This was significant (p<0001).Compared with normal subjects, patients withulcerative colitis also had significantly reducedsuppression to M avium (p<0-01), M fortuitum,andM nonchromogenecin (p<005). Patients withCrohn's disease had significantly reducedsuppression to Johnin antigen compared withnormal subjects and also to PPD, M fortuitum,and Kunin antigens (p<0O05). There was nocorrelation between the antigen stimulatedproliferative responses in the primary culturesand the subsequent antigen induced suppressionindex (Table II).

SKIN TESTSAll the patients with inflammatory bowel diseaseand normal controls had a negative skin testusing 0-1 ml 20 ttg/ml Johnin injected intra-dermally. Eight of 10 normal volunteers, 3 of 9patients with Crohn's disease, and 2 of 7 patientswith ulcerative colitis had a positive Mantouxtest (Fig 3a). Subjects with a positive Mantouxtest had a significantly higher PPD-inducedsuppressor index in PBMC compared withpatients who were Mantoux negative (p<005,Fig 3b).

PBMC CELL DEPLETION STUDIESSatisfactory depletion was achieved in all but oneexperiment. The CD8 depleted populationswere always <5% CD8+, and in two experi-ments <2% CD8+. After depletion ofCD8 cellsthe suppressor index fell in 10 of 14 assays. Inaddition, after the depletion of CD8 cells theamount of soluble CD8 recovered from the

culture supernatant fell to almost negligibleamounts (Table III). The effects of depletingmonocytes, CD2, and CD4 cells were not clearcut (data not shown).

PBMC SOLUBLE CD8From the above results if seemed that antigeninduced suppression requires CD8+ cells.Therefore, concentrations of soluble CD8 weremeasured in the supernatants of the antigendriven inducer phase, harvested at day 7. Ninetysix samples were analysed for soluble CD8,values ranging from 0 to 185 U/ml. There was adirect, although weak, correlation between thesolubleCD8 value in the culture supernatant andthe suppressor index (r=0-25, p=<005, Fig 4).This correlation was more pronounced inpatients with inflammatory bowel disease,compared with controls.

This is further evidence that CD8+ cells areinvolved in the phenomenon of antigen inducedsuppression.

LAMINA PROPRIA LYMPHOCYTESLmina propria lymphocytes from six patientswith ulcerative colitis, five patients with Crohn'sdisease, and six normal controls were assessedfor suppressor cell function using autologousPBMC as responder cells. There was no signifi-cant difference in the suppressor indices for any

100 -

x

0C,)Co0.

C)

0-

-100 -

A

0.0

9

do 090

00

0

80

Normal Crohn's Ulcerativesubjects disease colitis

* Mantoux +ve0 Mantoux -ve

TABLE I Median (range) suppressor indices (%) to the antigens in normal subjects, Crohn'sdisease and ukerative colitis.

Normal Crohn's Ulcerativesubjects disease colitis

Johnin +24-4 -27-4* +16-7(-212 to +24-1) (-13Oto +378) (-43-0to+31-6)

PPD +12-3 -20.3* +4-8(-50 4 to +43 5) (- 149 to + 13-6) (-57.7 to +25-6)

Mycobacterium avium +7-4 -9 3 -37.0**(- 10-1 to +45 8) (-426 to +22 4) (-51-5 to +24 6)

M kansasii +4-7 -21-6 -6-7(-32-2 to +45 0) (- 154 to +27 9) (-71-4 to +40 9)

Mnonchr +19-8 -0 5 -3 5*(-38-9 to +41-2) (-317 to +36 3) (-88-8 to +42-6)

Mvaccin +20-5 -8-1 -9-5(- 11-2 to +44 6) (-375 to +42 9) (-57 1 to +40 5)

Mfortuitum +12-0 -19.1* -7.9*(-20 5 to +55 8) (-371 to +32 0) (-28-8 to +40 8)

Kunin + 19-4 -20-4* -4443***(-44-2 to +35 6) (-528 to +21 0) (-60-0 to +4-4)

Data analysed by Mann-Whitney U test to analyse degree of reduced suppression in CD and UC:*p<0.05, **p<0-01 and ***p<0-001

100 -

xa,*0,

o

nCO0(A,aaCf)

0-

-100 -

-200 -

+ve> -ve p <0-05B

0ogoo-0-0 0

0

80

+ve -ve

Mantoux status

Figure 3: Tuberculin skin tests and purified protein derivative(PPD) induced suppression. (A) Mantoux status in normalsubjects, Crohn's disease, and ulcerative colitis compared withtheirPPD induced suppressor index (%). (B) Mantoux statusin all subjects compared with theirPPD induced suppressorindex (%).

-200

327

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from

Dalton, Hoang,Jewell

TABLE II Comparison ofmedian proliferative responses (PR) (dpm) of5 x 106 peripheralblood mononuclear cells (PBMC) to optimal dose ofantigen after seven days' culture and thesubsequent suppressor index (SI) (%) in normal subjects.

Subject I Subject 2 Subject 3

PR SI PR SI PR SI

PBMC alone 1509 - 731 - 1260Kunin 6955 +27-1 28901 +15-3 10603 + 9-8Johnin 12891 +34 5 6756 +16-0 - -

PPD 6435 +32-0 9637 -42-1 15037 +17-7Mycobacteriumavium 7236 +45 8 1112 - 3-3M kansasti 4298 +26-8 1706 -26-3M nonchr 3813 +28-1 780 + 41 - -

Mvaccin 3286 +32-7 1139 +15-5 -

Mfortuitum 2853 +42-4 1079 +13-9 -

p=NS

TABLE III Suppressor indices (%) induced by vanous antigens with undepleted peripheralblood mononuclear cells and a population ofCD8 depleted peripheral blood mononuclear cells.The supernatant soluble CD8 (Ulml) is shown in parentheses.

Experiment no Antigen Undepleted CD8 depleted

1 Johnin +10-5 - 1-7PPD -17-9 -29-4Kunin + 2-8 + 2-7Mycobacteriumfortuitum +10 1 -12-0

2 Johnin +33-1 +119PPD +20 9 -18-8Kunin + 5 2 +26-3Mfortuitum +16-1 +18 5

3 Johnin -10-5 - 3-8PPD + 5 4 -12-7

4 Johnin +26-3 (44) -15-4 (14)PPD +33 5 (22) - 9-9 (12)Kunin +44-3 (65) + 6-2 (0)M fortuitum +29-3 (22) -66-0 (0)

TABLE IV Median (range) suppressor indices (%) to the antigens in normal subjects, Crohn'sdisease, and ulcerative colitis with lamina propria lymphocytes (peripheral blood mononuclearcells).

Normal Crohn's Ulcerativesubjects disease colitis(n=6) (n=5) (n=6)

Johnin +26-7 +24-3 +37 9(-5 7 to +57 3) (+ 18-8 to +40 7) (+21-9 to +65 7)

PPD +17-3 +13-3 +24-8(- 18-7 to +39 0) (-7 7 to +37-1) (-6-8 to +82-1)

Mycobacteriumavium +12-2 -12-4 +24-6(-4 3 to +38 0) (-25 2 to + 14-7) (+5±5 to +45 3)

M kansasii +13-9 -2-9 +26-2(-39 9 to +63-5) (-44-6 to +40-1) (-3-8 to +39 3)

Mnonchr +22 5 -10-4 +20-4(- 19-8 to +31-8) (-31-2 to +15-3) (-2-9 to +58 6)

Mvaccin +17-5 +5 0 +28-7(-29-1 to +35 4) (-20-8 to +32 3) (-34-6 to +82 7)

Mfortuitum +19-4 +13-8 +13-2(- 10-3 to +44 8) (-47 3 to +23-5) (-30 7 to +74 4)

Kunin +4-1 -14-1 -23-7(- 1-6 to +35 2) (-455 to + 13-1) (-20-3 to +66-6)

There is no significant difference in suppression between normal controls and patients withinflammatory bowel disease (p>0 05 in all cases).

of the antigens used in any of these subgroups(Table IV). The lamina propria lymphocyteviability at day 7, as assessed by Trypan blueviability exclusion was 62% (range 45% to 72%).It was felt possible that the lack of differencein suppression between normal subjects andinflammatory bowel disease patients couldreflect either poor cellular viability by the endof the experiment (day 9) or some intrinsicdifference between the lamina proprialymphocytes (induced population) and PBMC(responder PBMC). For this reason furtherexperiments were performed using laminapropria lymphocytes that had been centrifugeddown a Ficoll-Paque gradient at the end of the

TABLE V Median suppressor indices (%) oflamina proprialymphocytes, using autologous lamina propria lymphocytes asresponder cells.

Normal Inflammatory bowelsubjects disease(n=3) (n =2)

Johnin -6-2 +16 4(- 174to +42 1) (+11-9to +20 9)

PPD +4-7 -2-0(- 1-9to +6 7) (-19-0to +15 1)

Kunin +11-2 -4-2(- 13-2to +35 6) (-440 to +35 6)

inducer phase (day 7). The cell viability in theseexperiments was increased to 88% (range 82 to98%). Autologous (non-induced) lamina proprialymphocytes were used as responder cells inthese experiments. There was still no differencein suppression between normal subjects andpatients with inflammatory bowel disease(Table V).

DiscussionThis study shows that there is reduced antigeninduced suppressor activity in PBMC of patientswith inflammatory bowel disease. Patients withinactive Crohn's disease have reduced suppres-sor cell activity to Johnin, PPD, Mfortuitum, andKunin antigens. Patients with ulcerative colitishave reduced suppression to the Kunin,M avium, M nonchromogenecin, and M fortuitumantigens. Subjects who have a positive Mantouxskin test have a significantly higher PPD inducedsuppressor cell activity in PBMC compared withMantoux negative subjects.The phenomenon of antigen induced suppres-

sion seems to be the result of a CD8 mediatedmechanism, as depleting CD8+ cells reduces theeffect and the level of soluble CD8 in the culturesupernatant is directly related to the degree ofsuppression.

Mycobacteria share common epitopes24 andthis may be one explanation for the reducedsuppressor cell activity to a range of myco-bacterial antigens observed in inflammatorybowel disease. However, since the same effectis seen in response to Kunin antigen, it is morelikely that this phenomenon is a more general-ised one.One possible explanation for the generalised

effect reported is that the assay employed is nottruly antigen restricted, as low dose lectin was

200 -

E 150-

00o 100-a)

o 50-U)

0

r= 025p < 0 05

*.U

1 U 9 U.

~~~~* Ftmm U.S

.w*.U U..,a

-150 -100 -50 0 50 100Suppressor index

Figure 4: Supernatant soluble CD8, after 7 days' culture withantigen, compared with the suppressor index (%).

328

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from

Antigen induced suppression in peripheral blood and lamina propria mononuclear cells in inflammatory bowel disease 329

used in the responder phase. An alternativeexplanation is that the antigens used in theprimary culture may have caused non-specificpolyclonal activation of the lymphocytes understudy. It is known that bacterial products, suchas the antigens used in this study, are capable ofpolyclonal activation, although not usually of thesame magnitude as lectins.25 This hypothesismay also provide an explanation for the quitehigh levels of stimulation seen after primaryculture with antigen. However, these argumentsdo not explain the relation between the PPDinduced suppressor index and the results of theMantoux testing, which imply that the antigeninduced suppressor cell assay used reflects, atleast in part, cell mediated responses to specificantigen.No correlation was observed between the

proliferative responses to antigen in the primaryculture and the subsequent SI to that antigen innormal subjects. It is known that there is nodifference in the proliferative responses ofPBMC from patients with inactive inflammatorybowel disease and normal subjects after stimu-lation with mycobacterial antigens. 1626 It istherefore reasonable to hypothesise that the lackof antigen induced suppression observed in thePBMC of patients with inflammatory boweldisease does not simply reflect differences in theprimary proliferative responses to the antigensbetween normal subjects and those with inflam-matory bowel disease. Using a slightly differentsystem, Ebert et al also observed no such cor-relation. 16These data disagree with those of Ebert et al,

who found that the addition ofM paratuberculosisantigen to a concanavalin-A proliferation assayresulted in reduced proliferation of PBMC frompatients with ulcerative colitis and Crohn'sdisease compared with control subjects. Thiswas interpreted as showing increased antigeninduced suppression. However, the assay systemused is not well designed to investigate antigeninduced suppression and therefore any dis-crepancy in the results may reflect differingimmunological phenomena.The present data are in keeping with most

studies of non-specific suppression which haveclaimed a defect in suppressor cells in ulcerativecolitis and Crohn's disease, although this defectin non-specific suppression is mainly seen inpatients with active disease. Since we did notstudy patients with active disease, to avoidpossible confounding effects of treatment ordietary deficiency, there are no data on howthis might have affected antigen induced sup-pression.The importance of the present data for the

pathogenesis of mucosal disease remains specu-lative. However, it is possible that they mayindicate a fundamental defect in the regulationof the immune response and may explain theheightened immune response to a wide varietyof dietary, bacterial, and self-antigens seen inthese diseases. Whether such a defect in im-munoregulation is under genetic control andtherefore predisposes an individual to develop achronic inflammatory response after an acuteinsult or whether it is secondary to the chronicinflammation, remains to be established.

It is interesting to note that the mostsignificant lack of suppression of PBMC inulcerative colitis was to Kunin antigen, and inCrohn's disease to Johnin antigen. The bacteriafrom which these antigens are derived havecreated considerable interest in their role in theaetiopathogenesis of these diseases over the pastfew years.' 18 20 21

If reduced antigen induced suppression inPBMC in inflammatory bowel disease is import-ant for pathogenesis, we would have expected tobe able to reproduce the results at mucosal level.However, differences in antigen induced sup-pression between normal subjects and patientswith inflammatory bowel disease in the laminapropria lymphocyte population could not beshown. Increasing the cell viability and usingan autologous lamina propria lymphocyteresponder system (instead of PBMC) made nodifference to this finding. There are severalpossible explanations for this. Firstly, in thelamina propria lymphocyte studies (unlike inthe PBMC studies) the most subjects withinflammatory bowel disease were taking corti-costeroid medication. It is known thatcorticosteroids affect suppressor cell activity invitro.2728 Secondly, the chemicals and enzymesused in the procedure to isolate the laminapropria lymphocytes could possibly haveaffected their function. A third explanation forthis is that the lamina propria lymphocytesconsist of a different subpopulation of cells,compared with PBMC. It is known thatCD4/CD8 ratio is similar in lamina proprialymphocytes and PBMC in normal and controlsubjects.2930 However, compared with PBMC,there is an increased proportion of CD45Ro Tcells ('memory'/helper-inducer T cells) in thelamina propria, particularly in areas of inflam-mation.31-34 It would, therefore, be reasonable topostulate that the 'resting tone' of the laminapropria lymphocytes is help rather than sup-pression, as originally suggested by Elson in1985.' If this is the case, it could explain the lackof difference in suppressor cell activity betweennormal subjects and patients with inflammatorybowel disease.

The authors thank Mr N J McC Mortensen and Mr M G WKettlewell for kindly providing the colonic resection specimens.This work was supported by Sandoz Pharmaceuticals.

1 Hodgson HJF, Wands JR, Isselbacher KJ. Decreased sup-pressor cell activity in inflammatory bowel disease. Clin ExpImmunol 1978; 32: 451-8.

2 Victorino RMM, Hodgson HJF. Spontaneous suppressor cellfunction in inflammatory bowel disease. Dig Dis Sci 1981;26: 801-6.

3 Knapp W, Smolen JS, Lanzer G, Berger R, Menzel EJ,Grabner G, et al. Con A induced suppressor cell activity inIBD and other inflammatory diseases. In: Pena AS,Weterman IT, Booth CC, Strobar W, eds. Recent advances inCrohn's disease. The Hague: Martinus Nijhoff, 1981: 380-9.

4 Auer IO, Roder A, Frohlich J. Immune status in Crohn'sdisease. VI Immunoregulation evaluated by multiple,distinct T-suppressor cell assays of lymphocyte prolifera-tion, and by enumeration of immunoregulatory T-lympho-cyte subsets. Gastroenterology 1984; 86: 1531-43.

5 Doldi K, Manger B, Koch B, Riemann J, Hermanek P,Kalden JR. Spontaneous suppressor cell activity in theperipheral blood of patients with malignant and chronicinflammatory bowel diseases. Clin Exp Immunol 1984; 55:655-63.

6 Kelleher D, Murphy A, Whelan CA, Feighery C, Weir DG,Keeling PWN. Defective suppression in the autologousmixed lymphocyte reaction in patients with Crohn's disease.Gut 1989; 30: 839-44.

7 Ginsburg CH, Falchuk ZM. Defective autologous mixed-lymphocyte reaction and suppressor cell generation in

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from

330 Dalton, Hoang,3'ewell

patients with inflammatory bowel disease. Gastroenterology1982; 83: 1-9.

8 Holdstock G, Chastenay BF, Krawitt EL. Functional suppres-sor T cell activity in Crohn's disease and the effects ofsulphasalazine. Clin Exp Immunol 1982; 48: 619-24.

9 Elson CO, Graeff AS, James SP, Strober W. Covert suppressorT cells in Crohn's disease. Gastroenterology 1981; 80:1513-21.

10 Davidsen B, Kristensen E. Lymphocyte subpopulations,lymphoblast transformation activity and Con A inducedsuppressor activity in patients with ulcerative colitis andCrohn's disease. ScandJr Gastroenterol 1987; 22: 785-90.

11 Shorter RG. Tg cells in non-specific concanavalin A-inducedsuppressor activity in vitro in colonic inflammatory boweldisease and in colorectal carcinoma. In: Pena AS, WetermanIT, Booth CC, Strobar W, eds. Recent advances in Crohn'sdisease. The Hague: Martinus Nijhoff, 1981: 448-55.

12 Fiocchi C, Battisto JR, Farmer RG. Gut mucosal lymphocytesin inflammatory bowel disease. Isolation and preliminaryfunctional characterisation. Dig Dis Sci 1979; 24: 705-17.

13 Goodacre RL, Bienenstock J. Reduced suppressor cell activityin intestinal lymphocytes from patients with Crohn'sdisease. Gastroenterology 1982; 82: 653-8.

14 Fiocchi C, Youngman KR, Farmer RG. Immunoregulatoryfunction of human intestinal mucosal lymphoid cells:evidence for enhanced suppressor cell activity in inflam-matory bowel disease. Gut 1983; 24: 692-701.

15 Elson CO, Machelski E, Weiserbs DB. T cell-B cellregulation in the intestinal lamina propria in Crohn'sdisease. Gastroenterology 1985; 89: 321-7.

16 Ebert EC, Bhatt BD, Liu S, Das KM. Induction of suppressorcells by Mycobacterium paratuberculosis antigen in inflam-matory bowel disease. Clin Exp Immunol 1991; 83: 320-5.

17 Hampson SJ, McFadden JJ, Hermon-Taylor J. Mycobacteriaand Crohn's disese. Gut 1988; 29: 1017-9.

18 Blaauwgeers HJLG, Mulder CJJ, Das PK, Haagsma J, TytgatGNJ. Crohn's disease, a mycobacterial aetiology? Eur 7Gastroenterol 1990; 2: 237-40.

19 Stanford JL, Dourmashkin R, Mcintyre G, Visuvanathan S.Do mycobacteria exist in alternative physical forms andwhat part do they play in the aetiology in inflammatorybowel disease. In: MacDermott RP, ed. Inflammatory boweldisease. Current status and future approach. Amsterdam:Elsevier Science, 1988; 503-8.

20 Lagercrantz R, Hammarstrom S, Perlmann P, Gustafsson BE.Immunological studies in ulcerative colitis IV origin ofautoantibodies. J Exp Med 1%8; 128: 1339-52.

21 Mee AS, McLaughlin JE, Hodgson HJF, Jewell DP. Chronicimmune colitis in rabbits. Gut 1979; 20: 1-5.

22 Mahida YR, Wu KC, Jewell DP. Respiratory burst activity ofintestinal macrophages in normal and inflammatory boweldisease. Gut 1989; 30: 1362-70.

23 Bull DM, Bookman MA. Isolation and functional character-isation of human intestinal mucosal lymphoic cells. J ClinInvest 1977; 59: %6-74.

24 Stanford JL. Immunologically important constituents ofmycobacteria: antigens. In: Routledge C, Stanford JL, eds.The biology of the mycobactena 2. London: Academic Press,1983: 85-127.

25 Vordermeier M, Stab K, Bessler WG. A defined fragment ofbacterial protein I (OmpF) is a polyclonal B-cell activator.Infect Immunol 1986; 51: 233-9.

26 Ibbotson JP, Lowes JR, Chahal H, Gaston H, KumararanteDS, Alexander-Williams J. Pathogenesis of inflammatorybowel disease: mucosal all mediated immunity to myco-bacterial and non-mycobacterial antigens. Gut 1990; 31:A1166.

27 Haynes BF, Fauci AS. Mechanisms of corticosteroid action onlymphocyte subpopulations. IV Effects of in vitro hydro-cortisone on naturally occurring and mitogen-induced sup-pressor cells in man. Cell Immunol 1979; 44: 157-68.

28 Smith CI, Sveigaard A. Concanavalin A-induced suppressorlymphocytes in normal individuals. Scandj Immunol 1981;13: 483-92.

29 Selby WS, Jewell DP. T lymphocyte subsets in inflammatorybowel disease: peripheral blood. Gut 1983; 24: 99-105.

30 Selby WS, Jannossy G, Bofill M, Jewell DP. Intestinallymphocyte subpopulations in inflammatory bowel disease:an analysis by immunohistological and cell isolation tech-niques. Gut 1984; 25: 32-40.

31 Moore K, Walters MT, Jones DB, Garvey E, Harvey J,Cawley MID, et al. An immunological study of CD4+lymphocyte subsets within inflammatory lesions with specialreference to rheumatoid arthritis and inflammatory boweldisease. Immunology 1988; 65: 457-63.

32 James SP, Fiocchi C, Graeff AS, Strober W. Phenotypicanalysis of lamina propria lymphocytes. Predominanceof helper-inducer and cytolytic T-cell phenotypes anddeficiency of suppressor inducer phenotypes in Crohn'sdisease. Gastroenterology 1986; 91: 1483-9.

33 Senju M, Wu KC, Mahida YR, Jewell DP. Two-colourimmunofluorescence and flow cytometric analysis of laminapropria lymphocyte subsets in ulcerative colitis and Crohn'sdisease. DigDis Sci 1991; 36: 1453-8.

34 Kanof ME, Strober W, Fiocchi C, Zeitz M, James SP. CD4positive Leu 8 negative helper-inducer cells predominate inthe human intestinal lamina propria. J Immunol 1988; 14:3029-36.

on Novem

ber 26, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.33.3.324 on 1 March 1992. D

ownloaded from