Thorax 1992;47:695-701

Generation of cytolytic T cells in individualsinfected by Mycobacterium tuberculosisand vaccinated with BCG

A D Pithie, M Rahelu, D S Kumararatne, P Drysdale, J S H Gaston, P B Iles, J A Innes,

C J Ellis

Department ofImmunology, MedicalSchool, University ofBirmingham,BirminghamA D PithieM RaheluD S KumararatneP DrysdaleDepartment ofCommunicable andTropical Diseases,East Birmingham,Hospital, BordesleyGreen East,Birmingham B9 SSTA D PithieJ A InnesC J EllisDudley Road Hospital,BirminghamD S KumararatneP B IlesDepartment ofRheumatology,Medical School,University ofBirmingham,BirminghamJ S H GastonReprint requests to:Dr A Pithie

Accepted 13 March 1992

AbstractBackground Macrophage activation bycytokines provides only a partial explana-tion of antimycobacterial immunity inman. Because cytolytic T lymphocyteshave been shown to contribute to immun-ity in animal models of intracellularinfection, the generation of myco-bacterial antigen specific cytotoxic T cellswas examined in the peripheral blood ofpatients with tuberculosis.Methods Subjects comprised 36patients with active tuberculosis (18newly diagnosed) and 32 healthy volun-teers, of whom 25 had had BCG vaccina-tion and seven were Mantoux negative.The ability of purified protein derivative(PPD) stimulated peripheral bloodlymphocytes to lyse autologous, myco-bacterial antigen bearing macrophageswas examined by using a chromium 51release assay.Results PPD stimulated lymphocytesfrom normal, Mantoux positive, BCGvaccinated subjects produced high levelsof PPD specific cytolysis, whereaslymphocytes from unvaccinated, un-infected subjects caused little or no cyto-lysis. The generation of cytolytic Tlymphocytes by patients with tuber-culosis was related to their clinical state.Those with cavitating pulmonary diseaseor lymph node tuberculosis generatedPPD specific lymphocytes with cytotoxicability similar to that of those fromMantoux positive control subjects,whereas lymphocytes from patients withnon-cavitating pulmonary infiltratesshowed poor antigen specific cytolysis.After seven days of stimulation with PPDin vitro, lymphoblasts contained bothCD4' and CD8' cells. Mycobacterialantigen specific cytolysis was restricted tothe CD4+ cell population and was blockedby monoclonal antibodies directedagainst major histocompatibility class II(MHC) antigens.Conclusion CD4 + cytolytic T cells canlyse autologous macrophages presentingmycobacterial antigen and were found inpatients with cavitating pulmonarytuberculosis or tuberculous lymph-adenitis and in normal, Mantoux positivecontrol subjects. The ability to generatethese T cell responses seems to be amarker for response to mycobacteria and

may contribute to tissue damage intuberculosis. These responses do notprovide protective immunity againstMycobacterium tuberculosis but mayhelp in disease localisation.

(Thorax 1992;47:695-701)

Tuberculosis remains a major global healthproblem, especially, though not exclusively, indeveloping countries. The global estimate ofthe number of people with tuberculosis isapproximately 50-60 million,' with eight to 10million new cases each year.2 Although mor-tality figures have been falling steadily over thepast three decades, tuberculosis is recognisedas a complication of T cell deficiency states, inparticular of HIV infection.3 Patients withAIDS are a hundred times more likely tocontract tuberculosis than people without thisdisease.4 This situation is of obvious impor-tance in countries where tuberculosis andAIDS are already prevalent.The causative agent, Mycobacterium tuber-

culosis, is a facultative intracellular pathogenthat can replicate within human monocytes andmacrophages. Protective immunity to tuber-culosis depends on the interaction of T lym-phocytes and mononuclear phagocytes,5 whilehypersensitivity to mycobacterial antigens(also mediated by T lymphocytes) contributesto the typical caseous tissue destruction.Protective immunity and hypersensitivity areclearly related but the detailed cellular mechan-isms have not been fully explained. As a resultof the work of Mackaness6 it has been generallyaccepted that activation of macrophage anti-mycobacterial mechanisms by T cell secretedlymphokines such as gamma interferon is thecornerstone of protective immunity. Mycobac-terial growth in mice is readily inhibited bylymphokine activated macrophages.78 Activa-tion of human macrophages with a variety ofrecombinant lymphokines, however, achievesmodest inhibition of mycobacterial growth atbest and is unlikely to fully explain antimyco-bacterial immunity.9'2The generation of cytolytic T cells in re-

sponse to intracellular bacterial pathogens,including mycobacteria, has been shown inanimals.'3 Mycobacterial antigen specificcytolytic T cells are able to lyse autologousantigen bearing macrophages in murine tuber-culosis. Cytolytic T cells have also been shownto inhibit the intracellular growth ofM tuber-culosis in a manner independent of interferon.'4

695

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

Pithie, Rahelu, Kumararatne, Drysdale, Gaston, Iles, Innes, Ellis

Furthermore, selective depletion of CD8+cytotoxic T cells increases the susceptibility ofmice to infection with virulentM tuberculosis."'In man, mycobacterial antigen specific cytolyticT cells have been generated from the peripheralblood of BCG vaccinated subjects,'6'7 frompatients with leprosy,'8 19 and from those withtuberculosis.0 There has been speculation,therefore, about the function of cytolytic Tcells in the immune response to tuberculosis.We have investigated the cytolytic activity ofpurified protein derivative (PPD) stimulatedlymphocytes from patients and normal controlsubjects.

MethodsPATIENTS AND CONTROL SUBJECTSThirty six patients with bacteriologically or

histopathologically proved tuberculosis were

recruited from East Birmingham and DudleyRoad Hospitals, Birmingham. The clinicaldetails were documented and radiographs re-

viewed by one individual (ADP). Eighteenpatients had newly diagnosed tuberculosis andhad received less than two weeks' treatmentwhen tested; the remaining 18 patients hadreceived treatment for more than two weeks.Seventeen patients had localised cavitatingpulmonary tuberculosis, eight had extra-thoracic lymph node tuberculosis, six hadpulmonary disease with subapical diffuse pul-monary involvement and little or no evidenceof localised cavitation (diffuse, non-cavitatingtuberculosis), and three had miliary tuber-culosis. One patient had abdominal tuber-culosis and one had pleural disease. One patientwho presented initially with diffuse pneumonicdisease but later developed diffuse lymph-adenopathy was tested on both occasions. Fiveof the patients had lymphadenopathy and fivedeveloped an abscess that required needleaspiration or surgical drainage. No patient hadknown risk factors for HIV infection or any

other disease associated with immunosuppres-sion. Three patients had diabetes (one insulindependent, two non-insulin dependent).

Thirty two healthy volunteers acted as con-

trol subjects. Twenty five had previouslyreceived BCG vaccination and were known tobe Mantoux positive; they were not retested forthis study. The size of the tuberculin reactionwas not known in most of the subjects. Sevensubjects had no history of BCG vaccination or

exposure to tuberculosis and were Mantouxnegative.

TISSUE CULTURE MEDIA

Complete medium (RPMI) consisted ofRPMI1640 (Gibco Biocult, Paisley) supplementedwith glutamine (2 mM), penicillin (100 U/ml),and gentamicin (50 U/ml).

SEPARATION OF PERIPHERAL BLOOD

MONONUCLEAR CELLSPeripheral blood cells were separated fromdefibrinated blood by Ficoll-Hypaque centri-fugation, washed three times in RPMI andresuspended at 106 mononuclear cells per ml inRPMI with 10% autologous serum.

GENERATION OF STIMULATED T BLASTS (EFFECTORCELLS)Isolated peripheral blood cells were incubatedin 2 ml volumes (1 x 106 cells/well) in 24 welltissue culture plates (Cell Cult) at 370C, in 5%CO2 in air for seven days. PPD (StatensSeruminstitute, Denmark) at 10 ,ug/ml finalconcentration was added at the beginning of theculture period to generate antigen specific effec-tor cells. Interleukin 2 stimulated cells wereprepared as above except that 20 U/ml recom-binant Interleukin 2 were added instead ofPPD at the beginning of culture Interleukin 2was used as a non-specific T cell stimulant.

SURFACE MARKER ANALYSISThe surface phenotype of lymphoblast culturesstimulated by antigen for seven days was deter-mined by a sensitive rosette assay using mono-clonal antibody coated indicator red cells'5 orby flow cytometry by Becton-DickinsonFACScan. Data were collected using the con-sort 30 program and analysed with the LYSYSprogram. The following monoclonal antibodieswere used, CD3 (OKT3 obtained fromAmerican Type Culture Collection, 12301Parklawn Drive, Rockville, Maryland 20852),CD4 (BMA 040, Behringwerke AG, Marburg,Germany), CD8 (B941 a kind gift of Dr CMawas, Marseilles, France), TCR-delta (T-cell Sciences, Boston, USA).

PREPARATION OF MACROPHAGE TARGETSFreshly isolated cells were distributed in 150 ,Mlvolumes (15 x 10' cells/well)in 96 well roundbottom microtitre tissue culture plates (Flow).We and others have previously shown thatapproximately 10% of the added mononuclearcells adhered as monocytes and this numberwas used to compute effector:target ratios incytotoxicity assays.202' After six days' incuba-tion, the non-adherent cells were washed offandthe adherent cells incubated with antigen (PPD25 ,ug/ml or streptokinase/streptodornase(SK/SD) 250 U/ml as control antigen andlabelled with chromium-51 (2 ,uCi/well,CJS-1, Amersham International) overnight ina final volume of 100 ,ul. After 16-24 hoursincubation the plates were washed three timeswith warm RPMI containing 5% heat inac-tivated, group A positive human serum afterwhich the wells were replenished with 50 pl ofRPMI containing 10% human serum.

CYTOTOXICITY ASSAYThis method has been described in detailelsewhere.'820 Briefly, effector cells stimulatedwith antigen for seven days were harvestedfrom the 24 well plates, washed once, andresuspended in RPMI supplemented with 10%human serum at 3 x 106/ml and7 5 x 10'/ml.One hundred pl of these suspensions wereadded to target macrophages in triplicate togive effector:target ratios ofapproximately 20:1and 5:1. Triplicate wells with 100 p1 ofmediumwere used to determine spontaneous isotoperelease.Some patients with tuberculosis were so

profoundly lymphopenic that very low num-bers of effector and target cells were available.

696

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

Generation of cytolytic T cells in individuals infected by Mycobacterium tuberclosis and vaccinated with BCG

For these patients only one effector:target cellratio was tested for each antigen combination.

After 18 hours' incubation the supernatantwas aspirated from individual wells into count-ing tubes. One hundred ,l of 1% Triton X- 100(Sigma) were added to each cell pellet andincubated for 45 minutes at 60°C to lyse theremaining adherent cells. The entire volumewas aspirated subsequently and transferred tofurther counting tubes. Activity of both thesupernatant and pellet was assessed bymeasurement of radioactive decay with agamma counter (LKB).The percentage isotope release for each well

was calculated by using the formula:

% Isotope release cpm supernatantcpm supernatant + cpm pellet

The level of cytotoxicity in each well wasdetermined as follows:

Cytotoxicity% isotope release in test well-

% isotope release in spontaneous release well.

Results are presented as the mean (SD) oftriplicate estimates of the percentage 51Crrelease. The SD between triplicates was notnormally greater than 5%. Spontaneous releasein these assays was usually no greater than 20%of the total uptake of isotope.

Cytolysis of macrophages which had notbeen incubated with antigen was termed non-specific cytotoxicity. Cytolysis of macrophagesincubated with PPD was termed antigenspecific cytotoxicity.

MONOCLONAL ANTIBODY BLOCKINGThe MHC restriction of PPD stimulated cellswas determined by using antibodies directedagainst MHC class I (W6/32, A, B, C specificW6/32, SeraLab, UK) and class II (BU26, DP,DQ, DR specific, Dr M Goodall, Departmentof Immunology, Birmingham University)antigens. The antibodies were preincubatedwith target macrophages for 30 minutes beforethe addition of the effector cells and werepresent throughout the assay.

LYMPHOCYTE PROLIFERATION ASSAYSLymphocytes were isolated as describedpreviously and added in triplicate to U bottommicrotitre plates at a concentration of 1 x 105well.2' 22 Antigen was added at the concentra-tions indicated previously and the plates wereincubated for six days before the addition oftritiated thymidine (Amersham, UK) at0 15 pCi/well for 16 hours. Cells were harves-ted using an automatic cell harvester (Skatron,Finland) by suction through glass fibre paperand radioactivity was assessed using a betacounter (LKB Rackbeta 2).The stimulatory capacity of the antigen is

expressed as a lymphocyte proliferation indexand is calculated as follows:

(mean dpm in cultures with antigen-mean dpm in cultures without antigen)mean dpm cultures without antigen

DEPLETION OF T CELL SUBSETSEffector cells stimulated with PPD for sevendays were generated from Mantoux positivecontrol subjects as described above. Thesewere used to obtain effector cells which weredepleted of CD4+ or CD8+ lymphocytes. Thisdepletion was performed by incubation withmouse monoclonal CD4 or CD8 antibodies(described in the surface marker section above)before a second incubation with goat anti-mouse immunoglobulin coated magnetic beads(Dynabeads, Dynal, Trondheim, Norway).For comparison, effector cells were used with-out any depletion procedure ("unmanipulatedeffector cells") or after sham depletion byincubation with anti-mouse immunoglobulincoated Dynabeads alone.

MAGNETIC SEPARATION AND CYTOTOXICITYASSAYSuspensions of effector cells and beads at a 20:1bead target cell ratio were incubated for 90minutes at 4°C in a glass test tube. Suspensionswere then adjusted to 2 ml with phosphatebuffered saline and placed next to a cobalt-sumarium magnet (Dynal, MPC-1) for two tothree minutes. Dynabeads, including thosebound to cells, adhered to the inside wall of thetube adjacent to the magnet. The unbound cellsuspension was removed by careful aspirationinto another tube. Depleted cells were pelletedby centrifugation for five minutes at 700 g,washed once with RPMI containing 10%autologous serum, and resuspended to3 x 106/ml in RPMI.The phenotype of the various populations

was assessed using antibody coated sheeperythrocytes, before and after the depletionprocedure, to ensure that CD4+ or CD8+ cellshad been adequately removed. The aboveprocedure normally resulted in >90% reduc-tion of either CD4 or CD8 cells. The lysismediated by unmanipulated, sham depleted,CD4 or CD8 depleted populations was testedusing the cytotoxicity assay detailed above.

STATISTICAL ANALYSISThe non-parametric Wilcoxon test was usedfor statistical analysis throughout. Measure-ments of statistical significance refer to com-parisons made using an effector:target ratio of20:1 unless otherwise indicated. Other testsused are quoted in the relevant results sections.Where ranges are quoted they correspond tomaximum and minimum values.

ResultsGENERATION OF CYTOLYTIC T CELLS BY CONTROL

SUBJECTS AND PATIENTS WITH TUBERCULOSISPeripheral blood lymphocytes from the 32normal donors and 36 patients with tuber-culosis, when stimulated in vitro with 10 Mg/mlPPD for seven days, lysed autologous macro-phage targets in a dose dependent manner(figure 1 and table 1; individual values availablefrom the author on request). There was a widevariation in the ability of lymphocytes to gen-erate lymphoblasts capable of killing auto-logous macrophages in both groups. Macro-

697

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

Pithie, Rahelu, Kumararatne, Drysdale, Gaston, Iles, Innes, Ellis

Table 1 Cytolysis of autologous macrophages produced by purifed protein derivative (PPD) stimulated effector cellsfrom patients with tuberculosisor normal control subjects (median (range))

Target antigen/E:T ratio

No Ag SK/SD PPD

Subject details 20:1 5:1 20:1 5:1 20:1 5:1

Diffuse non-cavitating/miliary 10 3 20 8 27 13(patient Nos 1-9) (2-28) (0-12) (6-30) (3-11) (14-41) (5-39)Cavitating pulmonary/lymph node 18 8 20 9 46 36(patient Nos 10-36) (5-39) (0-36) (10-40) (0-30) (35-62) (19-47)Normal controls 12 4 13 4 22 10(Mantoux negative, n 8) (0-15) (0-14) (7-13) (3-7) (15-46) (4-24)Normal controls 17 8 19 8 49 33(Mantoux positive, n 25) (5-49) (1-34) (5-49) (4-36) (28-60) (18-52)

E:T refers to the effector:target ratio. SK streptokinase, SD = streptodornase.

15F-

0

0

90

NoAg PPD

Mantoux negative

O Io0'e30

go8oo

NoAg PPD

Mantoux positive

0

Oil!Onot0R0 ;

000

to

NoAg PPD

Patients

Normal subjects and patients showed similarlevels of non-specific lysis (lysis of antigenunpulsed macrophages), which ranged from 10to 90% ofthe total lysis in both groups. Lysis ofmacrophages pulsed with control antigen(streptokinase/streptodornase) did not differfrom that seen with unpulsed macrophages ineither group. No relation was found betweenthe ability to generate antigen specific or non-specific cytotoxic T cells and age, sex, or race.

Interleukin 2 stimulated cells from Mantouxpositive control subjects were only poorlycytotoxic to autologous antigen pulsed andunpulsed macrophages (figure 2), even thoughthere was a good proliferative response to thiscytokine (lymphocyte proliferation index> 80).

Proliferative responses to mycobacterialantigen were measured in 22 normal subjectsand in 11 patients. There was a weak positivecorrelation between lymphocyte proliferationand cytolysis of antigen pulsed macrophages(r = 0 37, p > 0 04) (figure 3).

Figure 1 Percentage cytolysis of antigen unpulsed (open symbols) or purified proteinderivative (PPD) pulsed (closed symbols) macrophages by PPD stimulated effector cellsfrom Mantoux positive or Mantoux negative control subjects. Patients with diffuse non-

cavitating or miliary tuberculosis are shown as squares. The effector:target cell ratio was20:1.

phages pulsed with PPD were killed to a greaterextent than those not pulsed with PPD (un-pulsed) or macrophages pulsed with the controlantigen streptokinase/streptodornase (p <0 001 for all comparisons) in both groups.PPD stimulated lymphocytes from normal

BCG vaccinated subjects showed greaterantigen specific cytolytic capacity than thosefrom normal unvaccinated individuals (p <0-001). PPD stimulated lymphocytes frompatients with tuberculosis were less able to lysePPD pulsed macrophages than those fromnormal BCG vaccinated subjects (p < 0 03).When patients were stratified according toclinical features, those with cavitating pulmo-nary disease and extrathoracic lymphaden-opathy tended to generate higher amounts ofantigen specific cytotoxicity than those withnon-cavitatory pulmonary disease (p < 000 1).The cytolytic activity of PPD stimulated lym-phocytes from patients with cavitatory diseasedid not differ significantly from those of BCGvaccinated control subjects (p = 0 67).

LYMPHOCYTE SURFACE MARKER ANALYSIS

The surface phenotype of cells stimulated byPPD for seven days was characterised inparallel with cytotoxicity assays in 12 Mantouxpositive control subjects and nine patients(table 2). The lymphoblast population in allgroups consisted predominantly of CD3 cells,most of which were CD4' (table 2). Variableexpression ofHNK 1 and CD 1 lb (markers fornatural killer cells) was found (data not shown).Only a minority of cells expressed the gamma/delta T cell receptor. There was no significantdifference in the phenotype of cells frompatients and control subjects and no statisticalcorrelation between the CD4:CD8 ratio in thePPD stimulated cells and the level of specific ornon-specific cytotoxicity.

Surface marker analysis of unstimulatedlymphocytes from tuberculous pleural effusionsfrom two patients also showed a predominanceof CD3+ and CD4+ cells and low numbers ofgamma/delta cells.

EVIDENCE THAT MYCOBACTERIAL ANTIGENSPECIFIC LYSIS OF MACROPHAGES IS MHC CLASS II

RESTRICTEDAntibodies specific for class II MHC antigensinhibited antigen specific lysis in both patientsand normal control subjects (table 3). Themean inhibition observed at an antibody dilu-

75 r

60 H

45CO

04-ao

0- 30K

698

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

Generation of cytolytic T cells in individuals infected by Mycobacterium tuberclosis and vaccinated with BCG

stimulant IL-2 PPD

Figure 2 The lysis of autologous macrophages by peripheral blood lymphocytesstimulated in vitro with recombinant interleukin 2 (IL-2) (20 U) or with purifprotein derivative (PPD) for seven days. Data pooledfrom two experiments usinormal donors. E: T = effector:target ratio.

*. 0

* 0

**00

0

.

0

0

O Mantoux negative* Mantoux positiveO Miliary/diffuse* Cavitating/lymph n

40 60

Lymphocyte proliferation index80

tion of 1:400 was 45% (p < 0 01) in patientsand 65% (p < 0 001) in control subjects. Incontrast, no significant blocking or a modestenhancement of cytolysis was observed whenantibodies directed against class I determinantswere used.

T CELL SUBSET DEPLETIONPPD stimulated lymphocytes were depleted ofCD4+ or CD8+ cells (as described in methods)and the depleted cell fraction was resuspendedto give the final effector:target ratios indicated.Data pooled from four experiments are shownin figure 4. These experiments used cells fromnormal donors since only a limited number ofperipheral blood cells could be obtained frompatients. Effector cells which were treated with

ied magnetic beads alone (without antibody treat-ing ment beforehand) were capable of lysing

antigen pulsed or unpulsed target cells to thesame extent as the unmanipulated fraction;thus, use of magnetic beads alone did not itselfeffect cytolysis.

Depletion of CD4+ cells from PPDstimulated lymphocytes resulted in a decreasein antigen specific lysis of macrophages com-pared with unmanipulated or sham depletedfraction (p < 0 001 in each case). The residualkilling of antigen pulsed targets by CD4depleted cells was equivalent to that of un-pulsed targets (p = 0940; figure 4).The effector cell fraction, which was depleted

of CD8 bearing cells (and was therefore en-riched for CD4+ cells), showed an enhancedability to kill mycobacterial antigen bearingmacrophage targets (p < 0 02).

DiscussionWe have shown that the generation of sensit-ised CD4+ T cells can lyse autologous macro-

ode phages presenting mycobacterial antigen, andis thus an integral part of the immune responseto both BCG vaccination and infection with M

100 tuberculosis. This extends previous observa-tions by ourselves and others.'61720 Although

PPD) the patients with tuberculosis had significantlyless ability to generate these cytolytic T cells

ofPPD than did BCG vaccinated subjects, the dif-ference was a result of the relatively poor

Table 2 Surface marker analysis ofpurifed protein derivative (PPD) stimulated lymphoblasts and unstimulatedpleuralfluid populations: individual patient data and control group data

Surface marker (0% of cells)

Subject details CD3 CD4 CD8 -/6 TCR

Normal controls (median (range)) (n = 12) 82 48 49 5(68-98) (19-80) (27-77) (2-13)

Cavitating pulmonary/lymph node 90 69 14 6Cavitating pulmonary/lymph node 88 82 7 3Cavitating pulmonary/lymph node 92 82 17 2Cavitating pulmonary/lymph node 77 46 26 ndCavitating pulmonary/lymph node 95 31 40 ndDiffuse non-cavitating/miliary 86 38 50 ndDiffuse non-cavitating/miliary 72 50 30 ndPleural fluid 48 53 18 2Pleural fluid 93 69 19 3Patient summary (median (range)) 88 53 19 3

(48-95) (31-82) (7-50) (2-6)

TCR = T cell receptor, nd = not determined.

E:T ratio

10:15:1

- 2.5:1a; 40(jn

uz 30_

g 20

10 _

0Target antigen

75

60 _ * 0

0S0

* .1045 _

x00

-00o

30

0 .

00

00015 -

o0 20

Figure 3 Relation of lymphocyte proliferation to purified protein derivative (1(10 pg/ml) and the ability of these effector cells to kill PPD pulsed, autologousmacrophage targets. Lymphocyte proliferation index is plotted against cytolysispulsed targets at an effector:target ratio of 20:1.

699

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

Pithie, Rahelu, Kumararatne, Drysdale, Gaston, Iles, Innes, Ellis

Table 3 Effects of the addition of monoclonal anti-MHC class I or class II antibodies on the ability ofpurifed proteinderivative (PPD) stimulated effector cells to kill autologous macrophage targets. (Results are shown as mean (SEM)lysis values)

Antigen pulsed

Anti class I Anti class IINo antigen,

Subjects nio antibody No antibody 1:200 1:2000 1:400 1:4000

Patients (n = 7) 17 4 (3 3) 41.1 (3 3) 48-0 (2 9) 45-0 (2-9) 30 6 (3 8) 31 5 (4)Controls (n = 7) 15-0 (2 8) 45.7 (4 7) 51 7 (5 6) 53 2 (5-1) 25-7 (4 1) 37 1 (5 8)

60

50

a)> 40U,

na)a) 30

-0,0 20

10

0

Unmanipulated

Figure 4 The effect of dep,ability of PPD) stimulate(SEM) percentage lysis ofexperiments, using MantouAwere: PPD stimulated lymn

(unnmanipulated); shami dejdepleted, or CD8' cell deple

E:T ratio: low lymphocyte proliferation index. Otherevidence suggests that proliferation in response

20:1 to antigen and antigen specific cytotoxicity areZZ1 5:1 * essentially unrelated T cell functions.29 The

data presented here show that Interleukin 2stimulated T cells, which show good pro-liferative responses, are poorly cytolytic.We have confirmed previous data'8 20 21 show-

ing that the cytolytic activity of PPDT r T * stimulated lymphocyte populations has both

antigen specific and antigen non-specific com-

ponents. We have also shown that in addition tothe considerable variation in total cytolytic Tcell activity between individuals, the relative

+ + + contribution of antigen specific and antigenSham depleted CD4 depleted CD8 depleted non-specific components to the total cytolysis

leti'ng CD4 or CD8- lymphoblasts respectively, on the varies considerably between individuals,d effector cells to kill autologous macrophage targets. Mean whether patients or control subjects.PPD pulsed ( + ) or unpulsed ( - ) macrophages fromfour From previous studies of BCG vaccinatedx positive control subjects are shown. The effector populations)hoblasts used withoutfurther mtanipulation subects and patients with leprosy and tuber-pleted by incubation with mnagnetic beads alone; CD4' cell culosis there is good evidence that CD4 cellseted. restricted by MHC class II antigens are re-

sponsible for the cytolysis of antigen presentingmacrophages.' -'8 )21 This view is supported by

cytolytic activity of PPD stimulated lympho- our finding that depletion of CD4 but notcytes from patients with non-cavitatory pul- CD8 cells removed the antigen specific com-monary disease. In contrast, patients with ponent of cytotoxicity. Cytolysis of antigencavitatory disease and those with extrathoracic unpulsed macrophages was not significantlylymph node disease had an ability to generate affected by CD4 '/CD8' depletion. Antigencytolytic T cells similar to that of BCG vacci- non-specific cytolysis, however, is inhibited bynated subjects. Although only a small number coculture of macrophages with K562 cells as

of patients were studied, those with miliary competing target cells, suggesting that cellsdisease seemed to have only a moderate with natural killer like activity are re-capacity to generate cytolytic T cells. sponsible.2° An alternative suggestion is thatApproximately 20% of patients with exten- gamma/delta T cell receptor bearing lympho-

sive pulmonary tuberculosis and a greater cytes may be responsible for antigen non-

proportion of those with miliary disease are specific cytolysis.26 Gamma/delta T cells showanergic as shown by absent cutaneous tuber- broad cytolytic activity after culture withculin hypersensitivity and poor in vitro interleukin 2,27 proliferate to mycobacteriallymphocyte proliferative responses to antigens in both MHC restricted and unre-

mycobacterial antigen.2324 Most of the patients stricted manner,2829 and have been identified inwith poor cytolytic T cell activity described granulomatous lesions of human leprosy andhere also showed relatively poor in vitro leishmaniasis.30 In this study, however,lymphocyte proliferation (and absent cutan- gamma/delta cells accounted for a smalleous hypersensitivity), and were thus anergic. proportion only of the PPD stimulated popula-There was, however, only a poor correlation tions and are therefore unlikely to have con-

between cytolytic T cell activity and lympho- tributed greatly to the antigen specific or non-

cyte proliferation overall. This finding contrasts specific cytolysis shown. Furthermore, we

with that of Kaleab et al,'8 who found a close observed that lymphocyte populations fromcorrelation between cytolytic T cell activity tuberculous pleural fluid that had not beenand lymphocyte proliferation in patients with restimulated in vitro, showed low numbersleprosy and healthy control subjects. Whole only of gamma/delta cells.BCG was used to stimulate effector cells in their There has been considerable speculationstudy, however, while we used PPD. Further- recently about the role of cytolytic T cells in themore, we observed that mononuclear cell immune response to tuberculosis.26 Thispopulations from several of our patients results largely from the failure to achieveshowed spontaneous proliferation in the effective mycobacteriostasis with lymphokineabsence of antigen and this can give a falsely activated human macrophages in vitro.9"' In

700

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

Generation of cytolytic T cells in individuals infected by Mycobacteriumn tuberclosis and vaccinated with BCG

addition murine cytolytic T cells can inhibitmycobacterial growth independently oflymphokine activated macrophages.' Further-more, both adoptive transfer and selectivedepletion experiments in mice have shown thatCD8' cytolytic T cells provide protectionagainst M tuberculosis infection in vivo.'5 32

Mononuclear cells in tuberculous granulomatashow functional heterogeneity33 and tissuemacrophages seem to possess significantly lessantibacterial capacity than blood monocytes.34It has been suggested consequently thatcytolytic T cells may have a pivotal role inantimycobacterial immunity by releasingmycobacteria from the safe haven of poorlyactivated tissue macrophages so that they canbe phagocytosed by the more active, bloodderived monocytes. Extracellular mycobacteriawould also be exposed to toxic macrophageproducts released from dying macrophages.Some support for this hypothesis is providedby the observation that reconstitution ofcellular immunity in lepromatous leprosylesions by intralesional interleukin 2 isassociated with infiltration of CD4 and CD8cells. There is subsequent lysis of parasitisedmacrophages within these lesions followed byrephagocytosis of the released bacilli by freshlymigrated monocytes and a concomitant reduc-tion in the bacterial count.'9The finding of high levels of mycobacterial-

antigen specific, cytolytic T cells in certaingroups of patients with tuberculosis stronglysuggests that the ability to generate these cellsdoes not equate with protective immunity.There was, however, a strong association ofcytolytic activity and tissue destruction. Allpatients with pulmonary cavitation couldgenerate significant amounts of cytolyticactivity and high levels were also found in thosewith tuberculous lymphadenitis, many ofwhom formed cold abscesses that requiredopen or closed drainage. It is notable thatpatients with pneumonic tuberculosis andmiliary disease had a reduced capacity togenerate mycobacterial-antigen specific,cytolytic T cells. We therefore suggest that intuberculosis cytolytic T cells may be involvedin delayed type hypersentivity and contributeto tissue destruction.

Assaad F, Azuma I, Buchanan I, Collins TM, Curtiss FM,David JR, et al. Plan of action for research in immunologyof tuberculosis: memorandum from a WHO meeting. Bu.ilWHO 1983;61:779-85.

2 Styblo K, Riuillon A. Estimated global incidence of smearpositive pulmonary tuberculosis. Unreliability of officiallyreported figures on tuberculosis. Buill Ilt U?nioni tIberc1981;56:1 18-25.

3 Chaisson RE, Slutkin G. T uberculosis and humanimmunodeficiency virus infection. J Inifect Dis 1989;159:96-100.

4 Schmidt J. Opening remarks. Rez Infect Dis 1989;11(suppl2):335.

5 Hahn H, Kaufmann SHE. The role of cell-mediatedimmunity in bacterial infections. Ret! Inzfect Dis 1981;3:1221-50.

6 Mackaness GD. The immunological basis of acquiredcellular resistance. J Exp Med 1964;120:105-20.

7 Flesch IEA, Kaufmann SHE. Attempts to characterise themechanisms involved in mycobacterial growth inhibitionby gamma interferon activated bone marrow macro-phages. Infect Imnmun 1988;56:1464-9.

8 Walker L, Lowrie DB. Killing of Mycobacterium mnicroti byimmunologically activated macrophages. Natuire 1981;293:69-70.

9 Rook GA. Progress in the immunology of the mvco-bacterioses. Clin Exp Imm"iuniol 1987;69:1-9.

10 Rook GAW, Steele J, Ainsworth M, Champion BR. Activa-tion of macrophages to inhibit proliferation of MlYco-

bacteriumi tuberculosis: comparison of the effects of recom-binant gamma interferon on human monocytes andmurine peritoneal macrophages. Inmunology 1986;59:333-8.

11 Douvas GS, Looker DL, Vatter AE, Crowle AJ. Gammainterferon activates human macrophages to becometumoricidal and leishmanicidal but enhances replicationof macrophage-associated mycobacteria. Infect Ihumlun1985;50: 1-8.

12 Bermudez LE, Young LS. Recombinant tumour necrosisfactor alone or in combination with interleukin 2, but notgamma interferon is associated with killing of M avilOucomplex from AIDS patients. J Inuniunol 1988;140:3006-13.

13 Kaufmann SHE. CD8' T lymphocytes in intracellularmicrobial infections. Inuniunology Today 1988;9:168-74.

14 De Libero G, Flesch I, Kaufmann SHE. Mvcobacteriareactive Lvt-2 + T cell lines. Eur J Imnininol 1988;18:59-66.

15 Muller I, Cobbald SP, Waldmann H, Kaufmann SHE.Impaired resistance to Ml'cobacteriunm tuberculosis infec-tion after selective in vivo depletion of L3T4 + and Lyt2 +'I cells. Inifect Inmnmun 1987;55:2037-41.

16 Mustafa AS, Godal T. BCG induced CD4 + cytotoxicT-cells from BCG vaccinated healthy subjects: relationbetween cytotoxicity and suppression in vitro. Cli,, ExpImmunol 1987;69:255-62.

17 Hansen PW, Kristensen 1. Cell mediated PPD specificcytotoxicity against human monocytes targets: III.Cellular typing with CTLs restricted by class II HLAantigens. Tissue Antigens 1986;27:227-38.

18 Kaleab B, OttenhoffT, Converse P, Halapie E, Tadesse G,Rottenberg M, Kiessling R et al. Mycobacterial-inducedcytotoxic T cells as well as nonspecific killer cells derivedfrom healthy individuals and leprosy patients. Eur JImmniunol 1990,20:2651-9.

19 Kaplan G, Kiessling R, Teklemarium S, Hancock G, SheftelG, Job CK, et al. The reconstitution of cell-mediatedimmunity in the cutaneous lesions of lepromatous leprosyby recombinant interleukin 2. J Exp Med 1989;169:893-908.

20 Kumararatne DS, Pithie AD, Drvsdale P, Gaston JSH,Kiessling R, Iles PB, et al. Specific Iysis of mycobacterialantigen-bearing macrophages by class II MHC-restrictedpolvclonal T cell lines in healthy donors or patients withtuberculosis. Clin, Exp Immnunol 1990;80:314-23.

21 Ottenhoff lTHM, Kale B, Van Embden JDA, Thole JER,Kiessling R. The recombinant 65-Kd heat shock proteinof Mvcobacterimnu bovis bacillus Calmette-GuerinMycobacteriuni tuibercuilosis is a target molecule for CD4cvtotoxic T lymphocytes that Ivse human monocytes. JE.Yp Med 1988;,168:1947-52.

22 Oppenheim JJ, Schecter B. Lymphocyte transformation. In:Rose N, Friedman H, eds Mlanual of clinical inonnuologv2nd ed. Washington, DC: American Society ofMicrobiology, 1980:233-45.

23 Nash DR, Douglas JE. Anergy in active pulmonary tuber-culosis: a comparison between positive and negativereactors and an evaluation of 5 TIU and 250 TU skin testdoses. Chest 1980;77:32-7.

24 Ellner JJ. Suppressor adherent cells in human tuberculosis.J Inonunol 1 978;121 :2573-9.

25 Sekalv RP, MacDonald R, Zaech P, Glasebrook AL, Cerot-tini J-C, et al. Cytolvtic T Iymphocyte function isindependent of growth phase and position in the mitoticcycle. J Exp Med 1981;154:575-80.

26 Lowrie DB. Is macrophage death on the field of battleessential to victory, or a tactical weakness in immunityagainst tuberculosis? Clin, ExP Innnunol 1990;80:301-3.

27 Ferrina S, Zarcone D, Viale M, Cerruti G, Millo R, MorettaA, Grossi C. Morphologic and functional characterisationof human peripheral blood T cells expressing the T cellreceptor gamma delta. EFurJ Ininiunol 1989;19:1183-8.

28 Haregowain A, Soman G, Hom RC, Finberg RW. Humangamma delta T cells respond to mvcobacterial heat shockproteins. Nature 1989;340:309-12.

29 Kabelitz D, Bender A, Schondelmaier S, Schoel B,Kaufmann SHE. A large fraction of human peripheralblood gamma delta T cells is activated by MycobacteriuMtuberculosis but not by its 65kD heat shock protein. J ExpMed 1990;171:667-79.

30 Modlin RL, Pirmez C, Hofman FM, Torigian V, UyemuraK, Rea TH, et al. Lymphocytes bearing antigen specificgamma delta T cell receptors accumulate in humaninfectious disease lesions. Natuire 1989;339:544-8.

31 Rook GAW. Role of activated macrophages in the immu-nopathology of tuberculosis. British Medical Bulletin1988;44:61 1-23.

32 Orme IM. The kinetics of emergence and loss of mediator Tlymphocytes acquired in response to infection withMycobacteriunn tuibercutlosis. J Ininiunol 1987;138:293-8.

33 Suga M, Dannenberg AM, Higuchi S. Macrophage func-tional heterogeneity in vivo. Am2 J Path 1980;99:305-24.

34 Lepay DA, Nathan CF, Steinman RM, Murrav HW, CohnZA. Murine kupffer cells: mononuclear phagocytesdeficient in the generation of reactive oxygen inter-mediates. J Exp Med 1985;161:1079-96.

35 Ling NR, Richardson PR. A critical appraisal of the directantibody rosette test for the detection of cell surfaceantigens. J Ininumnol Meth 1981;47:265-74.

701

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from

vaccinated with BCG.infected by Mycobacterium tuberculosis and Generation of cytolytic T cells in individuals

Iles, J A Innes and C J EllisA D Pithie, M Rahelu, D S Kumararatne, P Drysdale, J S Gaston, P B

doi: 10.1136/thx.47.9.6951992 47: 695-701 Thorax 

http://thorax.bmj.com/content/47/9/695Updated information and services can be found at:

These include:

serviceEmail alerting

box at the top right corner of the online article. Receive free email alerts when new articles cite this article. Sign up in the

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on April 4, 2017 - Published by http://thorax.bmj.com/Downloaded from