Eur. J. Immunol. 1990.20: 369-377

Birhane Kale Abo, Rolf KiesslingO, Jan D. A.Van Embden., Jelle E. R. Thole., Dinakantha S. KumararatneA, Pave1 Pisan, Assefa WondimuO and Tom H. M. OttenhofP Armauer Hansen Research Instituteo, Addis Ababa, Laboratory of Bacteriology, National Institute of Public Health and Environmental Hygiene., Bilthoven, Department of Immunology, Medical SchoolA, Birmingham and Department of Immunology, Karolinska Institute”, Stockholm

hsp 65-induced specific and nonspecific cytotoxic effector cells

Induction of antigen-specific CD4+ HLA-DR-restricted cytotoxic T lymphocytes as well as nonspecific nonrestricted killer cells by the recombinant mycobacterial 65-kDa heat-shock protein* Acquired cell-mediated immunity to intracellular parasites like mycobacteria is dependent on antigen-specific T lymphocytes. We have recently found that mycobacteria not only induce helper Tcells but also cytotoxic CD4+ and/or CD8+ Tcells as well as nonspecific killer cells that lyse human macrophages in vitro. In addition, we have described that the recombinant heat-shock protein (hsp) 65 of Mycobacterium bovis BCGIM. tuberculosis is an important target antigen for CD4+CD8- cytotoxic T cells. We have now furthet investigated the cytotoxic effector cells that are induced by the hsp65 of BCG. Purified protein derivative of tuberculin (PPD)- or hsp65specific cytotoxic T cells specifically lysed PPD, hsp65 of BCG and hsp65 of M. leprae-pulsed macrophages in an HLA-DR-restricted manner. Nonpulsed macrophages were lysed to a much lower but still significant extent. hsp65induced effector cells expressed CD3, CD5, CD4, CD8 and CD56 markers. Depletion experiments showed that the antigen-specific HLA-DR-restricted killer cell was of the CD5+CD4+CD8-CD56- phenotype. Experiments using N-terminal truncated hsp65 fusion (cro-lacZ) proteins suggested that the N-terminal65 amino acid residues of the 540 amino acid molecule are critical for the expression of the cytotoxic target epitope(s) in two individuals tested. In addition to inducing antigen-specific cytotoxic effector cells, the hsp65 also triggered nonspecific nonrestricted effector cells with lytic activity against nonpulsed autologous or allogeneic macrophages as well as K-562 and Daudi tumor cells. hsp65-stimulated effector cells produced both interferon and tumor necrosis factor-a. An important finding was that hsp65-stimulated effector cells strongly inhibited colony-forming unit formation from live BCG-infected autologous macrophages.

1 Introduction Specific and protective immunity against mycobacteria is dependent on the cellular immune response [l,21. Infections with Mycobacterium tuberculosis and M. leprae chronically affect over 45 million people in developing countries. Attempts to define the type and function of the cells involved in the cellular immune response to these bacteria have focused mainly on antigen-specific CD4+ proliferative and IFN-y-producing T cells. Such lymphocytes have been arbitrarily defined as T h cells and have indeed been repeatedly isolated and cloned from exposed

[I 79071

*

369

This work was also supported by NIH Grants RO1 CA 44882-1 and A1 20198-3. The Armauer Hansen Research Institute is supported by the Norwegian and Swedish Save the Children Funds, and by the Swedish and Norwegian development agencies. It is affiliated with the All Africa Leprosy Rehabilitation and Training Centre (ALERT), Addis Ababa, Ethiopia.

Correspondence: Tom H. M. Ottenhoff, Department of Immunohematology and Bloodbank, University Hospital, PO. Box 9600, NL-2300 RC Leiden, The Netherlands Abbreviations: AK: Activated killer cell hsp: Heat-shock protein LTE Lymphocyte-transformation test 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990

individuals [3-lo]. Some of these Th cells specifically proliferated to the recombinant heat-shock protein (hsp) 65 of BCG [5-101 or to synthetic peptides of the M . leprae analogue [6, 7 , 111. It has been thought that Th cells, through the secretion of lymphokines like IFN-y [12], enable effector M a to eliminate intracellular mycobacteria [I,21. Recent experiments carried out in our laboratory have, however, shown that mycobacteria also induce cytotoxic cells, comprising of both antigen-specific HLA-restricted CD4+ or CD8+ cytotoxic T lymphocytes (T,) as well as of nonspecific nonrestricted killer cells with NWactivatedkiller (AK) activity. Cytotoxic cells were able to lyse human MQ, [13, 141 and, importantly, to inhibit the outgrowth of live BCG organisms from infected MQ, (D. S. Kumararatne et al., unpublished observations). Cytotoxic cells may, therefore, be pivotal effector cells in the protection against mycobacterial infections. In one other study, human BCGreactive CD4+ T cell clones were reported to display cytotoxic activity towards APC [15]. An increasing number of recent studies have suggested an important role for T, cells in the immune response to mycobacteria and other intracellular parasites in mice as well [2, 16-19]. If indeed T, cells are important as protective effector cells in the immune response to mycobacteria, the definition of the microhial components that trigger cytotoxic lymphocytes 0014-2980/90/0202-0369$02.50/0

370

Eur. J. Immunol. 1990. 20: 369-377

Birkhane Kale Ab, R. Kiesling, J. D. A.Van Embden et al.

would be an important step towards the development of protective mycobacterial subunit vaccines [l]. In search of such antigens, we have recently reported that the hsp65 of M . bovis BCG is an important target antigen for CD4+ T, lymphocytes in k 20% of in vitro BCG-responsive [standard lymphocyte-transformation test (LIT)] individuals [14]. In these same individuals, the hsp65 efficiently induced PPD/BCG-specific cytotoxic cells during the primary in vitro restimulation of fresh PBMC. The hsp65 of BCG [20]is identical to the hsp65 of M. tuberculosis, shares >95% homology with the M. leprue hsp65 [21-241 and cross-reacts extensively with the latter at the serological [8,25-301 and Tcell level [6-8, 111. In this report we have further investigated the cytotoxic effector cells that are induced by the mycobacterial hsp65 with regard to antigerdepitope specificity, membrane phenotype, HLA restriction, NK/AK activity, IFNRNF production and ability to inhibit bacillary growth.

2 Materials and methods 2.1 Antigens PPD and BCG antigens were purchased from Statens Serum Institute, Copenhagen, Denmark; tetanus toxoid was obtained from Statens Bakteriologiska Laboratorium, Stockholm, Sweden; M. leprue bacilli were armadilloderived and irradiated organisms provided by Dr. R. J. W. Rees and co-workers of the National Institute of Medical Research, Mill Hill, London, GB, through the WHO Immunology of Leprosy Programme. The recombinant hsp65 of M . leprue was a generous gift of Dr. M. J. Colston, Mill Hill, London, GB [31]. 2.2 Production of native hsp65 M . bovis BCG protein (MbaA) and cro-beta-galactosidase-MbaAfusion proteins m Escherichia coli E. coli strain M1070 [21] which contains the plasmid pCI857 expressing the temperature-sensitive h repressor C I S 7 was used as a host for plasmid pRIBlOOO and the plasmids pRIB1404, pRIB1418, pRIB1424, pRIB1425, pRIB1427, pRIB1426, pRIB1430, pRIB1444, pRLB1463, pRIB1453 and pEX2. The native r65-kDa protein of M . bovis BCG (MbaA) was purified as described [22].The cro-beta-galactosidase-65-kDa fusion proteins carrying the whole or part of MbaA were expressed by the plasmids pRIB1404-pRIB1463as indicated in Fig. 1.The N-terminal amino acid residue to cro-beta-galactosidase is indicated in the same figure. The construction of the MbaA fusion protein expressing plasmids using the expression vector pEX2 [32] is described in [8]. Production of the crobeta-galactosidase-MbaA fusion proteins in E. coli M1070 was obtained as described in [33].

2.3 Cytotoxicity assay The assay will be described in detail elsewhere [13]. In brief, PBMC were stimulated during 7 days with PPD (5 pg/ml), BCG (5 pg/ml), rIL2 (50 U/ml) or the hsp65 of BCG (5 pg/ml) in 2-ml volumes in 24-well tissue culture plates (Linbro, Hamden, CT; Flow Labs., Rockville, MD)

at a concentration of 1 x 106 cellslml. All cultures were set up in RPMI supplemented with penicillin (100U/ml), streptomycin (100 pglml), L-glutamine (2 m), and 5% normal human serum (complete medium). Target cells were day 6 adherent cells of PBMC that had been plated in 96-well U-bottom tissue culture plates (1.5 x lo5 cells/well; Titertek, Flow Labs). Approximately 10% of the cells would adhere as monocytes, and this number was used for calculating E/T ratios. On day 6, the cells were pulsed with antigen (PPD: 25 pg/ml; BCG: 25 pg/ml; tetanus toxoid: 1: 120 dilution; hsp65 of BCG: 20 pg/ml; truncated hsp65 derivatives: 20 pg of the total lysate/ml; hsp65 of M. leprue: 80 pg/ml; all given in final concentrations), labeled with 2 pCi (= 74 kBq) of Na2Wr04 (Amersham International, Amersham, GB) for 18 h and then washed three times with preheated medium. Effectors were then added to triplicate wells with target cells (150 pVwel1). One triplicate with medium only was used for the determination of spontaneous release. After 15 h the total SN content of each well was individually transferred into a corresponding detachable counting tube (Linbro) and 100 p1 of Triton X-100 was added into each harvested well to lyse the remaining cells during 3 h of incubation. The total volume of Triton X-100 was transferred to similar tubes and counted. (1270 Rack Gamma; LKB, Bromma, Sweden).The percentage specific (percent. sp.) killing was calculated as follows: percent. sp. lysis = [test cpm/(test cpm -t cpm after Triton X-100 treatment of the same well)] x 100% - percent. spont. release. The percentage spontaneous (spont.) release was calculated using the formula: [cpm in spont. release wellkpm in spont. release well cpm after Triton X-100 treatment of the same well)] x 100%. The percentage antigen-specific lysis was calculated by subtracting the percentage specific lysis of antigen nonpulsed monocytes from the percentage specific lysis of antigen-pulsed monocytes (Fig. 1).

+

2.4 Depletion of lymphocyte subsets with mAb attached to magnetic beads A minor modification of the technique described in [34]was used. Magnetic beads (Dynabeads, Prod. no. 14002; Dynal A. S., Oslo, Norway) were coated with rabbit anti-mouse Ig (Dakopatts 2 2 5 9 , Glostrup, Denmark) and then incubated with 50 p1 of the appropriate 2nd layer mAb [Leu-1 (CD5), Leu-2 (CD8), Leu-19 (CD56) (all Becton Dickinson, Sunnyvale, CA); BMA044 (CD4; Behring, Marliurg, FRG) or normal mouse serum (control beads)]. For some of the experiments, Dynabeads precoated with anti-CD8 or anti-CD4 mAb were used. Cells and beads were mixed in a 1:3 ratio, centrifuged for 5 min 4°C at lo00 rpm and then incubated for another 15 min at 4"C.The pellet was gently resuspended and depleted for marker-positive cells with a magnetic particle concentrator (Dynal, no. 12001). The procedure was repeated twice. Control bead-depleted cells were resuspended in complete medium to give an E/Tratio of 30 : 1 in the highest concentration. The other cell suspensions were adjusted to the same volume as the control bead-depleted cells. The effectiveness of depletion, as checked by immunofluorescence, varied slightly with the antibody used. Depletion with CD8 was usually 100%

Eur. J. Immunol. 1990. 20: 369-377

hsp 65-induced specific and nonspecific cytotoxic effector cells

effective, depletion with Leu-19 usually 90%-100% and depletion with CD4 usually 85%-100%.

371

2.7 IFN and TNF assay

IFN titers in the SN of cultures were determined as described in [35] using the WISH cell assay.TNF titers were measured using an ELISA as described in'[36] with minor modifications.

2.5 K-562 and Daudi cell lysis assay K-562 or Daudi cells were labeled with 200 pCi 51Cr for 60 min at 37"C, washed twice in RPMI, resuspended in RPMI 10% FCS for 60-90min to reduce background isotope release, pelleted and resuspended at a concentration of 105/ml. One hundred microliters of suspension was added to each well. Effector cells were added in triplicate in 96-well U-bottom plates (100 pYwe11). Control target cells were incubated with complete medium (spontaneous release) or 1% Triton X-100 (maximum release). After 4-6 h, 100 pl of SN was harvested from each well and counted in a gamma counter. Specific lysis ( W r release) was calculated by the formula [(test cpm - spont. release cpm)/(max. release cpm - spont. release cpm)] x 100%. 2.6 Immunotluorescence staining

A standard immunofluorescence assay was used, with the mAb indicated above at a 1/10 dilution as first layer, and a FITC-labeled rabbit anti-mouse Ig (Dakopats) at a 1/20 dilution as the second layer of antibody.

2.8 Mycobacterial growth inhibition assay This was modified from [37]. BCG was grown in 7H9 broth with 0.5% Tween 80 (Difco, Detroit, MI) for 7-10 days and then harvested by centrifugation at 400 x g for 10 min and resuspended in RPMI 1640 without antibiotics. Aliquots were stored at -70°C. MQ were obtained by seeding 0.5 x lo6 PBMC cells in 1mYwell of 24-well tissue culture plates (Linbro). The medium used was RPMI 1640 supplemented with ~-glutamine(2 mM) and 5% normal human serum (NHS) and an antibiotic mixture containing Vancomycin, Colistin, Nystatin and Tnmethoprim (VCNT). VCNT (Difco) was reconstituted as described by the manufacturer and 0.5ml was added per 100ml RPMI 1640/5% NHS. These antibiotics d o not inhibit the growth of M. bovis or M. tuberculosis [38]. On day 6 of culture the medium containing nonadherent cells was discarded and 1 ml RPMI 1640/5% NHSNCNTwas added. The thawed BCG suspension was centrifuged at 50 x g for

Table 1. Antigen-specific and nonspecific lysis of monocyte targets by PPD or hsp65 of BCG-induced cytotoxic effector lymphocytes % specific lysis of monocytes pulsed with

Exp. no.

1

Stimulating antigen

EfT ratio

PPD

PPD

25 : 1 8: 1 2: 1 25 :1 8: 1 2:l 25: 1 8: 1 2: 1 30: 1 10: 1 3: 1 30: 1 10: 1 3:l 25:l 8: 1 3: 1 25: 1 8: 1 3: 1 30: 1 6: 1 30: 1 6: 1

58 42 25 42 34 25 12 16 5 57 51 57 20 9 4 38 39 29 19 14 14 54 25 17 11

hsp65

rIL 2 PPD

hsp65

PPD

hsp65

5(BL)a)

PPD

hsp65

hsp65 hsp65 lktanus of BCG of M. leprae toxoid

ND ND ND ND ND ND ND ND ND

32 26 12 36 29 19 15 7 2 71 63 30 21 14 7

35 19 4 34 32 18 ND ND ND ND 59 44

ND ND ND\ ND ND ND

ND ND ND ND ND ND

ND ND ND ND

22 15 13 5

13 8 3 3

ND ND ND ND

ND ND ND

15 17 2 10 4 7

ND ND

No antigen 13 8 2 27 15 9 15 12 3 32 10 3 6 2 0 0 0 0 6 5

4 11 8 4 5

a) BL patient that responded to PPD and hsp65 of BCG but not to M ,leprae in an LTT.

372

Eur. J. Immunol. 1990.20: 369-377

Birkhane Kale Ab, R. Kiesling, J. D. A.Van Embden et al.

5 min to dispose of large clumps of bacteria and counted in a hemocytometer using phase-contrast microscopy. BCG was added to each MQ culture to give a bacteria to MQ ratio between 5 : 1 to 10: 1. The same volume of BCG suspension was added to a control well containing RPMI/5% NHSNCNT to determine survival of the bacteria in medium alone. On day 7 two infected M@-containing wells and wells containing bacteria only each received 1.5 ml of 5% serum supplemented RPMI. Duplicate wells with infected mycobacteria received effector cells in 1 ml of medium (E/Tratio 10 : 1).To each combination of effector cells and infected M@ the appropriate stimulating antigen was added at a final concentration of 25 pg/ml for BCG, 1/20 dilution of tetanus toxoid, and 20 pg/ml for hsp65 of BCG, respectively. All wells were then made up to a total volume of 2.5 ml with RPMI/5% NHSNCNT. After 96-h incubation the total SN from each M a culture was harvested and the remaining adherent cells were lysed with 0.2% Saponin in 0.5 ml RPMI/5% NHSNCNT.The SN and lysates were thoroughly vortexted and the BCG CFU in each was determined as described in [39]. In brief, 50-yl aliquots from each SN or cell lysate were serially diluted in flat-bottom 96-well microtiter plates (Flow) containing 200y1 of 7H9 medium without Tween80, in each well. Thus, each sample was diluted in fivefold steps through eight microtiter wells. The plates were incubated undisturbed at 37 "C in 5% COz/air. After 8-14 days, individual Table 2. Cell surface marker phenotype of PPD- or hsp65-induced cytotoxic lymphocytes (A) Cell surface marker expression of PPD- and hsp65-induced effector cells % of cells expressing

Exp. no. 1

Stimulating C D 3 CD5 0 4 CD8 antieen 78 83

ND ND

PPD

ND ND ND

hsp65

ND

86 70 56 60

PPD hsp65

2

PPD hsp65

3

57

CD56

82 60

22 52 25

70

38

41 46

26

16 49 21

50

54

hsp65 effector cells depleted for

Em ratio

PPD-pulsed monocytes

Nonpulsed monocytes

Control

25:l 5:l 25 :1 5: 1 25:1

28 18

0

cD4

CD8

Leu-19

5:l 25: 1 5: 1 25 :1 5:1

9

4 10 1 28 17 27 17

+

3 Results 3.1 hsp65 or PPD-stimulated cytotoxic effector cells lyse MCP pulsed with PPD, hsp65 of BCG or hsp of M. Zeprue as well as nonpulsed MCP

PPD or hsp65 of BCG-stimulated effector cells from donors that responded in a standard LTT to the hsp65 of BCG (k20% of the BCG responsive individuals; [14]) efficiently lysed PPD- or hsp65 of BCG-pulsed autologous monocytes in a dose-dependent manner (Table 1).Antigen-dependent lysis was always accompanied by lysis of nonpulsed target cells, albeit to a much lower extent, in accordance with our previous observations [131. Target cells pulsed with irrelevant antigen (tetanus toxoid) were lysed to the same extent as nonpulsed targets. Lysis of nonpulsed or irrelevant antigen-pulsed monocytes was higher in case of hsp65 effectors as compared with PPD effectors in the same experiment. Antigen-pulsed and nonpulsed target cells were equally sensitive to lysis by LAK cells, ruling out the possibility that (mycobacterial) antigen pulsing in itself would render monocytes more susceptible to nonspecific lysis by effector cells. PPD- and hsp65 of BCG-stimulated effector cells in Expt. 1, 2 and hsp65 of BCG-stimulated effector cells of another donor (not shown) also efficiently lysed monocytes that had been pulsed with the hsp65 of M. leprae. In one other experiment, PPD- or hsp65 of BCGstimulated effector cells of a borderline lepromatous patient (that did not proliferate to M. leprue but proliferated to PPD and hsp65 of BCG in a LIT) lysed both PPD and hsp65 of BCG but not hsp65 of M. leprue-pulsed monocytes (Table 1, Exp. 5 ) .

7

72

(B) Effect of depletion of specific lymphocyte subsets from hsp65-stimulated effector cells on lysis of monocyte targets

cD5

colonies were counted using an inverted microscope (Leitz, Giessen, FRG) as described [38].The total BCG CFU/culture = CFU in SN CFU in cell lysate.

6 0 0 0 1 4

2 2 0

3.2 Cell surface marker phenotype of PPD and hsp65 of BCG induced cytotoxic effector cells as determined by immunofluorescence and functional subset depletion studies

As shown in Table 2A, both PPD- and hsp65-stimulated cells mainly consisted of CD4+ T cells (CD3+CD5+). A remarkable difference was the much higher expression of the CD56 marker by hsp65-stimulated cells as compared to PPD raised ones. The CD56 marker is found on NK cells and activated Tcells [40]. In parallel with the higher 0 5 6 expression, there was also a higher expression of the CD8 marker among hsp65-induced cells. To determine the phenotype of the actual hsp65 of BCG-induced effector cytotoxic Tcell among the antigen-stimulated cells, depletion experiments were performed using mAb-coated magnetic beads. One representative experiment is shown in Table 2B: depletion of CD4+ or CDS+ but not CD8+ or CD56+ (or control) cells from the stimulated effector population significantly reduced or abolished lysis of hsp65-pulsed monocytes. This latter result confirms and extends our previous studies in which CD4+ CD8-CD56PPD-stimulated effector cells were found to be active in the killing of hsp65-pulsed monocytes [141.

Eur. J. Immunol. 1990. 20: 369-377

hsp 65-induced specific and nonspecific cytotoxic effector cells

373

Table 3. HLA restriction of hsp65-stimulated effector cells

(A) Inhibition of lysis of monocytes by hsp65-stimulated effector cells by mAb to HLA-DR but not to HLA-DQ or HLA-A, B, C determinantsa) % Specific lysis of monocytes pulsed or

Stimulating antigen

hsp65

mAb added

unpulsed with the hsp65 of BCG Em 25 :1 8:1 2:1 Pulsing: + - + -

mAb specificity

-

-

B8.11.2 SPV-L3 B9.12.1

DR

+

36 17 39 37

DQ A , B,C

27 8 17 18

(B) Lysis of autologous vs. allogeneic monocytes

29 1 39 37

15 19 4 6 6 23 16 17 35

9 2 17 16

a) mAbwereaddedto thecultures in saturating 1 :200 dilutions of filter-sterilized ascites.

3.4 Identification of T, cell epitopes on hsp65 with

recombinant truncated hsp65 derivatives hsp65 effector cell/ monocyte combination

Exp. 1

Autologous Allogeneic 1 Allogeneic 2

Exp. 2

Autologous Allogeneic 1 Allogeneic 2

Em ratio

25:1 5:1 25 : 1 5:1 25 : 1 5: 1 25 : 1 8:1 25 : 1 8: 1 25 : 1 8: 1

PPD-pulsed Nonpulsed monocytes rnonocytes (% specific lysis)

24 15 17 4 7 4 18 0 18 7 8 4

10 0 27 6 9 0 9 0 16 9 8 3

Given the apparent immunodominance of the hsp65 antigen for CD4+ HLA-DR-restricted T, cells, we attempted to define the epitopes recognized by these Tcells. Monocytes were pulsed with lysates of E. coli clones that expressed fusion proteins of lacZ and the entire hsp65 (pRIB1404) or N-terminal truncated hsp65 derivatives, as indicated in Fig. 1. PPD-stimulated effector cells of two hsp65 responders both appeared to lyse monocytes pulsed with pRIB1404 but not, or only to a low extent, monocytes pulsed with the truncated antigens. These results suggest that the N-terminal65 residues of the hsp65 fusion protein are critical for the expression of the epitope(s) that are recognized by CD4+ HLA-DR-restricted T, lymphocytes, at least in the two individuals tested. 3.5 hsp65 stimulates cytotoxic cells with NK/AK-Like

activity 3.3 HLA restriction phenotype of hsp65 of BCG-induced qtotoxic effector cells

Inhibition studies with HLA-specific mAb (mAb are described in [41J) were consistent with the antigen-specific CD3+CDS+CD4+CD8-CD56- phenotype of hsp65induced T, lymphocytes: only HLA-DR but not HLA-DQ or -A, B, C-specific mAb could inhibit the lysis of hsp65pulsed monocytes by such effector cells (Table 3A). Similar results have been obtained with PPD-stimulated effector cells [13]. Lysis of unpulsed monocytes was reduced by HLA-DR-specific mAb, whereas HLA-DQ and class I mAb either slightly reduced (25 : 1ratio), enhanced (2 : 1) or did not affect (8 : 1) lysis of unpulsed targets. Whereas hsp65 effector cells lysed PPD-pulsed autologous monocytes to a significantly higher extent that nonpulsed autologous targets, no such antigen-dependent difference in lysis was seen in the case of allogeneic monocyte targets: antigen-pulsed allogeneic, HLA-DR-mismatched monocytes were as efficiently lysed as nonpulsed ones (Table 3B). Monocyte targets of different donors in the same experiment appeared to vary in their sensitivity to lysis.

We then tested whether these effectors also were cytotoxic to K-562 (the conventional NK target) and Daudi (an NKresistant but AK-sensitive target) tumor cells. As shown in Table 4, PPD- and hsp65-stimulated effector cells lysed both target cells, indicating that they indeed display NK-like and AK activity. Seven-day hsp65-stimulated effector cells appeared to be as efficient as or more efficient than PPD-stimulated ones in lysing these targets. Nonstimulated cells exhibited very low or no NK/AK-like activity. As shown in Table 4, day-4 PPD effector cells were more effective than hsp65 ones in lysing K-562 and Daudi, whereas the reverse pattern was seen with day-7-stimulated effector cells. The results suggest that the stronger induction of day 7 NK/AK activity by the hsp65 as compared to PPD does not reflect an intrinsic difference between the two antigens with regard to their ability to induce of NK/AKlike activity, but probably reflects differences in kinetics. 3.6 The inhibition of growth of live M. bovis (BCG strain) in infected M a by hsp65-stimulated effector cells

A biologically important question is whether lysis of antigen-pulsed M a by hsp65-stimulated effector cells

Eur. J. Immunol. 1990. 20: 369-377

Birkhane Kale Ab, R. Kiesling, J. D. A. Van Embden et al.

374

Table 4. Kinetics of induction of NK/AK-like activity by PPD and hsp65 of BCG % of specific lysis of

K-562

Daudi

Day 4

E/T

PPD

hsp65

NoAg

PPD

52 54 30 3

39 27 18 0

11 0 1 0

51 17 8 5

Day 7 hsp65 N o A g

PPD

Day 4 hsp65 NoAg

PPD

Day 7 hsp65 NoAg

ratio

35 : 1 12: 1 4: 1 1:l

66 29 20 4

1 3 1 0

results in the inhibition of growth of intracellular bacteria. To answer this question we infected adherent MQ, cultures with small numbers of live BCG and added various autologous effector cells to such cultures. The initial effector to target-MQ, ratio used was 10: 1. The cultures were incubated for 96 h and the BCG CFU in each culture was determined. Under these conditions BCG bacilli do not grow extracellularly and only multiply intracellularly, as human serum inhibits extracellular growth of mycobacteria [39,42]. SN and cell lysates from each M a culture were cultured separately and the results added to obtain the total number of BCG CFU recovered per culture. Results of two representative experiments are shown in Fig. 2. Fig. 2A shows the results of a BCG growth inhibition assay done with cells from two healthy Caucasian donors. Donor 1 was a Mantoux-negative, BCG-unvaccinated individual with a poor proliferative T cell response to BCG or PPD, while donor 2 was a BCG-vaccinated strongly Mantoux-positive individual. The addition of rIL2-, BCG- or PPD-educated cells from donor 1did not cause a reduction in the BCG CFU within autologous MQ, cultures, but there was a 2 loglo reduction of BCG CFU in PPD or BCG effector cell-treated MQ, cultures from donor 2. Fig. 2B shows results of two independent assays using two healthy Ethiopian donors who had strong proliferative T

-

49 39 19 3

13 3 3 0

40 26 13 10

55

66 2 8 4 0 7 0

14 4

9 4 0 0

cell responses to the hsp65 of BCG. In both experiments no live BCG CFU could be detected in MQ, cultures treated with hsp65-stimulated effector cells. In the first assay (donor3) PPD-educated effector cells caused a 2 loglo reduction of BCG CFU,while in the second assay (donor 4) these cells caused only marginal reduction in bacteria. In the latter experiment (donor 4) BCG-educated effector cells caused a 1.5 loglo reduction of BCG CFU when additional BCG antigen (25 pg/ml, final concentration) was added to the cultures, but not in the absence of additional BCG antigen. Again rIL2-educated effector cells did not reduce BCG CFU to a significant degree. An interesting observation was the moderate (tenfold) reduction of BCG CFU in MQ, cultures receiving tetanus toxoid-educated effector cells and tetanus toxoid antigen, in this experiment. Taken together these experiments suggest that hsp65-stimulated effector cells possess the capability to inhibit the CFU formation of live BCG bacilli within infected MQ,. 3.7 Production of TNF and IF" by cytotoxic effector cells

As shown in Table 4, PPD-stimulated effector cells appeared to produce TNF-a than hsp65-stimulated effector cells. Both effectors produced IFW. Part of the IFN activity

X SPECIFIC LVSlS

-,'.:

H...'p RIB

140+.....

m....pRIB

1418

10

20

ma0

............. j -;-

I

. . I

....

RIB

D RIB I)...p

....

RIB

....'pRIB I)....pRIB ....

1428

..............

1427""""

..............

1428

........................... 303 -1 -01 1463 ................................. 0 1453'.................................. a

I I I

.............................................

I

1444

.p1

Figure 1. Specific lysis by PPD-stimulated effector cells of monocytes pulsed with truncated recombinant hsp65 derivatives fused to crolac Z. The numbers in the MbaA gene represent the amino acid residues in the MbaA protein sequence [22] which are fused to cro-beta-galactosidase. The percentage antigen-specific lysis was calculated as given in Sect. 2.3. Data are shown for two pooled experiments of two different donors at an Emratio of 5 : 1 .The individual two data points for the different recombinant fusion proteins from top to bottom were: 24,9;0,l; 2,O;5,O; 4,O;0,O; 0 , l ; 0,O; 0,O; 0,O; 0,O.

Eur. J. Immunol. 1990. 20: 369-377

-

I

1

hsp 65-induced specific and nonspecific cytotoxic effector cells

375

could be neutralized by an IFN-y-specific mAb (not shown). No TNF or IFN production was seen in nonstimulated cultures. Both lymphokines can be produced by Tcells [431.

I

4 Discussion 4.1 Antigemspecific (antigedepitope specificity and HLA restriction) cytotoxic activity

* *

* *

Type of Effector cells added

Influence of 65K effectars 011 Ianacellnlar BCG b5K

a U U RI

*

Nil IM

L 0 4 U

m

L L Y L 0

65K

112

a

a 0

€Idthy Domr 3

IW

v)

x

(61

m

U

The hsp65 of M. bovis BCGIM. tuberculosis is an important target antigen for BCG-specific CD4+ T, lymphocytes that lyse human McP [14]. Our present results confirm and extend these previous observations. More importantly, the hsp65 molecule also appears t o be an important antigen for restimulating BCG/PPD-specific cytotoxic precursor cells in vitro. It's immunodominance and it's remarkable capacity to trigger effector cells that inhibit the growth of live mycobacteria (see below) make this hsp a potentially interesting molecule for the design of a mycobacterial subunit design. The observed cross-reactivity between the BCGIM. tuberculosis and M. leprae hsp65 further strengthens its usefullness. Interestingly though, PPD- and hsp65 of BCG-stimulated effector cells of one lepromatous patient that did respond in a LTT to the hsp65 of BCG but not the M. leprae lysed monocytes pulsed with the BCG but not the M. leprae hsp65. One explanation for these findings would be that in lepromatous leprosy,Tcell responses are directed to BCG-specific rather than cross-reactive epitopes on the hsp65 antigen.

l-rx Bcc -&I

TheT, cell epitope(s) that we describe here at least for two individuals are most probably located on the N-terminal65 amino acid residues of the hsp65. This region is also frequently recognized by cloned human Tcells [6, 8, 111. It cannot be excluded, however, that the presence of residue 1-65 is critical for the expression of epitope(s) elsewhere on the molecule, or that the absence of this region alters intracellular antigen-processing events that could compromise the antigenicity of other parts of the protein [44]. C-terminal deletion mutants have not yet been systematically produced and thus could not be tested.

Nil

I-

*

IM

2

4 6 8 Log ,o BCG CFU per culture

10

Figure 2. Effect of mycobacterial antigen-reactivelymphocytes on intracellular BCG (see Sect. 2.3 for design of experiment). (A) Results for two healthy Caucasian donors. *IDO: Infecting dose of BCG added on day 0 (with no M a ) ; *ID4: infecting dose of BCG cultured (with no MO) in RPMI/5% NHSNCNT for 96 h; Nil: infected MO with no effector cells; IL2: rIL2-educated effector cells added to infected MQ; BCG: BCG-educated effector cells added to infected MQ in the presence of BCG 25 pglml; PPD: PPD-educated effector cells added to infected MQ in the presence of PPD 25 pg/ml. (B) Results for two healthy Ethiopian donors who were good responders to hsp65 of BCG. *IDO, ID4, Nil, PPD, IL 2: see (A); 65K: hsp65-educated effector cells in the presence of hsp65 (20 Fg/ml); BCG: BCG-educated effector cells added to MQ without extra BCG antigen; BCG Ag: BCG-educated effector cells added to MQ in the presence of BCG (25 pg/ml).

+

The hsp65-specific cytotoxicity was a function of CD5+CD4+CD8-CD56- Tcells, as determined by experiments in which effector cells were depleted for the different lymphocyte subsets ([14] and this study). Inhibition studies using HLA-specific mAb as well as allogeneic combinations of Emcells showed that antigen-specific lysis was HLA-DR restricted. 4.2 Nonspecific (NK/AK-like) cytotoxic activity Both PPD- and hsp65-stimulated effector cells displayed strong NK/AK-like activity. Induction of this nonspecific cytotoxicity was strictly antigen dependent, since virtually no lysis of K-562, Daudi or other (data not shown) tumor targets was seen with nonstimulated day 7 PBMC ([13] and this study). The fact that hsp65-educated effector cells display strong nonspecific cytotoxicity may be important with regard to the possible role of this hsp in autoimmunity (e.g. [33]); nonspecific lysis of tissue McP and other targets may lead to tissue damage, which is indeed a prominent

376

Eur. J. Immunol. 1990. 20: 369-377

Birkhane Kale Ab, R. Kiesling, J. D. A.Van Embden et al.

Table 5. Kinetics of TNF and IFN production by PPD or hsp65 of BCG-stimulated cytotoxic effector cellsa)

TNF production

TNF-a (U/ml) Stimulating antigen PPD

hsp65

No

24 770 150 < 10

48 900

130 12

Exp. 1 72 (h) 770 100