Fas/Fas Ligand-Independent Mechanism by a Perforin And

Home , Perforin

CD4+ and CD8+ T Cells Kill Intracellular Mycobacterium tuberculosis by a Perforin and Fas/Fas Ligand-Independent Mechanism

This information is current as David H. Canaday, Robert J. Wilkinson, Qing Li, Clifford V. of October 2, 2021. Harding, Richard F. Silver and W. Henry Boom J Immunol 2001; 167:2734-2742; ; doi: 10.4049/jimmunol.167.5.2734 http://www.jimmunol.org/content/167/5/2734 Downloaded from

References This article cites 56 articles, 30 of which you can access for free at: http://www.jimmunol.org/content/167/5/2734.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2001 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. CD4؉ and CD8؉ T Cells Kill Intracellular Mycobacterium tuberculosis by a Perforin and Fas/Fas Ligand-Independent Mechanism1

David H. Canaday,2* Robert J. Wilkinson,*§ Qing Li,* Clifford V. Harding,† Richard F. Silver,*‡ and W. Henry Boom*

Cytotoxic effector phenotype and function of MHC-restricted Mycobacterium tuberculosis (MTB)-reactive CD4؉ and CD8؉ T lymphocytes were analyzed from healthy tuberculin skin test-positive persons. After stimulation in vitro with MTB, both CD4؉ and CD8؉ T cells up-regulated mRNA expression for granzyme A and B, granulysin, perforin, and CD95L (Fas ligand). mRNA levels for these molecules were greater for resting CD8؉ than CD4؉ T cells. After MTB stimulation, mRNA levels were similar for both T cell subsets. Increased perforin and granulysin protein expression was conﬁrmed in both in CD4؉ and CD8؉ T cells Downloaded from by ﬂow cytometry. Both T cell subsets lysed MTB-infected monocytes. Biochemical inhibition of the granule exocytosis pathway in CD4؉ and CD8؉ T cells decreased cytolytic function by >90% in both T cell subsets. Ab blockade of the CD95-CD95L -interaction decreased cytolytic function for both T cell populations by 25%. CD4؉ and CD8؉ T cells inhibited growth of intra cellular MTB in autologous monocytes by 74% and 84%, respectively. However, inhibition of perforin activity, the CD95-CD95L interaction, or both CTL mechanisms did not affect CD4؉ and CD8؉ T cell mediated restriction of MTB growth. Thus, perforin ,and CD95-CD95L were not involved in CD4؉ and CD8؉ T cell mediated restriction of MTB growth. The Journal of Immunology http://www.jimmunol.org/ 2001, 167: 2734–2742.

cquired protective immunity to Mycobacterium tubercu- servation, together with the knowledge that both CD4ϩ and CD8ϩ losis (MTB)3 in humans is dependent on T cells and MTB-specific cytolytic cells are present in healthy infected sub- A involves multiple T cell subsets (1, 2). However, the jects (15–21), has encouraged investigation of the role of T cell mechanisms used by T cells to restrict the growth of MTB within cytolytic function as protective effector function against MTB. macrophages are poorly understood. Cytokine secretion or cyto- Two major mechanisms of lymphocyte-mediated cytotoxicity toxic effector function has been proposed as major effector mech- are recognized, exocytosis of cytotoxic granules containing pore- anisms for lymphocyte-mediated control of intracellular patho- forming perforin and serine esterase granzyme molecules, and in- by guest on October 2, 2021 gens. In mice, IFN-␥, TNF-␣, and IL-12 play obligate roles in duction of apoptosis by ligation of CD95 (Fas) by CD95 ligand mycobacterial containment (3–7). In humans, the protective role of (CD95L; Refs. 22 and 23). A study of MHC class II-restricted ␥ ϩ IFN- and IL-12 is illustrated by rare mutations in the genes for CD4 T cell clones specific for purified protein derivative (PPD) ␥ receptors for IFN- or IL-12, or for IL-12 itself. These mutations showed expression of both perforin and CD95L as measured by severely compromise the ability to control replication of ordinarily semiquantitative PCR (24). Cytolytic mechanisms of MHC class ␥ ϩ nonpathogenic mycobacteria (8–11). Paradoxically, IFN- treat- I-restricted MTB-reactive CD8 T cells have not been studied in ment of macrophages in vitro only moderately inhibits growth of detail but are likely to involve the granule exocytosis pathway MTB in human mononuclear phagocytes suggesting an indirect (25). In gene-disrupted murine models, neither perforin-mediated rather than direct effector role of this cytokine (12–14). This ob- lysis nor CD95 ligation alone is sufficient to provide early protection against MTB (26, 27). There are contradictory observations on ϩ *Division of Infectious Diseases, †Department of Pathology, and ‡Division of Pul- the ability of human CD4 cytolytic cells to inhibit growth of monary and Critical Care Medicine, Case Western Reserve University School of MTB in monocytes (MN) or macrophages. Although M. bovis- Medicine and University Hospitals of Cleveland, Cleveland, OH 44106; and §Well- come Center for Clinical Tropical Medicine, Imperial College School of Medicine, bacillus Calmette-Gue´rin (BCG)-infected macrophages were lysed London, United Kingdom. by mycobacterial-specific CD4ϩ T cell clones and polyclonal Received for publication April 21, 2000. Accepted for publication June 6, 2001. lines, growth of M. bovis BCG was not inhibited (28, 29). Recent ϩ The costs of publication of this article were defrayed in part by the payment of page studies with MTB-reactive CD8 T cells suggest that CTL func- charges. This article must therefore be hereby marked advertisement in accordance tion was associated with the inhibition of intracellular MTB with 18 U.S.C. Section 1734 solely to indicate this fact. growth (30–32). 1 This work was funded by grants from the National Institutes of Health to D.H.C. (K08 AI 01581), W.H.B. (AI 27243), the Tuberculosis Research Unit (AI 95383), and Recent studies in humans have suggested that granulysin, R.F.S. (HL 59858), and from Wellcome Trust Fellowship in Clinical Tropical Med- present in cytotoxic cells, may be a mediator of mycobacterial icine and British Lung Foundation to R.J.W. (049525). growth inhibition (32). Granulysin was initially identified in acti- 2 Address correspondence and reprint requests to Dr. David H. Canaday, Division of vated T cells (33) and subsequently found in lytic granules of CTL Infectious Disease, Case Western Reserve University and University Hospitals, 10900 Euclid Avenue, BRB 1010B, Cleveland, OH 44106-4984. E-mail address: and NK cells (34, 35). Granulysin has been recently described in [email protected]. mycobacterial Ag-stimulated ␥␦ϩ, CD1-restricted CD8ϩ T cells, 3 Abbreviations used in this paper: MTB, Mycobacterium tuberculosis; CD95L, and CD4ϩ T cells (32, 36, 37). CD95 ligand; PPD, purified protein derivative; MN, monocyte; MOI, multiplicity of In this study, we sought to extend our studies that demonstrated infection; TRAIL, TNF-related apoptosis-inducing ligand; CMA, concanamycin A; ϩ ϩ MFI, mean fluorescence index; BCG, bacillus Calmette-Gue´rin. cytotoxic ability by MTB-specific CD4 and CD8 T cells and

MTB growth restriction by peripheral blood lymphocytes from RNA extraction and the multiprobe RNase protection assay PPDϩ individuals (17, 18). In this study, we analyzed CD4ϩ and ϩ RNA extraction from T cells was performed using the RNeasy mini kit CD8 T cells to examine the role of the cytolytic molecules in (Qiagen, Valencia, CA). RNA was frozen at Ϫ80°C until use. RNA (1–5 CTL function and T cell-mediated control of MTB growth. ␮g) was used in the RNase protection assay (BD PharMingen) according to the manufacturer’s instructions, and as described previously (40). Ribo- probes specific for granzymes A and B, perforin, granulysin (two probes), Materials and Methods TNF-related apoptosis-inducing ligand (TRAIL), and CD95L were synthe- Cultivation and processing of mycobacteria sized by using appropriate templates. One riboprobe for granulysin was Broth cultures of MTB strains H37Rv and H37Ra (American Type Culture synthesized from DNA templates designed to detect both the spliced and Collection (ATCC), Manassas, VA) were grown in sterile Middlebrook unspliced transcript of the NKG5 gene and a second to specifically detect 7H9 medium (Difco, Detroit, MI) with 10% ADC (Difco) enrichment and unspliced message (35). The difference between the expression of these 0.2% glycerol as described previously in detail (38). Plated cultures were mRNA was taken to indicate the presence of mRNA specific for granulysin grown on Middlebrook 7H10 agar with 10% OADC (Difco) enrichment (9 kDa) rather than the unspliced message, which gives rise to a 15-kDa and 0.5% glycerol. In preparation for infection of MN, mycobacteria were product. At all points of detection, spliced and unspliced transcript was processed by mechanical disruption and centrifugation, based on the meth- 15–42% higher than the mRNA specific for the 15-kDa product, implying ods of Schlesinger (39) and as previously described in detail (38). These the presence of a minority mRNA encoding the 9-kDa granulysin molecule methods minimize clumping and provide accurate quantification of the form (41). All data are presented as the total (spliced and unspliced) mes- inoculum. sage. Riboprobes for the housekeeping gene L32 were included in every assay. Hybridized and RNase digested products were electrophoresed on a PBMC and MN preparation 5% denaturing polyacrylamide gel. After drying, the gel was exposed overnight in a Geldoc 1000 (Bio-Rad, Hercules, CA). The identity of the pro- PBMC were isolated from healthy PPDϩ donors by density centrifugation tected bands was confirmed by reference to the unhybridized probes and of whole blood diluted 1/1 with RPMI 1640 (BioWhittaker, Walkersville, quantitated by reference to bands for the housekeeping gene L32. A ratio Downloaded from MD) over Ficoll (Pharmacia, Uppsala, Sweden). MN were obtained by of the density of bands for each molecule in each cell sample and at each adherence purification on plastic plates (Falcon, Lincoln Park, NJ). The time point was generated compared with the housekeeping gene L32. Com- plates were washed extensively after1hofadherence, and MN were parison for induction of mRNA then were determined by calculating the scraped off with a cell scraper (Costar, Cambridge, MA). CTL molecule mRNA/L32 mRNA ratio at each time point and comparing each back to the ratio of the cell type at t ϭ 0 (unstimulated). Expansion and purification of MTB-specific CD4ϩ and CD8ϩ

T cells Flow cytometry for intracellular molecules http://www.jimmunol.org/ To expand MTB-specific T cells, PBMC (2 ϫ 106 cells per 2 ml well) from ϩ After 7 days of MTB stimulation, intracellular detection of cytolytic ef- healthy PPD subjects were cultured with MTB H37Ra as described fector molecules was performed on purified and bulk T cell populations. above. The culture medium was RPMI 1640 supplemented with 10% Bulk cells were first stained with PE conjugated anti-CD4 or anti-CD8 pooled human serum (Gemini Bio-Products, Calabasas, CA), 20 mM ␮ (Caltag, Burlingame, CA). After one wash, cells were fixed with the Per- HEPES, 2 mM L-glutamine, 100 U/ml penicillin, and 100 g/ml strepto- mWash kit (BD PharMingen) per the manufacturer’s instructions or for 15 mycin. IL-2 (10 U/ml; Chiron, Emeryville, CA) was added after 48–72 h min in 2% paraformaldehyde, then washed in 0.2 M lysine followed by of culture and every 3–4 days thereafter. After 7–10 days of bulk culture, RPMI 1640 before permeabilization (0.1% saponin) and blocking (10% T cells were either purified as described below or restimulated. Before ␥␦ ϩ pooled human serum in RPMI 1640) on ice for 15 min. Anti-perforin-FITC restimulation, the bulk culture underwent depletion of TCR cells by (Ancell, Bayport, MN) or isotype control was added to this buffer for an

␥␦ by guest on October 2, 2021 incubation with an anti- Ab (clone 5A6.E9; ATCC). After washing, cells additional 30 min. Cells were washed twice in 0.1% saponin, RPMI 1640, then were incubated with goat-anti-mouse IgG-conjugated magnetic beads 10% FCS (HyClone, Logan, UT) solution and fixed with 2% paraformal- (Dynal, Oslo, Norway) and negatively selected with a magnetic particle dehyde before flow cytometric analysis. At analysis, CD4ϩ and CD8ϩ cells ϫ 6 separator (Dynal). The resulting purified cells were cultured (1 10 /ml) were individually gated. Studies for granulysin protein expression used ϫ 6 with fresh irradiated autologous PBMC feeders (1 10 /ml and MTB). purified CD4ϩ and CD8ϩ T cells. Anti-granulysin polyclonal rabbit anti- Additional IL-2 was added 2 days after restimulation. T cells were purified serum (provided by A. Krensky, Stanford University, Stanford, CA) or 7–9 days after restimulation. nonimmunized control serum was incubated with cells in permeabilization Immunomagnetic separation was performed on both resting primary Ј ϩ buffer for 30 min and then washed twice before adding goat F(ab )2 anti- PBMC or PBMC stimulated in vitro with MTB. CD8 T cells were pos- rabbit-FITC (Caltag). itively selected with anti-CD8-conjugated magnetic beads (Dynal) and De- tachabead (Dynal) according to the manufacturer’s guidelines. The MTB- 51 stimulated CD8ϩ T cell population was further purified by depletion of ␥␦ Cr release assay TCRϩ cells (as described above) and CD4ϩ T cells with anti-CD4 conju- ϩ MN were infected overnight with MTB H37Ra (7.5:1 MOI) in IMDM gated magnetic beads (Dynal). Unstimulated positively selected CD8 T ϩ ϩ (BioWhittaker) and 10% autologous serum. Cells then were loaded with cells were depleted of CD56 CD8 contaminating NK cells with anti- ␮ 51 51 CD56 beads in a MACS magnetic column separation system (Miltenyi 100 Ci of Cr (ICN) for 2 h and washed four times. Cr-loaded target Biotech, Auburn, CA) according to the manufacturers instructions. CD4ϩ MN were added to round-bottom wells and preincubated with anti-CD95 (ZB4, 2 ␮g/ml; Immunotech, Marseille, France) or isotype control (IgG1; T cells were positively selected by anti-CD4 conjugated magnetic beads ␮ and detached by Detachabead. Further purification of this subset by neg- Zymed) for 20 min at room temperature. EGTA (2–5 M final concentration) was added at the same time as T cells. After addition of T cells at ative selection was not necessary. ϫ Cells were analyzed by two-color flow cytometry as described previ- varying E:T ratios, plates were centrifuged for 30 s at 200 g to maximize ϩ ϩ cell contact. After 4–5 h of incubation at 37°C, 50 ␮l of supernatant was ously (18). Primary and activated CD4 and CD8 T cells were at least ϩ ϩ ϩ ϩ harvested and counted in a Beckman counter. All culture conditions were 95% CD3 /CD4 or CD8 and confirmed to be ␣␤ TCR . performed in triplicate. Proliferation assays For control perforin and granule exocytosis inhibition experiments, a long-term MTB-reactive CD8ϩ T cell line from an HLA-A2ϩ donor was Purified CD8ϩ and CD4ϩ T cells (5 ϫ 104 per 200 ␮l) were incubated for preincubated with concanamycin A (CMA; Sigma, 10 nM) for2horstron- 2 days in a round-bottom 96-well plate with irradiated autologous MN (5 ϫ tium chloride (25 mM; Sigma) for 20 h. Lysis activity was analyzed under 104) and MTB (multiplicity of infection (MOI) 5:1) and then pulsed over- two conditions for strontium: either immediately after 20 h of preincuba- night with 1 ␮Ci of [3H]thymidine (ICN, Costa Mesa, CA). Plates were tion or during the last4hofa24-h chase period. The effect of CMA on harvested onto glass wool filter with a Filtermate-196 harvester (Packard, lysis was tested after 2 h preincubation and during the last4hof24h Meriden, CT) and cpm counted with Matrix 96 counter (Packard). In ex- continuous CMA exposure. Treated T cells were cultured with MTB-in- periments with blocking Abs, MN were preincubated at 4°C with mAbs for fected 51Cr-loaded THP1 targets (HLA-A2ϩ; ATCC). Maximum release 45 min before addition of T cells and Ags. Abs remained in the cultures was determined by incubating cells with 3% SDS, and spontaneous release during the entire assay. Anti-MHC class I Ab, W6/32 (ATCC), anti-MHC- was determined by incubating target cells in complete medium alone. The II, L243 (BD PharMingen, San Diego, CA), and IgG2a isotype control percentage of specific release was determined by the equation [(cpm ex- (Zymed, South San Francisco, CA) were used at 5 ␮g/ml final perimental Ϫ cpm spontaneous release)/(cpm maximum release Ϫ cpm concentration. spontaneous release)] ϫ 100%. 2736 CYTOTOXIC MECHANISMS OF MTB-SPECIFIC T CELLS

DNA fragmentation assay Stimulation with MTB results in increases in mRNA expression ϩ ϩ Jurkat cells were loaded for 4 h with 1 ␮Ci/ml of [3H]thymidine (ICN) and for cytolytic molecules in both CD4 and CD8 T cells washed. Jurkat were preincubated with anti-CD95 (5 ␮g/ml final concen- First, the kinetics of mRNA expression of secreted and surface- tration), and a cell line (KFL9), transfected to express high levels of associated cytotoxic effector molecules in PBMC from healthy CD95L (Ref. 42; a gift from D. Kaplan, Case Western Reserve University), ϩ was preincubated with anti-CD95L (NOK-1; 5 ␮g/ml final concentration; PPD donors were determined after MTB stimulation. Cells were BD PharMingen) for 30 min. Cells then were combined at a 1:1 ratio. At restimulated with MTB and autologous feeders on days 7 and 14. 24 h, cells were harvested onto a glass wool filter with a Filtermate-196 Cell lysates for RNA extraction were prepared on day 0, 3, 5, 7, 10, harvester (Packard). The cpm were counted on Matrix96 counter (Pack- 14, and 17. mRNA levels for cytotoxic effector molecules were ard). The percentage of specific killing was determined by the equation [(cpm Jurkat alone Ϫ cpm Jurkat and KFL9)/(cpm Jurkat)] ϫ 100%. measured by quantitative RNase protection assay and compared with mRNA levels for the housekeeping gene L32 (Fig. 2A). Intracellular MTB growth assay mRNA for granzyme A and B, granulysin, perforin, TRAIL, and Autologous MN (1 ϫ 105/well) were cultured in triplicate in round-bottom CD95L were selected for analysis because of their well-defined 96-well plates. After overnight culture, supernatants were removed and roles in cytotoxic effector mechanisms (43). A representative ␮ H37Rv added at a MOI of 1:1 in 100 l/well of IMDM plus 30% non- RNase protection assay gel is shown in Fig. 2B. Up-regulation of heat-inactivated autologous serum. After1hofincubation at 37°C, supernatants were aspirated and each well was washed three times to remove mRNA for these genes reached a steady state between days 10–20; ϩ ϩ noningested mycobacteria. Anti-CD95 and anti-CD95L at 5 ␮g/ml each or therefore, this time period was chosen to study CD4 and CD8 isotype IgG1 at 10 ␮g/ml were added to MN or T cells 20 min before T cells individually. addition of purified T cells. CMA was added at 10 nM to T cells for 2 h The next series of experiments determined whether cytotoxic

before addition to infected MN and maintained at this concentration after ϩ ϩ Downloaded from T cells were added. All T cells were resuspended in IMDM (no antibiotics) molecule mRNA expression in purified CD4 and CD8 T cells with 10% autologous serum and added so the total volume/well was 200 differed in response to MTB stimulation. PBMC from seven ␮l. After 1 and 24 h, wells were aspirated, and 100 ␮l of lysis buffer healthy PPDϩ donors were cultured in the presence of MTB. At (0.067% SDS in Middlebrook 7H9) added to each well. Plates were incu- the end of culture, MTB-activated CD4ϩ and CD8ϩ T cells were bated at 37°C for 10 min followed by neutralization of SDS with 300 ␮lof purified by positive selection and compared with resting CD4ϩ PBS with 20% BSA. Lysates from triplicates wells were pooled, and four ϩ 10-fold serial dilutions of the resultant 600 ␮l lysate in 7H9 were prepared. and CD8 T cells, purified from PBMC. Aliquots of each of these Similar serial dilutions of supernatants were prepared. Six 10-␮l aliquots of four cell populations were analyzed for mRNA expression by http://www.jimmunol.org/ each dilution of lysate and supernatant were plated onto Middlebrook 7H10 RNase protection assay. agar and incubated in 5% CO2 at 37°C until colonies were large enough to ϩ ϩ be counted. Results were expressed as CFU/ml of lysate (corresponding to Fig. 3 demonstrates ratios of mRNA in resting CD4 and CD8 CFU/106 cultured MN). T cells for granzyme A and B, granulysin, perforin, and CD95L compared with the housekeeping gene L32. Levels of mRNA were Statistics significantly higher in nonactivated CD8ϩ T cells than CD4ϩ T All statistics were determined by paired t test analysis. cells ( p Ͻ 0.05) for granzyme B, granulysin, and perforin. Dif- ferences for granzyme A did not reach statistical significance, but Results ϩ the trend suggested that CD8 T cells had higher baseline levels by guest on October 2, 2021 ϩ ϩ MHC restriction of MTB-reactive CD4 and CD8 T cell lines ( p ϭ 0.06). Overall, mRNA levels for cytolytic effector molecules ϩ ϩ CD4ϩ and CD8ϩ lines were generated by stimulating PBMC from were from 5.6 to 9.4 times higher in CD8 than in CD4 T cells healthy mycobacterial Ag-sensitized individuals with MTB as de- at baseline. Fig. 2B, lanes 1 and 2, illustrates these differences in ϩ ϩ scribed in Materials and Methods. CD4ϩ and CD8ϩ T cells were baseline mRNA levels between resting CD4 and CD8 T cells. purified by Ab-coated magnetic beads and tested for Ag specificity After stimulation with MTB, increased expression of mRNA for ϩ with autologous MN infected with MTB. As shown in Fig. 1, both granzymes A and B and granulysin in 7/7 CD4 T cell lines and ϩ T cell subsets proliferated to MTB-infected MN but were inhibited 5/5 CD8 T cell lines was measured (Fig. 3). Increased expression ϩ ϩ ( p Ͻ 0.05) by anti-MHC-I (W6/32) for CD8ϩ T cells and anti- of perforin mRNA was seen in 6/7 CD4 lines and all CD8 lines. MHC-II (L243) for CD4ϩ T cells. Consistent with our previous CD95L mRNA expression was up-regulated in 6/7 CD4ϩ lines studies, both T cell subsets lysed autologous MTB infected MN and 3/5 CD8ϩ lines, and the average increase was less than that (15, 17). observed for any of the granule exocytosis pathway molecules.

FIGURE 1. MTB speciﬁcity and MHC restriction of CD4ϩ and CD8ϩ T cell lines. MN (5 ϫ 104) were infected for 2 h with MTB (MOI 5:1) and then washed. MN were incubated with anti- MHC I (W6/32) or anti-MHC II (L243) or isotype at 5 ␮g/ml. 5 ϫ 104 MTB speciﬁc CD4ϩ or CD8ϩ T cells were added for a 3-day proliferation assay. Results are expressed as cpm and are representative of four experiments. Error bars indicate SEM. The Journal of Immunology 2737

FIGURE 2. A, Kinetics of mRNA expression for cytotoxic effector molecules in PBMC stimulated with MTB. PBMC were stimulated with MTB and IL-2 (20 U/ml) was added to cultures after 48 h. At indicated time points, cells were harvested for RNA extraction. mRNA expression was determined by RNase protection assay as described in Materials and Methods. All mRNA were normalized relative to housekeeping gene L32 mRNA expression, and fold induction was calculated compared with mRNA levels at time 0, which was set at 1. CD95 and TRAIL were measured only at days 0, 3, and 10, all others at days 0, 3, 5, 7, 10, 14, and 17. Results represent fold induction of two donors. B, Representative RNase protection assay. A Downloaded from gel is shown from one of the donors in this experiment. Lane 1, Resting CD8ϩ T cells. Lane 2, Resting CD4ϩ T cells. Lane 3, MTB stimulated PMBC at day 3. Lane 4, MTB stimulated PBMC at day 5. http://www.jimmunol.org/

The median fold induction of mRNA in CD4ϩ and CD8ϩ T cells achieve similar cytotoxic molecule mRNA levels to those in CD8ϩ by guest on October 2, 2021 after MTB stimulation for granzyme A was 12 (CD4ϩ) and 5.4 T cells after MTB activation. (CD8ϩ); for granzyme B, 11 (CD4ϩ) and 2.4 (CD8ϩ); for granu- Because IL-2 was added to all cultures after 48–72 h, it was lysin, 28 (CD4ϩ) and 3.8 (CD8ϩ); for perforin, 9.7 (CD4ϩ) and necessary to determine that IL-2 alone was not responsible for the 2.6 (CD8ϩ); and for CD95L, 2.8 (CD4ϩ) and 1.4 (CD8ϩ), respec- observed changes in gene expression. Resting CD4ϩ T cells were tively. After MTB stimulation, the median fold induction of positively selected from PBMC and cultured for 7 days with au- mRNA was smaller in CD8ϩ T cells than in CD4ϩ T cells. Higher tologous MN with MTB, IL-2, or both. Fig. 4 demonstrates that 20 levels of mRNA induction in CD4ϩ T cells allowed them to U/ml IL-2 alone did not signiﬁcantly increase mRNA expression

FIGURE 3. Up-regulation of cytotoxic molecule mRNA in CD4ϩ and CD8ϩ T cells by MTB. PBMC underwent two rounds of in vitro stimulation with MTB, except donors 5 and 7, who had one round. CD4ϩ and CD8ϩ T cells were puriﬁed by positive selection from these cultures and freshly isolated PBMC analyzed for mRNA expression by RNase protection assay. mRNA for each protein was determined by dividing the densitometry of that mRNA value by the housekeeping gene L32, which was set at 1. mRNA values are expressed as densitometry ratios. 2738 CYTOTOXIC MECHANISMS OF MTB-SPECIFIC T CELLS

analyzed after one round of in vitro stimulation by flow cytometry. After MTB stimulation, the mean fluorescence index (MFI) and percentage of positive cells increased for both perforin and granulysin above levels measured in resting CD4ϩ and CD8ϩ T cells (Fig. 5). In the four donors studied, the average MFI for perforin was 10-fold higher in unstimulated CD8ϩ than CD4ϩ T cells. After MTB stimulation, the average MFI for perforin was still sixfold higher in CD8ϩ than CD4ϩ T cells. A mean of 33% of CD4ϩ T cells and 72% CD8ϩ T cells expressed perforin after MTB stimulation. The average MFI for granulysin was 3.1-fold higher in unstimulated CD8ϩ than CD4ϩ T cells. After MTB stimulation, the average MFI was 1.7-fold higher in CD8ϩ than CD4ϩ T cells. A mean of 40% of CD4ϩ T cells and 66% CD8ϩ T cells expressed granulysin after MTB stimulation. These data demon- strate that perforin and granulysin protein expression is higher in both resting and MTB-activated CD8ϩ T cells compared with ϩ FIGURE 4. mRNA expression for mediators of cytotoxicity in MTB- CD4 T cells. These results indicate that up-regulation of mRNA stimulated, IL-2-stimulated, and MTB plus IL-2-stimulated CD4ϩ T cells. was associated with increased protein expression. Low levels of ϩ Positively selected CD4ϩ T cells from two donors were cultured for 7 days perforin in resting CD4 T cells have been observed by others (44, Downloaded from with autologous MTB-infected MN. mRNA expression was determined by 45). Western blotting of cell lysates from both CD4ϩ and CD8ϩ T RNase protection assay. As described in Materials and Methods, all cells with the polyclonal anti-granulysin antiserum confirmed that mRNA were normalized relative to the housekeeping gene L32 mRNA there were 15- and 9-kDa bands corresponding to the unprocessed expression and then fold induction calculated compared with mRNA levels and bioactive forms of granulysin (data not shown; Ref. 35). at time 0, which was set at 1. Error bars indicate SD. Comparison of cytotoxic mechanisms used by CD4ϩ and CD8ϩ http://www.jimmunol.org/ for the molecules tested, and more importantly, IL-2 did not fur- T cells ther enhance responses to MTB. Based on these findings, we con- Next we determined the mechanisms used by CD4ϩ and CD8ϩ T ϩ tinued to use IL-2 to expand CD8 T cells to achieve adequate cells to lyse MTB-infected MN. CD4ϩ and CD8ϩ T cell lines were numbers of cells to perform our studies. generated from PBMC from PPDϩ subjects (n ϭ 5). Fig. 6 dem-

ϩ ϩ onstrates the mean lytic activity of these puriﬁed T cells for MTB- Up-regulation of cytotoxic effector proteins in CD4 and CD8 infected MN. CD4ϩ and CD8ϩ T cells were equally cytolytic, T cells after stimulation with MTB consistent with our previous studies (17). Cytotoxicity by CD8ϩ T To determine whether increased mRNA expression induced by cells was partially inhibited by anti-CD95 Ab (25%; p Յ 0.02). ϩ MTB stimulation correlated with increased protein expression, Cytolytic activity of CD4 T cells was inhibited to a similar de- by guest on October 2, 2021 PBMC from four PPDϩ donors were stimulated with MTB and gree (26%; p Յ 0.01). The inhibitory activity of the anti-CD95 Ab,

FIGURE 5. Increased expression of perforin and granulysin protein in MTB activated CD4ϩ and CD8ϩ T cells. PBMC were cultured for 7 days with MTB. Intercellular staining for perforin (top) and granulysin (bottom) was performed as described in the Materials and Methods. The thin line represents unstimulated cells, and the thick line MTB-activated cells. Staining with isotype control or control rabbit anti-sera (data not shown) was performed for each condition and cell type, and used individually for each cell type and condition to calculate percent positive MFI. Control staining was slightly higher in activated T cells compared with unstimulated cells. This ﬁgure is representative of four experiments. The Journal of Immunology 2739

FIGURE 6. Inhibition of lytic activity of MTB- stimulated CD4ϩ and CD8ϩ T cell lines. Results represent mean cytotoxicity of MTB-infected MN by MTB- stimulated T cell lines from 5 PPDϩ donors. The highest E:T ratio for each donor is shown, which varied between 40:1 to 75:1. Targets were 51Cr-pulsed MTB-infected autologous MN, which were treated with EGTA or anti-CD95 Ab. Error bars indicate SEM.

veriﬁed by using Jurkat cells and CD95L transfected K562 cells (usually 3:1 in most experiments because of limited cell number) Downloaded from (KFL9), inhibited DNA fragmentation of Jurkat by 60–100% in reduced MTB CFU by 84% ( p Ͻ 0.001). Comparing growth in- 6 h (data not shown). EGTA, which inhibits granule exocytosis, hibition by both T cell subsets at the 3:1 ratio suggested that CD8ϩ completely inhibited cytotoxicity by CD4ϩ and CD8ϩ T cells T cells were more effective on a per cell basis than CD4ϩ T cells ( p Ͻ 0.005 for both T cell types). In addition, another perforin in controlling MTB growth. For both CD4ϩ and CD8ϩ T cells, the inhibitor, CMA, which induces depolymerization of perforin and mean number of organisms released into the supernatant at 24 h loss of lytic function, was used in selected experiments and was as was Ͻ10% of the total number of organisms present in the day effective as EGTA in inhibiting CTL activity (see Fig. 8B and 1-infected MN alone. Therefore, the observed decrease in intra- http://www.jimmunol.org/ Ref. 46). cellular MTB was not simply attributable to a shift to the extra- cellular compartment but represented true killing of MTB. Inhibition of intracellular growth of MTB in MN by CD4ϩ and CD8ϩ T cells CD4ϩ and CD8ϩ T cell-mediated MTB growth inhibition is independent of perforin and CD95-CD95L Next, we sought to determine whether cytotoxic MTB-speciﬁc CD4ϩ and CD8ϩ T cell lines inhibited intracellular growth of Next we determined whether inhibition of perforin or CD95- MTB by using a CFU assay. Autologous MN were infected with CD95L interactions affected MTB growth inhibition by CD4ϩ and ϩ

1:1 MOI virulent MTB H37Rv and incubated with no additional CD8 T cells. These experiments required that inhibition re- by guest on October 2, 2021 cells (negative control), or with MTB-activated CD4ϩ (seven do- mained effective during the 24-h period of the MTB growth inhi- nors) or CD8ϩ (six donors) T cell lines. Cells were harvested for bition assay. We considered two agents to inhibit perforin func- enumeration of CFU after 24 h of coculture. Fig. 7 demonstrates tion, strontium chloride, which induces degranulation, and CMA, that both CD4ϩ and CD8ϩ T cells significantly inhibited growth of which inhibits perforin polymerization to block perforin-mediated MTB. Addition of CD4ϩ T cells at the highest T cell-MN ratio CTL function. As shown in Fig. 8A, strontium and CMA both (usually 10:1) was associated with a 74% reduction in growth of inhibit lysis of MTB-infected targets by CD8ϩ T cells. EGTA was MTB at 24 h ( p Ͻ 0.001). CD8ϩ T cells at the highest ratio not suitable for the 24-h assay because of its toxicity. CTL function was inhibited after2hofCMApretreatment and remained inhibited even after 24 h of continuous exposure (Fig. 8A). Treatment of CD8ϩ or CD4ϩ T cells with CMA for up to 24 h was not associated with significant cell loss (trypan blue), and did not result in defective IFN-␥ production (Fig. 8B). Strontium was preincubated with CD8ϩ T cells for 20 h to induce degranulation. T cells were washed to remove strontium, and cells were added immediately or after a 24-h chase period to 51Cr-loaded targets. As shown in Fig. 8A, strontium initially completely inhibited CTL activity, but after a 24-h chase, CTL activity recovered. Recovery of perforin protein expression during the 24 h chase period after strontium treatment was confirmed also by intracellular flow cytometry (data not shown). Strontium treatment induced significant T cell death, which precluded its use continuously during the 24-h MTB growth inhibition assay. Fig. 8C demonstrates that a combination of anti-CD95 and anti-CD95L Abs resulted in 70% inhibition of apoptosis in a 24 h assay of Jurkat cells by a CD95L- FIGURE 7. Growth inhibition of MTB by CD4ϩ and CD8ϩ T cell lines. transfected cell line (KFL9) as measured by the DNA Autologous MN were infected with MTB H37Rv. After washing off non- phagocytosed bacilli, MN were cultured in the absence of additional cells fragmentation assay. ϩ ϩ CMA and anti-CD95-CD95L Abs were used to determine the or in the presence of purified MTB-activated CD4 or CD8 T cells. ϩ Results show the mean of eight experiments from seven different donors role of perforin and CD95-CD95L interaction on CD4 - and ϩ for CD4ϩ T cells and seven experiments with cells from six donors for CD8 -mediated MTB growth inhibition. Fig. 9 demonstrates that CD8ϩT cells. Error bars indicate SEM. CMA, despite its ability to inhibit cytotoxicity, had no effect on T 2740 CYTOTOXIC MECHANISMS OF MTB-SPECIFIC T CELLS

FIGURE 9. The role of perforin and CD95-CD95L in T cell-mediated MTB growth inhibition. CD4ϩ and CD8ϩ T cells were preincubated for 2 h with CMA, anti-CD95L Ab, or the combination of CMA and anti-CD95L Ab. MN were pretreated with anti-CD95 Ab in all experimental groups Downloaded from where T cells were treated with anti-CD95L. Inhibitors remained present throughout the 24-h assay. T cell-MN ratio was usually 10:1. The data represent the results for CD4ϩ and CD8ϩ T cells from four different donors. Error bars indicate SEM.

ciation between the lytic and growth inhibitory functions of MTB http://www.jimmunol.org/ speciﬁc CD4ϩ and CD8ϩ T cells.

Discussion Expression of granzyme A and B, granulysin, perforin, and CD95L was assessed by quantitative RNase protection assay after stimulation with MTB. Both CD4ϩ and CD8ϩ T cells up-regulated mRNA expression for these molecules. Both subsets were cyto-

lytic, and this action was more sensitive to inhibition of the granule by guest on October 2, 2021 exocytosis pathway than the CD95-CD95L pathway. CD4ϩ and CD8ϩ T cells significantly inhibited growth of intracellular MTB in MN. The effect of T cells on the growth of MTB was not sus- ceptible to inhibition of the granule exocytosis or CD95 pathways. Therefore, although MTB specific cytolytic T cells of both the CD4ϩ and CD8ϩ phenotype use similar and overlapping effector mechanisms to lyse infected targets, neither perforin nor signaling FIGURE 8. Stability of perforin and CD95-CD95L inhibitors for 24 h through CD95 is required for restricting early growth of MTB MTB growth inhibition assay. A, MTB-specific CD8ϩ T cells were incu- within MN. bated initially with strontium chloride (25 mM) for 20 h and then medium The fold increase of mRNA relative to the housekeeping gene for a 24-h chase period or with CMA (10 nM) for2hofpretreatment and L32 was higher in CD4ϩ T cells because of significantly lower then left in for the full 24 h. The cells were added immediately after the baseline levels of mRNA expression for the cytotoxic effector mol- 51 pretreatments or at the last4hofthe24-h incubation to Cr-loaded target ecules studied. Because of greater mRNA induction in CD4ϩ T cells. Results are expressed as percentage of lysis and are representative of cells, final mRNA levels for cytotoxic effector molecules were two experiments. B, Purified CD4ϩ or CD8ϩ T cells (5 ϫ 104) were in- similar for CD4ϩ but never higher than CD8ϩ T cells after MTB cubated with 5 ϫ 104 autologous MN Ϯ PHA (3 ␮g/ml) and Ϯ CMA. Supernatant was harvested at 24 h and IFN-␥ measured by ELISA as de- stimulation. Low-dose IL-2 was added to bulk cultures to increase scribed previously (18). This experiment is representative of two experi- T cell expansion to obtain sufficient numbers of cells to perform ments. Error bars indicate SD. C, CD95L-expressing cells (KFL9) were RNase protection assays and functional studies. IL-2 alone was not incubated with [3H]thymidine-pulsed Jurkat cells (CD95-sensitive target) responsible for the observed up-regulation of cytotoxic effector for 24 h in the presence of anti-CD95 and anti-CD95L Abs. DNA frag- molecule mRNA in CD4ϩ T cells. mentation results are expressed as percent specific killing and is represen- After MTB stimulation, the ability of both CD4ϩ and CD8ϩ T tative of two experiments. Error bars indicate SEM. cells to lyse MTB-infected targets was similar. Cytolysis of MTB infected MN in 4-h 51Cr release assay by CD4ϩ and CD8ϩ T cells was mediated primarily through the granule exocytosis pathway. cell-mediated control of intracellular MTB growth. Similarly, anti- The importance of CD95-related killing may be underrepresented CD95 and CD95L Abs had no effect. Combination of CMA and in the 4- to 5-h 51Cr release assay because cell lysis mediated by anti-CD95-CD95L Abs did not reduce the CD4ϩ-orCD8ϩ-me- the interaction of CD95-CD95L is slower than perforin-mediated diated inhibition of MTB growth. CMA and anti-CD95-CD95L lysis. Our data contradict the classical view that cell lysis by CD4ϩ Abs did not affect the growth of MTB in MN in the absence of T T cells is initiated primarily by the CD95-CD95L pathway (47). In cells (data not shown). These results suggest that there was disso- fact, up-regulation of CD95L was detected on MTB-activated T The Journal of Immunology 2741 cells from only one of four donors using a novel highly sensitive and Lowrie (31) found that only some murine CD4ϩ and CD8ϩ T flow cytometry enhancement technique that overcomes sensitivity cell clones inhibited mycobacterial growth and only apparently limitations for detection of CD95L (data not shown; Ref. 48). Fur- through cytotoxic granule release. They also used strontium to thermore, Oddo et al. (49) found that MTB infection of macro- inhibit the granule exocytosis pathway. They found a direct cor- phages caused down-regulation of CD95 on the cell surface, which relation between amount of granzyme A released and degree of also could explain why there was only a small contribution by the growth restriction. Their T cell lines were much less potent in CD95-CD95L pathway to cytolysis of infected MN. Our study controlling MTB growth because 50:1 T cell-APC ratios were nec- used polyclonal populations of T cells that may closely represent essary, in contrast to our studies where 5–10:1 ratios were suffi- the mixed population of T cells stimulated by MTB in vivo. Inhi- cient. Similar to Silva and Lowrie’s findings (31), our experiments bition of the CD95-CD95L pathway resulted in only partial loss of demonstrated that inhibition of the CD95-CD95L interaction also cytolytic activity by CD4ϩ and CD8ϩ T cells. This suggests that did not inhibit T cell-mediated control of MTB. This may be in neither subset of T cells initiated cytotoxicity primarily through the part because MTB-activated CD4ϩ and CD8ϩ T cells expressed CD95-CD95L pathway, or MN are less sensitive to CD95-CD95L insufficient levels of CD95L. Others have shown that high con- mediated lysis. Studies of PPD-specific CD4ϩ T cell clones, which centration of soluble CD95L can inhibit mycobacterial growth in found that use of the CD95-CD95L pathway depended on the tar- macrophages (49). get used, showed MN were less CD95L-sensitive targets, consis- Our experiments do not exclude a role for apoptosis through tent with our data (24). CD95-CD95L-independent mechanisms. ATP-induced apoptosis To inhibit granule exocytosis, we evaluated strontium chloride, in macrophages resulted in killing of M. bovis BCG (53, 54). In which has been used recently to explore cytotoxic and growth addition, mycobacterial infection itself can induce apoptosis in in- Downloaded from restriction mechanisms (31, 32). Strontium was initially shown to fected macrophages (55). Ingestion of infected apoptotic macro- induce degranulation in resting NK cells in nonactivated PBMC phages by uninfected macrophages results in increased control of (50). NK cells remained degranulated for up to 24 h. However, mycobacterial growth (56). Cytokine-secreting T cells could con- activated T cells rapidly repopulate their granules (51). In exper- tribute indirectly to this mechanism by activating these uninfected iments with MTB-activated T cells, we found that after strontium macrophages. Direct cell contact between T cells and macrophages treatment, lytic activity returned by 24 h. Therefore, we did not also is required for mycobacterial growth inhibition (14). Thus, feel that strontium was optimal for inhibition to evaluate the role combinations of cytokine secretion and cell contact may be nec- http://www.jimmunol.org/ of CTL activity in the ability of T cells to control MTB growth. We essary for optimal mycobacterial control by T cells. chose CMA, which inhibits perforin for 24 h and was not associ- Greater insight into the relative roles of cell contact, cytolytic ated with significant toxicity. These experiments demonstrated that effector function mechanisms, and cytokine secretion by T cells perforin had no effect on MTB growth inhibition by CD4ϩ and will determine how protective immunity successfully controls in- CD8ϩ T cells. tracellular growth of MTB in the majority of healthy individuals. These results contrast with findings where CD1- and MHC class I-restricted T cells were able to inhibit the intracellular growth of Acknowledgments

MTB through a granule exocytosis-dependent pathway (31, 32). We are grateful to Dr. A. Krensky of Stanford University for the kind gift by guest on October 2, 2021 They used strontium to inhibit the granule exocytosis pathway. of anti-granulysin Ab, Zahra Toossi and Fred Heinzel from Case Western There are at least three possible reasons for the differences in our Reserve University for reviewing the manuscript, and David Kaplan from results. First, CMA inhibits perforin only, whereas strontium Case Western Reserve University for helping us with flow cytometry anal- blocks the entire granule exocytosis pathway, which includes gran- ysis for CD95L and providing us with KFL9 and Jurkat cells. zymes and granulysin. These other granule contents may be able to compensate for the loss of perforin. Second, our polyclonal MHC- References restricted T cell lines may use different CTL mechanisms than the 1. Boom, W. H. 1996. The role of T-cell subsets in Mycobacterium tuberculosis infection. Infect. Agents Dis. 5:73. CD1-restricted T cells. Third, the possibility of residual strontium 2. Ellner, J. J. 1997. Review: the immune response in human tuberculosis-implica- toxicity to the T cells cannot be excluded. The absence of an in- tions for tuberculosis control. J. Infect. Dis. 176:1351. 3. Cooper, A. M., J. Magram, J. Ferrante, and I. M. Orme. 1997. Interleukin 12 hibitory effect of CMA on T cell ability to restrict MTB growth (IL-12) is crucial to the development of protective immunity in mice intrave- suggests that MN ability to present MTB was not significantly nously infected with Mycobacterium tuberculosis. J. Exp. Med. 186:39. effected. 4. MacMicking, J. D., R. J. North, R. LaCourse, J. S. Mudgett, S. K. Shah, and C. F. Nathan. 1997. Identification of nitric oxide synthase as a protective locus Granulysin expression does not appear restricted to unique T against tuberculosis. Proc. Natl. Acad. Sci. USA 94:5243. cell subsets as initially thought (32, 34, 36, 37). All T cell lines 5. Kindler, V., A. P. Sappino, G. E. Grau, P. F. Piguet, and P. Vassalli. 1989. The tested expressed granulysin. CMA inhibits perforin activity but inducing role of tumor necrosis factor in the development of bactericidal gran- ulomas during BCG infection. Cell 56:731. does not block degranulation. Granzymes in the absence of per- 6. Cooper, A., D. Dalton, T. Stewart, J. Griffin, D. Russell, and I. Orme. 1993. forin have very little lytic and apoptosis-inducing activity (52). In Disseminated tuberculosis in interferon-␥ gene-disrupted mice. J. Exp. Med. 178: 2243. addition, Stenger et al. (32) used purified perforin and granulysin 7. Flynn, J., J. Chan, K. Triebold, D. Dalton, T. Stewart, and B. Bloom. 1993. An to show that perforin was required for granulysin to kill mycobac- essential role for Interferon-␥ in resistance to Mycobacterium tuberculosis infec- teria. Granulysin alone incubated with infected macrophages did tion. J. Exp. Med. 178:2249. 8. Altare, F., D. Lammas, P. Revy, E. Jouanguy, R. Doffinger, S. Lamhamedi, not effect the growth of mycobacteria. Because perforin inhibition P. Drysdale, D. Scheel-Toellner, J. Girdlestone, P. Darbyshire, et al. 1998. In- in our studies did not effect T cell-mediated growth inhibition, the herited interleukin 12 deficiency in a child with bacille Calmette-Guerin and role of granulysin based on our studies is unclear. Salmonella enteritidis disseminated infection. J. Clin. Invest. 102:2035. 9. Dorman, S. E., and S. M. Holland. 1998. Mutation in the signal-transducing chain The dissociation between T cell-mediated cytolytic activity and of the interferon-␥ receptor and susceptibility to mycobacterial infection. J. Clin. growth inhibition was an unexpected finding. We were able to Invest. 101:2364. effectively inhibit perforin-mediated cell lysis of MN without af- 10. Altare, F., A. Durandy, D. Lammas, J. F. Emile, S. Lamhamedi, F. Le Deist, P. Drysdale, E. Jouanguy, R. Doffinger, F. Bernaudin, et al. 1998. Impairment of fecting intracellular growth of MTB. These results are consistent mycobacterial immunity in human interleukin-12 receptor deficiency. Science with studies in perforin and granzyme knockout mice, which in- 280:1432. 11. Newport, M. J., C. M. Huxley, S. Huston, C. M. Hawrylowicz, B. A. Oostra, dicated that loss of these granule exocytosis molecules alone were R. Williamson, and M. Levin. 1996. A mutation in the interferon-␥ receptor gene not required for early restriction of MTB growth (26, 27). Silva and susceptibility to mycobacterial infection. N. Engl. J. Med. 335:1941. 2742 CYTOTOXIC MECHANISMS OF MTB-SPECIFIC T CELLS

12. Rook, G. A., J. Steele, L. Fraher, S. Barker, R. Karmali, J. O’Riordan, and 34. Pena, S. V., and A. M. Krensky. 1997. Granulysin, a new human cytolytic gran- J. Stanford. 1986. Vitamin D3, ␥ interferon, and control of proliferation of My- ule-associated protein with possible involvement in cell-mediated cytotoxicity. cobacterium tuberculosis by human monocytes. Immunology 57:159. Semin. Immunol. 9:117. 13. Hirsch, C. S., T. Yoneda, L. Averill, J. J. Ellner, and Z. Toossi. 1994. Enhance- 35. Pena, S. V., D. A. Hanson, B. A. Carr, T. J. Goralski, and A. M. Krensky. 1997. ment of intracellular growth of Mycobacterium tuberculosis in human monocytes Processing, subcellular localization, and function of 519 (granulysin), a human by transforming growth factor-␤1. J. Infect. Dis. 170:1229. late T cell activation molecule with homology to small, lytic, granule proteins. 14. Silver, R. F., Q. Li, W. H. Boom, and J. J. Ellner. 1998. Lymphocyte-dependent J. Immunol. 158:2680. inhibition of growth of virulent Mycobacterium tuberculosis H37Rv within hu- 36. Spada, F. M., E. P. Grant, P. J. Peters, M. Sugita, A. Melian, D. S. Leslie, ϩ man monocytes: requirement for CD4 T cells in purified protein derivative- H. K. Lee, E. van Donselaar, D. A. Hanson, A. M. Krensky, et al. 2000. Self- positive, but not in purified protein derivative-negative subjects. J. Immunol. recognition of CD1 by ␥␦ T cells: implications for innate immunity. J. Exp. Med. 160:2408. 191:937. ϩ 15. Tsukaguchi, K., K. N. Balaji, and W. H. Boom. 1995. CD4 ␣␤ T cell and ␥␦ 37. Ochoa, M. T., S. Stenger, P. A. Sieling, S. Thoma-Uszynski, S. Sabet, S. Cho, T cell responses to Mycobacterium tuberculosis: similarities and differences in A. M. Krensky, M. Rollinghoff, E. Nunes Sarno, A. E. Burdick, T. H. Rea, and Ag recognition, cytotoxic effector function, and cytokine production. J. Immunol. R. L. Modlin. 2001. T-cell release of granulysin contributes to host defense in 154:1786. leprosy. Nat. Med. 7:174. 16. Balaji, K. N., S. K. Schwander, E. A. Rich, and W. H. Boom. 1995. Alveolar 38. Silver, R. F., Q. Li, and J. J. Ellner. 1998. Expression of virulence of Mycobac- macrophages as accessory cells for human ␥␦ T cells activated by Mycobacterium terium tuberculosis within human monocytes: virulence correlates with intracel- tuberculosis. J. Immunol. 154:5959. lular growth and induction of tumor necrosis factor ␣ but not with evasion of 17. Tan, J. S., D. H. Canaday, W. H. Boom, K. N. Balaji, S. K. Schwander, and lymphocyte-dependent monocyte effector functions. Infect. Immun. 66:1190. E. A. Rich. 1997. Human alveolar T lymphocyte responses to Mycobacterium ϩ ϩ 39. Schlesinger, L. S. 1993. Macrophage phagocytosis of virulent but not attenuated tuberculosis antigens: role for CD4 and CD8 cytotoxic cells and relative re- strains of Mycobacterium tuberculosis is mediated by mannose receptors in ad- sistance of alveolar macrophage to lysis. J. Immunol. 159:290. dition to complement receptors. J. Immunol. 150:2920. 18. Canaday, D. H., C. Ziebold, E. H. Noss, K. A. Chervenak, C. V. Harding, and 40. Wilkinson, R. J., P. Patel, M. Llewelyn, C. S. Hirsch, G. Pasvol, G. Snounou, ϩ ␣␤ ϩ W. H. Boom. 1999. Activation of human CD8 TCR cells by Mycobacte- R. N. Davidson, and Z. Toossi. 1999. Influence of polymorphism in the genes for rium tuberculosis via an alternate class I MHC antigen-processing pathway. the interleukin 1 receptor antagonist and interleukin IL-1␤ on tuberculosis. J. Immunol. 162:372. J. Exp. Med. 189:1863. Downloaded from 19. Agrewala, J. N., and R. J. Wilkinson. 1999. Influence of HLA-DR on the phe- 41. Hanson, D. A., A. A. Kaspar, F. R. Poulain, and A. M. Krensky. 1999. Biosyn- ϩ ␣ notype of CD4 T lymphocytes specific for an epitope of the 16-kD -crystallin thesis of granulysin, a novel cytolytic molecule. Mol. Immunol. 36:413. antigen of Mycobacterium tuberculosis. Eur. J. Immunol. 29:1753. 42. Sieg, S., D. Smith, and D. Kaplan. 1999. Differential activity of soluble versus 20. Lalvani, A., R. Brookes, R. J. Wilkinson, A. S. Malin, A. A. Pathan, P. Andersen, cellular Fas ligand: regulation by an accessory molecule. Cell Immunol. 195:89. H. M. Dockrell, G. Pasvol, and A. V. S. Hill. 1998. Human cytolytic and inter- ␥ ϩ 43. Trapani, J. A. 1998. Dual mechanisms of apoptosis induction by cytotoxic lym- feron- secreting CD8 T lymphocytes specific for Mycobacterium tuberculosis. phocytes. Int. Rev. Cytol. 182:111. Proc. Natl. Acad. Sci. USA 95:270. 44. Rukavina, D., G. Laskarin, G. Rubesa, N. Strbo, I. Bedenicki, D. Manestar, 21. Ottenhoff, T. H. M., B. Kaleab, J. D. A. Van Embden, J. E. R. Thole, and M. Glavas, S. E. Christmas, and E. R. Podack. 1998. Age-related decline of R. Kiessling. 1988. The recombinant 65-kD heat shock protein of Mycobacterium http://www.jimmunol.org/ perforin expression in human cytotoxic T lymphocytes and natural killer cells. bovis Bacillus Calmette-Guerin/M. tuberculosis is a target molecule for Blood 92:2410. CD4superϩ cytotoxic T lymphocytes that lyse human monocytes. J. Exp. Med. 45. Rutella, S., C. Rumi, M. B. Lucia, B. Etuk, R. Cauda, and G. Leone. 1998. Flow 168:1947. cytometric detection of perforin in normal human lymphocyte subpopulations 22. Nagata, S. 1999. Fas ligand-induced apoptosis. Annu. Rev. Genet. 33:29. defined by expression of activation/differentiation antigens. Immunol. Lett. 60:51. 23. Griffith, T. S., and D. H. Lynch. 1998. TRAIL: a molecule with multiple receptors 46. Kataoka, T., N. Shinohara, H. Takayama, K. Takaku, S. Kondo, S. Yonehara, and and control mechanisms. Curr. Opin. Immunol. 10:559. K. Nagai. 1996. Concanamycin A, a powerful tool for characterization and es- 24. Lewinsohn, D. M., T. T. Bement, J. Xu, D. H. Lynch, K. H. Grabstein, timation of contribution of perforin- and Fas-based lytic pathways in cell-medi- S. G. Reed, and M. R. Alderson. 1998. Human purified protein derivative-specific ϩ ated cytotoxicity. J. Immunol. 156:3678. CD4 T cells use both CD95-dependent and CD95-independent cytolytic mechanisms. J. Immunol. 160:2374. 47. Ka¨gi, D., B. Ledermann, K. Bu¨rki, R. Zinkernagel, and H. Hengartner. 1996. 25. Smith, S. M., A. S. Malin, T. Pauline, Lukey, S. E. Atkinson, J. Content, Molecular mechanisms of lymphocyte-mediated cytoxicity and their role in im- K. Huygen, and H. M. Dockrell. 1999. Characterization of human Mycobacte- munological protection and pathogenesis in vivo. Annu. Rev. Immunol. 14:207. by guest on October 2, 2021 rium bovis bacille Calmette-Guerin-reactive CD8ϩ T cells. Infect. Immun. 67: 48. Kaplan, D., and D. Smith. 2000. Enzymatic amplification staining for flow cy- 5223. tometric analysis of cell surface molecules. Cytometry 40:81. 26. Cooper, A. M., C. D’Souza, A. A. Frank, and I. M. Orme. 1997. The course of 49. Oddo, M., T. Renno, A. Attinger, T. Bakker, H. R. MacDonald, and Mycobacterium tuberculosis infection in the lungs of mice lacking expression of P. R. Meylan. 1998. Fas ligand-induced apoptosis of infected human macro- either perforin- or granzyme-mediated cytolytic mechanisms. Infect. Immun. 65: phages reduces the viability of intracellular Mycobacterium tuberculosis. J. Im- 1317. munol. 160:5448. 27. Laochumroonvorapong, P., J. Wang, C. C. Liu, W. Ye, A. L. Moreira, 50. Neighbour, P. A., H. S. Huberman, and Y. Kress. 1982. Human large granular K. B. Elkon, V. H. Freedman, and G. Kaplan. 1997. Perforin, a cytotoxic mol- lymphocytes and natural killing ultrastructural studies of strontium-induced de- ecule which mediates cell necrosis, is not required for the early control of my- granulation. Eur. J. Immunol. 12:588. cobacterial infection in mice. Infect. Immun. 65:127. 51. Isaaz, S., K. Baetz, K. Olsen, E. Podack, and G. M. Griffiths. 1995. Serial killing 28. Pithie, A. D., D. A. Lammas, N. Fazal, M. Rahelu, R. Bartlett, J. S. Gaston, and by cytotoxic T lymphocytes: T cell receptor triggers degranulation, re-filling of D. S. Kumararatne. 1995. CD4ϩ cytolytic T cells can destroy autologous and the lytic granules and secretion of lytic proteins via a non-granule pathway. Eur. MHC-matched macrophages but fail to kill intracellular Mycobacterium bovis- J. Immunol. 25:1071. BCG. FEMS Immunol. Med. Microbiol. 11:145. 52. Browne, K. A., E. Blink, V. R. Sutton, C. J. Froelich, D. A. Jans, and 29. Fazal, N., D. A. Lammas, M. Rahelu, A. D. Pithie, J. S. Gaston, and J. A. Trapani. 1999. Cytosolic delivery of granzyme B by bacterial toxins: evi- D. S. Kumararatne. 1995. Lysis of human macrophages by cytolytic CD4ϩ T dence that endosomal disruption, in addition to transmembrane pore formation, is cells fails to affect survival of intracellular Mycobacterium bovis-bacille an important function of perforin. Mol. Cell Biol. 19:8604. Calmette-Guerin (BCG). Clin. Exp. Immunol. 99:82. 53. Molloy, A., P. Laochumroonvorapong, and G. Kaplan. 1994. Apoptosis, but not 30. Cho, S., V. Mehra, S. Thoma-Uszynski, S. Stenger, N. Serbina, R. J. Mazzaccaro, necrosis, of infected monocytes is coupled with killing of intracellular bacillus J. L. Flynn, P. F. Barnes, S. Southwood, E. Celis, et al. 2000. Antimicrobial Calmette-Guerin. J. Exp. Med. 180:1499. activity of MHC class I-restricted CD8ϩ T cells in human tuberculosis. Proc. 54. Lammas, D. A., C. Stober, C. J. Harvey, N. Kendrick, S. Panchalingam, and Natl. Acad. Sci. USA 97:12210. D. S. Kumararatne. 1997. ATP-induced killing of mycobacteria by human mac- 31. Silva, C. L., and D. B. Lowrie. 2000. Identification and characterization of murine rophages is mediated by purinergic P2Z(P2 ϫ 7) receptors. Immunity 7:433. cytotoxic T cells that kill Mycobacterium tuberculosis. Infect. Immun. 68:3269. 55. Balcewicz-Sablinska, M. K., H. Gan, and H. G. Remold. 1999. Interleukin 10 32. Stenger, S., D. A. Hanson, R. Teitelbaum, P. Dewan, K. R. Niazi, C. J. Froelich, produced by macrophages inoculated with Mycobacterium avium attenuates my- T. Ganz, S. Thoma-Uszynski, A. Melian, C. Bogdan, et al. 1998. An antimicro- cobacteria-induced apoptosis by reduction of TNF-␣ activity. J. Infect. Dis. 180: bial activity of cytolytic T cells mediated by granulysin. Science 282:121. 1230. 33. Jongstra, J., T. J. Schall, B. J. Dyer, C. Clayberger, J. Jorgensen, M. M. Davis, 56. Fratazzi, C., R. D. Arbeit, C. Carini, M. K. Balcewicz-Sablinska, J. Keane, and A. M. Krensky. 1987. The isolation and sequence of a novel gene from a H. Kornfeld, and H. G. Remold. 1999. Macrophage apoptosis in mycobacterial human functional T cell line. J. Exp. Med. 165:601. infections. J. Leukocyte Biol. 66:763.