Its Effects on Survival and Proliferation T Cells Independently of +

Home , Granzyme

IL-2 Regulates Perforin and Granzyme Gene Expression in CD8 + T Cells Independently of Its Effects on Survival and Proliferation

This information is current as Michelle L. Janas, Penny Groves, Norbert Kienzle and Anne of October 1, 2021. Kelso J Immunol 2005; 175:8003-8010; ; doi: 10.4049/jimmunol.175.12.8003 http://www.jimmunol.org/content/175/12/8003 Downloaded from

References This article cites 59 articles, 33 of which you can access for free at: http://www.jimmunol.org/content/175/12/8003.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 1, 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 © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

IL-2 Regulates Perforin and Granzyme Gene Expression in CD8؉ T Cells Independently of Its Effects on Survival and Proliferation1

Michelle L. Janas,2 Penny Groves, Norbert Kienzle, and Anne Kelso3

Perforin and the serine protease granzymes are key effectors of CD8؉ T cell granule-mediated cytotoxicity, but the requirements for their expression remain largely undeﬁned. We show in this study that IL-2 increased the expression of perforin and granzyme A, B, and C mRNA; intracellular granzyme B protein levels; and cytolytic function in a dose-dependent manner during primary activation of murine CD8؉ T cells in vitro. Two approaches showed that these responses were not a consequence of the effects of IL-2 on cell survival and proliferation. First, IL-2 enhancement of perforin and granzyme expression was equivalent in CD8؉ T cells from wild-type and bcl-2 transgenic mice, although only the latter cells survived in low concentrations or the absence of added

IL-2. This property of bcl-2 transgenic T cells also allowed the demonstration that induction of granzyme A, B, and C mRNA and Downloaded from granzyme B protein required exogenous IL-2, whereas induction of perforin and IFN-␥ expression did not. Second, analysis of perforin and granzyme mRNA levels in cells separated according to division number using the dye CFSE showed that the effects of IL-2 were unrelated to division number. Together, these findings indicate that IL-2 can directly regulate perforin and granzyme ,gene expression in CD8؉ T cells independently of its effects on cell survival and proliferation. The Journal of Immunology, 2005 175: 8003–8010. http://www.jimmunol.org/ ranule-mediated cytotoxicity is one of the major mech- the impaired ability of mice deficient in both these granzymes to anisms used by CD8ϩ T cells to eliminate harmful or control poxvirus infections (11). Recombinant granzyme C has foreign bodies, such as virus-infected cells, tumors, and been reported to induce an apoptotic pathway that is distinct from G ϩ allografts. After Ag recognition, activated CD8 T cells release the pathways activated by granzymes A and B (12). However, its the contents of their cytotoxic granules into the extracellular space, enzymatic targets have yet to be determined, and its biological where they are taken up by the target cell, and apoptosis is initiated relevance in vivo is unknown. Nevertheless, granzyme C mRNA (1). The cytotoxic granules contain a number of molecules, includ- can be expressed at comparable levels as granzyme B mRNA in ing the pore-forming protein, perforin, and serine proteases, known activated CD8ϩ T cells in vitro (13), and it has been suggested that by guest on October 1, 2021 as granzymes. Perforin was originally thought to cause cell lysis by this and other orphan granzymes may provide a fail-safe mecha- penetrating the target cell membrane (2), but recent work favors nism against immune evasion by pathogens (14). the theory that perforin functions by enabling the granzymes to Although the perforin and granzyme genes are known to be escape from endosomes into the cytosol of the target cell (3, 4). inducible, because a T cell must be activated before the cytolytic Whatever its exact role, perforin is essential, because Ag-specific molecules are expressed at the mRNA or protein levels (13, 15), granule-mediated cytotoxicity is absent in perforin-deficient CD8ϩ the signals responsible for regulating gene expression have yet to T cells and NK cells (5). ϩ be identified. The exceptions are studies examining the role of the Murine CD8 T cells can express at least three granzymes, A, cytokine, IL-2. IL-2 has been shown to up-regulate perforin and B, and C, which initiate distinct apoptotic pathways. Substrates for granzyme B include procaspase-3 (6) and Bid (7), whereas one of granzymes A and B in human PBL (16), and binding sites for the the primary targets of granzyme A is the SET complex (8). Al- IL-2-induced transcription factor, STAT-5, have been located in though mice deficient for either granzyme A or B do not display the perforin promoter region (17, 18). However, it is not known the severity of phenotype observed in perforin-deficient mice (9, whether regulation of perforin and granzyme genes by IL-2 is di- 10), a key role for these cytotoxic molecules is demonstrated by rect or a consequence of its other properties. Although IL-2 can initiate the expression of effector genes, such as IFN-␥ (19), it also controls T cell growth and survival (20, 21). The signaling cas- Cooperative Research Center for Vaccine Technology and Queensland Institute of cades that lead to these effects are not distinct, but consist of a Medical Research, Brisbane, Australia complex network of kinases and transcription factors (22). For Received for publication January 18, 2005. Accepted for publication September example, IL-2 activation of STAT-5a/b leads to an increase in 29, 2005. the expression of the antiapoptotic gene bcl-2 (23) as well as The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance the expression of cyclins, which are essential for cell cycle with 18 U.S.C. Section 1734 solely to indicate this fact. progression (24). 1 This work was supported by the National Health and Medical Research Council of It is possible that the responses of perforin and the granzymes to Australia and the Australian Government’s Cooperative Research Centers Program. IL-2 are due to enhanced cell viability or proliferation and not to 2 Current address: The Babraham Institute, Babraham Research Campus, Cambridge, direct induction of these effector genes. To determine how IL-2 U.K. CB24AT. regulates cytolytic gene expression, this study examined the con- 3 Address correspondence and reprint requests to Dr. Anne Kelso, Queensland Insti- tributions of survival and proliferation to the transcription of per- tute of Medical Research, Post Office Royal Brisbane Hospital, Queensland 4029, ϩ Australia. E-mail address: [email protected] forin and granzymes A, B, and C in naive CD8 T cells.

Table I. Real-time PCR primer and probe sequences

Gene Oligonucleotide Sequence

Murine ␤2M 5Ј Primer 5Ј-TCTTTCTGGTGCTTGTCTCAC-3Ј 3Ј Primer 5Ј-GTTCGGCTTCCCATTCTC-3Ј Probe 5Ј-JOE-CGGCCTGTATGCTATCCAGAAAACC-BHQ-1–3Ј Murine CD3␧ 5Ј Primer 5Ј-AAGTAATGAGCTGGCTGC-3Ј 3Ј Primer 5Ј-CTACACTGGTTCCTGAGATG-3Ј Probe 5Ј-JOE-CAGGACGATGCCGAGAACATT-BHQ-1–3Ј Murine IFN-␥ 5Ј Primer 5Ј-GGATGCATTCATGAGTATTGC-3Ј 3Ј Primer 5Ј-GTGGACCACTCGGATGAG-3Ј Probe 5Ј-FAM-TGAGGTCAACAACCCACAGGTCC-BHQ-1–3Ј Murine perforin 5Ј Primer 5Ј-GAGAAGACCTATCAGGACCA-3Ј 3Ј Primer 5Ј-AGCCTGTGGTAAGCATG-3Ј Probe 5Ј-FAM-AACCTCCACTCCACCTTGACTTCA-BHQ-1–3Ј Murine granzyme A 5Ј Primer 5Ј-TTTCATCCTGTAATTGGACTAA-3Ј 3Ј Primer 5Ј-GCGATCTCCACACTTCTC-3Ј Probe 5Ј-FAM-CAGCCCTCTGCTATGTGATGGTATT-BHQ-1–3Ј Murine granzyme B 5Ј Primer 5Ј-CCTCCTGCTACTGCTGAC-3Ј 3Ј Primer 5Ј-GTCAGCACAAAGTCCTCTC-3Ј Probe 5Ј-FAM-CCCTACATGGCCTTACTTTCGATCA-BHQ-1–3Ј Murine granzyme C 5Ј Primer 5Ј-TTCTCCTGACCCTACTTCTG-3Ј Downloaded from 3Ј Primer 5Ј-TGTTAGCACGAATTTGTCTC-3Ј Probe 5Ј-FAM-ATGTTCTGCGGAGGCTTCCTG-BHQ-1–3Ј

Materials and Methods incubated at 65°C for 2 min. cDNA synthesis mix was added to samples, Mice giving the following final concentrations: buffer (50 mM KCl, 10 mM Tris-HCl (pH 8), and 2 mM MgCl2), 6 mM MgCl2, 2.4 nM oligo(dT) http://www.jimmunol.org/ Specific pathogen-free female C57BL/6 mice were obtained from the An- (Roche), 500 ␮M dNTPs (Promega), 1 mM DTT, 20 U of RNaseOUT imal Resource Center and used at 6–9 wk of age. Bcl-2-36 mice (25) were (Invitrogen Life Technologies), and5Uofavian myeloblastosis virus re- provided by Dr. A. Harris (The Walter and Eliza Hall Institute of Medical verse transcriptase (Promega). Samples were incubated for2hat42°C. ␮ Research, Parkville, Australia) and were maintained on a C57BL/6 back- Samples were then diluted to a final volume of 100 linH2O and stored ground at the Queensland Institute of Medical Research. All animal studies at Ϫ20°C. were approved by the Queensland Institute of Medical Research animal Real-time PCR analysis ethics committee. cDNA was quantified using real-time PCR analysis. PCR mix (10 ␮l) was Cell preparation and culture added to cDNA samples (5 ␮l; equivalent to cDNA from 500 cells) to give ␮

the following final concentrations: 5 mM MgCl2; 200 M dNTPs; 50 nM by guest on October 1, 2021 Cell suspensions were obtained by passing axillary, brachial, inguinal, and Ј Ј lumbar lymph nodes (LN)4 through stainless steel mesh, followed by Fi- 5 primer, 3 primer, and internal probe; and 0.3 U of platinum Taq (In- coll-Paque (Amersham Biosciences) separation. Cells were incubated with vitrogen Life Technologies). Primers and probes were designed using the PE-conjugated anti-CD8␣ Ab (53.6) and FITC-conjugated anti-CD44 Ab VectorNTI program (InfoMax) or Primer3 online software (Whitehead In- (IM7.81; BD Pharmingen), then resuspended in balanced salt solution with stitute for Biomedical Research). PCR products spanned intron-exon junc- 5% heat-inactivated FCS (CSL) and 1 ␮g/ml propidium iodide (Calbio- tions to avoid genomic DNA amplification (see Table I). All probes were labeled with Black Hole Quencher-1 at the 3Ј end, and either FAM for chem). Viable cells were purified by FACS (MoFlo; DakoCytomation) ␥ based on CD8ϩ and CD44low (lowest 30%) expression; cells were Ͼ97% IFN- , perforin, granzyme A, granzyme B, and granzyme C or 2,7-dime- ϩ ␤ ␤ thoxy-4,5-dichloro-6-carboxyfluorescein for 2-microglobulin ( 2M) and CD8 on reanalysis. For some experiments, purified cells were labeled ⑀ Ј ␤ ⑀ with CFSE (Molecular Probes), as described previously (26). CD8ϩ T cells CD3 , at the 5 end. 2M and CD3 were coamplified with genes using were stimulated in an accessory cell-free system (27, 28). Briefly, 24-well FAM-labeled probes. Products were amplified using a Rotor-Gene 3000 plates (Falcon; BD Biosciences) were incubated overnight with protein (Corbett Research) under the following conditions: 95°C for 2 min and G-purified hamster anti-CD3⑀ (145-2C11; 10 ␮g/ml), rat anti-CD8␣ (53.6; 95°C for 5 s, followed by 60°C for 30 s for 40 cycles. Known copy num- 10 ␮g/ml), and rat anti-CD11a (I21/7.7; 5 ␮g/ml) mAb. Plates were then bers of cloned cDNA were used to generate a standard curve for each gene. washed three times in PBS. Purified CD8ϩ T cells (2 ϫ 104) were cultured Absolute cDNA copy numbers for each gene were extrapolated from the in 2 ml of growth medium (modified DMEM supplemented with 50 ␮M respective standard curve and then expressed as a ratio to the number of ␤2M cDNA molecules detected in the same sample (␤2M units ϭ (target 2-ME, 216 mg/l L-glutamine, and 10% heat-inactivated FCS) (27) contain- ␤ ϫ ing various concentrations of human rIL-2 (National Institutes of Health). gene copy number/ 2M copy number) 1000). Triplicate cDNA samples were generated for each culture condition, and each cDNA sample was In some experiments anti-murine IL-2 mAb (protein G purified from S4B6 ␤ supernatant) was added to growth medium at 10 ␮g/ml. Growth medium then assayed in duplicate. The results were reported in 2M units and was changed every 24 h, which included rIL-2 and anti-murine IL-2 mAb represent the mean of all replicates. where relevant. At 3 or 4 days of culture, cells were harvested, and viable Intracellular granzyme B detection cells were purified by Ficoll-Paque separation. Cells previously stained with CFSE were separated by FACS according to division peaks. ModFit Cells were washed with PBS and incubated in 0.2% saponin in PBS on ice (Verity Software House) was used to determine the number of cells in each for 10 min, then with a previously optimized concentration of R-PE-con- division peak. jugated mouse anti-human granzyme B Ab GB12 or control R-PE-conjugated murine IgG1 (Caltag Laboratories) for an additional 50 min on ice. Cytoplasmic RNA extraction and cDNA synthesis Cells were then washed, resuspended in balanced salt solution with 2% heat-inactivated FCS, and analyzed using a FACSCalibur and CellQuest RNA was isolated from FACS-purified or Ficoll-purified samples of 1 ϫ 4 Pro version 5.1.1 software (BD Biosciences) with forward and side scatter 10 cells by mixing with TRIzol reagent (Invitrogen Life Technologies) gates set to include both small lymphocytes and blasts. and freezing on dry ice, followed by chloroform extraction, isopropanol precipitation, and washing with ethanol according to the manufacturer’s 51Cr release assay ␮ instructions. Isolated RNA pellets were dissolved in 20 lofH2O and ϩ 51 Cells of the FcR mastocytoma line P815 were labeled with Na CrO4 (Amersham Biosciences) for 60 min at 37°C and washed twice. Labeled 4 ␤ ␤ ϫ 3 Abbreviations used in this paper: LN, lymph node; 2M, 2-microglobulin; Tg, target cells (2–5 10 ) were incubated for 4–5 h at 37°C with serial transgenic; WT, wild type. dilutions of T cells and 1 ␮g/ml anti-CD3 Ab (redirected assay) in 200 ␮l The Journal of Immunology 8005 of growth medium in round-bottom, 96-well plates. Harvested supernatants were dried onto 96-well solid Lumaplates (Packard), and radioactivity was counted in a TopCount microplate scintillation counter (Packard). Sponta- neous lysis of target cells was typically Ͻ10%, and differences in sample release, performed in triplicate, were within 5%. Total 51Cr release from target cells was obtained by lysis in 1% SDS. The percent specific lysis was calculated by the following formula: 100 ϫ ((sample cpm Ϫ spontaneous release cpm)/(total release cpm Ϫ spontaneous release cpm)). Results Perforin and granzyme A, B, and C genes are responsive to IL- 2 in a dose-dependent manner Previous studies examining the responses of perforin and the granzyme genes to IL-2 have used semiquantitative Northern blot analysis, often with cell lines or heterogeneous leukocyte populations (16, 18, 29). To investigate whether IL-2 can induce perforin and granzyme gene expression in naive lymphocytes, CD8ϩ LN cells of naive (CD44low) phenotype were cultured in an accessory cell- free culture system with immobilized mAb to CD3, CD8, and CD11a and varying concentrations of human rIL-2. This well- defined culture system was used to limit the influence of other Downloaded from cytokines and costimulatory molecules on cytolytic gene expression. Culture medium was replaced daily to reduce the accumula- tion of potentially stimulatory endogenous products, and IL-2 was supplemented daily to overcome IL-2 exhaustion in cultures where the cytokine concentration was limiting. On days 3 and 4, live cells were analyzed for the expression of a panel of genes by real-time http://www.jimmunol.org/ PCR under conditions that allowed reliable quantification of one cDNA copy per cell and detection of one cDNA copy per 10 cells. Identity of the PCR products was confirmed using fluorescent probes, which was particularly important for the granzyme genes because they share a high degree of sequence similarity at the cDNA level (14). Dose-response curves for perforin and granzymes A, B, and C are shown in Fig. 1A. Quantification was achieved by comparison with by guest on October 1, 2021 known copy numbers of cloned standards and normalization against the housekeeping gene, ␤2M. To ensure that ␤2M was a suitable reference gene, results were compared with a second gene, CD3⑀. Fig. 1 shows that CD3⑀ expression levels were constant at both time points and all IL-2 concentrations. IFN-␥ gene expression was used as a FIGURE 1. Perforin and granzyme A, B, and C genes are responsive to IL-2 in a dose-dependent manner. A, CD8ϩCD44low LN cells were cultured positive control in this study, because IL-2 is a known regulator of ␥ ␥ with immobilized mAb to CD3, CD8, and CD11a and varying concentra- IFN- in cytolytic T cells (19, 30). IFN- expression levels increased tions of rIL-2 for 3 (E)or4(f) days before cDNA isolation. Gene ex- Ͼ 100-fold across the IL-2 concentration gradient, and the dose-re- pression was assayed by real-time PCR. The threshold of assay detection sponse curves for both time points were similar. is indicated by the broken line. ND, not detected. Due to low cell viability, Minimal up-regulation of perforin gene expression by IL-2 was no data were obtained for cells grown in the absence of IL-2. B, Cells observed on day 3. However, after 4 days of culture, perforin dis- cultured in 20 (f),2(‚), or 0.2 (F) U/ml IL-2 for 4 days were assayed for played a 10-fold increase in expression levels across the IL-2 gra- cytolytic activity against anti-CD3 mAb-coated, 51Cr-labeled P815 target dient, consistent with previous reports that demonstrated sensitiv- cells. Cytolytic activity in the absence of the bridging Ab was at back- ity of perforin to IL-2 (16, 18). All three granzyme genes ground levels (data not shown). C, The number of viable cells harvested on E f responded to a rise in IL-2 concentration, particularly granzyme B, days 3 ( )and4( ) was determined using trypan blue exclusion. Data represent the average from two to six pooled culture wells. whose expression increased 10,000-fold. Granzyme B was previously shown to reach maximum expression levels by day 3 in this culture system, whereas the induction of granzymes A and C was terpretation of these results is confounded by the poor survival and delayed (13). This result was reproduced in this study, with gran- proliferation of cells cultured in low IL-2 concentrations. Fig. 1C zyme B expression rising 5- to 10-fold from days 3 to 4, whereas shows that the numbers of recovered cells were significantly lower granzyme A and C expression levels rose 50- to 100-fold. To as- at IL-2 concentrations Ͻ2 U/ml, particularly by day 4. These cul- sess the functional effects of varying IL-2 concentrations, CD8ϩ T tures contained a high proportion of dying cells, and the remaining cells were tested for their cytolytic activity using a redirected 51Cr live cells were small in size (data not shown). release assay. Fig. 1B shows that although T cells cultured in 20 U/ml IL-2 demonstrated significant cytolytic activity, cells cul- Regulation of perforin and granzyme genes by IL-2 is tured in lower IL-2 concentrations had poor or no detectable lytic independent of IL-2-induced survival ability. Together, the results demonstrate that the perforin and Effects of IL-2 on cell viability were overcome by using T cells granzyme A, B, and C genes in naive CD8ϩ T cells all respond to transgenic (Tg) for bcl-2, one of the anti-apoptotic molecules up- IL-2 in a dose-dependent manner, and that these responses are regulated by signaling through IL-2R␤ (31, 32). T cells overex- associated with marked effects on cytolytic activity. However, in- pressing Bcl-2 are factor independent in vitro (33). Fig. 2A shows 8006 PERFORIN AND GRANZYME REGULATION BY IL-2

these cells. Therefore, no data were obtained for WT cells cultured in 0.2 U/ml IL-2 with the anti-murine IL-2 mAb. A comparison of the bcl-2 Tg and WT dose-response curves showed an equivalent response for all genes (Fig. 3). Results from the bcl-2 Tg cultures show that IL-2 is not necessary for the expression of either IFN-␥ or perforin, because significant cDNA levels were detectable when cells were stimulated in the absence of IL-2, consistent with previous findings (18, 34). However, granzymes A, B, and C all required IL-2 for gene induction, because cDNA was undetectable in T cells cultured without IL-2. The addition of the anti-murine IL-2 mAb further highlighted the sensi- tivity of the granzyme genes to IL-2. In these cultures, granzyme mRNA from bcl-2 Tg T cells was not detectable in the presence of 0.2 U/ml IL-2, presumably because the Ab lowered levels of endogenous cytokine below the threshold concentration required to induce gene expression. Similar results were obtained when intracellular granzyme B protein expression was measured by flow cytometry (Fig. 4). Both the frequency of granzyme B-containing WT and bcl-2 Tg T cells and the average granzyme B level per positive cell declined with decreasing IL-2 concentration. Because Downloaded from perforin and granzyme expression was no higher in bcl-2 Tg T cells than in WT T cells at limiting IL-2 concentrations, we con- clude that the up-regulation of perforin and granzyme expression by IL-2 is not due its anti-apoptotic effects.

Regulation of perforin and the granzyme genes by IL-2 is http://www.jimmunol.org/ independent of IL-2-induced proliferation It has been reported that there is a direct relationship between cell cycle progression and the initiation of cytokine gene transcription in T cells (35, 36). It is therefore possible that IL-2 regulation of the perforin and granzyme genes is directly linked to IL-2-induced proliferation. The relationship between gene expression and division number was examined by labeling cells with CFSE before FIGURE 2. The bcl-2 Tg T cells have enhanced survival, but not en- by guest on October 1, 2021 hanced proliferation, in limiting IL-2 concentrations. A,WT(E) and bcl-2 culture and then purifying them on the basis of division number Tg (f) T cells were cultured as described in Fig. 1. Cell viability was before RNA isolation (Fig. 5). The proliferation profiles of these measured on days 3 and 4 by propidium iodide uptake. B,WTandbcl-2 Tg cultures were similar to those shown in Fig. 2B, with an increase T cells were stained with CFSE, and the percentage of cells in each division in the IL-2 concentration correlating to an increase in the prolif- peak was determined on days 3 and 4. The broken lines indicate the major erative rate (data not shown). With the exception of the reference division peak of Bcl-2 Tg T cells cultured in the absence of IL-2. gene, CD3⑀, the average expression level of each gene increased with the rise in IL-2 concentration, as shown in Fig. 1. However, these increases in expression levels did not correlate with division progression. Instead, the expression levels of some genes, partic- that although the viability of wild-type (WT) T cells was Ͻ50% by ularly granzyme C, increased with time. One exception is the di- day 4 in cultures with Ͻ0.6 U/ml IL-2, 80–90% of bcl-2 Tg T cells vision-dependent decline in the low levels of granzyme C mRNA remained viable. At higher IL-2 concentrations, the viabilities of seen on day 4 in cultures with 0.2 U/ml IL-2; this result is likely WT and bcl-2 Tg cells were comparable. to have reflected impaired viability in the most rapidly dividing To confirm that the bcl-2 transgene only enhanced survival and cells at this limiting IL-2 concentration. Overall, these results dem- did not affect proliferation, T cells were stained with CFSE, and onstrate that the responses of perforin and the granzyme genes to their division profile was measured after culture with varying IL-2 IL-2 were not due to differences in division rates between cultures. concentrations (Fig. 2B). Only the bcl-2 Tg T cells were cultured in the absence of IL-2 due to the inability of WT cells to survive. In the presence of IL-2, the proliferative rate of bcl-2 Tg T cells Discussion was equivalent to that of WT cells. Increasing IL-2 concentrations We show in this study that IL-2 is a potent regulator of perforin increased the proliferative rate for bcl-2 Tg and WT T cells, al- and the granzyme genes during primary activation of CD8ϩ T though these differences were minimal at IL-2 concentrations cells. Previously, the pleiotropic properties of IL-2 made it difficult Ն2 U/ml. to dissociate the effects of IL-2 on T cell survival and proliferation To determine whether the perforin and granzyme genes in bcl-2 from its effects on the expression of genes for cytolytic mediators Tg T cells displayed the same dose-dependent response to IL-2 as and other products. This issue was addressed in the present study WT cells, CD8ϩCD44low bcl-2 Tg and WT T cells were cultured in two ways, demonstrating that the IL-2 signaling pathway re- as described above. An anti-murine IL-2 mAb (S4B6) was added sponsible for the regulation of these genes is distinct from the to some cultures to inhibit the effects of endogenous IL-2 secre- pathways responsible for cell cycle progression and protection tion. The poor viability of WT cells grown in limiting IL-2 con- from apoptosis. centrations was exacerbated by the anti-murine IL-2 mAb, pre- First, we used T cells from the bcl-2 Tg mouse strain. Despite sumably because it inhibited the small amount of IL-2 secreted by their lack of dependence on IL-2 for survival in vitro, these Tg T The Journal of Immunology 8007

FIGURE 3. The regulation of perforin and the granzyme genes by IL-2 is independent of IL-2-induced sur- Downloaded from vival. WT (E) and bcl-2 Tg (f) T cells were cultured as described in Fig. 1. Anti-murine IL-2 mAb (S4B6) was added to the indicated cultures. On days 3 and 4, live cells were harvested, and cDNA was isolated and analyzed by real-time PCR. The threshold of assay detection is indicated by the dashed line. ND, not detected.

Due to low cell viability, no data were obtained for WT http://www.jimmunol.org/ cells that had been cultured in 0.2 U/ml IL-2 with the anti-murine IL-2 mAb. by guest on October 1, 2021

cells exhibited marked IL-2 dose-dependent increases in the ex- after the first division. Other work in our laboratory has shown that pression of perforin and granzyme A, B, and C mRNA and gran- most of the mRNA species assayed in this study are detected in zyme B protein comparable to those observed in WT cells, indi- both undivided and divided cells by day 2 in CD8ϩ T cells acti- cating that cytolytic gene induction is independent of IL-2- vated in this culture system, indicating that the onset of expression mediated cell survival. The use of bcl-2 Tg T cells also revealed does not require division; by comparison, the IL-4-induced induc- that the addition of IL-2 was essential for the induction of detect- tion of IL-4 expression is highly division dependent as described able levels of granzyme A, B, and C mRNA, whereas TCR-me- by others (35, 36) (our unpublished observations). The division diated signaling alone was sufficient for perforin and IFN-␥ ex- independence of perforin, granzyme, and IFN-␥ expression in the pression. Effects on granzyme B were particularly notable; present study, therefore, mainly reflects the early induction of although undetectable in the absence of IL-2, granzyme B mRNA these genes under the strong activation conditions used in this reached levels 106-fold above the detection threshold at optimal study. The data also show that once these genes are induced, the IL-2 concentrations. Strong enhancement of cytolytic activity was enhancement of their expression by IL-2 does not depend on other seen in assays with the perforin/granzyme B-sensitive target division-linked events. cell, P815. The upstream regulatory sequences of perforin and granzyme B Second, using the dye CFSE to separate cells on the basis of contain a number of sites that could mediate regulation by IL-2 in division number, we found that the enhancement of perforin and CD8ϩ T cells directly rather than as an indirect consequence of granzyme expression by IL-2 was unrelated to cell proliferation survival or proliferation. Phosphorylation of IL-2R␤ activates the 8008 PERFORIN AND GRANZYME REGULATION BY IL-2 Downloaded from http://www.jimmunol.org/

FIGURE 4. IL-2 enhances the expression of granzyme B protein in WT and bcl-2 Tg T cells. CD8ϩCD44low LN T cells from WT (left panels) and bcl-2 Tg (right panels) mice were cultured as described in Fig. 1. After 4 days, cells were harvested, permeabilized, incubated with anti-granzyme B Ab (shaded histograms) or isotype control Ab (open histograms), and analyzed by FACS. The percentage of granzyme B-positive cells is indicated

within each panel. Due to low cell viability, no data were obtained for WT by guest on October 1, 2021 cells cultured in the absence of IL-2.

FIGURE 5. The regulation of perforin and the granzyme genes is inde- ϩ low transcription factors STAT5a/b, for which binding sites have been pendent of IL-2-induced proliferation. CD8 CD44 LN cells were identified upstream of the human and mouse perforin genes (17, stained with CFSE and then cultured under the conditions described in Fig. 1. Cells were purified from CFSE peaks corresponding to division cycles 18); IL-2-activated Stat5 has been shown to induce the expression E f ␥ ␣ 2–5 in day 3 cultures ( ) and division cycles 5–8 in day 4 cultures ( ). of a number of genes, including perforin, IFN- , and IL-2R , di- cDNA was isolated from the purified cells and analyzed by real-time PCR. rectly after IL-2 stimulation (18, 37, 38). In addition, IL-2 activa- The threshold of assay detection is indicated by the dashed line. ND, not tion of NF-␬B leads to binding of this transcription factor to the detected. Due to insufficient cell numbers, no data were obtained for cells upstream enhancer of the perforin promoter in NK cells (39), and that had undergone five divisions from day 3 cultures with 0.2 U/ml IL-2. both the perforin and granzyme B promoters contain sequences that can bind AP-1 (40, 41), which is also phosphorylated upon ute to CTL responses under some conditions in vitro and in vivo signaling through IL-2R␤ (22). Recently, the novel transcription (46–48). However, although the common ␥-chain receptor-shar- factor, eomesodermin, was identified in activated CD8ϩ T cells ing cytokine IL-4 can enhance CTL activity in some circumstances and was shown to drive perforin, granzyme B, and IFN-␥ expres- (47), its effects on perforin and granzyme gene expression in the sion when ectopically expressed in Th2 CD4ϩ cells; dominant system used in this study are distinct from those of IL-2. We found negative eomesodermin impaired granzyme B expression and cy- that activation of naive CD8ϩ T cells in the presence of IL-2 and tolytic function in CD8ϩ T cells (42). It will be important to assess IL-4 leads to generation of poorly cytolytic CD8low effectors in whether IL-2 acts upstream or downstream of this proposed master which levels of perforin, granzyme B, and granzyme C mRNA and regulator of CD8ϩ effector T cell differentiation. perforin and granzyme B protein are markedly lower than in cells Our finding that IL-2 is required for granzyme A, B, and C activated without IL-4 (49) (our unpublished observations). expression under defined conditions in vitro does not rule out the Many studies have demonstrated beneficial effects of IL-2 on possibility that other molecules can serve this function in vitro and CTL responses in vivo, for example, in mouse models of virus or in vivo. Mice deficient in IL-2 retain the ability to raise CTL re- intracellular bacterial infection where CTL play roles in direct sponses against many viruses, tumors, and allografts (43, 44), al- elimination of infected cells and control of pathogen spread be- though impairment has been reported in certain conditions (45). tween cells (50–52). In humans, IL-2 has been used extensively Candidates to compensate for the absence of IL-2 include cyto- with positive results as an immunotherapeutic agent in the treat- kines whose receptors share the common ␥-chain with the IL-2R, ment of malignancy and some infections by both direct adminis- particularly IL-15, whose receptor also shares the IL-2R ␤-chain; tration and expression in cellular vaccines (53–55). However, al- several of these cytokines have been shown to enhance or contrib- though leukocyte numbers and phenotypes are usually measured, The Journal of Immunology 8009 few human trials assess the effects of IL-2 on CTL activity. Even 9. Ebnet, K., M. Hausmann, F. Lehmann-Grube, A. Mullbacher, M. Kopf, when CTL are assayed, it is difficult in both human and animal M. Lamers, and M. M. Simon. 1995. Granzyme A-deficient mice retain potent cell-mediated cytotoxicity. EMBO J. 14: 4230–4239. studies to distinguish direct effects of IL-2 on cytolytic function 10. Heusel, J. W., R. L. Wesselschmidt, S. Shresta, J. H. Russell, and T. J. Ley. 1994. from its effects on lymphocyte expansion. This is particularly per- Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA frag- tinent to studies of IL-2 treatment in HIV infection. Because IL-2 mentation and apoptosis in allogeneic target cells. Cell 76: 977–987. 11. Mu¨llbacher, A., P. Waring, R. Tha Hla, T. Tran, S. Chin, T. Stehle, is generally used to reconstitute the immune system, most reports C. Museteanu, and M. M. Simon. 1999. Granzymes are the essential downstream from clinical trials in HIV infection focus on its effect on CD4ϩ T effector molecules for the control of primary virus infections by cytolytic leuko- cell numbers (56, 57). However, CD8ϩ T cells are activated in this cytes. Proc. Natl. Acad. Sci. USA 96: 13950–13955. ϩ 12. Johnson, H., L. Scorrano, S. J. Korsmeyer, and T. J. Ley. 2003. Cell death in- infection, and Lieberman et al. (58) have reported that CD8 T duced by granzyme C. Blood 101: 3093–3101. cells kill HIV-infected lymphocytes predominantly by granule-me- 13. Kelso, A., E. O. Costelloe, B. J. Johnson, P. Groves, K. Buttigieg, and D. R. Fitzpatrick. 2002. The genes for perforin, granzymes A-C and IFN-␥ are diated cytotoxicity. It has also been reported that although freshly ϩ ϩ differentially expressed in single CD8 T cells during primary activation. Int. isolated CD8 T cells were poorly cytolytic against HIV-infected Immunol. 14: 605–613. targets, overnight exposure to IL-2 restored cytolytic function to a 14. Grossman, W. J., P. A. Revell, Z. H. Lu, H. Johnson, A. J. Bredemeyer, and T. J. Ley. 2003. The orphan granzymes of humans and mice. Curr. Opin. Im- level that could not be explained by clonal expansion of virus- munol. 15: 544–552. specific T cells (59). Our data raise the possibility that this re- 15. Garcia-Sanz, J. A., H. R. MacDonald, D. E. Jenne, J. Tschopp, and M. Nabholz. sponse reflected a direct effect of IL-2 on the expression of cyto- 1990. Cell specificity of granzyme gene expression. J. Immunol. 145: 3111–3118. 16. Liu, C. C., S. Rafii, A. Granelli-Piperno, J. A. Trapani, and J. D. Young. 1989. lytic mediator genes. Perforin and serine esterase gene expression in stimulated human T cells: kinet- We have shown in this study that IL-2 regulates perforin and ics, mitogen requirements, and effects of cyclosporin A. J. Exp. Med. 170: granzyme gene expression directly and independently of its effects 2105–2118. ϩ 17. Yu, C. R., J. R. Ortaldo, R. E. Curiel, H. A. Young, S. K. Anderson, and Downloaded from on CD8 T cell survival and proliferation. Although it is well P. Gosselin. 1999. Role of a STAT binding site in the regulation of the human documented that IL-2 enhances cytolytic function and the expres- perforin promoter. J. Immunol. 162: 2785–2790. sion of some cytolytic mediators, many previous studies could not 18. Zhang, J., I. Scordi, M. J. Smyth, and M. G. Lichtenheld. 1999. Interleukin 2 receptor signaling regulates the perforin gene through signal transducer and ac- determine whether these effects were simply the consequence of T tivator of transcription (Stat)5 activation of two enhancers. J. Exp. Med. 190: cell dependence on IL-2 for survival or expansion. Our data indi- 1297–1308. cate that IL-2 may mediate its immunotherapeutic role not only by 19. Farrar, W., H. Johnson, and J. Farrar. 1981. Regulation of the production of immune interferon and cytotoxic T lymphocytes by interleukin 2. J. Immunol. ϩ ϩ http://www.jimmunol.org/ expanding activated CD8 and CD4 T cell numbers, but also by 126: 1120–1125. directly increasing perforin and granzyme gene expression in cy- 20. Morgan, D. A., F. W. Ruscetti, and R. Gallo. 1976. Selective in vitro growth of tolytic CD8ϩ T cells. The ability of IL-2 to enhance perforin and T lymphocytes from normal human bone marrows. Science 193: 1007–1008. 21. Miyazaki, T., Z. J. Liu, A. Kawahara, Y. Minami, K. Yamada, Y. Tsujimoto, granzyme expression directly may be physiologically important in E. L. Barsoumian, R. M. Permutter, and T. Taniguchi. 1995. Three distinct IL-2 circumstances where other signals counteract the effects of IL-2 on signaling pathways mediated by bcl-2, c-myc, and lck cooperate in hematopoietic lymphocyte expansion or would otherwise down-regulate cytolytic cell proliferation. Cell 81: 223–231. 22. Gaffen, S. L. 2001. Signaling domains of the interleukin 2 receptor. Cytokine 14: gene expression. 63–77. 23. Kelly, J., R. Spolski, K. Imada, J. Bollenbacher, S. Lee, and W. J. Leonard. 2003. A role for Stat5 in CD8ϩ T cell homeostasis. J. Immunol. 170: 210–217.

Acknowledgments 24. Moriggl, R., D. J. Topham, S. Teglund, V. Sexl, C. McKay, D. Wang, by guest on October 1, 2021 We thank Paula Hall and Grace Chojnowski for their skillful assistance A. Hoffmeyer, J. van Deursen, M. Y. Sangster, K. D. Bunting, et al. 1999. Stat5 is required for IL-2-induced cell cycle progression of peripheral T cells. Immunity with FACS, Dr. Alan Harris (Walter and Eliza Hall Institute of Medical 10: 249–259. Research) and Dr. Christian Engwerda (Queensland Institute of Medical 25. Strasser, A., A. W. Harris, D. L. Vaux, E. Webb, M. L. Bath, J. M. Adams, and Research) for generously providing Bcl-2-36 mice, and the National Insti- S. Cory. 1990. Abnormalities of the immune system induced by dysregulated tutes of Health AIDS Research and Reference Reagent Program for the gift bcl-2 expression in transgenic mice. Curr. Top. Microbiol. Immunol. 166: of rIL-2. 175–181. 26. Lyons, A. B., and C. R. Parish. 1994. Determination of lymphocyte division by flow cytometry. J. Immunol. Methods 171: 131–137. Disclosures 27. Maraskovsky, E., A. B. Troutt, and A. Kelso. 1992. Co-engagement of CD3 with LFA-1 or ICAM-1 adhesion molecules enhances the frequency of activation of The authors have no financial conflict of interest. single murine CD4ϩ and CD8ϩ T cells and induces synthesis of IL-3 and IFN-␥ but not IL-4 or IL-6. Int. Immunol. 4: 475–485. 28. Fitzpatrick, D. R., and A. Kelso. 1995. Dissociated expression of granulocyte- References macrophage CSF and IL-3 in short-term T cell clones from normal mice. J. Im- 1. Lieberman, J. 2003. The ABCs of granule-mediated cytotoxicity: new weapons in munol. 155: 5140–5150. the arsenal. Nat. Rev. Immunol. 3: 361–370. 29. Lu, P., J. A. Garcia-Sanz, M. G. Lichtenheld, and E. R. Podack. 1992. Perforin 2. Podack, E. R., J. D. Young, and Z. A. Cohn. 1985. Isolation and biochemical and expression in human peripheral blood mononuclear cells: definition of an IL-2- ϩ functional characterization of perforin 1 from cytolytic T-cell granules. Proc. independent pathway of perforin induction in CD8 T cells. J. Immunol. 148: Natl. Acad. Sci. USA 82: 8629–8633. 3354–3360. 3. Pinkoski, M. J., M. Hobman, J. A. Heibein, K. Tomaselli, F. Li, P. Seth, 30. Arad, G., R. Nussinovich, and R. Kaempfer. 1995. Interleukin-2 induces an early C. J. Froelich, and R. C. Bleackley. 1998. Entry and trafficking of granzyme B in step in the activation of interferon-␥ gene expression. Immunol. Lett. 44: target cells during granzyme B-perforin-mediated apoptosis. Blood 92: 213–216. 1044–1054. 31. Otani, H., M. Erdos, and W. J. Leonard. 1993. Tyrosine kinase(s) regulate apo- 4. Metkar, S. S., B. Wang, M. Aguilar-Santelises, S. M. Raja, L. Uhlin-Hansen, ptosis and bcl-2 expression in a growth factor-dependent cell line. J. Biol. Chem. E. Podack, J. A. Trapani, and C. J. Froelich. 2002. Cytotoxic cell granule-me- 268: 22733–22736. diated apoptosis: perforin delivers granzyme B-serglycin complexes into target 32. Akbar, A. N., N. J. Borthwick, R. G. Wickremasinghe, P. Panayoitidis, D. Pilling, cells without plasma membrane pore formation. Immunity 16: 417–428. M. Bofill, S. Krajewski, J. C. Reed, and M. Salmon. 1996. Interleukin-2 receptor 5. Kagi, D., B. Ledermann, K. Burki, P. Seiler, B. Odermatt, K. J. Olsen, common ␥-chain signaling cytokines regulate activated T cell apoptosis in re- E. R. Podack, R. M. Zinkernagel, and H. Hengartner. 1994. Cytotoxicity medi- sponse to growth factor withdrawal: selective induction of anti-apoptotic (bcl-2, ated by T cells and natural killer cells is greatly impaired in perforin-deficient bcl-xL) but not pro-apoptotic (bax, bcl-xS) gene expression. Eur. J. Immunol. 26: mice. Nature 369: 31–37. 294–299. 6. Darmon, A. J., D. W. Nicholson, and R. C. Bleackley. 1995. Activation of the 33. Strasser, A., A. W. Harris, and S. Cory. 1991. Bcl-2 transgene inhibits T cell apoptotic protease CPP32 by cytotoxic T-cell-derived granzyme B. Nature 377: death and perturbs thymic self-censorship. Cell 67: 889–899. 446–448. 34. Heckford, S. E., E. P. Gelmann, C. L. Agnor, S. Jacobson, S. Zinn, and 7. Heibein, J. A., I. S. Goping, M. Barry, M. J. Pinkoski, G. C. Shore, D. R. Green, L. A. Matis. 1986. Distinct signals are required for proliferation and lymphokine and R. C. Bleackley. 2000. Granzyme B-mediated cytochrome c release is reg- gene expression in murine T cell clones. J. Immunol. 137: 3652–3663. ulated by the Bcl-2 family members Bid and Bax. J. Exp. Med. 192: 1391–1402. 35. Gett, A. V., and P. D. Hodgkin. 1998. Cell division regulates the T cell cytokine 8. Lieberman, J., and Z. Fan. 2003. Nuclear war: the granzyme A-bomb. Curr. repertoire, revealing a mechanism underlying immune class regulation. Proc. Opin. Immunol. 15: 553–559. Natl. Acad. Sci. USA 95: 9488–9493. 8010 PERFORIN AND GRANZYME REGULATION BY IL-2

36. Bird, J. J., D. R. Brown, A. C. Mullen, N. H. Moskowitz, M. A. Mahowald, 48. Ye, W., J. D.-E. Young, and C.-C. Liu. 1996. Interleukin-15 induces the expres- J. R. Sider, T. F. Gajewski, C. R. Wang, and S. L. Reiner. 1998. Helper T cell sion of mRNAs of cytolytic mediators and augments cytotoxic activities in pri- differentiation is controlled by the cell cycle. Immunity 9: 229–237. mary murine lymphocytes. Cell. Immunol. 174: 54–62. 37. Lin, J. X., and W. J. Leonard. 2000. The role of Stat5a and Stat5b in signaling by 49. Kienzle, N., K. Buttigieg, P. Groves, T. Kawula, and A. Kelso. 2002. A clonal IL-2 family cytokines. Oncogene 19: 2566–2576. culture system demonstrates that IL-4 induces a subpopulation of noncytolytic T 38. Bream, J. H., D. L. Hodge, T. Gonsky, R. Spolski, W. J. Leonard, S. Krebs, cells with low CD8, perforin, and granzyme expression. J. Immunol. 168: S. Targan, A. Morinobu, J. J. O’Shea, and H. A. Young. 2004. A distal region in 1672–1681. the interferon-␥ gene is a site of epigenetic remodeling and transcriptional reg- 50. Osorio, Y., and H. Ghiasi. 2003. Comparison of adjuvant efficacy of herpes sim- ulation by interleukin-2. J. Biol. Chem. 279: 41249–41257. plex virus type 1 recombinant viruses expressing TH1 and TH2 cytokine genes. 39. Zhou, J., J. Zhang, M. G. Lichtenheld, and G. G. Meadows. 2002. A role for J. Virol. 77: 5774–5783. NF-␬B activation in perforin expression of NK cells upon IL-2 receptor signal- 51. Haak-Frendscho, M., and C. J. Czuprynski. 1992. Use of recombinant interleu- ing. J. Immunol. 169: 1319–1325. kin-2 to enhance adoptive transfer of resistance to Listeria monocytogenes in- 40. Wargnier, A., S. Legros-Maida, R. Bosselut, J. F. Bourge, C. Lafaurie, fection. Infect. Immun. 60: 1406–1414. C. J. Ghysdael, M. Sasportes, and P. Paul. 1995. Identification of human gran- 52. Fayad, R., H. Zhang, D. Quinn, Y. Huang, and L. Qiao. 2004. Oral administration zyme B promoter regulatory elements interacting with activated T-cell-specific with papillomavirus pseudovirus encoding IL-2 fully restores mucosal and sys- proteins: implication of Ikaros and CBF binding sites in promoter activation. temic immune responses to vaccinations in aged mice. J. Immunol. 173: Proc. Natl. Acad. Sci. USA 92: 6930–6934. 2692–2698. 53. Gaffen, S. L., and K. D. Liu. 2004. Overview of interleukin-2 function, produc- 41. Babichuk, C. K., B. L. Duggan, and R. C. Bleackley. 1996. In vivo regulation of tion and clinical applications. Cytokine 28: 109–123. murine granzyme B gene transcription in activated primary T cells. J. Biol. Chem. 54. Liu, M., B. Acres, J. M. Balloul, N. Bizouarne, S. Paul, P. Slos, and P. Squiban. 271: 16485–16493. 2004. Gene-based vaccines and immunotherapeutics. Proc. Natl. Acad. Sci. USA 42. Pearce, E. L., A. C. Mullen, G. A. Martins, C. M. Krawczyk, A. S. Hutchins, 101: 14567–14571. V. P. Zediak, M. Banica, C. B. DiCioccio, D. A. Gross, C. A. Mao, et al. 2003. 55. Tomioka, H. 2004. Adjunctive immunotherapy of mycobacterial infections. Curr. Control of effector CD8ϩ T cell function by the transcription factor eomesoder- Pharm. Des. 10: 3297–3312. min. Science 302: 1041–1043. 56. Katlama, C., G. Carcelain, C. Duvivier, C. Chouquet, R. Tubiana, M. De Sa, 43. Kundig, T. M., H. Schorle, M. F. Bachmann, H. Hengartner, R. M. Zinkernagel, L. Zagury, V. Calvez, B. Autran, and D. Costagliola. 2002. Interleukin-2 accel- and I. Horak. 1993. Immune responses in interleukin-2-deficient mice. Science erates CD4 cell reconstitution in HIV-infected patients with severe immunosup- Downloaded from 262: 1059–1061. pression despite highly active antiretroviral therapy: the ILSTIM study: ANRS 44. Steiger, J., P. W. Nickerson, W. Steurer, M. Moscovitch-Lopatin, and 082. AIDS 16: 2027–2034. T. B. Strom. 1995. IL-2 knockout recipient mice reject islet cell allografts. J. Im- 57. Farel, C. E., D. G. Chaitt, B. K. Hahn, J. A. Tavel, J. A. Kovacs, M. A. Polis, munol. 155: 489–498. H. Masur, D. A. Follmann, H. C. Lane, and R. T. Davey, Jr. 2004. Induction and 45. Kramer, S., C. Mamalaki, I. Horak, A. Schimpl, D. Kioussis, and T. Hung. 1994. maintenance therapy with intermittent interleukin-2 in HIV-1 infection. Blood Thymic selection and peptide-induced activation of T cell receptor-transgenic 103: 3282–3286. CD8 T cells in interleukin-2-deficient mice. Eur. J. Immunol. 24: 2317–2322. 58. Shankar, P., Z. Xu, and J. Lieberman. 1999. Viral-specific cytotoxic T lympho-

46. Bertagnolli, M., and S. Herrmann. 1990. IL-7 supports the generation of cytotoxic cytes lyse human immunodeficiency virus-infected primary T lymphocytes by the http://www.jimmunol.org/ T lymphocytes from thymocytes. Multiple lymphokines required for proliferation granule exocytosis pathway. Blood 94: 3084–3093. and cytotoxicity. J. Immunol. 145: 1706–1712. 59. Shankar, P., M. Russo, B. Harnisch, M. Patterson, P. Skolnik, and J. Lieberman. 47. Widmer, M. B., and K. H. Grabstein. 1987. Regulation of cytolytic T-lymphocyte 2000. Impaired function of circulating HIV-specific CD8ϩ T cells in chronic generation by B-cell stimulatory factor. Nature 326: 795–798. human immunodeficiency virus infection. Blood 96: 3094–3101. by guest on October 1, 2021