Cooperation of TNF Family Members CD40 , Receptor Activator of NF- κB Ligand, and TNF- α in the Activation of Dendritic Cells and the Expansion of Viral This information is current as Specific CD8+ Memory Responses in of September 24, 2021. HIV-1-Infected and HIV-1-Uninfected Individuals Qigui Yu, Jenny X. Gu, Colin Kovacs, John Freedman, Elaine K. Thomas and Mario A. Ostrowski Downloaded from J Immunol 2003; 170:1797-1805; ; doi: 10.4049/jimmunol.170.4.1797 http://www.jimmunol.org/content/170/4/1797 http://www.jimmunol.org/

References This article cites 36 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/170/4/1797.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 24, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

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 © 2003 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Cooperation of TNF Family Members CD40 Ligand, Receptor Activator of NF-␬B Ligand, and TNF-␣ in the Activation of Dendritic Cells and the Expansion of Viral Specific CD8؉ T Cell Memory Responses in HIV-1-Infected and HIV-1-Uninfected Individuals1

Qigui Yu,* Jenny X. Gu,* Colin Kovacs,* John Freedman,† Elaine K. Thomas,‡ and Mario A. Ostrowski2*†

Members of the TNF superfamily have been shown to be instrumental in enhancing cell-mediated immune responses, primarily Downloaded from through their interactions with dendritic cells (DCs). We systematically evaluated the ability of three TNF superfamily molecules, CD40 ligand (CD40L), receptor activator of NF-␬B ligand (RANKL), and TNF-␣, to expand ex vivo EBV-specific CTL responses in healthy human individuals and ex vivo HIV-1-specific CTL responses in HIV-1-infected individuals. In both groups of indi- viduals, we found that all three TNF family molecules could expand CTL responses, albeit at differing degrees. CD40L treatment alone was better than RANKL or TNF-␣ alone to mature DCs and to expand CTL. In healthy volunteers, TNF-␣ or RANKL could cooperate with CD40L to maximize the ability of DCs to expand virus-specific CTL responses. In HIV-1 infection, cooperative http://www.jimmunol.org/ effects between TNF-␣ or RANKL in combination with CD40L were variable. TNF-␣ and RANKL cooperated with CD40L via differing mechanisms, i.e., TNF-␣ enhanced IL-12 production, whereas RANKL enhanced survival of CD40L-stimulated DCs. These findings demonstrate that optimal maturation of DCs requires multiple signals by TNF superfamily members that include CD40L. In HIV-1 infection, DCs may only require CD40L to maximally expand CTL. Finally, CTL responses were higher in CD4؉ T cell-containing conditions even in the presence of TNF family molecules, suggesting that CD4؉ T cells can provide help .to CD8؉ T cells independently of CD40L, RANKL, or TNF-␣. The Journal of Immunology, 2003, 170: 1797Ð1805

ϩ

vigorous and sustained CD8 T cell cytotoxic (CTL) tion and activation (12, 13), as manifested by the increased surface by guest on September 24, 2021 response in the host is required to control infections expression of MHC class I and II molecules, the costimulatory A caused by persistent viruses such as lymphocytic cho- molecules CD80 (-1) and CD86 (B7-2), and the adhesion mol- riomeningitis virus, CMV, EBV, and possibly HIV-1 and SIV (1– ecules CD54 (ICAM-1) and CD58 (LFA-30). In addition, CD40 ϩ 7). In this regard, CD4 T cell help has been shown to be instru- ligation of mature DCs stimulates the production of proinflamma- mental in maintaining effective CTL memory responses, through at tory IL-1, IL-6, and TNF-␣, and T cell growth and dif- ϩ least two ways. In murine models, activated Ag-specific CD4 T ferentiation factors, IL-12 and IL-15, as well as increasing DC cells have been shown to help CTL by local IL-2 production and survival (12, 14–20). These phenotypic changes confer an en- more importantly through the expression of CD40 ligand hanced capacity of DCs to induce CD8ϩ T cell proliferation and 3 (CD40L) (8–11). CD40L is a type II membrane of the cytotoxicity. Thus, CD4ϩ T cells help CTL responses via their ϩ TNF superfamily predominantly expressed on activated CD4 T effects on DC activation and maturation. cells. CD40 is constitutively expressed on immature dendritic cells Other members of the TNF receptor/ligand superfamily have (DCs), and CD40 ligation of immature DCs induces their matura- also been proposed to play a role in either enhancing or bypassing CD4ϩ T cell help in eliciting potent antiviral CTL responses. The receptor activator of NF-␬B ligand (RANKL)-RANK (also called *Clinical Sciences Division, University of Toronto, Toronto, Canada; †St. Michael’s TRANCE (TNF-related activation-induced )-TRANCE Hospital, University of Toronto, Toronto, Canada; and ‡Immunex Corporation, Se- attle, WA 98101 receptor) pathway has recently been shown in murine models to Received for publication May 17, 2002. Accepted for publication December 6, 2002. have similar functions to that of the CD40L-CD40 pathway in the ϩ The costs of publication of this article were defrayed in part by the payment of page generation of antiviral CD4 T-specific immune responses (21– charges. This article must therefore be hereby marked advertisement in accordance 23). RANKL is a member of the TNF superfamily sharing greatest with 18 U.S.C. Section 1734 solely to indicate this fact. amino acid similarity (28%) with CD40L (CD154). RANKL is 1 This study was funded through grant support supplied by American Foundation for expressed on both activated CD4ϩ T and CD8ϩ T cells (24). AIDS Research and Aventis-Pasteur. Q.Y. and M.A.O. are career scientists of The Ontario HIV Treatment Network. RANK, the receptor for RANKL, is constitutively expressed on DCs (22); however, RANK expression is greater on mature DCs or 2 Address correspondence and reprint requests to Dr. Mario A. Ostrowski, Clinical Sciences Division, Room 6271, Medical Sciences Building, University of Toronto, on CD40-ligated DCs (24). Similar to CD40L, conditioning of Toronto, Ontario, Canada, M5S 1A8. E-mail address: [email protected] murine DCs with RANKL has been shown to elicit production of 3 Abbreviations used in this paper: CD40L, CD40 ligand; B-LCL, B-lymphoblastoid proinflammatory cytokines, IL-1 and IL-6, and T cell growth and cell line; CD40LT, trimeric CD40L; DC, ; iMDDC, immature MDDC; MDDC, monocyte-derived DC; RANK, receptor activator of NF-␬B; RANKL, differentiation factors, IL-12 and IL-15. RANKL also dramatically RANK ligand; RANKLT, trimeric RANKL. inhibits DC via increased Bcl-xL expression (22–25).

Copyright © 2003 by The American Association of Immunologists, Inc. 0022-1767/03/$02.00 1798 COOPERATION BETWEEN CD40L, RANKL, AND TNF-␣

RANKL, however, does not markedly up-regulate DC surface ex- Generation of MDDCs, cell cultures, and stimulation pression of MHC molecules, costimulatory molecules (CD80 and MDDCs were generated by a modification of a method previously de- CD86), CD40, or adhesion molecules (CD54), which is observed scribed (19, 30). Briefly, PBMCs obtained by the Ficoll-Paque gradient with CD40L conditioning. It has been postulated that the primary centrifugation (Amersham Pharmacia Biotech, Uppsala, Sweden) were effect of RANKL on DCs is by enhancing DC survival in tissues, separated on multistep Percoll gradients (Sigma-Aldrich, St. Louis, MO). thus allowing greater opportunity for DCs to prime T cells (26). The recovered monocytes were depleted of contaminating B cells, T cells, NK cells, and granulocytes using Ab-conjugated magnetic beads from the The role of the RANKL-RANK pathway in induction of murine or Monocyte Negative Isolation Kit (Dynal, Oslo, Norway). Purified mono- human CTL responses, however, has not yet been evaluated. cytes were cultured at 1 ϫ 106 cells/ml in medium consisting of RPMI TNF-␣ is produced and secreted early in the inflammatory re- 1640 plus 10% FCS, 2 mM glutamine, 25 mM HEPES, and antibiotics in sponse by activated monocytes, NK cells, and Th1 T cells. TNF-␣ the presence of 50 ng/ml human rGM-CSF and 100 ng/ml human rIL-4 ϩ (PeproTech, Rocky Hill, NJ). GM-CSF and IL-4 were added again on days has been shown to differentiate CD34 progenitor cells into DCs. 3 and 5 with the fresh complete RPMI 1640 medium. After 7 days of In addition, TNF-␣ has been shown to up-regulate adhesion and culture, more than 50% of the cells were CD1ahigh, MHC class IIϩ, costimulatory molecules of immature monocyte-derived DCs CD80low, and CD14Ϫ, which represents an immature DC phenotype. The (iMDDCs), but not to the extent as that seen with CD40L (14). immature DCs were then aliquoted and cultured for 72 h in the following Thus, DC activation and maturational signals can originate from conditions: 1) medium alone; 2) soluble recombinant human trimeric CD40L (CD40LT) at 1 ␮g/ml; 3) soluble recombinant human trimeric multiple sources. The relative importance of each of these signals, RANKL (RANKLT) at a final concentration of 1 ␮g/ml (CD40LT and either alone or in combination, has not been systematically eval- RANKLT were obtained as a gift from Immunex, Seattle, WA); 4) soluble uated, although knowledge of this would be important for vaccine human rTNF-␣ at a final concentration of 10 ng/ml (PeproTech); 5) ϩ ϩ ␣ ϩ ␣ design. Lapointe et al. (27) recently showed that human DCs most CD40LT RANKLT; 6) CD40LT TNF- ; 7) RANKLT TNF- ;or Downloaded from 8) CD40LT ϩ RANKLT ϩ TNF-␣. After 72-h stimulation, cells were optimally stimulated antitumor immune responses if they received harvested and supernatants were saved for cytokine analysis. combined signals from LPS and CD40L, rather than alone. Simi- larly, Morel et al. (28) showed that the TNF superfamily member Induction of peptide-specific CTL LIGHT (a cellular ligand for herpes virus entry mediator and lym- The protocol for expanding circulating memory CTL ex vivo was de- photoxin receptor) could enhance the ability of CD40L to scribed previously (19). MDDCs stimulated for 3 days with CD40LT, ␣ mature DCs. RANKLT, or TNF- , alone or in combination, were pulsed with the spe- http://www.jimmunol.org/ cific HLA class I-restricted peptide at 40 ␮g/ml for1hat37°C, then plated The current study evaluates the ability of three TNF superfamily ϫ 5 ␣ in 24-well plates (5 10 pulsed or nonpulsed MDDCs/well) in RPMI molecules, CD40L, RANKL, and TNF- , either alone or in com- 1640 medium supplemented with 10% FCS, 2 mM glutamine, 25 mM bination, to mature human MDDCs, and to elicit antiviral CTL HEPES, and antibiotics. Freshly isolated or thawed autologous PBMCs responses. We wanted to determine in a systematic fashion, the were prepared in both unfractionated (CD4ϩ T cell-containing) and CD4ϩ T cell-depleted conditions and then added to MDDCs at a 1:10 ratio (5 ϫ optimal signals required to maximally expand virus-specific mem- 6 ϩ ϩ 10 cells/well in 2 ml medium). CD4 T cells were depleted from PBMCs ory CTL responses both in the presence or absence of CD4 T cell using two rounds of magnetic bead depletion (Dynal); the purity of deple- help. We also compared the effects of these TNF family members tion as tested by FACS analysis was always Ͻ0.1% CD4ϩ T cell con- on DC maturation and CTL induction in cells taken from healthy tamination. The percentage of CD8ϩ T cells within total unfractionated ϩ HIV-1-uninfected and HIV-1-infected individuals. PBMCs and CD4 T cell-depleted PBMCs was determined by FACS anal- by guest on September 24, 2021 ysis so that equal input of CD8ϩ T cells could be plated in both unfrac- tionated (total PBMCs) and CD4ϩ T cell-depleted conditions. Thus, the Materials and Methods following conditions were included in all experiments using both unfrac- Study subjects tionated PBMCs or CD4ϩ T cell-depleted PBMCs: 1) MDDCs not pulsed with peptides; 2) MDDCs pulsed with peptides; and 3) MDDCs stimulated Two HIV-1-uninfected individuals, who were HLA-A*0201 positive and ϩ with CD40LT, RANKLT, and TNF-␣, alone or in combination, then had detectable EBV-specific CD8 T cell IFN-␥ responses by ELISPOT pulsed with peptides. On days 3, 5, and 7, the medium was changed. On assay (data not shown), were recruited for leukopheresis to obtain large day 10, duplicate wells were pooled and cells were harvested and tested for amounts of PBMCs. Three HIV-1-seropositive individuals at different CTL activity by intracellular IFN-␥ staining and, in some experiments, also stages of disease were also studied; this included a long-term nonprogres- 51 ϩ ϩ by standard Cr release assay. The percentages of CD8 T cells in both sor (patient 1, who was HIV-1 infected for 10 years, had stable CD4 T ϩ unfractionated and CD4 -depleted conditions were again determined by cell counts Ͼ500 ␮l, and a viral load Ͻ50 copies/ml by bDNA), a chronic ϩ ϩ FACS analysis before CTL assays to assure for equal inputs of CD8 T progressor (patient 2, who was HIV-1 infected for 5 years, CD4 T cell cells for all conditions. Experiments were repeated in HIV-1-negative sub- count 410, viral load 45,000 copies/ml, asymptomatic), and a recent sero- ϩ ject 2 and HIV-1-positive patient 1 with similar results. converter (patient 3, HIV-1 infected for 4 mo, CD4 T cell count 400, viral load 1000 copies/ml). All HIV-1-infected individuals were not taking an- Intracellular staining tiretroviral drugs during the study. Before the study, individuals were class I HLA typed and screened for EBV- or HIV-1-specific CTL by culturing Intracellular staining was performed to enumerate the number of IFN-␥-or PBMC with HLA-restricted EBV or HIV-1 peptides and detecting IFN-␥- IL-12-producing cells, as previously described (31). Briefly, for peptide- producing cells by ELISPOT assay, as previously described (data not specific IFN-␥ staining, 0.25 ϫ 106 cells were cultured in U-bottom 96- shown) (29). Informed consent was obtained from participants in accor- well plates in the presence of peptide-pulsed (1–10 ␮M) autologous B- dance with the guideline for conduction of clinical research at the Univer- lymphoblastoid cell lines (B-LCL) or autologous T cell-depleted PBMCs sity of Toronto and St. Michael’s Hospital. All investigational protocols as stimulator cells; nonpeptide-pulsed stimulator cells were used as back- were approved by the University of Toronto and St. Michael’s Hospital ground controls. Positive control cells were stimulated with the bacterial institutional review boards. superantigen staphylococcal enterotoxin B (1 ␮g/ml) (Sigma-Aldrich). Cells were incubated with peptide-pulsed or nonpeptide-pulsed stimulator Peptide synthesis cells for6hat37°Cin6%CO2. Monensin was added for the duration of the culture period to facilitate intracellular cytokine accumulation. Cell Peptides were synthesized by F-moc chemistry using a Zinnser Analytical surface and intracellular cytokine staining was performed using the Cyto- Synthesizer (Research Genetics, Huntsville, AL), and purity was estab- fix/Cytoperm (BD PharMingen, San Diego, CA) in accordance with the lished by HPLC. Peptides were dissolved in RPMI medium, and concen- manufacturer’s recommendations. The following Abs were used for the trations were determined using the Bio-Rad protein assay kit (Bio-Rad, ϩ intracellular staining: anti-IFN-␥ (clone 4S.B3) and anti-IL-12 (p40/p70, Hercules, CA). The following peptides were used for CD8 T cell expan- sion: 1) the HLA-A *0201-restricted BMLF1 region of EBV (GL clone C11.5). All Abs were obtained from BD PharMingen. CLVAML); and 2) the HIV-1-specific peptides HLA-B 27-restricted P24 Flow cytometry (KRWIIGLNK) for patient 1, HLA-A 24-restricted P17 (KYKLKHIVW) for patient 2, and HLA-A *0201-restricted P17 (SLYNTVATL) for PBMCs or MDDCs were stained in PBS/1% FCS/0.02% NaN3 using flu- patient 3. orochrome-conjugated Abs obtained from BD PharMingen. The Abs used The Journal of Immunology 1799

were anti-CD80, anti-CD83, anti-CD1a, anti-CD14, anti-HLA-DR, anti- Statistical analysis CD4, anti-CD8, anti-CD3, anti-IL-12, and anti-IFN-␥. Events were ac- quired using FACSCalibur flow cytometer (BD Biosciences, San Diego, Data were compared using the Wilcoxon signed test for paired CA). Dead cells were excluded on the basis of forward and side samples. scatter. For intracellular IFN-␥ assay, a total of 50,000–100,000 events were collected for each sample, and CD8ϩ T cells were enumerated after gating on CD3-positive cells. Data were analyzed using CellQuest (BD Results Biosciences) or FlowJo (San Carlos, CA) software. Induction of phenotypical maturation of MDDCs by TNF superfamily members CD40L, RANKL, and TNF-␣, or with Cytotoxicity assay combinations thereof Autologous B-LCL were labeled by incubating in 100 ␮Ci sodium 51Cr iMDDCs generated from healthy volunteers or HIV-1-infected in- chromate and pulsed with the specific peptide at 10 ␮Mfor1hat37°C. dividuals were conditioned for 72 h with soluble trimeric CD40L Control B-LCL were either pulsed with an irrelevant peptide or cultured in (CD40LT, 1 ␮g/ml), trimeric RANKL (RANKLT, 1 ␮g/ml), complete RPMI 1640 alone. Labeled target cells and serial dilutions of ␣ effector cells in triplicate were incubated in complete RPMI 1640 medium TNF- (10 ng/ml), or the following combinations: CD40LT with for 4 h. Supernatants were then collected and analyzed in a microplate RANKLT, CD40LT with TNF-␣, RANKLT with TNF-␣,orall scintillation counter (TopCount; Packard Instrument, Meriden, CT). Back- three. In comparison with medium alone, conditioning with either Ͻ ground chromium release was always 10%. Percentage of lysis was cal- TNF-␣ or RANKLT alone induced small cell clusters; TNF-␣ plus culated from the formula: 100% ϫ (E Ϫ M)/(T Ϫ M), in which E is experimental release, M is the release in the presence of complete RPMI RANKLT induced greater clustering, whereas CD40LT alone or in 1640 medium, and T is the release in the 5% Triton X-100 detergent. combination with either TNF-␣ or RANKLT or both induced the Downloaded from Specific lysis was determined by subtracting the lysis of control targets greatest degree of clustering (data not shown). Similar morpho- from the lysis of peptide-pulsed targets. logic effects were seen in MDDCs derived from both HIV-1-un- infected and HIV-1-infected individuals (data not shown). The cell RNA isolation and RT-PCR surface phenotypes of these differentially matured DCs were com- pared by flow cytometry (Fig. 1). As shown in Fig. 1, the non- For RT-PCR analysis, total RNA was extracted from MDDCs cultured in low treated DCs were HLA-DR and expressed high levels of CD1a http://www.jimmunol.org/ complete RPMI 1640 medium treated for 16 h with CD40LT, RANKLT, TNF-␣, or the combinations thereof, using TRIzol reagent (Life Technol- and low levels of CD80 and CD83. CD40LT was the most potent ogies, Grand Island, NY). In some experiments, to prime MDDCs to en- inducer of DC maturation, as demonstrated by decreased CD1a hance IL-12 production after stimulation, MDDCs were also cultured in the expression and increased CD83 expression (12). Although TNF-␣ ␥ presence of 100 ng/ml IFN- . A total of 500 ng of total RNA extracted and RANKLT were mild inducers of DC maturation, as shown by from conditioned MDDCs was used for cDNA synthesis using SuperScript First-Strand Synthesis System (Invitrogen, Carlsbad, CA) in a total volume CD1a down-regulation, additive effects were not seen when used of 20 ␮l, and 4 ␮l of cDNA was then subjected to RT-PCR with the PCR in combination. Addition of either TNF-␣ or RANKLT to SuperMix High Fidelity kit (Life Technologies). The final RT-PCR vol- CD40LT minimally altered CD1a and CD83 expression. CD40LT ␮ ume was 50 l, and the conditions were: 1 min at 94°C, 1 min at 55°C, induced the greatest expression of HLA-DR and of the costimu- and 2 min at 72°C. PCR was performed in a 4800 DNA thermal cycler ␣ by guest on September 24, 2021 (PerkinElmer/Cetus, Wellesley, MA). The following IL-12 primers latory molecule CD80. TNF- and RANKLT also induced mod- were used, resulting in RT-PCR products of 637 bp for IL-12 p35 and erate levels of CD80 and HLA-DR, which were additive in 818 bp for IL-12 p40 (Coronel, 2001, 680): IL-12 p35, sense, 5Ј-CCCT combination. The addition of TNF-␣ or RANKLT to CD40LT GCAGTGCCGGCTCAGCATGTG-3Ј, and antisense, 5Ј-GCCCGAAT increased the levels of CD80 or HLA-DR when compared with TCTGAAAGCATGAAG-3Ј; IL-12 p40, sense, 5Ј-TCTCTGCAGAG AGTCAGAGGG-3Ј, and antisense, 5Ј-ACGGATCCTGATGGATCAG only CD40LT-treated DCs. The combination of all three mol- GTCATAAGAG-3Ј. GAPDH primers were sense, 5Ј-TCACCATCTTCC ecules, CD40LT, RANKLT, and TNF-␣, induced the most ex- AGGAGGG-3Ј, and antisense, 5Ј-CTGCTTCACCACCTTCTTGA-3Ј. tensive expression of HLA-DR and CD80. Thus, although A total of 6 ␮l of PCR product from each sample was size fractionated CD40LT is the most potent inducer of DC maturation, the ad- using ethidium-stained 1.5% agarose gel electrophoresis. Band intensity dition of either TNF-␣ or RANKLT further enhances the acti- on gels was quantitated by an Image Analyzer (Bio-Rad). vation state of DCs.

MDDC survival analysis Role of conditioned MDDC by CD40L, RANKL, and TNF-␣, Duplicate wells containing 0.5ϳ1 ϫ 105 MDDCs were cultured in flat- or their combinations in the expansion of EBV-specific bottom 96-well plates in complete RPMI 1640 medium for 24 h in the ϩ presence or absence of CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), TNF-␣ memory CTL in vitro in both CD4 -containing and (10 ng/ml), or the combinations thereof. Cell viability was assessed by -depleted conditions trypan blue exclusion. Cell apoptosis was determined by detecting the phosphatidylserine on the outer leaflet of apoptotic cell membranes using To determine the role that the TNF superfamily molecules CD40L, annexin V-fluorescein staining (Boehringer Mannheim, Indianapolis, IN), RANKL, and TNF-␣ play in the expansion of virus-specific mem- according to the manufacturer’s protocol. Apoptotic MDDCs were ana- ory CTL responses, we used an in vitro coculture method in which lyzed by a FACSCalibur using CellQuest, and the data were analyzed using peptide-pulsed conditioned MDDCs stimulate CD8ϩ T cells in the FlowJo software, as previously described (19). absence of exogenous cytokines, as previously described (19). We wanted to determine the optimal signals provided by TNF super- Cytokine measurements family molecules that would induce the most effective Ag-specific

6 CTL response. To determine whether these molecules could com- MDDCs were cultured at a concentration of 10 /ml in 24-well plates and ϩ stimulated for 72 h with CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), and pletely substitute for CD4 T cell help of CTL responses, we also TNF-␣ (10 ng/ml), alone or in combinations. Supernatants were harvested performed experiments in both CD4ϩ T cell-containing and -de- and stored at Ϫ80°C until analysis. IL-10 and IL-15 production was mea- pleted conditions. Two healthy, HIV-1-uninfected individuals who sured using ChemiKine human IL-10 and IL-15 sandwich ELISA kits had previously demonstrated an HLA-A*0201-restricted res- (Chemicon International, Temecula, CA), according to the manufacturer’s specifications. The detection limits of these ELISA tests were 1.6 and 2.7 ponse to the EBV epitope of BMLF1 (lytic cycle Ag) were studied. pg/ml for IL-10 and IL-15, respectively. MDDCs from these two HIV-1-uninfected individuals were condi- 1800 COOPERATION BETWEEN CD40L, RANKL, AND TNF-␣ Downloaded from

FIGURE 1. TNF superfamily members induce the maturation and activation of MDDCs. iMDDCs were treated for 72 h with medium (negative control), CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), TNF-␣ (10 ng/ml), or combinations thereof. Cells were harvested, and expression of surface molecules was analyzed by flow cytometry using PE-conjugated mAbs against CD1a, CD80, CD83, and HLA-DR (BD PharMingen). Values represent the mean fluo- http://www.jimmunol.org/ rescence intensity subtracted from the value of matched isotype control mouse mAbs (dotted histogram). Data are analyzed using FlowJo software. Data from HIV-1-infected patient 1 are shown and are representative of five experiments performed with MDDC obtained from two HIV-1-uninfected blood donors and three HIV-1-infected individuals.

tioned for 3 days by the TNF superfamily members CD40L, Role of conditioned MDDC by CD40L, RANKL, and TNF-␣,or RANKL, and TNF-␣, alone or in combinations, then pulsed with their combinations in the expansion of HIV-1-specific memory EBV-specific peptide, and cocultured with autologous CD8ϩ T CTL in vitro in both CD4ϩ-containing and -depleted PBMCs by guest on September 24, 2021 cells in CD4ϩ T cell-containing or -depleted conditions. After 10 Three HIV-1-seropositive individuals with different rates of dis- days of coculture, CTL effector activity was assessed by two as- ease progression were examined in this study: a long-term non- says: direct cytolysis of peptide-pulsed targets and intracellular progressor (patient 1), a chronic progressor (patient 2), and a re- IFN-␥ production after exposure to peptide-pulsed autologous B- cent seroconverter (patient 3) (see Materials and Methods). LCL or T cell-depleted PBMC. A representative experiment mea- iMDDCs derived from these individuals were treated with TNF suring CTL by intracellular IFN-␥ flow cytometry from subject 1 family molecules for 3 days, then pulsed with an HLA class I-re- and chromium release assay from subject 2 is illustrated in Fig. 2, stricted HIV-1-specific epitope (determined by standard IFN-␥ a and b, respectively. Summary IFN-␥ flow cytometric data from ELISPOT assays obtained from that individual; see Materials and subject 2 are illustrated in Fig. 2c. iMDDC treated with TNF-␣, Methods), and cocultured with autologous PBMC in CD4ϩ T cell- RANKLT, or CD40LT significantly enhanced EBV-specific mem- ϩ ory CTL responses when compared with medium-treated iMDDCs containing or CD4 T cell-depleted conditions. After 10-day cul- ␥ ( p Ͻ 0.05) (Fig. 2). Of the three, DC treated with CD40LT gen- ture, CTL responses were measured by intracellular IFN- pro- erally gave the most potent induction of CTL (see Fig. 2, a and c). duction after exposure to peptide-pulsed autologous B-LCL or T When TNF superfamily molecules were combined, MDDCs con- cell-depleted PBMC. Summary data from all three individuals are ditioned with CD40LT combined with either TNF-␣ or RANKLT illustrated in Fig. 3. induced greater CTL than with CD40LT alone or TNF-␣ plus Similarly to HIV-1-uninfected individuals, iMDDC that were ␣ RANKLT ( p Ͻ 0.05) (Fig. 2). Treatment with all three molecules conditioned with either TNF- , RANKLT, or CD40LT signifi- provided even greater enhancement of CTL in subject 1, but not in cantly expanded HIV-1-specific CTL compared with iMDDCs subject 2. Thus, CD40LT was a stronger inducer of CTL than alone ( p Ͻ 0.05), with CD40LT inducing the greatest responses RANKLT or TNF-␣. However, the addition of either TNF-␣ or (Fig. 3). In contrast to HIV-1-uninfected individuals studied, the RANKLT could further enhance CTL responses induced by addition of RANKLT or TNF-␣ to CD40LT to conditioned CD40LT. CTL responses were always higher in CD4ϩ T cell- iMDDCs did not always enhance CTL responses when compared containing conditions even in those containing CD40LT ( p Ͻ with CD40LT alone ( p ϭ NS). For example, addition of TNF-␣ to 0.05) or a combination of TNF superfamily molecules (Fig. 2). CD40LT treatment of iMDDCs did not enhance CTL compared Thus, although addition of TNF family molecules can dramatically with CD40LT treatment alone in patient 3, an acute seroconverter. enhance CTL responses in CD4ϩ T cell-depleted conditions, the RANKLT minimally enhanced CTL with CD40LT treatment of ϩ addition of CD4ϩ T cells can still help CTL responses further. This iMDDCs in three of six experiments (patient 1, CD4 T cell de- suggests that CD4ϩ T cells may also provide help to CD8ϩ T cells pleted; patients 2 and 3, CD4ϩ T cell containing). Similarly, treat- independently of CD40L, RANKL, or TNF-␣ in our culture ment of iMDDCs with a combination of all three TNF family system. molecules did not always enhance CTL responses above those The Journal of Immunology 1801 Downloaded from http://www.jimmunol.org/

FIGURE 2. Effects of TNF superfamily members CD40L, RANKL, and TNF-␣, alone or in combination, on the induction of EBV-specific antiviral CTL activity. CD4ϩ T cell-containing or -depleted PBMC from an EBV-infected healthy donor were cocultured with HLA-A*0201-restricted peptide-pulsed or nonpulsed autologous MDDCs that were previously conditioned for 72 h with CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), and TNF-␣ (10 ng/ml), alone or in combinations thereof. On day 10, EBV-specific CTL activity was assessed by intracellular flow cytometric analysis of IFN-␥-producing cells, and specific

CTL lysis directed against EBV epitope-pulsed autologous target cells was measured. a, Representative intracellular IFN-␥ flow cytometric data obtained by guest on September 24, 2021 from subject 1 are shown. For intracellular cytokine flow cytometric analysis, cells were gated for CD3 and CD8 to enumerate IFN-␥-producing CD8ϩ T cells only. b, CTL as measured by chromium release assay from subject 2. Corresponding cocultures were tested at E:T ratios ranging from 40:1 to 2.5:1 in a standard 4-h 51Cr release assay on autologous B-LCL in the presence or absence of HLA-A *0201-restricted epitope of the EBV. Results were expressed as percentage of specific lysis determined by subtracting lysis percentage of control from the lysis percentage of peptide-pulsed targets. c, Summary data from intracellular IFN-␥ flow cytometry of subjects 2 are graphically depicted. The experiment from subject 2 was repeated with similar results. Statistical comparisons of data pooled from both subjects were performed between the following culture conditions: DCP vs CD40LT, p Ͻ 0.05; DCP vs TNF-␣, p Ͻ 0.05; DCP vs RANKLT, p Ͻ 0.05; CD40LT vs CD40LT ϩ RANKLT or CD40LT ϩ TNF-␣, p Ͻ 0.05.; CD40LT vs CD40LT ϩ RANKLT ϩ TNF␣, p ϭ NS; CD4ϩ T cell containing vs CD4ϩ T cell depleted, p Ͻ 0.05. DC ϭ nonpeptide-pulsed DC; DCP ϭ DC pulsed with peptides; 40LT ϭ CD40 ligand trimer. with CD40LT treatment alone (Fig. 3). Thus, CD40LT, TNF-␣,or differences in CTL responses from various TNF family combina- RANKLT conditioning of iMDDCs could expand HIV-1-specific tions may be related to differential cytokine production. We thus memory CTL with the greatest expansions observed with compared the ability of TNF superfamily molecules to regulate the CD40LT. In contrast to EBV memory-specific responses in HIV- production of IL-12, IL-15, and IL-10 in MDDCs. For IL-12, 1-uninfected individuals, the addition of TNF-␣ or RANKLT to iMDDCs were stimulated for 16 h with CD40L, RANKL, and CD40LT gave inconsistent effects, with enhancement in some TNF-␣, and combinations thereof, and IL-12 induction was mea- cases and no enhancement in others. sured both by RNA and protein expression. Analysis of IL-12 p35 For most culture conditions, CTL responses were higher in and p40 transcripts by RT-PCR revealed that CD40LT-containing CD4ϩ T cell-containing conditions, including those containing conditions induced the greatest amounts of p35 and p40 in MDDCs, CD40LT ( p Ͻ 0.05). (Fig. 3), suggesting that CD4ϩ T cells can followed by TNF-␣, and then by RANKLT (Fig. 4a). This finding provide help to CD8ϩ T cells independently of CD40L, RANKL, correlated with intracellular IL-12 expression (Fig. 4b), as assessed by or TNF-␣ in our culture system. intracellular flow cytometry. For example, in subject 2, CD40LT stimulation induced 98-fold intracellular IL-12 p40/p70 production, The role of TNF family members CD40L, RANKL, and compared with unstimulated MDDCs (Fig. 4b). RANKLT alone ␣ TNF- in the regulation of IL-10, IL-12, and IL-15 failed to induce stainable IL-12 secretion, whereas TNF-␣ alone production from MDDCs induced an 11-fold increase of IL-12 secretion. TNF-␣ cooperated IL-12 is a central regulator of Th1 responses (32). IL-15 plays an with CD40L to increase IL-12 p40/p70 production, whereas RANKL important role in long-term maintenance of Ag-specific memory failed to enhance the effect of CD40L stimulation on the IL-12 CD8ϩ T cells, and has been shown to be induced by DCs after p40/p70 production. The combination of all three TNF family mem- CD40 ligation (19). IL-10, an immunoregulatory cytokine, is also bers, however, gave the strongest induction of IL-12 p40/p70 produc- induced after DC activation (19, 27). Possible explanations for tion. Thus, CD40LT is the strongest inducer of IL-12 in iMDDCs, 1802 COOPERATION BETWEEN CD40L, RANKL, AND TNF-␣ Downloaded from http://www.jimmunol.org/

FIGURE 3. Effects of TNF superfamily members CD40L, RANKL, and FIGURE 4. Effects of TNF family members CD40L, RANKL, and TNF-␣, alone or in combination, on the induction of HIV-1-specific anti- TNF-␣ on IL-12 expression by MDDCs. MDDCs were stimulated for 16 h viral CTL activity. CD4ϩ T cell-containing or -depleted PBMC from three with CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), and TNF-␣ (10 ng/ml), HIV-1-infected individuals were cocultured with autologous MDDCs that alone or in the combinations. a, Expression of mRNA for IL-12 p35 and were either not pulsed or pulsed with HLA-resticted epitopes of HIV-1 p40 subunits and for human housekeeping GAPDH was determined (see Materials and Methods). MDDCs were previously condi- on reverse-transcribed mRNA from the conditioned MDDCs described, tioned for 72 h with CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), and TNF-␣

followed by PCR amplification using specific primer sets. The resulting by guest on September 24, 2021 (10 ng/ml), alone or in combinations thereof. On day 10, HIV-1-specific cDNA fragments were visualized after electrophoresis on an ethidium- CTL activitiy was assessed by intracellular flow cytometric analysis of stained 1.5% agarose gel. The amplified cDNA fragments with lengths IFN-␥-producing cells. Summary data from intracellular IFN-␥ flow cy- corresponding to transcripts of p35 and p40 and GAPDH are shown. Lane tometry of patients 1–3 are graphically depicted. The experiment from M, A 100-bp DNA size marker (Life Technologies); signal intensities of patient 1 was repeated with similar results. Statistical comparisons on data p35 and p40 bands were quantitated by image analysis and compared with pooled from all patients were performed between the following culture medium control. Ratios of respective p35 and p40 band intensity over conditions: DCP vs CD40LT, p Ͻ 0.05; DCP vs TNF-␣, p Ͻ 0.05; DCP medium control are indicated in the parentheses: Lanes 1Ð8, medium (1, vs RANKLT, p Ͻ 0.05; CD40LT vs CD40LT ϩ RANKLT or TNF-␣, p ϭ 1), CD40LT (98, 12), RANKLT (1, 6), CD40LT ϩ RANKLT (99, 13), NS; CD40LT vs CD40LT ϩ RANKLT ϩ TNF-␣, p ϭ NS; CD4ϩ T cell TNF-␣ (6, 6), TNF-␣ ϩ CD40LT (121, 13), TNF-␣ ϩ RANKLT (8, 6), containing vs CD4ϩ T cell depleted, p Ͻ 0.05. DC ϭ nonpeptide-pulsed TNF-␣ ϩ CD40LT ϩ RANKLT (155, 13). b, Protein expression was de- DC; DCP ϭ DC pulsed with peptides; 40LT ϭ CD40 ligand trimer. termined by IL-12 intracellular staining using anti-IL-12 (p40/p70) Ab (clone C11.5) (BD PharMingen) after overnight incubation with Monensin. The numbers in upper right corner represent percentage of IL-12 p40/p70- which can be further enhanced by TNF-␣, or with TNF-␣ in combi- positive MDDCs per total MDDCs. Data are taken from HIV-1-uninfected nation with RANKLT. Similar findings were observed in iMDDCs subject 2 and are representative of experiments with MDDCs derived from obtained from both HIV-1-uninfected and HIV-1-infected individuals HIV-1-infected and HIV-1-uninfected individuals. Experiments were per- (data not shown). Detection of IL-15 and IL-10 in the supernatants of formed with or without priming MDDCs with 100 ng/ml IFN-␥ to further MDDCs cultured for 72 h with CD40L, RANKL, and TNF-␣, and enhance IL-12 production, with similar trends observed. combinations thereof showed that CD40L-containing conditions in- duced high levels of both IL-10 and IL-15 in MDDCs derived from HIV-1-uninfected and HIV-1-infected individuals (see Table I). IL-10 vented from apoptosis, as determined by annexin V expression and IL-15 production from CD40L-stimulated cells was not signifi- compared with untreated cells (4-fold inhibition), whereas cantly enhanced by addition of either RANKL or TNF-␣ or both CD40LT and TNF-␣ were mild inhibitors of MDDC apoptosis (Table I). (Fig. 5). Addition of RANKLT also inhibited apoptosis in CD40LT- or TNF-␣-stimulated conditions, but not to the level ␣ The role of TNF family members CD40L, RANKL, and TNF- seen with RANKLT alone. These findings also correlated with in MDDC survival viable cell counts, as determined by trypan blue exclusion (data not Inhibition of DC apoptosis with resultant DC survival has been shown). Similar results were obtained when MDDC from HIV-1- postulated to be one mechanism whereby memory CTL responses infected individuals were studied, regardless of their viral load can be maintained (12). The effects of TNF family members (data not shown). Thus, the TNF family members have effects on CD40L, RANKL, and TNF-␣ on MDDC survival were studied DC apoptosis. RANKLT alone is the greatest inhibitor of DC (Fig. 5). MDDCs treated with RANKLT were remarkably pre- apoptosis, and the addition of RANKLT can inhibit apoptosis in The Journal of Immunology 1803

specific CTL. This is also the first demonstration of the ability of RANKL to help virus-specific CD8ϩ T cell responses. RANKL could help virus-specific CTL in both HIV-1-infected and HIV-1- uninfected individuals. RANKL is expressed on activated CD4ϩ and CD8ϩ T cells. This observation indicates that activated CD4ϩ or CD8ϩ T cells may provide signals to DCs to allow greater activation of Ag-specific CD8ϩ T cells. Of TNF-␣, RANKL, and CD40L, CD40L, however, was consistently the most potent stim- ulus to mature and activate DCs and conferred the greatest ability to expand virus-specific CTLs in PBMCs taken from both HIV- 1-uninfected and HIV-1-infected individuals. These findings clearly establish CD40L and thus CD4ϩ T cells as playing a cen- tral role in DC maturation and expansion of memory CTL. We also demonstrated particularly in PBMCs taken from HIV- 1-uninfected individuals, that TNF-␣ or RANKL can cooperate FIGURE 5. Regulation of MDDC survival by TNF family members CD40L, RANKL, and TNF-␣, alone or in combination. MDDCs were with CD40L to induce maximal maturation and expansion of ␣ incubated with CD40LT (1 ␮g/ml), RANKLT (1 ␮g/ml), and TNF-␣ (10 EBV-specific CTL. Given that the effects of TNF- or RANKLT ng/ml), or the combinations thereof. MDDCs were then assessed for the alone were minimal to moderate on DC activation suggests that Downloaded from phosphatidylserine on the outer leaflet of apoptotic cell membranes using they may have more of a cooperative and synergistic role in DC annexin V-fluorescein staining (Boehringer Mannheim). MDDCs were also maturation and CTL induction. These observations also suggest counterstained for HLA-DR. The numbers in upper right corner represent that multiple signals are required to properly induce DC to be percentage of annexin V-positive MDDCs per total viable MDDCs. Data optimally activated. For example, a DC could encounter TNF-␣ are taken from HIV-1-uninfected subject 2 and are representative of ex- from activated monocytes or RANKL from activated T cells in periments with MDDCs derived from HIV-1-infected and HIV-1-unin- combination with CD40L from activated CD4ϩ T cells to opti- fected individuals. Representative data of three independent experiments http://www.jimmunol.org/ are shown. mally prime CTL responses. The mechanisms whereby TNF-␣ or RANKL can synergize with CD40L were investigated in this study. Although CD40L MDDCs treated with CD40LT or TNF-␣ from both HIV-1-unin- induced the greatest degree of costimulatory molecule and HLA fected and HIV-1-infected individuals. expression on MDDCs, both TNF-␣ and RANKL could further increase expression of these molecules, which would allow a more ϩ Discussion intense activation of virus-specific CD8 T cells. TNF-␣ enhanced A number of stimuli have been shown to induce DC maturation the ability of CD40L to induce IL-12 in DCs, which was not ob- ϩ and activation, including LPS, influenza virus, and activated CD4 served with RANKL. Thus, TNF-␣ may cooperate with CD40L by by guest on September 24, 2021 T cells (33). As DCs encounter various stimuli during their mi- also synergizing IL-12 production. RANKL was a potent inhibitor gration from the peripheral tissues to the lymph nodes, it is likely of MDDC apoptosis, even in the presence of other TNF family that multiple signals allow optimal maturation of DCs to provide molecules. This suggests that RANKL may enhance the effects of sustained CTL responses (27). Members of the TNF superfamily CD40L on CTL responses by increasing the survival of CD40L- have been shown to directly activate and mature DCs, suggesting stimulated DCs. Cytokines including IL-15 and IL-10 produced by that there is considerable redundancy in the for MDDCs may also play important roles in expanding or down- DC maturation. In this study, we show that CD40L, RANKL, and regulating CTL responses, respectively. Interestingly, we observed TNF-␣ can all to differing degrees mature DCs and expand virus- that CD40L induced high levels of both IL-10 and IL-15 from

Table I. IL-10 and IL-15 production in the supernatant of cultured MDDCs stimulated with TNF family members CD40L, RANKL, and TNF-␣, alone or in combinationsa

TNF-␣ ϩ TNF-␣ ϩ TNF-␣ ϩ CD40L ϩ CD40L ϩ IL-10 (pg/ml) Medium TNF-␣ RANKL RANKL CD40L CD40L RANKL RANKL

HIV-1-uninfected Subject 1 3.2 9.5 7.8 6.1 144.5 142.5 166.5 157.4 Subject 2 68.9 6.4 44.9 6.6 382.1 392.6 419.4 386.1 HIV-1-infected Patient 1 108.3 54 175.6 20.1 393.1 399.9 421 372.6 Patient 2 189.1 119.1 94 97.4 149.9 268.6 463.6 385.7 Patient 3 0 12.6 5.2 4.4 14.4 19.5 12.6 32.9 IL-15 (pg/ml) HIV-1-uninfected Subject 1 0 0.3 0 1.3 26.8 25.8 27.6 29.4 Subject 2 0 0 1.5 3.4 34.6 25.7 26.1 31.2 HIV-1-infected Patient 1 0 0 0.3 0.3 29.1 26.3 26.3 28.1 Patient 2 0 0 0 0 33.8 26.8 25.3 27.1 Patient 3 0 0 0 0 1.6 0.5 2.2 1.9

a iMDDCs were resuspended at a concentration of 106/ml in 24-well plates and stimulated for 72 h with CD40LT (1 mg/ml), RANKLT (1 mg/ml), and TNF-␣ (10 ng/ml), alone or in combinations. Supernatant from each condition was harvested and stored at Ϫ80°C until use. 1804 COOPERATION BETWEEN CD40L, RANKL, AND TNF-␣

MDDCs. We were, however, unable to demonstrate further IL-15 always higher in CD4ϩ T cell-containing conditions (Fig. 3). We production or a decrease in IL-10 production in conditions in have previously observed this phenomenon in PBMC from long- which TNF-␣ or RANKLT or both were added to CD40LT. In term nonprogressors (19), which suggests that CTL responses in summary, the mechanisms whereby TNF-␣ and RANKL cooper- some long-term nonprogressors may expand well without CD4ϩ T ate with CD40L may differ in part, but have the same overall effect cell help. Further study in this regard is clearly warranted. by enhancing CTL responses. Elucidating the signaling pathways Our findings have important implications for vaccine design and in which CD40 and TNFR ligation induce IL-12 production and immunotherapeutics. For example, DCs conditioned in the pres- RANK ligation inhibits apoptosis will be important for future ence of CD40L plus TNF-␣ or RANKL may be more effective studies. than DC conditioned with CD40L alone when designing autolo- Surprisingly, when studying PBMCs taken from HIV-1-infected gous DC-based vaccines. In addition, the inclusion of CD40L plus individuals, TNF-␣ or RANKL did not always cooperatively ex- RANKL or TNF-␣ in the design of DNA-based vaccines may pand HIV-1-specific CTL responses in the presence of CD40L, allow for greater immunogenicity than a vaccine expressing despite observing enhanced expression of costimulatory molecules CD40L alone. In the case of therapeutic DNA vaccines targeted on DCs. These findings suggest either dysregulation of activated toward HIV-1-infected individuals, CD40L expression alone may DC or an inability of HIV-1-specific CD8ϩ T cells to be primed by be sufficient to augment vaccine-induced immune responses. maximally activated DCs. We have previously noted a defect in HIV-1-specific CD8ϩ T cells in some HIV-1-infected individuals Acknowledgments to maximally respond to matured DCs (19). In the current study, We thank our patients for their time and commitment. we were unable to correlate a lack of synergy between CD40LT Downloaded from and TNF-␣ or RANKLT with IL-10 or IL-15 production. Thus, References aberrant or deficient cytokine production by MDDCs cannot ex- 1. Battegay, M., D. Moskophidis, A. Rahemtulla, H. Hengartner, T. W. Mak, and R. M. Zinkernagel. 1994. Enhanced establishment of a virus carrier state in adult plain these findings. An alternative possibility is that circulating ϩ ϩ CD4 T-cell-deficient mice. J. Virol. 68:4700. HIV-1-specific CD8 T cell memory cells in HIV-1 infection are 2. Rickinson, A. B., and D. J. Moss. 1997. Human cytotoxic T lymphocyte re- already activated due to constant virus production, and may not sponses to Epstein-Barr virus infection. Annu. Rev. Immunol. 15:405. 3. Reusser, P., G. Cathomas, R. Attenhofer, M. Tamm, G. Thiel, M. J. Reddehase, require fully activated MDDCs to be expanded. This hypothesis http://www.jimmunol.org/ and U. H. Koszinowski. 1999. Cytomegalovirus (CMV)-specific T cell immunity will require further study. after renal transplantation mediates protection from CMV disease by limiting the We occasionally observed paradoxical effects if DCs were con- systemic virus load: significance of herpesvirus immediate early ␣ in cellular immunity to cytomegalovirus infection. J. Infect. Dis. 180:247. ditioned with all three TNF family molecules, TNF- , RANKL, 4. Schmitz, J. E., M. J. Kuroda, S. Santra, V. G. Sasseville, M. A. Simon, and CD40L, on CTL responses in PBMCs taken from certain in- M. A. Lifton, P. Racz, K. Tenner-Racz, M. Dalesandro, B. J. Scallon, et al. 1999. ϩ dividuals (subject 2, patient 1, and patient 3), either HIV-1 infected Control of viremia in simian immunodeficiency virus infection by CD8 lym- phocytes. Science 283:857. or uninfected. That is, we did not always see greater expansion of 5. Jin, X., D. E. Bauer, S. E. Tuttleton, S. Lewin, A. Gettie, J. Blanchard, CTL when DCs were maximally stimulated by all three TNF fam- C. E. Irwin, J. T. Safrit, J. Mittler, L. Weinberger, et al. 1999. Dramatic rise in ϩ ily molecules together if compared with CD40L stimulation alone. plasma viremia after CD8 T cell depletion in simian immunodeficiency virus- infected macaques. J. Exp. Med. 189:991. by guest on September 24, 2021 This suggests that TNF family molecules may also induce inhib- 6. Pantaleo, G., J. F. Demarest, T. Schacker, M. Vaccarezza, O. J. Cohen, itory effects on immune responses at certain stages of DC matu- M. Daucher, C. Graziosi, S. S. Schnittman, T. C. Quinn, G. M. Shaw, et al. 1997. ration. Excess costimulation has been shown to induce activation- The qualitative nature of the primary immune response to HIV infection is a ϩ prognosticator of disease progression independent of the initial level of plasma induced nonresponsiveness in CD8 T cells (34). In the latter viremia. Proc. Natl. Acad. Sci. USA 94:254. situation, activation-induced nonresponsiveness was characterized 7. Pantaleo, G., H. Soudeyns, J. F. Demarest, M. Vaccarezza, C. Graziosi, ϩ S. Paolucci, M. Daucher, O. J. Cohen, F. Denis, W. E. Biddison, et al. 1997. by an inability of CD8 T cells to produce IL-2 and subsequently Evidence for rapid disappearance of initially expanded HIV-specific CD8ϩ T cell proliferate. Other potential mechanisms may include increased clones during primary HIV infection. Proc. Natl. Acad. Sci. USA 94:9848. IL-10 production by MDDCs or to the induction of counterregu- 8. Bennett, S. R., F. R. Carbone, F. Karamalis, R. A. Flavell, J. F. Miller, and ϩ ϩ W. R. Heath. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 latory CD4 and/or CD8 T cells (35). We were unable to cor- signalling. Nature 393:478. relate paradoxical decreases in CTL with IL-10 levels in superna- 9. Hermans, I. F., D. S. Ritchie, A. Daish, J. Yang, M. R. Kehry, and F. Ronchese. 1999. Impaired ability of MHC class IIϪ/Ϫ dendritic cells to provide tumor pro- tants. Further studies are warranted to investigate this phenomenon tection is rescued by CD40 ligation. J. Immunol. 163:77. in our coculture system. 10. Ridge, J. P., F. Di Rosa, and P. Matzinger. 1998. A conditioned dendritic cell can ϩ CD4ϩ T cells have been shown to help CTL responses by at be a temporal bridge between a CD4 T-helper and a T-killer cell. Nature 393: 474. least two mechanisms: 1) expression of CD40L (8, 10, 11), and 2) 11. Schoenberger, S. P., R. E. Toes, E. I. van der Voort, R. Offringa, and C. J. Melief. production of soluble , namely IL-2 (36). In this 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L study, we have compared the effects of TNF family-conditioned interactions. Nature 393:480. ϩ ϩ ϩ 12. Caux, C., C. Massacrier, B. Vanbervliet, B. Dubois, C. Van Kooten, I. Durand, DCs on CD8 T cells in both CD4 T cell-containing and CD4 and J. Banchereau. 1994. Activation of human dendritic cells through CD40 T cell-depleted conditions. Interestingly, we generally observed cross-linking. J. Exp. Med. 180:1263. ϩ 13. Mackey, M. F., J. R. Gunn, C. Maliszewsky, H. Kikutani, R. J. Noelle, and greater expansion of virus-specific CTL in CD4 T cell-containing R. J. Barth, Jr. 1998. Dendritic cells require maturation via CD40 to generate conditions, even in the presence of exogenous TNF family mole- protective antitumor immunity. J. Immunol. 161:2094. cules. Because the addition of exogenous trimeric CD40L or 14. Cella, M., D. Scheidegger, K. Palmer-Lehmann, P. Lane, A. Lanzavecchia, and ϩ G. Alber. 1996. Ligation of CD40 on dendritic cells triggers production of high RANKL did not equalize the CTL responses between CD4 T levels of -12 and enhances T cell stimulatory capacity: T-T help via ϩ cell-containing and -depleted conditions, it is likely that CD4 T APC activation. J. Exp. Med. 184:747. cell help was also supplied independently of these TNF family 15. Guo, Y., Y. Wu, S. Shinde, M. S. Sy, A. Aruffo, and Y. Liu. 1996. Identification of a costimulatory molecule rapidly induced by CD40L as CD44H. J. Exp. Med. molecules in our coculture system. In this regard, Lu et al. (36) 184:955. recently demonstrated using a similar in vitro murine model that 16. Kennedy, M. K., K. S. Picha, W. C. Fanslow, K. H. Grabstein, M. R. Alderson, CD4ϩ T cell help comprised at least three components: CD40- K. N. Clifford, W. A. Chin, and K. M. Mohler. 1996. CD40/CD40 ligand inter- actions are required for T cell-dependent production of interleukin-12 by mouse dependent DC sensitization, CD40-independent DC sensitization, . Eur. J. Immunol. 26:370. and direct -dependent CD4ϩ-CD8ϩ T cell communi- 17. Koch, F., U. Stanzl, P. Jennewein, K. Janke, C. Heufler, E. Kampgen, N. Romani, and G. Schuler. 1996. High level IL-12 production by murine dendritic cells: cation. Interestingly, in HIV-1-infected patient 1, who was a long- up-regulation via MHC class II and CD40 molecules and down-regulation by term nonprogressor, we observed that CTL responses were not IL-4 and IL-10. J. Exp. Med. 184:741. The Journal of Immunology 1805

18. Kuniyoshi, J. S., C. J. Kuniyoshi, A. M. Lim, F. Y. Wang, E. R. Bade, R. Lau, 27. Lapointe, R., J. F. Toso, C. Butts, H. A. Young, and P. Hwu. 2000. Human E. K. Thomas, and J. S. Weber. 1999. Dendritic cell secretion of IL-15 is induced dendritic cells require multiple activation signals for the efficient generation of by recombinant huCD40LT and augments the stimulation of -specific cy- tumor antigen-specific T lymphocytes. Eur. J. Immunol. 30:3291. tolytic T cells. Cell. Immunol. 193:48. 28. Morel, Y., A. Truneh, R. W. Sweet, D. Olive, and R. T. Costello. 2001. The TNF 19. Ostrowski, M. A., S. J. Justement, L. Ehler, S. B. Mizell, S. Lui, J. Mican, superfamily members LIGHT and CD154 (CD40 ligand) costimulate induction of B. D. Walker, E. K. Thomas, R. Seder, and A. S. Fauci. 2000. The role of CD4ϩ dendritic cell maturation and elicit specific CTL activity. J. Immunol. 167:2479. T cell help and CD40 ligand in the in vitro expansion of HIV-1-specific memory 29. Ostrowski, M. A., J. X. Gu, C. Kovacs, J. Freedman, M. A. Luscher, and cytotoxic CD8ϩ T cell responses. J. Immunol. 165:6133. K. S. MacDonald. 2001. Quantitative and qualitative assessment of human im- 20. Yang, L. P., C. E. Demeure, D. G. Byun, N. Vezzio, and G. Delespesse. 1995. munodeficiency virus type 1 (HIV-1)-specific CD4ϩ T cell immunity to gag in Maturation of neonatal human CD4 T cells. III. Role of B7 co-stimulation at HIV-1-infected individuals with differential disease progression: reciprocal in- priming. Int. Immunol. 7:1987. terferon-␥ and interleukin-10 responses. J. Infect. Dis. 184:1268. 21. Bachmann, M. F., B. R. Wong, R. Josien, R. M. Steinman, A. Oxenius, and 30. Sallusto, F., and A. Lanzavecchia. 1994. Efficient presentation of soluble antigen Y. Choi. 1999. TRANCE, a family member critical for by cultured human dendritic cells is maintained by granulocyte/ col- CD40 ligand-independent activation. J. Exp. Med. 189:1025. ony-stimulating factor plus interleukin-4 and down-regulated by tumor necrosis 22. Green, E. A., and R. A. Flavell. 1999. TRANCE-RANK, a new signal pathway factor-␣. J. Exp. Med. 179:1109. involved in lymphocyte development and T cell activation. J. Exp. Med. 189: 31. Prussin, C., and D. D. Metcalfe. 1995. Detection of intracytoplasmic cytokine 1017. using flow cytometry and directly conjugated anti-cytokine . J. Immu- 23. Wong, B. R., R. Josien, S. Y. Lee, B. Sauter, H. L. Li, R. M. Steinman, and nol. Methods 188:117. Y. Choi. 1997. TRANCE (tumor necrosis factor (TNF)-related activation-in- 32. McDyer, J. F., T. J. Goletz, E. Thomas, C. H. June, and R. A. Seder. 1998. CD40 duced cytokine), a new TNF family member predominantly expressed in T cells, ligand/CD40 stimulation regulates the production of IFN-␥ from human periph- is a dendritic cell-specific survival factor. J. Exp. Med. 186:2075. eral blood mononuclear cells in an IL-12- and/or CD28-dependent manner. J. Im- 24. Josien, R., B. R. Wong, H. L. Li, R. M. Steinman, and Y. Choi. 1999. TRANCE, munol. 160:1701. a TNF family member, is differentially expressed on T cell subsets and induces 33. Bachmann, M. F., R. M. Zinkernagel, and A. Oxenius. 1998. Immune responses cytokine production in dendritic cells. J. Immunol. 162:2562. in the absence of costimulation: viruses know the trick. J. Immunol. 161:5791. 25. Anderson, D. M., E. Maraskovsky, W. L. Billingsley, W. C. Dougall, 34. Deeths, M. J., R. M. Kedl, and M. F. Mescher. 1999. CD8ϩ T cells become M. E. Tometsko, E. R. Roux, M. C. Teepe, R. F. DuBose, D. Cosman, and nonresponsive (anergic) following activation in the presence of costimulation. Downloaded from L. Galibert. 1997. A homologue of the TNF receptor and its ligand enhance T-cell J. Immunol. 163:102. growth and dendritic-cell function. Nature 390:175. 35. Trinchieri, G. 2001. Regulatory role of T cells producing both ␥ and 26. Josien, B. R., H. L. Li, E. Ingulli, S. Sarma, B. R. Wong, M. Vologodskaia, in persistent infection. J. Exp. Med. 194:F53. R. M. Steinman, and Y. Choi. 2000. TRANCE, a tumor necrosis factor family 36. Lu, Z., L. Yuan, X. Zhou, E. Sotomayor, H. I. Levitsky, and D. M. Pardoll. 2000. member, enhances the longevity and adjuvant properties of dendritic cells in vivo. CD40-independent pathways of T cell help for priming of CD8ϩ cytotoxic T J. Exp. Med. 191:495. lymphocytes. J. Exp. Med. 191:541. http://www.jimmunol.org/ by guest on September 24, 2021