CpG-DNA Activates In Vivo T Cell Epitope Presenting Dendritic Cells to Trigger Protective Antiviral Cytotoxic T Cell Responses This information is current as of September 29, 2021. Ramunas M. Vabulas, Hanspeter Pircher, Grayson B. Lipford, Hans Häcker and Hermann Wagner J Immunol 2000; 164:2372-2378; ; doi: 10.4049/jimmunol.164.5.2372 http://www.jimmunol.org/content/164/5/2372 Downloaded from

References This article cites 51 articles, 22 of which you can access for free at:

http://www.jimmunol.org/content/164/5/2372.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 by guest on September 29, 2021

*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 © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. CpG-DNA Activates In Vivo T Cell Epitope Presenting Dendritic Cells to Trigger Protective Antiviral Cytotoxic T Cell Responses1

Ramunas M. Vabulas,*† Hanspeter Pircher,‡ Grayson B. Lipford,* Hans Ha¨cker,* and Hermann Wagner2*

MHC class I-restricted T cell epitopes lack immunogenicity unless aided by IFA or CFA. In an attempt to circumvent the known inflammatory side effects of IFA and CFA, we analyzed the ability of immunostimulatory CpG-DNA to act as an adjuvant for MHC class I-restricted peptide epitopes. Using the immunodominant CD8 T cell epitopes, SIINFEKL from OVA or KAVYN- FATM (gp33) from lymphocytic choriomeningitis virus glycoprotein, we observed that CpG-DNA conveyed immunogenicity to

these epitopes leading to primary induction of peptide-specific CTL. Furthermore, vaccination with the lymphocytic choriomen- Downloaded from ingitis virus gp33 peptide triggered not only CTL but also protective antiviral defense. We also showed that MHC class I-restricted peptides are constitutively presented by immature dendritic cells (DC) within the draining lymph nodes but failed to induce CTL responses. The use of CpG-DNA as an adjuvant, however, initiated peptide presenting immature DC progression to professional licensed APC. Activated DC induced cytolytic CD8 T cells in wild-type mice and also mice deficient of Th cells or CD40 ligand. CpG-DNA thus incites CTL responses toward MHC class I-restricted T cell epitopes in a Th cell-independent manner. Overall, these results provide new insights into CpG-DNA-mediated adjuvanticity and may influence future vaccination strategies for http://www.jimmunol.org/ infectious and perhaps tumor diseases. The Journal of Immunology, 2000, 164: 2372–2378.

ytotoxic CD8 T lymphocytes (CTL) recognize antigenic of the innate immune system, DC express pattern recognition re- peptides presented by MHC class I molecules, whereas ceptors that on binding of pathogen-derived ligands initiate sig- C CD4 Th cells detect peptides presented by MHC class II naling cascades activating the DC to provide T cell costimulation (1). CTL priming usually requires active involvement of Th cells and cytokine release (11). Prototypic for pattern recognition are (2–4). T cell help is principally mediated indirectly by Th-den- responses to LPS, which triggers a signaling complex when toll- dritic cell (DC)3 interactions via CD40-CD40 ligand (CD40L) (5). like receptor 4 or toll-like receptor 2 interacts with LPS bound to Cellular engagement allows CD40 ligand (CD154) of Ag-reactive CD14 (14, 15). Professional APC function of DC may thus be by guest on September 29, 2021 CD4 Th cells to cross-link CD40 on Ag-presenting DC. Cross- induced indirectly via CD40-CD40L engagement with Th cells or linking of CD40 activates immature DC to transit to mature DC directly via recognition of pathogen-derived molecular pattern that up-regulate surface costimulatory molecules and secrete ef- ligands. fector cytokines such as IL-12 (6, 7). Once matured and activated, CpG-DNA serves as a profound Th1-polarizing adjuvant for Ag-presenting DC directly activate CD8 CTL precursors (8–10). proteinaceous Ags (16–19). In addition to strongly promoting hu- Alternatively, a CD4 Th-independent pathway of DC activation moral responses, CpG-DNA was shown to allow priming of CTL exists that operates via recognition of molecular patterns associ- specific for exogenously given Ags (20, 21). The pronounced ated with microbial pathogens. Pathogen derived molecular pattern adjuvanticity of CpG-DNA may be, at least in part, explained ligands include bacterial cell wall components like LPS (11), dou- by its ability to directly activate immature DC to transit to ble-stranded RNA (12), and bacterial CpG-DNA (13). As member professional APC (22, 23). CpG-DNA triggers DC to up- regulate costimulatory molecules (CD40, CD80, CD86), MHC *Institute of Medical Microbiology, Immunology and Hygiene, Technical University class II, and to produce cytokines such as IL-12 both in vitro of Munich, Munich, Germany; †Institute of Immunology, Vilnius, Lithuania; and (22) and in vivo (23).4 ‡Department of Immunology, Institute of Medical Microbiology and Hygiene, Uni- Here we addressed the question of whether MHC class I-re- versity of Freiburg, Freiburg, Germany stricted CD8 T cell peptides may be rendered immunogenic when Received for publication September 13, 1999. Accepted for publication December 17, 1999. presented in vivo by CpG-DNA-activated DC. By infecting im- munized H-2b mice with lymphocytic choriomeningitis virus 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 (LCMV), we tested whether induced CTL responses are paralleled with 18 U.S.C. Section 1734 solely to indicate this fact. by protective antiviral immunity. To clarify the mechanisms of 1 This work was supported by SFB 456 (Project B4), Bundesministerium fu¨r Bildung, CpG-DNA adjuvanticity, we analyzed the fate of OVA-derived Wissenschaft, Forschung und Technologie (Germany) Grant 0311675 and CpG-Im- peptide SIINFEKL in vivo and report here the strikingly exclusive munopharmaceutical (Hilden, Germany). R.M.V. was supported by a fellowship from the Boehringer Ingelheim Stiflung. ability of DC from draining lymph nodes (LN) to present the in- 2 Address correspondence and reprint requests to Dr. H. Wagner, Institute of Medical jected peptide. Finally, we analyzed cellular requirements for Microbiology, Immunology and Hygiene, Technical University of Munich, Troger- CpG-DNA-mediated CTL induction and describe that CpG-DNA strasse 9, 81675 Munich, Germany. E-mail address: [email protected] 3 Abbreviations used in this paper: DC, dendritic cells; CD40L, CD40 ligand; B3Z, the SIINFEKL/Kb-specific T cell hybridoma; LCMV, lymphocytic choriomeningitis 4 T. Sparwasser, R. M. Vabulas, B. Villmow, G. B. Lipford, and H. Wagner. Bacterial virus; gp33, LCMV peptide KAVYNFATM; LN, lymph nodes; ODN, oligodeoxy- CpG-DNA activates dendritic cells in vivo: T helper cell independent cytotoxic T cell nucleotides. responses to soluble proteins. Submitted for publication.

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00 The Journal of Immunology 2373

␮ allows Th cell-independent CTL responses to MHC class I-re- After4hofincubation at 37°C and 5% CO2, 100 l of culture supernatant stricted T cell epitopes. were removed from each well and ␥-irradiation was measured. Specific lysis was calculated according to the formula: % specific lysis ϭ [cpm (sample) Ϫ cpm (spontaneous release)/cpm (maximal release) Ϫ cpm Materials and Methods (spontaneous release) ] ϫ 100. Spontaneous release ranged between 5 Reagents and 15%.

L-Phosphatidylcholine (type XI-E), L-phosphatidyl-DL-glycerol, choles- Preparation of DC terol, and IFA were purchased from Sigma (Munich, Germany). OVA pep- tide 257–264 (SIINFEKL) and LCMV glycoprotein 33–41 peptide For preparation of DC, draining LN were removed and collected into ice- (KAVYNFATM ϭ gp33) were custom synthesized by Research Genetics cold HBSS (Life Technologies, Karlsruhe, Germany). LN were digested (Huntsville, AL). For technical reasons (to prevent dimer formation), the for1hat37°C using collagenase type Ia purchased from Sigma. Single- original cysteine at the anchor position 41 in the gp33 had been replaced by cell suspensions were prepared, and clumps were removed using a 100-␮m methionine. Phosphorothioate-stabilized oligodeoxynucleotides (ODN) pore size filter (Becton Dickinson, Heidelberg, Germany). Immediately were synthesized by TibMolBiol (Berlin, Germany). CpG-ODN (ϭODN after collagenase treatment LN cells were washed in Ca2ϩ-free HBSS; 1668, containing a CG-motif marked with bold letters: 5Ј-TCC-ATG- from this point onwards the cell suspension was always handled in buffers ACG-TTC-CTG-ATG-CT) and control GpC-ODN (inverted CG ϭ ODN containing 2 mM EDTA. For positive selection of dendritic cells, single- 1720: 5Ј-TCC-ATG- AGC-TTC-CTG-ATG-CT) were taken from Ref. 24. cell suspensions were incubated with magnetic beads coated with the anti- Human rIL-2 was donated by Eurocetus (Amsterdam, The Netherlands). CD11c Ab (Miltenyi, Bergisch Gladbach, Germany). Selection of DC was performed using MSϩ separation columns and MiniMACS according to Mice the protocol supplied by the manufacturer (Miltenyi). The selection ensued ϩ Female C57BL/6 mice were purchased from Harlan Winkelmann (Borchen, Ͼ80% pure CD11c cell population as judged by FACS analysis.

Germany). CD40L knockout mice (C57BL/6J-Tnfsf5Ͻtm1ImxϾ) were Downloaded from bought from The Jackson Laboratory (Bar Harbor, ME) and MHC class II- Presentation assay deficient (B6 A␣0/A␣0) mice (25) were kindly provided by Dr. I. Fo¨rster Presentation of SIINFEKL was assayed by measurement of induced lacZ (Technical University Munich, Munich, Germany). All animals were activity in SIINFEKL/Kb-specific T cell hybridoma (B3Z) transfected with housed under specific pathogen-free conditions and were used at 8–12 wk a LacZ reporter under the transcriptional control of IL-2 gene promoter of age. elements (26) after coincubation with subpopulations of LN cells. Briefly, 12 and 24 h after SIINFEKL injection, the draining LN were harvested, and Cell lines and in vitro culture medium ϩ Ϫ

LN cells were separated into CD11c and CD11c fractions as described http://www.jimmunol.org/ EL-4 (H-2b) thymoma cells were purchased from American Type Culture above. Graded numbers of fractionated cells were incubated with 104 B3Z

Collection (Rockville, MD). B3Z, a somatic T cell hybrid generated by cells/well in 96-well plate at 37°C/5% CO2 overnight. On the next day, fusing the OVA/Kb-specific cytotoxic clone, B3, with a lacZ-inducible individual cultures were washed with 100 ␮l PBS and lysed by addition of ␮ derivative of BW5147 fusion partner (26), was kindly provided by Dr. 100 l Z buffer (100 mM 2-ME, 9 mM MgCl2, 0.125% Nonidet P-40, 0.15 B. L. Kelsall (National Institutes of Health, Bethesda, MD). Cells were mM chlorophenol red ␤-galactoside (Calbiochem, San Diego, CA) in cultured in Clicks/RPMI (Biochrom, Berlin, Germany) supplemented with PBS). After 4 h incubation at 37°C, 50 ␮l stop buffer (300 mM glycine and Ϫ 10% (v/v) heat-inactivated FCS (Biochrom), 5 ϫ 10 5 M 2-ME, and an- 15 mM EDTA in water) were added to each well, and absorption of indi- tibiotics (100 IU/ml penicillin G and 100 IU/ml streptomycin sulfate) at vidual wells was read using a 96-well Emex plate reader (Molecular De- 37°C and 5% CO2. vices, Sunnyvale, CA). The absorption wavelength was 570 nm, with 650 nm as the reference wavelength. Induction index was calculated by divid- Peptide-liposome formation ing induced activity by background. by guest on September 29, 2021 Liposomes were manufactured by a rehydration entrapment method (27). Briefly, 18 mg phosphatidylcholine, 2 mg phosphatidylglycerol, and 5 mg Flow cytometry cholesterol (2:0.2:1) were suspended in 5 ml chloroform in a 250-ml round Cells were washed in PBS containing 2% FCS (PBS/FCS) and first incu- bottom flask. The mixture was rotary evaporated under reduced pressure bated for 10 min at 4°C with anti-Fc␥RII/III mAb to block unspecific until a thin lipid film formed on the flask wall. Residual chloroform was binding of the following Ab reagents. All mAbs used were purchased from removed by vacuum desiccation. The peptide (3 mg) was solubilized in a PharMingen (Hamburg, Germany). FITC- and PE-labeled mAbs (used at minimal volume of water and diluted to 1 ml with PBS. This solution was 5–20 ␮g/ml) included Abs against MHC class II (clone 2G9), CD11c slowly added to the dried lipid and hand shaken until the lipids were re- (clone N418), CD80/B7-1 (clone 1G10), CD86/B7-2 (clone GL1), and ␮ suspended. The mixture was then filter extruded through a 0.2- m pore CD40 (clone 3/23). Isotype controls included purified rat IgG2a and ham- size Anotop 10 Plus syringe mount filter (Whatman, Maidstone, England). ster IgG. After incubation with mAbs for 30 min at 4°C, cells were washed Immunization and 51Cr release assay with PBS/FCS. FACS analysis was performed on a flow cytometer FAC- SCalibur (Becton Dickinson), acquiring 10,000 events. FACS data were For priming of cytotoxic T cells, 0.3 mg of respective peptide entrapped in analyzed using CellQuest FACS software. liposomes or solubilized in PBS were injected into both fore footpads of mice. As an adjuvant CpG-ODN was coinjected (10 nmol/mouse). After Protection of mice from replication of LCMV 4–5 days, draining LN were removed and a single-cell suspension was prepared by pressing the LN through a screen. LN cells were cultured at Mice were immunized once s.c. at the tail base with 0.3 mg gp33 plus 10 1–2 ϫ 106/ml in 24-well plates in medium conditioned with 10 U/ml rIL-2 nmol CpG-ODN 1668, gp33 emulsified in IFA, CpG-ODN 1668 alone, or for additional 4–5 days, and then 51Cr release assay was performed. For in gp33 alone or were mock treated. Two weeks later, mice were infected i.v. vitro restimulation of primed lymphocytes 2 wk after immunization, spleen with 200 PFU of LCMV, and viral titers in the spleens were determined cell suspensions were prepared (5 ϫ 106/ml) and cocultured with SIIN- after 4 days in a virus plaque assay, as described previously (28). LCMV FEKL-labeled (0.5 ␮M) irradiated (15 Gy) spleen cells (2 ϫ 106/ml) with was quantified with an immunological focus assay in 24- or 96-well plates. 10 U/ml rIL-2 for 7 days, and then 51Cr release assay was performed. In The LCMV-WE virus strain used in this study was originally obtained some experiments, CD4- or CD8-positive T cells were depleted from the form Rolf Zinkernagel (University Hospital Zurich, Zurich, Switzerland) expanded lymphocyte population before measuring lytic activity. This was and was grown on L929 fibroblast cells with a low multiplicity of infection. done by labeling lymphocytes with magnetic beads coated with the anti- CD4 or anti-CD8 Abs (Dynal, Hamburg, Germany) allowing negative se- Results ϩ ϩ lection of the CD4 or CD8 T cell subset, respectively. The efficiency of CpG-DNA renders CD8 T cell peptides immunogenic negative selection was Ͼ90% as judged by FACS analysis. The 51Cr re- lease assay was performed as follows. EL-4 target cells (2 ϫ 106) were The use of CpG-DNA as an adjuvant appears to function by virtue ␮ 51 labeled with 200 Ci Na2 CrO4 (Amersham Pharmacia Biotech, Freiburg, of its direct DC-activating qualities (22, 23). We were interested to 51 Germany) for1hat37°C and washed; one-half of Cr-labeled cells was determine whether CpG-ODN-mediated activation of DC in vivo incubated with peptide solution (0.1 ␮M SIINFEKL or 1 ␮M KAVYN- FATM) for an additional 30 min. Peptide-untreated cells served as a spec- would render MHC class I-restricted peptides immunogenic for ificity control. After extensive washing, 100 ␮l (103 cells) of target cells CD8 CTL precursors. C57BL/6 mice were challenged with SIIN- were added to the same volume of replicate serial dilutions of effector CTL. FEKL, either encapsulated in liposomes or as an aqueous solution, 2374 BACTERIAL CpG-DNA RENDERS T CELL EPITOPES IMMUNOGENIC

FIGURE 1. Priming of peptide-specific CD8ϩ CTL. A, C57BL/6 mice were immunized in the fore footpads with 0.3 mg SIINFEKL either entrapped in liposomes (a and b) or free in PBS (c–e). In a and c, 10 nmol CpG-ODN 1668 were admixed as adjuvant. In b and d, SIINFEKL alone was used for immunization. In e, 10 nmol nonimmunostimulatory GpC-ODN 1720 were used. After 4 days, lymphocytes from the draining LN were harvested and Downloaded from cultured in vitro with 10 U/ml rIL-2 for an additional 4 days. Thereafter, lytic activity against EL-4 cells (E, …) or EL-4 cells pulsed with the SIINFEKL (F, ) was scored in a standard 51Cr release assay. Symbols represent CTL derived from individual mice. B. Before testing lytic activity in 51Cr release assay, SIINFEKL plus CpG-ODN-induced CTL were depleted of CD4ϩ or CD8ϩ cells. Targets were EL-4 cells (E, …) or EL-4 cells pulsed with the SIINFEKL (F, ). Symbols represent CTL derived from individual mice. A is representative of at least four experiments performed; B is representative of two experiments. http://www.jimmunol.org/ together with the CpG-ODN 1668 or the control GpC-ODN 1720 draining LN in mice challenged locally with Ag plus 10–1 nmol (inverted CG). Mice immunized with either SIINFEKL or CpG- CpG-ODN 1668 (data not shown). ODN 1668 alone served as controls. Fig. 1A depicts a representa- tive experiment of at least four performed to show that immuni- Immature DC constitutively present SIINFEKL whereas CpG- zation with SIINFEKL, encapsulated in liposomes or even as an DNA activates Ag-presenting DC aqueous solution, generated peptide-specific CTL in the draining In situ, DC are believed to initiate primary CTL responses in a LN provided the inoculum contained CpG-ODN as an adjuvant. two-step process. After sampling and processing Ag as immature LN cells of mice immunized with SIINFEKL alone, SIINFEKL DC, they must subsequently mature into professional APC via ac- by guest on September 29, 2021 plus GpC-ODN 1720 or with CpG-ODN 1668 alone failed to tivation (7). DC activation may be brought about by engagement display CTL responses. Fig. 1B depicts a representative experi- with CD40L-positive Th cells (8–10) or via direct recognition of ment of two performed to show that SIINFEKL-specific cytolytic activity is mediated by CD8ϩ T cells. These data suggested that in draining LN of wild-type mice, the peptide SIINFEKL activates CD8 CTL precursors in the presence of the immunostimulatory CpG-ODN 1668. Next we analyzed the time window required to activate peptide- specific CTL precursors in vivo. Activation of CTL precursors within draining LN was scored over a period from day 1 until day 9 postchallenge with SIINFEKL plus CpG-ODN 1668. As shown in Fig. 2A, within 24 h LN draining the injection site developed peptide-specific CTL which could be detected after in vitro IL-2- driven expansion. Activated CTL precursors were present in drain- ing LN for up to 9 days. Based on comparison of lytic units gen- erated per culture, activation of CTL precursors peaked at days 3–4 (data not shown). We also tested whether activation of CTL precursors is re- stricted to the draining LNs or becomes disseminated to a systemic FIGURE 2. A, Early activation of CTL precursors. SIINFEKL, 0.3 mg level. To probe for the generation of primed T cells in the spleen, dissolved in PBS, plus 10 nmol CpG-ODN 1668 were injected in the fore mice were challenged in the fore footpads with SIINFEKL (Ag) footpads; 24 h after immunization, lymphocytes from the draining LN were plus CpG-ODN 1668 (adjuvants). After 2 wk, their splenocytes harvested and cultured in vitro with 10 U/ml rIL-2 for an additional 4 days. were prepared and restimulated in vitro with the peptide SIIN- Thereafter, lytic activity against EL-4 cells (E, …, ᮀ) or EL-4 cells pulsed F  ■ 51 FEKL. As shown in Fig. 2B, strong SIINFEKL-specific lytic ac- with the SIINFEKL ( , , ) was scored in a standard Cr release assay. Symbols represent CTL derived from individual mice. The experiment has tivity was generated, implying systemic dissemination of activated been repeated twice. B, Systemic dissemination of primed CTL. Mice were CTL. This conclusion is also supported by the induction of pro- immunized with 0.3 mg SIINFEKL plus 10 nmol CpG-ODN 1668 in the tective immunity toward LCMV after immunization with an MHC fore footpads; 2 wk after immunization, splenocytes were isolated and class I peptide from LCMV (see below). restimulated in vitro for 7 days as described in Materials and Methods. Finally, we tested the effective dose range of the adjuvant CpG- Lytic activity against EL-4 cells (E) or EL-4 cells pulsed with the ODN 1668. SIINFEKL-specific CTL could be generated in the SIINFEKL (F) was scored in a standard 51Cr release assay. The Journal of Immunology 2375

FIGURE 3. SIINFEKL is presented in vivo by DC with similar effi- Downloaded from ciency in CpG-ODN-treated and untreated mice. C57BL/6 mice were in- jected in the footpads with 0.15 mg SIINFEKL free in PBS (E, F) per footpad, SIINFEKL plus 5 nmol CpG-ODN 1668 (…, ) per footpad, or CpG-ODN 1668 alone (ᮀ, ■); 12 and 24 h after injection draining LN were harvested and digested with collagenase for 1 h, and DC were isolated using CD11c-MicroBeads. CD11c-positive (E, …, ᮀ) and CD11c-negative http://www.jimmunol.org/ (F, , ■) cell populations were serially diluted and tested for SIINFEKL presentation by incubating them overnight with 104 B3Z cells (the SIINFEKL/Kb complex-specific T cell hybridoma transfected with NF- AT-LacZ reporter). Induced ␤-galactosidase activity was measured in du- plicates with CPRG substrate as described in Materials and Methods. In- duction index was calculated by dividing induced activity through background. Pooled cells from two mice per group were used. Individual points represent means from duplicate wells Ϯ SD. Data are representative of three experiments performed. by guest on September 29, 2021

FIGURE 4. Activation of DC by CpG-DNA in draining LN. C57BL/6 pathogen-derived molecular pattern ligands (11). We analyzed mice were injected in the footpads with 0.15 mg SIINFEKL free in PBS whether immature DC within draining LN take up SIINFEKL and (solid histograms) per footpad or SIINFEKL plus 5 nmol CpG-ODN 1668 b ϩ present the peptide in a H-2K -restricted fashion. CD11c DC per footpad (thick lines); 20 h after injection DC from draining LN were from LN draining a local site of SIINFEKL challenge were posi- isolated using CD11c-MicroBeads, stained for CD40, B7.1 (CD80), B7.2 tively selected and probed for H-2Kb-restricted peptide presenta- (CD86), and MHC class II molecules. The expression of surface markers tion. This was accomplished by using the SIINFEKL/Kb-reactive were analyzed with FACS acquiring 10,000 events. Thin lines indicate T cell hybridoma B3Z, which is transfected with a LacZ reporter isotype controls. Data are representative of two experiments performed. under the transcriptional control of IL-2 gene promoter elements (26). Within 12 h after s.c. challenge with soluble SIINFEKL, positively selected DC from draining LN effectively presented SI- cules and MHC class II (Fig. 4). DC activation is CpG-motif de- INFEKL, in that the T cell hybridoma became activated (Fig. 3). pendent, because ODN lacking a CpG-motif fail to activate DC The DC minus LN cell population, however, was strongly impov- (22, 23). Together, these data suggested that within the draining erished for the ability to activate the hybridoma. By 24 h after LN immature DC constitutively take up and present SIINFEKL. It injection, SIINFEKL presentation by DC from the draining LN is the CpG-DNA inoculation, however, that causes their activation began to decline. After injection of CpG-ODN 1668 alone, DC and transit to professional APC. were unable to activate the hybridoma. On the other hand, chal- lenge with SIINFEKL plus CpG-ODN 1668 generated slightly re- duced SIINFEKL presentation when tested after 12–24 h. Whether CpG-DNA bypasses the requirement for T help in generating this reflects activation of DC known to curtail the capacity of DC CTL responses to T cell peptides to take up extracellular ligands (29) needs to be analyzed. These Activation of Ag-presenting immature DC can be brought about by data implied that SIINFEKL presentation was transient, DC de- CD40Lϩ Th cells or by direct DC recognition of pathogen derived rived, and CpG-ODN independent. ligands. We analyzed whether CTL induction by peptides plus In a second step, the activation phenotype of the CD11cϩ DC in CpG-ODN was independent of T cell help. CTL induction by SI- LN was scored by FACS analysis. Costimulatory molecules INFEKL plus CpG-ODN 1668 was tested in wild-type C57BL/6 (CD80, CD86, CD40) and MHC class II expression by DC from mice and compared with that of haplotype matched CD40LϪ/Ϫ LN draining a SIINFEKL challenge remained immature and were mice or Th cell-deficient MHC class IIϪ/Ϫ mice. Fig. 5 shows indistinguishable from mock challenged DC (data not shown). that the magnitudes of SIINFEKL specific CTL responses in Challenge with a mixture of SIINFEKL plus CpG-ODN, however, CD40LϪ/Ϫ and MHC class IIϪ/Ϫ mice were similar to those in greatly augmented DC surface expression of costimulatory mole- H-2Kb wild-type mice. These data suggests that CpG-DNA 2376 BACTERIAL CpG-DNA RENDERS T CELL EPITOPES IMMUNOGENIC

FIGURE 5. CpG-DNA mediates Th cell-independent activation of pre- cursor CTL. C57BL/6 mice (ᮀ, ■), CD40LϪ/Ϫ and MHC class IIϪ/Ϫ mice (E, F, …, ) were injected in the fore footpads with 0.3 mg SIINFEKL free in PBS plus 10 nmol CpG-ODN 1668. After 4 days, lymphocytes from the

draining LN were harvested and cultured in vitro with 10 U/ml rIL-2 for an Downloaded from additional 4 days. Thereafter, lytic activity against EL-4 cells (E, …, ᮀ)or EL-4 cells pulsed with the SIINFEKL (F, , ■) was scored in a standard FIGURE 7. CpG-ODN allows induction of peptide-specific antiviral 51Cr release assay. The experiments were repeated twice. immunity. Mice were immunized s.c. at the base of the tail with 0.3 mg gp33 plus 10 nmol CpG-ODN 1668 (n ϭ 6), gp33 emulsified in IFA (n ϭ 6), PBS (n ϭ 3), CpG-ODN 1668 alone (n ϭ 3), and gp33 alone (n ϭ 3). allowed induction of Th cell-independent CTL responses to- Two weeks after immunization mice were infected i.v. with 200 PFU ward MHC class I-restricted T cell peptides. LCMV, and 4 days later viral titers in spleens were determined. Pooled http://www.jimmunol.org/ results for two independent experiments are shown. E, viral loads in in- O ⅐⅐⅐⅐⅐ CpG-DNA allows T cell peptide to induce protective antiviral dividual mice. , average virus titer in each group. , virus detection CTL responses limit in the assay. Murine LCMV infections are controlled by LCMV-specific CD8 CTL via perforin-dependent cytolysis of virus-infected cells (30). These data demonstrate that CpG-DNA used as an adjuvant con- A major MHC class I-restricted T cell epitope for H-2b mice is the veyed immunogenicity to the MHC class I-restricted T cell aa 33–41 peptide (gp33) derived from LCMV glycoprotein (31). epitope gp33.

This T cell epitope injected in IFA has been shown to induce To analyze whether activation of gp33-reactive CTL by CpG- by guest on September 29, 2021 protective immunity (32). We used this model system to determine DNA is paralleled by protective immunity against LCMV infec- whether CpG-ODN 1668 renders the MHC class I-restricted pep- tion, C57BL/6 mice were immunized once with gp33 plus CpG- tide gp33 immunogenic for CTL and whether the immunized mice ODN 1668. Control mice were immunized with gp33 or with CpG- are protected against LCMV infection. Draining LN of C57BL/6 ODN 1668 alone. For comparison, a group of mice was challenged mice challenged s.c. with an aqueous solution of gp33 plus CpG- with gp33 emulsified in IFA, an immunization protocol known to ODN 1668 generated gp33-specific CTL (Fig. 6). The draining LN protect mice from LCMV infection (33). Two weeks later, mice of mice immunized with gp33 alone, however, failed to do so. were infected i.v. with 200 PFU of LCMV, and 4 days later viral titers in spleens were analyzed. As shown in Fig. 7, mice immu- nized once with gp33 plus CpG-ODN 1668 were protected from LCMV infection similar in magnitude to mice immunized once with gp33 in IFA. Mice immunized with gp33 alone or CpG-ODN 1668 alone, however, remained unprotected. These data demon- strate that like IFA, CpG-DNA acts as an adjuvant for the T cell peptide gp33 allowing for priming of CTL and induction of pro- tective antiviral immunity.

Discussion Adjuvants, such as CFA or IFA allow the induction of protective immunity against infections using defined T cell epitopes as vac- cines (32, 34). The severe side effects caused primarily by the paraffin oil contained in these adjuvants precluded a detailed anal- ysis of their mode of action (35) and their use in humans. There is accumulating evidence that CpG-DNA acts as a potent vaccine FIGURE 6. Priming of LCMV-specific CTL. C57BL/6 mice were im- adjuvant for promoting Th-1-like immune responses (16–19, 36). munized in the fore footpads with 0.3 mg gp33 peptide without or with 10 In contrast to CFA and IFA, the side effects of CpG-DNA, such as nmol CpG-ODN 1668 as adjuvant. After 5 days, lymphocytes from the draining LN were harvested and cultured in vitro with 10 U/ml rIL-2 for an toxic cytokine syndrome, appear to be transient (37). We therefore additional 4 days. Thereafter, lytic activity against EL-4 cells (E, … ᮀ)or considered CpG-ODN useful as a model to study the mechanisms EL-4 cells pulsed with the gp33 was tested in a standard 51Cr release assay. of adjuvant action. Here we demonstrate that CpG-DNA activates Symbols represent CTL derived from individual mice. The experiment was in situ immature Ag-presenting DC to transit to professional APC done twice. that are able to activate CTL precursors in the absence of T help. The Journal of Immunology 2377

In the case of LCMV peptide gp33, activation of CTL was paral- ing the completion of these studies Oxenius et al. (51) arrived at a leled by the induction of protective antiviral immunity. These data similar conclusion by priming and boosting H-2b mice with CpG- implicate CpG-DNA as a potential powerful adjuvant in vaccina- ODN admixed with gp33. tion protocols toward infectious and perhaps tumor-associated The nature of the Ag is believed to control the induction of CD8 diseases. or CD4 T cell immunity, in that endogenous Ags are presented by MHC class I-restricted T cell epitopes (Ag) admixed with CFA MHC class I molecules whereas exogenous and endocytosed Ags or IFA have been shown to prime peptide-specific CTL responses are presented by MHC class II molecules (1). DC, however, have (32, 34), implying presentation of T cell epitopes by APC. To the capacity to generate MHC class I-restricted T cell epitopes address the question of whether upon s.c. challenge soluble T cell from soluble proteins via TAP-dependent pathways. High rate epitopes become presented by DC within draining LN, DC from macropinocytosis and cell type-specific routes may deliver exog- such LN were scored ex vivo for H-2Kb-restricted peptide presen- enous material to the cytosol and may account, at least in part, for tation. Within 12 h after peptide challenge, positively selected im- the cross-priming capacity of DC (38–41). Here we show that DC mature DC presented the OVA peptide SIINFEKL, activating the from LN draining a site of soluble SIINFEKL-peptide challenge SIINFEKL-specific, H-2Kb-restricted T cell hybridoma, whereas retain their immature phenotype but effectively present SIINFEKL the DC negative population did not (Fig. 3). Thus, it appears that in a H-2Kb-restricted fashion. DC-depleted LN cells, however, did uptake and presentation of peptide are constitutive functions of not present SIINFEKL. The exclusivity of cell type for the ability phenotypically immature DC (38). Others have pointed out a MHC to present SIINFEKL contradicts a simple loading of empty MHC class I presentation pathway for cytosolic delivery of exogenous class I heavy chains. Whether DC indeed are specialized to take up

Ags in DC (39–41). Injection of peptide alone failed to induce and to sample, via macropinocytosis, extracellular peptides to de- Downloaded from productive CTL responses. Whether T cell peptides presented by liver them in the MHC class I presentation pathway needs to be immature DC are ignored by Ag-reactive T cells or induce toler- clarified. ance (34, 42) needs to be analyzed. In contrast, coadministration of The data described here define the adjuvanticity of CpG-ODN CpG-DNA activated Ag-presenting immature DC to transit to by their ability to activate Ag-presenting immature DC to transit to professional APC as defined by up-regulation of costimulatory professional APC. Operationally, similar to CD40 ligation by Th

molecules (CD80, CD86, CD40) and MHC class II (Fig. 4). The cells or by anti-CD40 mAb, activated Ag-presenting DC initiate http://www.jimmunol.org/ evidence for DC maturation is also supported by priming of pep- CTL responses in the absence of T help. Our data endorse CpG- tide-reactive precursor CTL (Fig. 1) and by induction of cytokine DNA as an inexpensive and promising adjuvant for vaccination production.4 The effects of CpG-ODN on immature DC within LN protocols directed toward infectious and perhaps tumor-associated were CpG-specific, because control ODN were inactive. diseases. Recently, a pivotal role of CD40 ligation in DC activation has been identified (6, 7, 43). CD40 ligation by CD40L-positive Th Acknowledgments cells or cross-linking by anti-CD40 mAb caused Ag-presenting We thank Drs. N. Shastri (University of California, Berkeley, CA) and immature DC to transit to professional APC. Once matured DC B. L. Kellsall (National Institutes of Health, Bethesda, MD) for providing by guest on September 29, 2021 could activate CTL precursors in a Th cell-independent fashion the B3Z hybridoma, Dr. S. Bauer (Technical University of Munich, Mu- (8–10). Here we describe that similar to CD40 ligation, CpG-DNA nich, Germany) for valuable suggestions about work with peptides, and also activates in situ Ag-presenting immature DC to transit to pro- M. Mayer and B. Villmow for excellent technical assistance. fessional APC. There are additional parallels between DC CD40 ligation and pathogen pattern recognition receptor engagement. References Pathogen-derived ligands like CpG-DNA and LPS or Th cell-de- 1. Germain, R. N., and D. H. Margulies. 1993. The biochemistry and cell biology pendent CD40L-CD40 interaction initiate in APC signal transduc- of antigen processing and presentation. Annu. Rev. Immunol. 11:403. tion via activation of c-Jun NH -terminal kinase and p38, but not 2. Keene, J. A., and J. Forman. 1982. Helper activity is required for the in vivo 2 generation of cytotoxic T lymphocytes. J. Exp. Med. 155:768. via the extracellular signal-related kinase (44–46). We propose 3. Husmann, L. A., and M. J. Bevan. 1988. Cooperation between helper T cells and that CD40 or the CpG-DNA receptor, despite being different re- cytotoxic T lymphocyte precursors. Ann. NY Acad. Sci. 532:158. 4. Bennett, S. R., F. R. Carbone, F. Karamalis, J. F. Miller, and W. R. Heath. 1997. ceptors, trigger similar intracellular signal pathways. These simi- ϩ Induction of a CD8 cytotoxic T lymphocyte response by cross-priming requires larities therefore mimic each other in functional outcome, i.e., in cognate CD4ϩ T cell help. J. Exp. Med. 186:65. the activation/maturation of immature DC. One would thus predict 5. Foy, T. M., A. Aruffo, J. Bajorath, J. E. Buhlmann, and R. J. Noelle. 1996. Immune regulation by CD40 and its ligand GP39. Annu. Rev. Immunol. 14:591. a number of potential uses for CpG-ODN as adjuvants or immu- 6. Cella, M., F. Sallusto, and A. Lanzavecchia. 1997. Origin, maturation and antigen nomodulating agents (47–50). As shown here, CpG-ODN renders presenting function of dendritic cells. Curr. Opin. Immunol. 9:10. T cell epitopes immunogenic by activating Ag-presenting imma- 7. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392:245. ture DC. Because activation of Ag-presenting DC by CpG-DNA 8. Ridge, J. P., F. Di Rosa, and P. Matzinger. 1998. A conditioned dendritic cell can bypasses the need for Th cells, productive CTL responses can be be a temporal bridge between a CD4ϩ T-helper and a T-killer cell. Nature 393: initiated merely by MHC class I-restricted T cell epitopes. This 474. 9. Bennett, S. R., F. R. Carbone, F. Karamalis, R. A. Flavell, J. F. Miller, and attribute could be helpful in cases where the Th cell compartment W. R. Heath. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 is compromised. Additionally, CpG-ODN represents a single, eas- signalling. Nature 393:478. 10. Schoenberger, S. P., R. E. Toes, E. I. van der Voort, R. Offringa, and C. J. Melief. ily produced, and chemically defined reagent. 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L To determine whether the potential benefits may be of practical interactions. Nature 393:480. significance, we vaccinated H-2b mice with a mixture of MHC 11. Hoffmann, J. A., F. C. Kafatos, C. A. Janeway, and R. A. Ezekowitz. 1999. Phylogenetic perspectives in innate immunity. Science 284:1313. class I-restricted LCMV-derived gp33 peptide plus CpG-ODN. 12. Cella, M., M. Salio, Y. Sakakibara, H. Langen, I. Julkunen, and A. Lanzavecchia. LCMV infections are controlled by LCMV-specific CTL (30). 1999. Maturation, activation, and protection of dendritic cells induced by double- This model system allowed us to analyze whether CpG-ODN not stranded RNA. J. Exp. Med. 189:821. 13. Wagner, H. 1999. Bacterial CpG DNA activates immune cells to signal infectious only aids peptide-specific CTL induction but also confers antiviral danger. Adv. Immunol. 73:329. protection. Indeed, mice challenged once with gp33 peptide ad- 14. Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. V. Huffel, X. Du, D. Birdwell, E. Alejos, M. Silva, C. Galanos, M. Freudenberg, P. Ricciardi-Castagnoli, mixed with CpG-ODN 1668 generated primary CTL responses B. Layton, and B. Beutler. 1998. Defective LPS signaling in C3H/HeJ and and were also protected from a subsequent LCMV infection. Dur- C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282:2085. 2378 BACTERIAL CpG-DNA RENDERS T CELL EPITOPES IMMUNOGENIC

15. Kirschning, C. J., H. Wesche, A. T. Merrill, and M. Rothe. 1998. Human toll-like 33. Schulz, M., R. M. Zinkernagel, and H. Hengartner. 1991. Peptide-induced anti- receptor 2 confers responsiveness to bacterial lipopolysaccharide. J. Exp. Med. viral protection by cytotoxic T cells. Proc. Natl. Acad. Sci. USA 88:991. 188:2091. 34. Aichele, P., K. Brduscha-Riem, S. Oehen, B. Odermatt, R. M. Zinkernagel, 16. Roman, M., E. Martin-Orozco, J. S. Goodman, M. D. Nguyen, Y. Sato, H. Hengartner, and H. Pircher. 1997. Peptide antigen treatment of naive and A. Ronaghy, R. S. Kornbluth, D. D. Richman, D. A. Carson, and E. Raz. 1997. virus-immune mice: antigen-specific tolerance versus immunopathology. Immu- Immunostimulatory DNA sequences function as T helper-1-promoting adjuvants. nity 6:519. Nat. Med. 3:849. 35. Warren, H. S., F. R. Vogel, and L. A. Chedid. 1986. Current status of immuno- 17. Chu, R. S., O. S. Targoni, A. M. Krieg, P. V. Lehmann, and C. V. Harding. 1997. logical adjuvants. Annu. Rev. Immunol. 4:369. CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th1) immunity. J. Exp. Med. 186:1623. 36. Zimmermann, S., O. Egeter, S. Hausmann, G. B. Lipford, M. Rocken, 18. Lipford, G. B., T. Sparwasser, M. Bauer, S. Zimmermann, E. S. Koch, K. Heeg, H. Wagner, and K. Heeg. 1998. CpG oligodeoxynucleotides trigger protective and H. Wagner. 1997. Immunostimulatory DNA: sequence-dependent production and curative Th1 responses in lethal murine leishmaniasis. J. Immunol. 160:3627. of potentially harmful or useful cytokines. Eur. J. Immunol. 27:3420. 37. Sparwasser, T., T. Miethke, G. Lipford, A. Erdmann, H. Hacker, K. Heeg, and 19. Sun, S., H. Kishimoto, and J. Sprent. 1998. DNA as an adjuvant: capacity of H. Wagner. 1997. Macrophages sense pathogens via DNA motifs: induction of insect DNA and synthetic oligodeoxynucleotides to augment T cell responses to tumor necrosis factor-␣ -mediated shock. Eur. J. Immunol. 27:1671. specific antigen. J. Exp. Med. 187:1145. 38. Yewdell, J. W., C. C. Norbury, and J. R. Bennink. 1999. Mechanisms of exog- 20. Lipford, G. B., M. Bauer, C. Blank, R. Reiter, H. Wagner, and K. Heeg. 1997. enous antigen presentation by MHC class I molecules in vitro and in vivo: im- CpG-containing synthetic oligonucleotides promote B and cytotoxic T cell re- plications for generating CD8ϩ T cell responses to infectious agents, tumors, sponses to protein antigen: a new class of vaccine adjuvants. Eur. J. Immunol. transplants, and vaccines. Adv. Immunol. 73:1. 27:2340. 39. Rock, K. L., L. Rothstein, S. Gamble, and C. Fleischacker. 1993. Characteriza- 21. Davis, H. L., R. Weeratna, T. J. Waldschmidt, L. Tygrett, J. Schorr, A. M. Krieg, tion of antigen-presenting cells that present exogenous antigens in association and R. Weeranta. 1998. CpG DNA is a potent enhancer of specific immunity in with class I MHC molecules. J. Immunol. 150:438. mice immunized with recombinant hepatitis B surface antigen. J. Immunol. 160: 40. Mitchell, D. A., S. K. Nair, and E. Gilboa. 1998. Dendritic cell/macrophage 870. precursors capture exogenous antigen for MHC class I presentation by dendritic 22. Sparwasser, T., E. S. Koch, R. M. Vabulas, K. Heeg, G. B. Lipford, J. W. Ellwart,

cells. Eur. J. Immunol. 28:1923. Downloaded from and H. Wagner. 1998. Bacterial DNA and immunostimulatory CpG oligonucle- otides trigger maturation and activation of murine dendritic cells. Eur. J. Immu- 41. Brossart, P., and M. J. Bevan. 1997. Presentation of exogenous protein antigens nol. 28:2045. on major histocompatibility complex class I molecules by dendritic cells: path- 23. Jakob, T., P. S. Walker, A. M. Krieg, M. C. Udey, and J. C. Vogel. 1998. way of presentation and regulation by cytokines. Blood 90:1594. Activation of cutaneous dendritic cells by CpG-containing oligodeoxynucleoti- 42. Stockinger, B. 1999. T lymphocyte tolerance: from thymic deletion to peripheral des: a role for dendritic cells in the augmentation of Th1 responses by immuno- control mechanisms. Adv. Immunol. 71:229. stimulatory DNA. J. Immunol. 161:3042. 43. Grewal, I. S., and R. A. Flavell. 1998. CD40 and CD154 in cell-mediated im- 24. Krieg, A. M., A. K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale, munity. Annu. Rev. Immunol. 16:111.

G. A. Koretzky, and D. M. Klinman. 1995. CpG motifs in bacterial DNA trigger http://www.jimmunol.org/ 44. Hacker, H., H. Mischak, T. Miethke, S. Liptay, R. Schmid, T. Sparwasser, direct B-cell activation. Nature 374:546. K. Heeg, G. B. Lipford, and H. Wagner. 1998. CpG-DNA-specific activation of 25. Kontgen, F., G. Suss, C. Stewart, M. Steinmetz, and H. Bluethmann. 1993. Tar- antigen-presenting cells requires stress kinase activity and is preceded by non- geted disruption of the MHC class II Aa gene in C57BL/6 mice. Int. Immunol. specific endocytosis and endosomal maturation. EMBO J. 17:6230. 5:957. 26. Karttunen, J., S. Sanderson, and N. Shastri. 1992. Detection of rare antigen- 45. Sakata, N., H. R. Patel, N. Terada, A. Aruffo, G. L. Johnson, and E. W. Gelfand. presenting cells by the lacZ T-cell activation assay suggests an expression cloning 1995. Selective activation of c-Jun kinase mitogen-activated protein kinase by strategy for T-cell antigens. Proc. Natl. Acad. Sci. USA 89:6020. CD40 on human B cells. J. Biol. Chem. 270:30823. 27. Lipford, G. B., H. Wagner, and K. Heeg. 1994. Vaccination with immunodom- 46. Yi, A. K., and A. M. Krieg. 1998. Rapid induction of mitogen-activated protein inant peptides encapsulated in Quil A-containing liposomes induces peptide-spe- kinases by immune stimulatory CpG DNA. J. Immunol. 161:4493. ϩ cific primary CD8 cytotoxic T cells. Vaccine 12:73. 47. Diehl, L., A. T. den Boer, S. P. Schoenberger, E. I. van der Voort, 28. Battegay, M., S. Cooper, A. Althage, J. Banziger, H. Hengartner, and T. N. Schumacher, C. J. Melief, R. Offringa, and R. E. Toes. 1999. CD40 acti-

R. M. Zinkernagel. 1991. Quantification of lymphocytic choriomeningitis virus vation in vivo overcomes peptide-induced peripheral cytotoxic T-lymphocyte tol- by guest on September 29, 2021 with an immunological focus assay in 24- or 96-well plates. J. Virol. Methods erance and augments anti-tumor vaccine efficacy. Nat. Med. 5:774. 33:191. 48. Sotomayor, E. M., I. Borrello, E. Tubb, F. M. Rattis, H. Bien, Z. Lu, S. Fein, 29. Sallusto, F., M. Cella, C. Danieli, and A. Lanzavecchia. 1995. Dendritic cells use S. Schoenberger, and H. I. Levitsky. 1999. Conversion of tumor-specific CD4ϩ macropinocytosis and the mannose receptor to concentrate macromolecules in the T-cell tolerance to T-cell priming through in vivo ligation of CD40. Nat. Med. major histocompatibility complex class II compartment: downregulation by cy- 5:780. tokines and bacterial products. J. Exp. Med. 182:389. 49. Dullforce, P., D. C. Sutton, and A. W. Heath. 1998. Enhancement of T cell- 30. Kagi, D., B. Ledermann, K. Burki, R. M. Zinkernagel, and H. Hengartner. 1996. independent immune responses in vivo by CD40 antibodies. Nat. Med. 4:88. Molecular mechanisms of lymphocyte-mediated cytotoxicity and their role in immunological protection and pathogenesis in vivo. Annu. Rev. Immunol. 14:207. 50. Gurunathan, S., K. R. Irvine, C. Y. Wu, J. I. Cohen, E. Thomas, C. Prussin, 31. Pircher, H., U. H. Rohrer, D. Moskophidis, R. M. Zinkernagel, and N. P. Restifo, and R. A. Seder. 1998. CD40 ligand/trimer DNA enhances both H. Hengartner. 1991. Lower receptor avidity required for thymic clonal deletion humoral and cellular immune responses and induces protective immunity to in- than for effector T-cell function. Nature 351:482. fectious and tumor challenge. J. Immunol. 161:4563. 32. Aichele, P., K. Brduscha-Riem, R. M. Zinkernagel, H. Hengartner, and 51. Oxenius, A., M. M. Martinic, H. Hengartner, and P. Klenerman. 1999. CpG- H. Pircher. 1995. T cell priming versus T cell tolerance induced by synthetic containing oligonucleotides are efficient adjuvants for induction of protective peptides. J. Exp. Med. 182:261. antiviral immune responses with T-cell peptide vaccines. J. Virol. 73:4120.