The first two authors contributed equally to this work.
Testing mouse mammary tumor virus superantigen as adjuvant in cytotoxic T-lymphocyte responses against a melanoma tumor antigen
Article first published online: 12 MAR 2002
Copyright © 2002 Wiley-Liss, Inc.
International Journal of Cancer
Volume 99, Issue 2, pages 201–206, 10 May 2002
How to Cite
Wirth, S., Bille, F., Koenig, S., Wehrli, N., Miconnet, I., Lévy, F., Diggelmann, H., Romero, P. and Acha-Orbea, H. (2002), Testing mouse mammary tumor virus superantigen as adjuvant in cytotoxic T-lymphocyte responses against a melanoma tumor antigen. Int. J. Cancer, 99: 201–206. doi: 10.1002/ijc.10237
- Issue published online: 11 APR 2002
- Article first published online: 12 MAR 2002
- Manuscript Accepted: 16 NOV 2001
- Manuscript Revised: 12 NOV 2001
- Manuscript Received: 29 AUG 2001
- Giorgio Cavaglieri Foundation
- Swiss National Science Foundation. Grant Number: 31-59165.99
- vaccinia virus;
- tumor immunology;
- mouse mammary tumor virus;
- cytotoxic T lymphocyte
Cytotoxic T cells represent a powerful strategy for antitumor treatment. Depending on the route of injection, an important role for CD4 T cell–mediated help was observed in the induction of this response. For this reason, we investigated whether induction of a CTL response to the HLA-A2–restricted immunodominant peptide melanoma antigen Melan-A was improved by using rVVs expressing the CTL-defined epitope alone or in combination with an SAg. In the latter case, the few infected dendritic cells simultaneously presented an SAg and an antigen, i.e., peptide. Here, we show that the anti-Melan-A response was efficiently induced but not significantly improved by coexpression of the SAg. © 2002 Wiley-Liss, Inc.
The identification of CTL-defined peptide tumor antigens has opened the possibility for the development of cancer vaccines.1 Since tumor cells are often poorly immunogenic, much effort has been directed toward developing vaccination strategies that induce or increase the CTL response against these weak, most often self-derived antigens. Several approaches are currently being investigated. In the case of unidentified tumor antigens, strategies are employed that either increase the immunogenicity of tumor cells by transducing them with genes coding for cytokines or costimulatory molecules2 or enhance the delivery of tumor antigens to professional APCs by loading whole-tumor lysates on dendritic cells.3 Isolated dendritic cells are also used to introduce identified tumor antigens either as synthetic peptides or as peptides expressed from recombinant viral vectors into the lymphoid system.3 While these approaches have in some cases induced strong and efficient T-cell responses, they require handling the patient's cells in vitro. Attempts were thus made to design cell-free vaccines with identified tumor antigens. Vaccination with free tumor peptides has shown limited success due to their rapid degradation by serum proteases4 or their low binding to the presenting MHC class I molecules.5 The use of synthetic peptide analogues with increased affinity for MHC class I6, 7 or TCR molecules8 greatly enhanced specific CTL responses that were able to control tumor cells in vivo, especially in combination with appropriate adjuvants or recombinant viral vectors, allowing efficient antigen delivery and induction of a nonspecific inflammatory response.9, 10
Another important factor that can influence the induction and maintenance of antitumor CTLs is the simultaneous activation of CD4+ helper T cells. Depending on the route of antigen administration, CD4 T-cell responses become very important.11 These T cells provide local cytokines, such as IL-2, or trigger costimulatory signals in APCs. Since CD4+ helper cell–related epitopes are generally not present in peptide vaccines, they must be added to the adjuvant or expressed by the recombinant viral vector. Coexpression of the tumor epitope for CD8+ CTLs and the helper epitope for CD4+ T-helper cells by a recombinant virus may be superior to the mixture of free peptides because this strategy ascertains that epitopes for CD4+ and CD8+ T cells are presented by the same APC.
In the present work, we addressed this question by generating and evaluating a tumor vaccine in the form of a recombinant viral vector that targets the delivery of a CTL-defined peptide melanoma antigen analogue with enhanced immunogenicity and a potent, tumor-unrelated adjuvant/helper epitope for CD4+ T cells to the same professional APCs in vivo. The HLA-A2–restricted immunodominant peptide derived from the human melanoma-associated protein Melan-A/MART-1 was selected in our study.12, 13 Melan-A, a cell lineage-specific protein of unknown function, is found in nearly all melanoma tumor samples and in 60% of melanoma cell lines, as well as in normal melanocytes. It contains 4 epitopes for CD8+ CTLs restricted by HLA-A*0201, the 2 overlapping epitopes 27–35 and 26–35 being immunodominant.14–16 Despite its immunodominance in vitro, the epitope 26–35 (EAAGIGILTV) is poorly immunogenic in vivo. However, a single AL substitution, introducing a favorable anchor residue at position 2 of the antigenic peptide sequence, greatly increased the binding to HLA-A2, the stability of HLA-A2–peptide complexes and its immunogenicity in vivo.7 CD8+ CTLs generated against this Melan-A26–35A27L (ELAGIGILTV) peptide analogue cross-react with the natural Melan-A antigen and lyse melanoma cell lines in vitro.17 We therefore used the Melan-A26–35A27L peptide analogue as the optimized immunizing antigen in our melanoma-specific vaccine. As a potent adjuvant/helper molecule for CD4+ T cells, we included the MMTV orf gene product, which has SAg activity. reviewed in 18 SAg proteins bind to MHC class II molecules at the cell surface and interact with the variable domain of the β chain (Vβ) of the TCR. They show little MHC restriction and can be recognized by up to 4 of 25 mouse Vβ elements, thereby inducing SAg-reactive CD4+ T cells with frequencies that reach one-third of the total T-cell repertoire in naive mice. SAg interaction not only induces deletion and anergy but can also prime efficient secondary responses.19 Moreover, in the presence of antigen, SAg can have strong adjuvant effects.20, 21
The MMTV SAg can also have an important adjuvant effect on the generation of envelope-specific anti-MMTV antibodies, as previously reported by our group.22 MMTV SAg can be presented by human MHC class II molecules and recognized by human CD4+ T cells.23 In contrast to bacterial SAgs, MMTV SAg lacks function in the soluble form and needs to be produced by MHC class II+ APCs for efficient stimulation of SAg-reactive CD4+ T cells. We explored this feature of the MMTV SAg to target it together with the Melan-A26–35A27L peptide to the same professional APCs using rVV. This expression system efficiently stimulates MMTV SAg-reactive T cells in vivo without inducing their deletion.24 As the CD4+ T cells remain fully active to challenge with the same MMTV, the host T-cell repertoire remains unchanged (S. Wirth, unpublished observation). rVVs induce strong immune responses to inserted antigens, as well as proinflammatory cytokines, costimulatory signals and bystander activation that may further enhance antigen-specific responses.25 They can infect macrophages and, more importantly, dendritic cells,26, 27 which play a crucial role in the initiation of T cell-dependent immune responses.28 We thus generated an rVV expressing both the mtv-7 SAg and the Melan-A26–35A27L peptide analogue and compared it with an rVV expressing the Melan-A26–35A27L peptide analogue only for the induction of Melan-A-specific CTLs in HLA-A*0201/Kb transgenic mice.
MATERIAL AND METHODS
BALB/c (H-2d) and C57BL/6 (H-2b) mice were purchased from Harlan-Olac (Bicester, UK). HLA-A*0201/Kb transgenic mice (line 6, H-2b, mtv-8+mtv-9+) were obtained from Harlan Sprague-Dawley (Indianapolis, IN) and maintained as breeding pairs in our animal facility.
BSC-40 cells, derived from BSC-1 cells (ATCC, Manassas, VA; CCL-26), huTk-143B cells (ATCC CRL-8303) and 293T cells,29 were grown in DMEM supplemented with 10% FCS and antibiotics. The NA8-MEL human melanoma cell line (HLA-A*0201+MelanA–12) was maintained in RPMI-1640 supplemented with 10% FCS and antibiotics. Mouse EL-4 cells transfected with the HLA-A*0201/Kb gene, EL-4.A2/Kb,30 were kindly provided by Dr. L. Sherman (Scripps Research Institute, La Jolla, CA) and maintained in DMEM supplemented with 10% FCS, antibiotics and 0.5 mg/ml G418.
Construction of plasmids and rVV
The recombinant vector pKT/GFP-Ub-M26-35A27L (construct IV), hereafter called pKT401/IV, as well as the rVV GFP-Ub-M26-35A27L (v-ELA, Copenhagen strain) have been previously described.10 This rVV encodes a linear fusion product in which ubiquitin is located between the GFP and the Melan-A26–35 immunodominant peptide analogue A27L (ELAGIGILTV) under the control of a mutant vaccinia virus early P 7.5 promoter (p1). The mtv-7 SAg coding sequence was amplified by PCR from a pHβApr plasmid (provided by Dr. A. Terskikh, Stanford University, Stanford, CA). The following primers containing the AflII or ApaI endonuclease site (underlined) and the canonical Kozak consensus sequence (italic) in front of the ATG were used: 5′-GATCCTTAAGCGCCACCATGCCTCGCCTGCAGCAGAAA-3′ and 5′-GATCGGGCCCTTAAAAGGGATCGAAGCCAA-3′. To avoid a potential strong hairpin, a G was replaced by T (bold), which does not change the amino acid sequence. The PCR product was cloned into the kanamycin-resistant pCR-Blunt vector (Invitrogen, Groningen, the Netherlands) upon digestion with AflII and ApaI and sequenced. After digestion with AflII and ApaI, the resulting 986 bp fragment was inserted into pKT401/IV, yielding expression of the mtv-7 SAg under the control of a mutant vaccinia virus early thymidine kinase promoter (p3). Recombination into the I4L locus of vaccinia virus (Copenhagen strain) was achieved following the transient dominant gpt selection protocol on CV-1 and BSC-40 cells, as described.10 After plaque purification, rVVs (named v-ELA/SAg) were screened by PCR and amplified. The control rVV, v-MAGE-3 (Copenhagen strain) expressing the Melan-A-irrelevant nonapeptide MAGE-3271–279, has been previously described.10
Quantitation of GFP Melan-A26–35A27L expression by rVV in 293T cells in vitro
Human embryonic kidney cells (HEK293T) seeded in 24-well plates at a density of 1 × 106 cells/well were infected with 1 multiplicity of infection of wild-type vaccinia virus (v-cpn, Copenhagen strain), v-ELA or v-ELA/SAg for 1 hr at room temperature. After addition of medium (DMEM supplemented with 10% FCS and antibiotics), cultures were incubated at 37°C for different periods. Cells were then harvested and fixed with 0.5% paraformaldehyde/PBS and GFP expression was determined by FACScan analysis.
Induction of Melan-A26–35A27L-specific CTLs
Different protocols were used for the immunization of HLA-A*0201/Kb transgenic mice with v-ELA, v-ELA/SAg or v-MAGE-3. Saturating doses for induction of the SAg response, corresponding to either 2 × 106,6 5 × 106 or 1 × 107 PFU, were injected i.p., or a total of 4 × 106 PFU were injected in 2 equal doses s.c. at the base of the tail and i.p. Two weeks later, mice were killed and spleen and draining lymph node cells (4 × 106) were cultured with 2 × 105 irradiated (10 krad) EL-4 A2/Kb cells, prepulsed with 1 × 10–6 M Melan-A26–35A27L peptide for 1 hr at 37°C in 24-well plates in 2 ml of DMEM supplemented with 10% FCS, antibiotics, 5 × 10–5 M β-mercaptoethanol, 10 mM HEPES, 1 mM sodium pyruvate, 1% MEM nonessential amino acids and 30 U/ml IL-2 in the form of EL-4 supernatant. After 1 or 2 rounds of weekly stimulation, cultured cells were tested for Melan-A26–35A27L-specific CTLs using tetramer staining and 51Cr-release assay.
Flow-cytometric immunofluorescence analysis
Fluorescent tetramers of HLA-A*0201–β2-microglobulin–peptide complexes were synthesized as described10 using the Melan-A26–35A27L peptide analogue (ELAGIGILTV) as antigen and the influenza peptide (GILGFVFTL) as control. Cells were stained with tetramers at room temperature for 20 min and then with a mixture of anti-CD8Cy5, anti-TCRCy and anti-CD44FITC MAbs (all from Pharmingen, San Diego, CA) at 4°C for 30 min. Cells were washed once and analyzed immediately in a FACSCalibur (Becton Dickinson, San Jose, CA). Data were analyzed using CellQuest software (Becton Dickinson).
Cytolytic activity was assessed in a standard 4 hr 51Cr-release assay using EL-4 A2/Kb as target cells. Melan-A26–35A27L peptide was added to a final concentration of 5 × 10–7 M where indicated. Percent specific 51Cr release was calculated as follows: (exp. cpm – spontaneous cpm)/(total cpm – spontaneous cpm) × 100.
NA8-MEL cells (HLA-A*0201+Melan-A–) were infected or not with rVV at 100, 30 or 10 multiplicities of infection. After 3 hr of incubation at 37°C, cells were irradiated (3 krad) and washed and 2 × 104 cells were cocultured with an equal number of cells of a Melan-A26–35A27L-specific mouse CTL line (H-2b, CD8+) in a 96-well round-bottomed plate in 200 μl of DMEM with 10% FCS, antibiotics and β-mercaptoethanol. Melan-A26–35A27L peptide was added where indicated. After 24 hr incubation at 37°C, IFN-γ levels in culture supernatants were measured by ELISA, as described.10
RESULTS AND DISCUSSION
Generation and characterization of the v-ELA/SAg rVV coding for the Melan-A26–35A27L peptide analogue and the mtv-7 SAg
To address the question of whether the MMTV SAg can act as an adjuvant for antitumor CTL responses, we constructed an rVV expressing both the mtv-7 SAg and the Melan-A26–35A27L peptide analogue. The previously described ubiquitin/protein/reference vector (pKT1401/IV10), containing the minigene for the Melan-A26–35A27L peptide analogue under the control of the mutant promoter P7.5, was modified by insertion of the mtv-7 SAg coding sequence downstream of the Pkt13 early promoter.31 After recombination into the I4L locus of vaccinia virus of the intermediately attenuated Copenhagen strain, the recombinant virus, v-ELA/SAg, was plaque-purified, amplified and titrated. v-ELA/SAg was first tested for its ability to induce a SAg response in MHC class II I-E-positive BALB/c and I-E-negative C57BL/6 mice. Titration experiments showed maximal responses with viral doses of 2 × 106 PFU or more viral particles in either strain (data not shown) within 2 days of infection, as previously reported with an rVV expressing the SAg of MMTV (GR).24 As shown in Figure 1, v-ELA/SAg efficiently stimulated mtv-7 SAg-reactive Vβ6+CD4+ T cells but not CD4+ T cells bearing irrelevant Vβ 14+ TCRs. Both mouse strains showed a specific 4-fold increase in the percentages of SAg-reactive T cells at the peak of the reaction (Fig. 1). These results demonstrate that I-E-negative mouse strains maximally respond to the mtv-7 SAg expressed by rVV. As expected, the control virus v-ELA, harboring the minigene for the Melan-A26–35A27L peptide analogue, alone did not stimulate mtv-7 SAg-specific Vβ6+CD4+ T cells. These data indicate that infection with v-ELA/SAg leads to expression of a functional mtv-7 SAg at the surface of APCs.
We next compared v-ELA/SAg with the existing rVV v-ELA for expression of the Melan-A26–35A27L peptide analogue and for its presentation by HLA-A*0201 to Melan-A-specific CD8+ CTLs. v-ELA and v-ELA/SAg use the ubiquitin/protein/reference system to generate the Melan-A epitope.10, 32, 33 Both encode the same linear fusion product, in which ubiquitin is located between the reference GFP and the Melan-A26–35 immunodominant peptide analogue A27L. Cotranslational cleavage by ubiquitin-specific proteases after the last residue of ubiquitin produces equimolar amounts of Melan-A peptide and GFP, which can be measured directly by flow cytometry. The results presented in Figure 2a demonstrate that v-ELA/SAg and v-ELA generate similar amounts of Melan-A peptide as the percentage of GFP-positive 293T cells did not significantly differ between these 2 rVVs at most time points of infection. Only at 4 hr of infection was a slightly but significantly lower percentage of GFP-positive cells detected for v-ELA/Sag, as determined with a 2-tailed Student's t-test (unpaired, p < 0.05). To ensure that the Melan-A26–35A27L peptide analogue was not only expressed but also correctly presented by HLA-A*0201 and recognized by CD8+ T cells, we determined the presentation of the Melan-A26–35A27L peptide analogue by rVV-infected HLA-A*0201+ NA8-MEL cells to a Melan-A-specific CD8+ CTL line. NA8 cells infected with v-ELA or v-ELA/SAg efficiently stimulated CD8+ CTLs for IFN-γ secretion and the response observed with v-ELA/SAg was even superior to that of v-ELA (Fig. 2b). CTL activation was antigen-specific since v-MAGE-3–infected NA8 cells did not induce detectable levels of IFN-γ. Furthermore, IFN-γ levels varied only slightly when different multiplicities of infection of rVV were used, suggesting that the Melan-A peptide was produced in amounts sufficient for CTL activation at the lowest viral dose tested. Comparable levels of CTL activation were achieved when 1 to 3 × 10–7 M peptide was added to CTL cultures with uninfected NA8 cells. These results demonstrate that v-ELA and v-ELA/SAg produce the Melan-A26–35A27L peptide in similar amounts and in a form that is recognized by Melan-A-specific CTLs in vitro.
Induction of Melan-A-specific CD8+ CTL upon immunization of HLA-A*0201/Kb transgenic mice with v-ELA/Sag
To evaluate the ability of the 2 rVVs to induce Melan-A-specific CTL responses in vivo, we immunized HLA-A2/Kb transgenic mice by a single i.p. rVV injection. Spleen cells were analyzed for Melan-A-specific CD8+ CTLs either ex vivo or after 1 and 2 rounds of restimulation in vitro with 2 different assays: (i) fluorescent tetramers of HLA-A2–peptide complexes were used for direct visualization of CD8+ T cells carrying a TCR specific for Melan-A and (ii) the classical 51Cr-release assay was used to assess the antigen-specific cytolytic function of the cell cultures. Fluorescent tetramer staining clearly revealed CD8+ T cells with a Melan-A-specific TCR in cultures from mice immunized with Melan-A epitope–encoding rVVs (v-ELA and v-ELA/SAg, Fig. 3). Whereas tetramer+ CD8+ T cells were barely detectable after 1 round of stimulation in vitro (Fig. 3a), strong Melan-A-specific responses with frequencies of tetramer+ CD8+ T cells up to 40% were observed after 2 rounds of stimulation in vitro (Fig. 3b). However, no significant difference in the percentage of tetramer+ cells was found between v-ELA- and v-ELA/Sag-immunized mice, which would have suggested an adjuvant effect of the mtv-7 SAg. Mice immunized with rVV expressing a Melan-A-unrelated melanoma antigen (v-MAGE-3) showed low to undetectable frequencies of tetramer+ CD8+ T cells, indicating that the Melan-A response was due to T-cell priming in vivo. Parallel measurements of cytolytic function also detected similar Melan-A26–35A27L-specific cytolytic activity in v-ELA- and v-ELA/SAg-infected mice but not in v-MAGE-3–infected control mice (Fig. 4), in agreement with the results obtained with fluorescent tetramers. Again, 2 rounds of stimulation in vitro were necessary. Only 1 mouse of the vELA and the v-ELA/SAg group showed a strong antigen-specific CTL response after 1 round of stimulation in vitro (data not shown). As no significant difference from the v-ELA group was determined after 2 rounds of stimulation in vitro, v-ELA and v-ELA/SAg are equivalent for the induction of antigen-specific CD8+ T cells with cytolytic function in vivo. Different immunizing regimens with respect to rVV dose (2 × 106, 5 × 106vs. 1 × 107 PFU i.p.) and site of injection (i.m. at tail base vs. i.p.) were subsequently tested. In response to 2 × 106 PFU of rVV, cytolytic activities after the second round of stimulation at an E:T ratio of 20:1 in the v-ELA and v-ELA/SAg groups were 25%, 12.5%, 1.4% and 3.4%, 10.5%, 1.6%, respectively. Cells from draining lymph nodes tested in parallel gave the following results at an E:T ratio of 20:1: 28.8% and 27.3% for v-ELA and 49.7%, 33.2% and 22.1% for v-ELA/SAg. This led to the same conclusion that Melan-A-specific CTLs, but no increasing effect of the mtv-7 SAg, could be detected.
Altogether, the present data clearly indicate that in the presence of the mtv-7 SAg a Melan-A-specific CD8+ T-cell response with cytolytic function could be induced, though not stronger than the one triggered by the rVV encoding the Melan-A26–35A27L peptide analogue alone. This failure to detect an adjuvant effect of the viral SAg was unexpected for several reasons. First, we chose rVVs and their early promoters for the simultaneous expression of Melan-A CTL epitope and mtv-7 SAg. Early vaccinia promoters are the most suitable for antigen expression in dendritic cells, leading to efficient CTL priming in vivo.26,27 Furthermore, we have ascertained that the mtv-7 SAg is produced in a functional form and within a time frame that should allow the interaction of SAg-reactive CD4+ T cells or their lymphokines with Melan-A-specific CD8+ T cells in the vicinity of the APC. Expression of MMTV SAg by rVV leads to a detectable increase in the percentage of SAg-reactive CD4+ T cells within 6 hr of infection. This SAg-specific T-cell response reaches its maximum after 12 hr, remains high until day 5 but returns to normal levels at day 6 of infection. No deletion of SAg-reactive CD4+ T cells is observed thereafter.24 This temporal expression of the mtv-7 SAg should be adequate for CTL priming since the Melan-A26–35A27L peptide analogue is generated within the same time window (Fig. 2a) and the CTL priming appears to require only a single short contact of 2 hr with the APC.34 Concerning the quantitative aspects of SAg expression, our v-ELA/SAg produced sufficient Melan-A26–35A27L peptide analogue to induce CTLs in vivo, which has higher requirements for antigen than the SAg response. In summary, we have shown that concomitant expression of SAg and tumor antigens by the same APC did not significantly increase the antitumor CTL response.
We greatly acknowledge Ms. A. Vassaz and Mr. L. Scarpellino for expert technical help and Dr. P. Majcherczyk for critical reading of the manuscript.