Bioconjugate Chem. 2006, 17, 1330−1338
1330
Chemical Conjugate TMV-Peptide Bivalent Fusion Vaccines Improve Cellular Immunity and Tumor Protection Alison A. McCormick,* Tina A. Corbo, Sherri Wykoff-Clary, Kenneth E. Palmer, and Gregory P. Pogue Large Scale Biology Corporation, 3333 Vacavalley Parkway, Suite 1000, Vacaville, California 95688. Received May 12, 2006; Revised Manuscript Received July 17, 2006
Chemical conjugation of CTL peptides to tobacco mosaic virus (TMV) has shown promise as a molecular adjuvant scaffold for augmentation of cellular immune responses to peptide vaccines. This study demonstrates the ease of generating complex multipeptide vaccine formulations using chemical conjugation to TMV for improved vaccine efficacy. We have tested a model foreign antigen targetsthe chicken ovalbumin-derived CTL peptide (Ova peptide), as well as mouse melanoma-associated CTL epitopes p15e and tyrosinase-related protein 2 (Trp2) peptides that are self-antigen targets. Ova peptide fusions to TMV, as bivalent formulations with peptides encoding additional T-help or cellular uptake via the integrin-receptor binding RGD peptide, showed improved vaccine potency evidenced by significantly enhanced numbers of antigen-reactive T cells measured by in vitro IFNγ cellular analysis. We measured the biologically relevant outcome of vaccination in protection of mice from EG.7-Ova tumor challenge, which was achieved with only two doses of vaccine (∼600 ng peptide) given without adjuvant. The p15e peptide alone or Trp2 peptide alone, or as a bivalent formulation with T-help or RGD uptake epitopes, was unable to stimulate effective tumor protection. However, a vaccine with both CTL peptides fused together onto TMV generated significantly improved survival. Interestingly, different bivalent vaccine formulations were required to improve vaccine efficacy for Ova or melanoma tumor model systems.
INTRODUCTION Peptides derived from cancer antigens have been used for more than a decade as cheap, safe, and easy-to-manufacture vaccines that stimulate target-specific immunogenicity (1-6). Peptide vaccines, typically given as single agents mixed with adjuvant, are taken up by antigen-presenting cells (APC) at the site of injection, and, by processes not well-understood, are cross presented on MHC class I to naive T cells (7, 8). However, peptide vaccines have generally shown limited efficacy in a series of disappointing clinical trials. Although immune responses may be detectible in patients immunized with peptidebased vaccines (9-12), the functional activation of antigenpresenting cells, induction of antigen-specific CD4+ and CD8+ cells (13, 14), is not sufficient to mediate subsequent tumor clearance or disease regression in the majority of treated patients (13, 15, 16). Much effort has been made to understand why peptide vaccines have limited efficacy, and new generations of peptide vaccine formulations are being tested with the hopes that they will overcome the limitations of previous peptideadjuvant formulations. The importance of the context of antigen presentation to achieve appropriate immune stimulation, especially for selfantigens and breaking tolerance, has led to a recent understanding that vaccines composed of peptides or whole protein antigens generally do not sufficiently stimulate APC activation (recently reviewed in (17)). Resulting T-cell responses are often only transient and subject to T-regulatory mechanisms that eliminate antigen-specific reactivity (18, 19). There has been significant effort, therefore, to improve peptide antigen-specific responses by a number of different mechanisms that improve antigen delivery to APCs. First, by using a pathogen to deliver the antigen, by bacterial (20, 21), yeast (22), viral vector, or viruslike particle delivery (23-26). These approaches take advantage * E-mail:
[email protected]. Office: 707-469-2384. Fax: 707-446-3917.
of particulate antigen uptake and improved cellular targeting and activation of professional APCs, such as dendritic cells (DCs). Virus expression in mammalian cells and VLP manufacturing, however, are still production-limited processes (27). Direct DC pulsing has also proven successful but requires generation of patient-specific DC, which remains costly (28). Another method to improve immunogenicity is by association of poorly immunogenic antigens to toxins (29-31), cytokines (32), T-helper epitopes (33-35), or peptides that improve uptake by APCs (36). We have melded two approaches to improve peptide antigen efficacy, by chemically conjugating peptides to the surface of a virus to create a chimeric virus particle, as well as by formulating bivalent fusions with toxin-derived peptides, T-helper epitopes, or peptides that improve uptake. We have explored the use of the tobacco mosaic virus as an ideal peptide antigen carrier, because it provides an easily manipulated scaffold of >2100 uniform coat proteins that are 17 KDa in size. TMV also contains a +strand RNA that only transcribes in plants, but which may still engage innate antiviral immunity and activate important signaling molecules like Tolllike receptor 7, and thereby stimulate immune activation (3739). Many studies have shown that fusions of B-cell epitopes to TMV stimulate high levels of antibody titers, and more recently high levels of tumor-protective CTL responses (40). TMV is easily produced on the kilogram scale from infected plants, with low cost of goods, and has shown an excellent safety profile in preclinical tests because it is produced without animal or bacterial byproducts. Using TMV as a CTL peptide antigen carrier by chemical conjugation, we have improved vaccine efficacy by bivalent fusions with toxin-derived peptides, Thelper epitopes, or peptides that improve uptake. We report that, in two different murine tumor models, representing foreign (Ova) and self (melanoma) antigens, bivalent vaccine preparations stimulate significantly improved immune responses, as measured by interferon gamma (IFNγ) secreting cells and survival from lethal tumor challenge. Interestingly, different
10.1021/bc060124m CCC: $33.50 © 2006 American Chemical Society Published on Web 08/22/2006
Bioconjugate Chem., Vol. 17, No. 5, 2006 1331
TMV Bivalent Fusion Vaccines Table 1. TMV Conjugation Reactions: Theoretical and Observed Masses in Daltons conjugate name
peptide sequence
1295.10 SPDP mass
peptide mass
expected mass
observed MALDI-TOF
1295.10 + LC-SPDP 1295.10 + LC-SPDP + Ova 1295.10 + LC-SPDP + E7 1295.10 + LC-SPDP + TT 1295.10 + LC-SPDP + RGD 1295.10 + LC-SPDP + p15e 1295.10 + LC-SPDP + trp2
SIINFEKLc DRAHYNIVTFc QYIKANSKFIGITELKKc SGRGDSGc KSPWFTTLc SVYDFFVWLc
18353.5 18353.5 18353.5 18353.5 18353.5 18353.5
1065.3 1337.6 1827.8 736.7 1082.6 1279.5
18353.5 19418.8 19691.1 20181.3 19090.2 19436.1 19633.0
18356.0 19423.9 19690.0 20185.0 19091.6 19431.4 19629.6
formulations are required for Ova or melanoma-effective tumor protective immunity. Peptide antigen delivery on TMV, using less that 1 µg per dose, stimulates these protective immune responses in murine models indicating that this approach has promise for improving the efficacy of clinically relevant peptide vaccines without sacrificing safety.
MATERIALS AND METHODS Vaccine Production. Chemical conjugation of peptides to TMV required the introduction of a surface-reactive lysine at the N-terminus. In these experiments, a well-characterized lysine-modified TMV (TMV-lys) was used (30; M1 EPMK S2YS) that expresses a single reactive lysine (K) residue on the surface of the TMV virion (1295.10; (41)). The virus was expressed by transient infection of plants and purified as previously described (40-42). Purified 1295.10 TMV-lys virus was primed with a 20-fold molar excess of the heterobifunctional linker sulfo-LC-SPDP (hereafter SPDP; Pierce, Rockfold, IL) in PBS/0.05 M EDTA. After dialysis to remove free SPDP, Ova, melanoma p15e, or melamona Trp2 peptide (>95% pure; BioMer Technology, Concord, CA) was added at a 10-fold molar excess and reacted at room temperature overnight (sequences are listed in Table 1). Resulting vaccines, Ova, p15e, or Trp2, were subsequently dialyzed and analyzed by matrixassisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectroscopy using a Voyager-DE STR Biospectrometry Workstation (Applied Biosystems, Foster City CA) (40, 41). Mass calculations were based on the combined amino acid masses of peptide and TMV-SPDP, determined by General Protein/Mass Analysis for Windows software (GPMAW) version 5.0 (Lighthouse Data, Denmark). Differences between expected and observed masses were within MALDI-TOF error of measurement (0.05%; Table 1). Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (10-20% TrisGlycine; Invtrogen) and densitometry analysis (BioRad, Hercules, CA; Quantity One software) were used to confirm the extent of purity and >90% peptide conjugation efficiency. Bivalent vaccine formulations were prepared by two methods. First, for most conjugations where peptide sizes were distinguishable by either SDS-PAGE or MALDI-TOF (TT, RGD, E7), 50/50 peptide mixtures were added directly to SPDPactivated TMV and evaluated for approximately 40-45% conjugation of each peptide. In some cases, peptide ratios were adjusted to balance conjugation ratios. For p15e bivalent formulation with Trp2 peptide, where peptide sizes are nearly identical, p15e peptide was added in limiting amounts to generate a TMV conjugate at 50% loading efficiency. Then, Trp2 peptide was added, and on the basis of the reduction of unconjugated TMV coat protein to ∼5%, Trp2 peptide conjugation at ∼45% was confirmed by SDS-PAGE (by the reduction in unconjugated TMV band) and by MALDI-TOF analyses (Table 1). Vaccination. C57b/6 mice (Harlan Sprague Dawley, Indianapolis, IN) were housed at Antibodies, Inc. (Davis, CA), according to guidelines established in the Care and Use of Animals, and protocols were performed according to IACUC approval. Typically, 10 to 15 mice per group were given a 100-
200 µL subcutaneuous (s.c.) injection of 10 or 15 µg protein with or without 25 µg of unmethylated CpG oligonucleotide (TCTCCCAGCGTGCGCCAT; (43)) having a stabilized phosphorothioate backbone (Oligo’s Etc., Inc, Watsonville OR). Vaccines were typically administered at two-week intervals. ELISpot Analysis of IFNγ T-Cell Secretion. Interferon-γ enzyme-linked immunospot (IFNγ ELISpot) analysis was carried out using an antibody kit, according to manufacturer’s instructions (BD/Pharmingen) using membrane filter plates (Millipore, Bedford, MA), and as previously described in detail (40). Briefly, single cell suspensions of 2 × 105 spleen cells from each group were plated in 200 µL RPMI complete media with 10-5 M Ova SIINFEKL peptide, or 10-6 M Ova SIINFEKL peptide, or 10-5 M p15e (KSPWFTTL) or Trp2 (SVYDFFVWL) peptides, using a β-gal control (DAPIYTNV (44)) for all experiments. PMA/Ionomycin (2.5/250 ng/mL; Sigma, St. Louis, MO) treatment was used as a positive control. Cells were incubated for 22 h at 37 °C in a 5% CO2 incubator, stained with an IFNγ-horseradish peroxidase (HRP) conjugate, and then developed for HRP reactivity according to the antibody manufacturer’s instructions. IFNγ secreting cells were measured as positive spots (intensity, 7-255; size, 22-5000; gradient, 1-90) that were counted on an ELISpot reader (AID, Strasbourg, Germany), and normalized to 106 cells after background subtraction (irrelevant peptide stimulation, typically less than 10 spots per 106 cells). Tumor Challenge. EG.7-Ova cells (ATTC, Manassas, VA (45)) were grown from a master cell bank for one week and maintained between 0.2 and 2 × 106 cells per milliliter in RPMI complete media. On the day of the tumor challenge, cells were washed three times in Hanks buffered saline solution (HBSS; Invtirogen/Gibco), counted (Vicell XR; Beckman Coulter, Fullerton, CA), and diluted to 2 × 106 cells/mL. 2 × 105 cells in 100 µL final volume (stored on ice until administration) were used to inoculate mice s.c. on the back. All mice showed a single subcutaneous bubble of fluid under the skin, without leakage. B16F1 cells (46) were grown in DMEM high glucose with 10% FCS, Glutamax, Pen/strep, Hepes, pyruvate, and nonessential amino acids (Invitrogen/Gibco). B16F1 cells were maintained for one week prior to challenge at ∼20-80% confluence, removed by Trypsin/EDTA treatment for 3 min at room temperature, washed extensively, counted, and resuspended at 5 × 105 cells/mL for delivery of 5 × 104 cells s.c. in 100 µL final volume. Tumor appearance was monitored by eye for 8-10 days until visible signs of tumor formation were evident. Then, calipers were used to measure the longest and widest point until tumor size was in excess of 2.0 cm2 or 1.5 cm2, for Ova or B16 tumors, respectively, at which point animals were euthanized by CO2 asphyxiation according to IACUC protocol. Tumor volume and survival curve comparison statistics were plotted in GraphPad Prism 4.0 using Kaplan-Meier log rank analysis.
RESULTS Chemical conjugate vaccines were prepared using Ova SIINFEKL peptide alone, or as bivalent vaccines combining Ova with T-helper epitopes HPV E7 (47, 48)) or Tetanus toxoid
1332 Bioconjugate Chem., Vol. 17, No. 5, 2006
Figure 1. Qualification of TMV Ova conjugate vaccines. (A) SDSPAGE analysis of SPDP conjugates of Ova peptides, and helper or uptake peptides, to TMV. Monovalent formulation (Ova) or bivalent formulations (::) with E7, tetanus (TT), or RGD uptake peptides are shown as final vaccine formulations, with approximately 45% Ova and 45% additional peptides in bivalent vaccines, compared to 90% conjugation of Ova alone. Single (*) and double (90% of the coat proteins conjugated with peptide and, to the extent possible, were composed of equal mixtures of Ova and additional epitopes. SDS-PAGE analysis indicates the degree of conjugation and relative percentage of Ova (Figure 1A). Ova peptide conjugation alone results in a clear size shift of TMV coat protein migration compared to TMV-SPDP alone under nonreducing conditions (*). Small amounts of Ova peptide dimers (500 IFN secreting cells, reported as spots per 106 spleen cells; Figure 2A) with a nearly 50% increase in the Ova::E7 group at >800 IFNγ spots per 106 cells and almost 400% more activated cells in the Ova::RGD bivalent group at
Figure 2. In vivo analysis of vaccine Ova TMV vaccine potency (with CpG DNA adjuvant) by IFNγ ELISpot and EG.7-Ova tumor challenge after three or two doses. (A) Nine days after 3 biweekly s.c. doses of 10 µg Ova TMV vaccines given with 25 µg CpG adjuvant, mice were sacrificed, and 2 × 105 single suspensions of spleen cells were tested on antibody-coated membranes for Ova-specific peptide reactivity (SIINFEKL; 10-6 M) after 22 h stimulation by IFNγ ELISpot. Error bars represent variation between 12-well replicates. (B) One day later, the remaining 10 animals were challenged with a lethal dose of 2 × 105 EG.7-Ova tumor cells given s.c., and the tumor area was monitored by the longest and widest caliper measurements after the onset of visible tumor. Mice were euthanized when the area exceeded 2.0 cm2, and the time of death after tumor challenge recorded; long-term survivors were tumor-free. GraphPad Prism 4.0 software was used to plot survival curves, and Kaplan-Meier analysis was used to determine significant difference (log-rank, p ) 0.05). (C) EG.7-Ova survival results after two doses of the same vaccines, given with CpG adjuvant. Only Ova::E7 was statistically improved in comparison to the Ova monovalent control (p ) 0.05); both groups were better than the control (p < 0.0001). (D) Tumor regression was observed in one Ova::E7 mouse, compared to Ova alone.
TMV Bivalent Fusion Vaccines
Bioconjugate Chem., Vol. 17, No. 5, 2006 1333
Figure 3. In vivo analysis of vaccine Ova TMV vaccine potency (no adjuvant) by IFNγ ELISpot and EG.7-Ova tumor challenge after two doses. (A) Nine days after two biweekly s.c. doses of 10 µg Ova TMV vaccines given without adjuvant, mice were sacrificed and tested for Ova-specific peptide reactivity (SIINFEL; 10-6 M) after 22 h stimulation by IFNγ ELISpot. (B-D) One day later, the remaining 12 animals were challenged with a lethal dose of 2 × 105 EG.7-Ova tumor cells given s.c., and the tumor area was monitored by the longest and widest caliper measurements after the onset of visible tumor and analyzed as in Figure 2. (B) Ova::RGD vaccine stimulated significantly improved survival compared to Ova TMV alone (p ) 0.0169). (C) Ova::TT vaccine stimulated similar survival compared to Ova TMV alone (p ) 0.19). (D) Ova::RGD vaccine stimulated significantly improved survival compared to Ova TMV alone (p ) 0.0049).
nearly 2000 IFNγ spots per 106 cells. Ova::TT group had slightly lower numbers of activated cells. When the remaining mice in these groups were challenged with a lethal dose of Ovalbumin expressing EG.7-Ova tumor cells, all vaccine groups had significantly improved survival compared with the control (p < 0.0001; Figure 2B), with approximately 50-75% long-term survivors that were tumor-free. Previous experiments confirmed that animals immunized with TMV alone or TMV mixed with an equivalent dose of peptide (unlinked) showed similar survival kinetics compared with PBS controls (data not shown). Although Ova::RGD had the highest IFNγ activation in vitro and the highest percentage of survivors in vivo, this group was statistically similar to all other vaccine groups, including Ova-TMV given as a monovalent vaccine. Such data suggest an immune response threshold has been achieved after vaccination that protects the majority of mice from tumor lethality. To discriminate between vaccine compositions, we administered only two immunizations normalized to deliver 600 ng of Ova peptide/dose, given with 25 µg CpG DNA adjuvant. All vaccine groups had high levels of IFN secretion by ELISpot analysis (data not shown), and all vaccine groups showed significantly improved tumor protection compared to the PBS control (p > 0.0001). Only one group, Ova::E7 CpG, showed improved survival compared with the monovalent Ova-TMV conjugate vaccine (Figure 2C; p ) 0.05). Also of note, one animal in the Ova::E7 CpG group formed a visible tumor that regressed and was eliminated (Figure 2D), unlike tumors in mice vaccinated with the monovalent Ova-TMV. The figure shows tumor growth curves that are typically observed in three animals that received the Ova monovalent vaccine (triangles), in comparison with the single Ova:E7 vaccine recipient (closed circles), in which we observed clear regression of the tumor. In
our experience, and on the basis of a lack of similar findings using the EG.7-Ova tumor model in the literature, peptidevaccine-induced regression of an established tumor is unprecedented. To further discriminate between vaccine potencies, we increased the stringency of the challenge. The same groups (as above) were vaccinated in parallel two times without CpG adjuvant. As shown in Figure 3A, in vitro analysis of cellular immunity by IFNγ ELISpot still demonstrated high levels of peptide-stimulated cells in all groups after only two immunizations, with greater than 500 IFNγ secreting (spot forming) cells per 106 splenocytes in the Ova-TMV group. Again, there were significantly higher numbers of activated cells in both the Ova:: RGD and Ova::E7 groups compared with the monovalent vaccine, and fewer activated cells in the Ova::TT group. These results were remarkably similar to what we measured in an independent analysis after three doses with adjuvant (Figure 2A). A subsequent challenge by Ova expressing EG.7-Ova tumor cells indicates that all vaccine groups showed significantly improved tumor survival patterns in comparison with the control (p < 0.0001). However, two doses of the Ova-TMV monovalent conjugate vaccine only generated a significant delay in survival compared to the control, but no long-term survivors. The bivalent Ova::RGD vaccine group, with the highest level of IFNγ secreting cell numbers, performed better than OvaTMV alone (p ) 0.0169), with 25% long-term survivors. Ova:: TT bivalent vaccine, that stimulated the lowest numbers of IFNγ secreting cells, performed no better than Ova-TMV 2× (p ) 0.19) even though there were 25% long-term survivors. Ova:: E7 outperformed Ova-TMV (p ) 0.0049) with 50% long-term survivors. The responses of animals immunized with the Ova:: RGD and Ova::E7 in the absence of adjuvant confirms the
1334 Bioconjugate Chem., Vol. 17, No. 5, 2006
Figure 4. In vivo testing of p15e bivalent vaccine formulation efficacy by IFNγ ELISpot and by B16 tumor challenge. Three doses of 15 µg monovalent p15e vaccine, or peptide equivalent, was given on a biweekly schedule with 25 µg CpG adjuvant. (A) Nine days after the last vaccine, mice were sacrificed, and 2 × 105 single suspensions of spleen cells were tested on antibody-coated membranes for p15e specific peptide reactivity (KSPWFTTL; 10-5 M) after 22 h stimulation, by IFNγ ELISpot. (B) One day later, the remaining 10 animals were challenged with a lethal dose of 5 × 104 B16F1 melanoma tumor cells given s.c., and the tumor area was monitored by the longest and widest caliper measurements after the onset of visible tumor. Animals were euthanized when tumor area exceeded 1.5 cm2; long-term survivors were tumor-free. Survival analysis was completed as in Figure 2.
improved efficacy associated with bivalent vaccine formulations when Ova peptide is presented on TMV. In order to broaden the application of bivalent vaccine formulations, we also tested similar combinations of helper and uptake antigens in combination with a melanoma self-antigen peptide, p15e (56). Vaccines were formulated and qualified essentially as described for Ova bivalent vaccines, using SDSPAGE and densitometry to gauge conjugation efficiency and MALDI-TOF mass analysis to confirm the presence of epitopes (data not shown). After three 15 µg biweekly s.c. vaccinations (∼900 ng of peptide) with 25 µg CpG adjuvant, mouse splenocytes were tested for IFNγ secretion by ELISpot. Functionality of the immune response was measured by monitoring survival after challenging vaccinated animals with B16 melanoma tumor. Interestingly, for p15e bivalent vaccinations, p15e::RDG stimulated the lowest number of IFNγ secreting cells, while p15e::E7 and p15e::TT bivalents stimulated high levels of peptide-activated cell numbers after 22 h stimulation with 10-5 M p15e peptide (Figure 4A). After s.c. challenge with a lethal dose of B16F1 melanoma tumor cells, there were one or two tumor-free long-term survivors in all vaccine groups, but surprisingly, none of the p15e vaccine formulations provided improved protection compared to the control (Figure 4B). In addition to testing bivalent formulations with helper or uptake peptides, we also tested the concept of delivering bivalent vaccines prepared from two different melanoma tumor-associated antigens, p15e and Trp 2 protein (56). Formulations of monovalent vaccines (p15e or Trp2 alone), mixtures of monovalent vaccines (p15e + Trp2), or bivalent conjugates (p15e::Trp2)
McCormick et al.
Figure 5. Qualification of TMV melanoma antigen conjugate vaccines. (A) SDS-PAGE analysis of SPDP conjugates of 15e or Trp2 peptides, and helper or uptake peptides, to TMV. Monovalent formulation (p15e or Trp2) or bivalent formulations (::) of p15e with Trp2 peptides combined are shown as final vaccine formulations, with approximately 45% p15e and 45% additional peptides in bivalent vaccines, compared to 90% conjugation of p15e or Trp2 alone. Single (*) and double (
