Vaccination with Human HER-2/neu (435-443) CTL Peptide Induces Effective Antitumor Immunity against HER-2/neu-Expressing Tumor Cells In vivo

CTLs recognize antigenic peptides, 8 to 12 amino acids long, presented to them by antigen-presenting cells in the context of MHC class I molecules (1). CTL peptides have been identified in several proteins, including those expressed by tumor cells (2). CTL sensitized in vitro against tumor-associated peptides have been shown to lyse cancer cell lines expressing both the appropriate MHC class I allele and the relevant peptide (3, 4).

HER-2/neu is a 185-kDa transmembrane glycoprotein with tyrosine-specific kinase activity that has a similarity in structure and sequence to the epidermal growth factor receptor (5). It is reportedly overexpressed, mostly via gene amplification, in a large proportion of aggressive cancers (6, 7). The HER-2/neu protein seems to be an ideal tumor-associated antigen for immunotherapy, because CTL responses specific for MHC class I epitopes have been observed in some cancer patients (3, 8). Moreover, tumor-reactive CTL and helper T-cell responses have been induced in vitro using various MHC class I and II binding synthetic peptides derived from the HER-2/neu sequence (9–17).

The HER-2/neu epitope spanning amino acids 435 to 443 [HER-2(9₄₃₅)] has been shown by us (17) and others (12) to bind with high affinity to HLA-A2.1 molecules and to elicit CTLs from tumor-associated lymphocytes of patients with ovarian cancer. We were also able to show increased frequencies of CTL precursors specific for HER-2(9₄₃₅) among peripheral blood mononuclear cells (PBMC) from patients with HER-2/neu⁺ tumors (13). PBMC from those patients showed specific lysis of autologous tumor cells, suggesting that HER-2(9₄₃₅) is naturally processed and expressed by HER-2/neu⁺ carcinomas. Thus far, no HER-2(9₄₃₅)-based vaccination studies have been reported in preclinical or clinical models.

As most tumor-associated antigens are self-antigens, the antigen-specific CTL repertoires may be significantly reduced during the processes of negative selection, leaving a T-cell repertoire that is poorly effective at mounting a productive antitumor response. Efforts to use unmodified forms of peptide immunogens to elicit antitumor responses in vivo have been largely unsuccessful, leading investigators to modify naturally occurring tumor antigens. Indeed, several peptide analogues capable of eliciting enhanced immunity to tumor-associated epitopes have been developed (18–20). In the present study, we have tested the capacity of hHER-2(9₄₃₅), which differs from its murine homologue by one amino acid substitution at position 4, to function both as a protective and as a therapeutic vaccine in HHD mice inoculated with transplantable ALC tumor cells. We have cotransfected ALC with HLA-A2.1 cDNA and hHER-2/neu cDNA (ALC.A2.1.hHER) to enable expression of hHER-2(9₄₃₅) epitope. Cotransfection of ALC was also done with a rHER-2/neu cDNA (ALC.A2.1.rHER) because (a) rHER-2/neu differs from mHER-2/neu in <6% of the amino residues (21) and (b) rHER-2(9₄₃₅) and mHER-2(9₄₃₅) nonamers have identical sequences (http://ncbi.nlm.nih.gov/BLAST). hHER-2(9₄₃₅) and mHER-2(9₄₃₅) are predicted to bind to HLA-A2.1 or H-2K^b molecules with equivalent efficacy [see ref. 22 and SYFPEITHI (http://www.uni-tuebingen.de/uni.kxi.database.html)]. Therefore, we also tested the potency of these two peptide vaccines in C57BL/6 mice inoculated with HER-2/neu (human or rat) single transfectants of ALC (ALC.hHER and ALC.rHER, respectively). We found that both HHD and C57BL/6 mice developed significant antitumor responses when prophylactically immunized with hHER-2(9₄₃₅). Moreover, treatment with hHER-2(9₄₃₅) of established syngeneic ALC.A2.1.rHER tumors showed an impressive therapeutic effect. Finally, CD8⁺ T cells from mice immunized with hHER-2(9₄₃₅) when adoptively transferred in naive syngeneic recipients highly protected them against the growth of ALC transfectants. Our data provide the basis for clinical use of xenogeneic HER-2/neu CTL epitopes in vaccination protocols.

Animals. HHD mice are β_2m^−/−,D^b−/− and express a HLA-A*0201 monochain composed of a chimeric heavy chain (α1 and α2 domains of HLA-A*0201 and the α3 intracellular domain of D^b) linked by its NH₂ terminus to the COOH terminus of the human β_2m (23). These mice were generated by crossing HHD transgenic C57BL/6xSJL mice with H-2b^−/−β_2m^−/− C57BL/6 double-knockout mice. HHD mice were kindly provided by Prof. Francois Lemonnier (Unite d'Immunite Cellulaire Antivirale, Institut Pasteur, Paris, France).

All mice were maintained in pathogen-free conditions in the animal facilities of our center; all protocols were reviewed by the St. Savas Cancer Hospital competent authority for compliance with the Greek and European regulations on Animal Welfare and with Public Health Service recommendations.

Synthetic peptides. The mHER-2(9₄₃₅) (ILHDGAYSL) and hHER-2(9₄₃₅) (ILHNGAYSL) were synthesized by standard solid-phase chemistry on a multiple peptide synthesizer and analyzed by mass spectrometry. The same approach was also applied for the human gp100-derived peptide hgp(9₁₅₄), which was used as a control. All peptides were >95% pure as indicated by analytic high-performance liquid chromatography. Lyophilized peptides were stored at −20°C.

Transfection of ALC and SKOV3 cell lines. The parental SKOV3 cell line was kindly donated by Dr. C.G. Ioannides (Department of Gynecologic Oncology and Immunology, M. D. Anderson Cancer Center, Houston, TX). The murine lymphoma cell line ALC (kindly provided by Dr. R. Kiessling, Microbiology and Tumor Biology Center, Karolinska Institute, Stockholm, Sweden) was grown in vivo as ascites by serial passages in C57BL/6 syngeneic mice. Both cell lines were maintained in vitro in RPMI 1640 containing 10% fetal bovine serum (FBS), 50 μmol/L 2-mercaptoethanol, 2 mmol/L l-glutamine, 100 units/mL penicillin, and 100 μg/mL streptomycin (all purchased from Life Technologies, Gaithersburg, MD). To transfect the ALC cell line, we used the Amaxa Nucleofection System (Amaxa Biosystems, Koln, Germany) according to the manufacturer's instructions. Cells were tested at 48 hours post-transfection by fluorescence-activated cell sorting analysis (FACScalibur, BD Biosciences, Mountain View, CA) using monoclonal antibody (mAb) 7.16.4 (Oncogene Research Products, Cambridge, MA) for hHER-2/neu and rHER-2/neu expression. The mAb BB7.2 (a kind gift from Prof. H.G. Rammensee, Department of Immunology, University of Tuebingen, Tuebingen, Germany) was used to detect HLA-A2.1 expression. Cells were stained in a standard indirect immunofluorescence procedure with primary antibody followed by FITC-conjugated anti-mouse IgG (Dako, Glostrup, Denmark). Cells were subsequently cloned by limiting dilution in the presence of selective medium (1,000 μg/mL neomycin, Sigma-Aldrich, St. Louis, MO). Transfections were done with pCMV-c-erbB-2 constructs encoding full-length hHER-2/neu cDNA (provided by Dr. Wei, Department Immunology and Microbiology, Karmanos Cancer institute, Wayne State University, Detroit, MI) or full-length rHER-2/neu cDNA (provided by Dr. Ostrand-Rosenberg, Department Biological Sciences, University of Maryland, Baltimore, MD) and the pcDNA3.1-A2-neo^R construct (provided by Prof. Jotereau, Institut de Biologie, Nantes, France). Following this procedure, stable clones were established expressing either HER-2/neu or HLA-A2.1. To establish cells expressing both antigens, we used stable clones expressing either hHER-2/neu or rHER-2/neu protein, transfected them with the HLA-A2.1 cDNA following the same nucleofection procedure, and at 48 hours post-transfection sorted the double-positive cells on an ALTRA cell sorter (Coulter, Hialeah, FL). The sorted cells were cloned in the presence of selective medium (RPMI 1640 containing 10% FBS and 1,000 μg/mL neomycin) and double-positive clones were established (Fig. 1A and B). To transfect the SKOV3 cell line with pcDNA3.1-A2-neo^R, we used the LipofectAMINE 2000 (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. Cells were tested for HLA-A2.1 expression 48 hours post-transfection and positive clones were established after cloning.

Immunization protocol. Mice were immunized thrice every 5 days (starting on day 0) by s.c. injections at the base of the tail with 100 μg peptide emulsified in 200 μL incomplete Freund's adjuvant (IFA). During this period (i.e., 10 days), murine recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF; Endogen, Boston, MA) was given i.p. at 100 ng in 0.3 mL PBS every other day (five injections in total).

Protective tumor model. HHD or C57BL/6 mice were immunized with the appropriate peptides along with IFA and GM-CSF as above. One day following the last injection (i.e., day 11), mice were inoculated by s.c. injection on the right flank with 2 × 10⁴ transfected syngeneic ALC tumor cells in 200 μL PBS. Total observation was >150 days. The observation was terminated with the euthanasia of mice when the tumor mass grew up to 200 to 250 mm² mean diameter. Every group consisted of 10 mice and the experiments were repeated twice (Fig. 3).

Therapeutic tumor model. HHD mice were inoculated with syngeneic ALC.A2.1.rHER tumor cells (2 × 10⁴, in 200 μL PBS injected s.c. in the right flank). Once tumors of ∼10 to 15 mm² (palpable tumors) were established (days 22-26), peptides in IFA were given along with GM-CSF as described in detail above. Tumor size was monitored regularly every 4 days and was expressed as the product (in mm²) of the perpendicular diameters of individual tumors. Two experiments with five mice per group were done.

Adoptive cellular therapy model. HHD or C57BL/6 mice were immunized with peptides as described above. One day after the end of the immunization protocol, CD8⁺ T cells were isolated from total immune splenocytes by negative selection using the CD8⁺ T-cell isolation kit (Miltenyi Biotec, Berghisch Gladbach, Germany). Freshly isolated CD8⁺ HHD splenocytes were expanded in vitro, as described below, tested for specificity in cytotoxicity assays, and then given i.p. (4 × 10⁷ cells in 0.5 mL PBS) in syngeneic mice carrying palpable ALC tumors transfected with rHER-2/neu. Tumor growth was monitored as described above. Each group consisted of five mice and experiments were done twice.

CTL expansion procedures. For adoptive transfer experiments, CTLs from immunized animals were expanded in cultures following a method described by Riddell et al. (24) with some modifications. A total of 1 × 10⁵ CD8⁺ cells were resuspended in 25 mL complete medium with 25 × 10⁶ peptide-pulsed, syngeneic, irradiated (3,000 rads) splenocytes. One day after initiating the cultures, 100 IU/mL recombinant interleukin (IL)-2 (Proleukin; Chiron BV, Amsterdam, the Netherlands) was added to the medium. Usually on days 10 to 12, cultures were split because T-cell concentration reached numbers >1.5 × 10⁶/mL. After this time point, cultures were restimulated with IL-2 and 30 ng/mL anti-CD3 mAb (CD3 ε-chain, clone 145-2C11; BD PharMingen, San Diego, CA) to achieve higher expansion. On average, ∼4 × 10⁷ to 6 × 10⁷ CTL were obtained by days 15 to 18. Before transferring into syngeneic HHD mice, CTLs were tested for specificity against various appropriate tumor targets or peptide-pulsed T2 targets.

Cytotoxicity assay. The cytotoxicity assay was done as described previously (25). Briefly, effector CTLs (1 × 10⁶/mL) in 100 μL medium were placed in 96-well V-bottomed plates (Costar, Cambridge, MA). As targets, tumor cell lines were labeled with sodium chromate (Radiochemical Centre, Amersham, the Netherlands; 100-200 μCi isotopes per 1 × 10⁶-2 × 10⁶ target cells) and added to effectors at the indicated E:T ratios. For peptide recognition, T2 cells were incubated overnight at 26°C together with 20 μg/mL peptide, washed, and then labeled. Incubation was done for 6 hours in CO₂ incubators. Supernatants (100 μL) were collected from each well (all determinations done in triplicates) and radioactivity was measured in a gamma counter (Packard, Downers Grove, IL). The percentage of cytotoxicity was calculated according to the formula: % lysis = 100 × (test ⁵¹Cr release − spontaneous ⁵¹Cr release) / (maximum ⁵¹Cr release − spontaneous ⁵¹Cr release). Cytotoxicity values were considered to indicate significant recognition of a target when the differences between mean values (from triplicate analyses) for % lysis of the particular target (e.g., pulsed T2 cells or transfected tumor cell lines) and unloaded T2 cells (or nontransfected tumor cell lines) were ≥10% at an E:T ratio of 20:1 and statistically significant (P < 0.05). Wherever indicated, mAb were added throughout the 6 hours incubation period (at 10 μg/mL final concentration).

T2 binding assay. Each peptide was tested for concentration-dependent binding to T2 cells in a HLA-A* 0201 stabilization assay (26). T2 (TAP)–deficient cells, which have been incubated at room temperature the previous night to increase cell-surface MHC class I molecule expression, were then incubated overnight without or with each peptide over a range of peptide concentrations from 1 nmol/L to 10 μmol/L. Stability of HLA-A* 0201 was assayed by flow cytometry after staining the cells with the BB7.2 anti-HLA-A2 mAb and phycoerythrin (PE)–conjugated anti-mouse immunoglobulins (Dako). The HLA-A*0201 strongly binding Flu matrix peptide (amino acids 58-66; EZBiolab, Inc. Westfield, IN) was used as a positive control. Results are presented as fold of increase of mean fluorescence intensities (MFI) in the presence of peptides relative to MFI without peptides.

Measurement of peptide/HLA-A*0201 complex stability. This was done as described (27). Briefly, T2 cells were incubated overnight at 37°C without peptides or with 1 μmol/L of each peptide in serum-free RPMI 1640 supplemented with β2-microglobulin at 100 ng/mL. Next, cells were incubated with brefeldin A (10 μg/mL; Sigma-Aldrich) for 1 hour to block cell-surface expression of newly synthesized HLA-A*0201 molecules, washed, and further incubated for 0, 2, 4, and 6 hours. Subsequently, cells were stained with mAb BB7.2 followed by anti-mouse PE. MFI measured at 0 hours was considered as 100%. MFIs at all other time points are expressed relative to MFI at 0 hours and calculated as follows: [MFI (0 hour) − MFI (2, 4, or 6 hours) / MFI (0 hour)] × 100.

ELISpot assay. Splenocytes from every single immunized mouse were plated at 500,000 per well in quadruplicates in 96-well flat-bottomed plates. Irradiated (3,000 rads) HHD splenocytes pulsed with 500 nmol/L mHER-2(9₄₃₅) were added to the syngeneic responder splenocytes at a cell ratio of 1:1 in a total volume of 200 μL/well RPMI 1640 supplemented with 10% FBS, murine recombinant IL-7 (10 ng/mL), and murine recombinant IL-12 (100 pg/mL; both purchased from R&D Systems, Abington, United Kingdom). Control cultures were stimulated with syngeneic, irradiated splenocytes pulsed with 500 nmol/L hgp(9₁₅₄). Cultures were incubated at 37°C in a CO₂ incubator. On day 3, 100 μL culture supernatant was replenished with fresh medium supplemented with 20 ng/mL IL-7 and 200 pg/mL IL-12. Growing microcultures were restimulated on day 7 with T2 cells pulsed with 500 nmol/L of either mHER-2(9₄₃₅) or hgp(9₁₅₄) in control cultures. Twenty-four hours later, IFN-γ production was estimated using the IFNγ ELISpot Assay kit (BD PharMingen) according to the manufacturer's protocol. Spots were counted under a stereomicroscope (Zeiss, Jena, Germany) using the Image ProPlus software (Media Cybernetics, San Diego, CA). Specific spots were calculated by subtracting the mean number of spots obtained from the control cultures.

Statistics. Tumor sizes among the respective groups were compared with Wilcoxon's rank-sum test. Kaplan-Meier curves were plotted for survival analyses. All Ps were two tailed and considered significant at P < 0.05. Significance of differences among groups in the cytotoxicity and ELISpot assays was assessed with the two-tailed Student's t test. Statistically significant differences were considered at P < 0.05.

ALC tumor cells transfected to coexpress HLA-A*0201 and hHER-2/neu or rHER-2/neu, process, and present HLA-A*0201-restricted HER-2/neu peptides. The capacity of hHER-2(9₄₃₅) and mHER-2(9₄₃₅) to bind to HLA-A*0201 and to form stable peptide/HLA-A*0201 complexes was tested in standard T2 binding and MHC peptide stability assays, respectively. As shown in Fig. 1A and B, both peptides exhibited a comparable strong binding and stabilization capacity.

To evaluate the ability of HLA-A*0201-restricted, HER-2/neu-derived epitopes to control tumor growth, we first cotransfected the ALC leukemic cell line (C57BL/6 origin) with the genes for the MHC class I HLA-A2.1 molecule and the full-length hHER-2/neu or rHER-2/neu. Flow cytometric evaluation of HER-2/neu and HLA-A2.1 expression on a stable double transfectant (called ALC.A2.1.hHER or ALC.A2.1.rHER) maintained under neomycin selection is shown in Fig. 1C and D. HLA-A2.1 and HER-2/neu expression by the transfected ALC remained constant after 8 to 9 weeks in culture without neomycin selection (data not shown), the anticipated duration of in vivo tumor growth in naive animals.

We next examined whether the ALC expressing both HLA-A2.1 and HER-2/neu could function as a target recognized by HER-2/neu peptide-specific effectors. To address this, we used a CTL clone, which was elicited previously from an ovarian patient's T lymphocytes with pooled peptide extracts from HLA-A2.1⁺, HER-2/neu-overexpressing autologous tumor cells (17). This particular clone was shown to specifically lyse T2 cells pulsed with peptide hHER-2(9₄₃₅) as well as the HLA-A2.1 transfectant of the HER-2/neu-overexpressing SKOV3 human ovarian cell line (SKOV3.A2), which naturally expresses this epitope (17, 28). As shown in Fig. 1E, ALC.A2.1.hHER targets, but not the wild-type ALC or the ALC.A2.1 and ALC.hHER single transfectants, were lysed by our CTL clone. In addition, as shown in Fig. 1F, the same CTL clone also lysed T2 cells loaded with the mHER-2(9₄₃₅) as well as the ALC.A2.1.rHER although to a lesser extent compared with the cytotoxicity levels against T2 loaded with hHER(9₄₃₅) and ALC.A2.1.hHER, respectively. Such differences could be attributed to the one amino acid difference at position 4 between hHER-2(9₄₃₅) (ILHNGAYSL) and m,rHER-2(9₄₃₅) (ILHDGAYSL), which may cause less effective recognition of the respective targets by our CTL clone. These data suggested that our ALC double transfectants stably expressed the HLA-A2.1-restricted human or rat CTL epitope HER-2(9₄₃₅), making these cell lines suitable for evaluating the antitumor efficacy of immune responses to this antigen.

To evaluate the effect of HLA-A2.1 and hHER-2/neu or rHER-2/neu expression in ALC on its immunogenicity and tumorigenicity, we compared the growth of wild-type or single-transfected ALC and double-transfected ALC in naive C57BL/6 and HHD mice. ALC.A2.1.hHER and ALC.A2.1.rHER as well as ALC.A2.1 were rejected by normal C57BL/6 mice as a xenogeneic target consistent with the expression of HLA-A2.1 (Fig. 2A). In contrast, ALC.hHER and ALC.rHER as well as the wild-type ALC [Fig. 2A; or a mock transfectant of ALC (data not shown)] grew and killed all C57BL/6 animals. Apparently, the expression of hHER-2/neu or rHER-2/neu, both showing great homology with mHER-2/neu, was insufficient to induce any antitumor responses. Moreover, the growth kinetics of double-transfected ALC in HHD mice were identical with those of wild-type ALC (or ALC-mock; data not shown) growing in either C57BL/6 or HHD mice (Fig. 2A and B). These results showed that neither HLA-A2.1 nor HER-2/neu (or HER-2/neu peptide antigens presented by HLA-A2.1) function as effective rejection antigens in unimmunized HHD mice. Therefore, changes in kinetics of tumor progression on immunization with a HLA-A*0201-restricted HER-2/neu peptide should be interpreted as an antitumor response directed toward this particular peptide presented on ALC in the context of HLA-A2.1 molecules.

Active immunization with hHER-2(9₄₃₅) in the presence of GM-CSF delays tumor outgrowth. We next sought to investigate whether active immunization against these HLA-A2.1-restricted peptides could augment protection against the growth of ALC double transfectants in HHD mice. No protection against ALC.A2.1.hHER was observed when HHD mice were immunized thrice with mHER-2(9₄₃₅) in IFA alone. In contrast, immunization with hHER-2(9₄₃₅) induced a low, although significant, effect (mice survived up to day 95 versus day 65 in the other groups; P < 0.05; Fig. 3A). By coinjecting HHD mice with GM-CSF, the peptide-induced antitumor effect was greatly enhanced: 50% of mice injected with hHER-2(9₄₃₅) survived up to day 140, whereas in total from three independently done experiments there were 10 of 30 (33%) long-term survivors (Fig. 3B). The latter resisted rechallenge with ALC.A2.1.hHER but not with wild-type ALC or single-transfected ALC (data not shown). Combined treatment with mHER-2(9₄₃₅) plus GM-CSF resulted also in a substantial delay of ALC.A2.1.hHER outgrowth, which however was inferior compared with that induced by hHER-2(9₄₃₅): there were no long-term survivors and 50% survival was already achieved by day 85 (P < 0.01; Fig. 3B). Injections with GM-CSF alone or with an irrelevant HLA-A2.1-restricted peptide [i.e., melanoma hgp(9₁₅₄) peptide] plus GM-CSF did not have any effect (Fig. 3B). No antitumor effects could be detected on peptide immunization against the growth of wild-type ALC or ALC.A2.1 (data not shown).

More importantly, immunization with hHER-2(9₄₃₅) plus GM-CSF greatly enhanced in vivo immunity also against the ALC.A2.1.rHER: 8 of 30 (26.6%) mice became long-term survivors, whereas 50% of them survived up to day 120 (pooled results from three independently done experiments; see also Fig. 3C). Mice that were fully protected against ALC.A2.1.rHER growth also resisted rechallenge with this tumor (data not shown). In contrast, in the group of mice immunized with mHER-2(9₄₃₅), there were no long-term survivors and half of the mice survived up to day 90 [P < 0.01, compared with vaccination with hHER-2(9₄₃₅); Fig. 3C].

To further substantiate the observed difference in the potency of the two vaccines [i.e., hHER-2(9₄₃₅) versus mHER-2(9₄₃₅)], we compared ALC growth in C57BL/6 mice actively immunized with either of the peptides. Also in this model, hHER-2(9₄₃₅) induced significantly higher protection compared with mHER-2(9₄₃₅). The results from this type of experiments, which were repeated twice, showed that vaccination with hHER-2(9₄₃₅) induced long-term survivors in mice inoculated with either ALC.hHER [7 of 30 (23.3%); Fig. 3D] or ALC.rHER [9 of 30 (30%); Fig. 3E], whereas in both groups 50% of the animals survived up to day 140. In contrast, on immunization with mHER-2(9₄₃₅), 50% of the mice survived up to day 80, although there were no long-term survivors [Fig. 3D and E; P < 0.001, compared with vaccination with hHER-2(9₄₃₅)].

Treatment of established tumors by immunization with HER-2/neu peptides plus GM-CSF. To test the efficacy of peptides in the potential treatment of established tumors, we inoculated HHD mice with ALC.A2.1.rHER tumor cells, and once tumors were palpable (∼22-26 days after administration), peptides were injected with rat GM-CSF. As shown in Fig. 4C, when hHER-2(9₄₃₅) was given, three mice were totally cured, and in the remainders (n = 7), tumor growth was greatly delayed. A delay in tumor growth was also observed when immunizations were done with mHER-2(9₄₃₅) (Fig. 4B), which however was significantly less pronounced (P < 0.01). Nontreated mice (data not shown) or mice injected with control peptide hgp(9₁₅₄) (Fig. 4A) were sacrificed by euthanasia already by days 60 to 65, at which time point their tumors had already grown >200 mm² mean diameter.

Therapeutic immunity induced by peptide immunization is mediated by CD8⁺ CTL. To confirm that CD8⁺ CTLs were responsible for in vivo therapeutic protection against the ALC transfectants tumors, we studied the ability of adoptively transferred CD8⁺ T lymphocytes from immunized mice to cause tumor regression in syngeneic, naive hosts, carrying established tumors. CD8⁺ T cells were first expanded with IL-2 and anti-CD3-mAb and then tested for cytotoxicity against the transfected ALC and SKOV3 targets before adoptive transfer. Lysis of ALC.A2.1.rHER and SKOV3.A2 was more profound with CD8⁺ CTL from HHD mice immunized with hHER-2(9₄₃₅) (HHDCD8-hHER₄₃₅) as opposed to killing with HHDCD8-mHER₄₃₅ (P < 0.001; Fig. 5A). Similarly, CD8 CTL from C57BL/6 immunized with hHER-2(9₄₃₅) (C57BL/6CD8-hHER₄₃₅) mice exhibited significantly increased lytic activity against ALC.rHER targets compared with syngeneic C57BL/6CD8-mHER₄₃₅ (P < 0.001; Fig. 5B). Cytotoxicity against the ALC.A2.1.rHER or SKOV3.A2 targets was abolished in the presence of anti-HLA-A2.1 mAb, whereas lysis of ALC.rHER was abolished by anti-H-2K^b mAb (data not shown).

In response to hHER-2(9₄₃₅), HHD-derived CTL with profoundly higher functional activity compared with mHER-2(9₄₃₅)-primed HHD mice were induced in all five HHD mice tested. Thus, HHDCD8-hHER₄₃₅ lysed T2 targets coated with 10 nmol/L mHER-2(9₄₃₅) at 12.4 ± 5.22%, which is not statistically different compared with 11.5 ± 5.0% lysis of T2 pulsed with 100 nmol/L of this peptide by HHDCD8-mHER₄₃₅ effectors (Fig. 5C and D). The increased avidity of HHDCD8-hHER₄₃₅ versus HHDCD8-mHER₄₃₅ for mHER-2(9₄₃₅) was evident at each peptide concentration used for coating T2 targets: 18 ± 2.96% versus 11.5 ± 5.0% (P < 0.05), 12.4 ± 5.22% versus 5.8 ± 3.9% (P < 0.005), and 1.33 ± 0.57% versus 0.46 ± 0.11% (P < 0.05) at 100, 10, and 1 nmol/L, respectively [pooled data (n = 5) from the individual results shown in Fig. 5C and D]. Pulsing of T2 cells with doses of peptide higher than 100 nmol/L (i.e., 1,000 nmol/L) was inappropriate to uncover differences in the avidity between these two CTL effectors (comparable levels of lysis were achieved; Fig. 5C and D).

When adoptively transferred, HHDCD8-hHER₄₃₅ cells efficiently protected all syngeneic mice with established ALC.A2.1.rHER tumors: 2 mice became long-term survivors, and in the remainders (n = 8), tumor size reached 200 to 250 mm² between days 100 and 130 (Fig. 5E). In contrast, only 4 of 10 mice treated with HHDCD8-mHER₄₃₅ cells were protected. In these mice, tumor growth reached 200 to 250 mm² between days 70 and 90 (Fig. 5F; P < 0.001). No protection was achieved with CD8⁺ T cells from HHD mice immunized under the same protocol but with the control hgp(9₁₅₄) peptide (tumor size of 200-250 mm² was already achieved by days 50-55; data not shown). Adoptively transferred C57BL/6CD8-hHER₄₃₅ induced a high degree of protection in syngeneic hosts inoculated with ALC.rHER. In all 10 mice, tumor growth reached an approximate size of 225 to 245 mm² between days 95 and 130 (Fig. 5G). However, C57BL/6CD8-mHER₄₃₅ were less efficient, protecting only 30% of mice to a significantly lesser extent: tumor size reached 200 to 225 mm² between days 85 and 100 (Fig. 5H; P < 0.05). C57BL/6 mice treated with control CTL recognizing hgp(9₁₅₄) were sacrificed by euthanasia by days 45 to 50 (data not shown).

Immunization with hHER-2(9₄₃₅) increases the frequency of CTL precursors recognizing mHER-2(9₄₃₅). We next tested the capacity of the two vaccines to modulate the frequency of CTL specific for self-mHER-2(9₄₃₅) in HHD mice. After immunizing HHD mice (n = 11) with either of the peptides along with GM-CSF, as above, the freshly isolated murine splenocytes were cultured for 7 days in the presence of syngeneic APC pulsed with mHER-2(9₄₃₅) and then tested, in the presence of T2 cells pulsed with mHER-2(9₄₃₅), in the IFN-γ-based ELISpot assay. As shown in Fig. 6, the mean mHER-2(9₄₃₅)-specific CTL precursors in HHD mice immunized with hHER-2(9₄₃₅) was 39.45 ± 15.73 (range, 19-65) compared with 16.18 ± 6.67 (range, 8-30) in HHD mice immunized with mHER-2(9₄₃₅) (P < 0.01) and 6.18 ± 2.4 (range, 2-10) in nonimmunized mice (P < 0.001). All mice tested were equally immunocompetent, as they perform high proliferative responses to concanavalin A (data not shown). In addition, CTL precursor frequencies to the control peptide hgp(9₁₅₄) were negligible in all three groups (data non shown).

In the present study, we report the development of a preclinical model based on human HLA-A2.1 transgenic HHD mice and a transplantable ALC tumor cell line transfected to coexpress HLA-A*0201 and hHER-2/neu or rHER-2/neu molecules. This model can be used to evaluate candidate antigens to control tumor outgrowth and immunization strategies for human use. We used this model to estimate the significance of a HLA-A2.1-restricted HER-2/neu CTL peptide epitope in active immunization studies. hHER-2(9₄₃₅) is important not only for its potential use in the immunotherapy of hHER-2/neu-overexpressing tumors but also because HER-2/neu is a self-protein and thus is subjected to self-tolerance. hHER-2(9₄₃₅) has been recently shown by us (13, 17, 24) and others (12) to represent a naturally processed and presented epitope on the surface of primary tumor cells or tumor cell lines constitutively expressing HER-2/neu, such as SKOV3.A2 and SW-626.

As a basis for studies of antitumor immunization efficacy, we created an ALC (H-2^b) double transfectant. These cells express HER-2/neu and HLA-A2.1 and are efficiently lysed by a CTL clone specifically recognizing HER-2(9₄₃₅) in the context of HLA-A2.1 (17, 28). The growth of ALC transfectants in HHD mice is identical with that of parental ALC in HHD or in C57BL/6 wild-type mice. These data indicate that the expression of both HLA-A2.1 and HER-2/neu (human or rat) molecules in ALC does not create significantly immunogenic antigens, which is in keeping with its generally poor immunogenicity. The advantage of this is that it allows evaluation of immune responses to specific antigens in the absence of such responses to the parental tumor. Additionally, considering the fact that (a) rHER-2/neu shows a great homology (>94%) to mHER-2/neu (21) and (b) mHER-2(9₄₃₅) is identical to rHER-2(9₄₃₅) in amino acid composition and sequence, it is fair to conclude that the rHER-2/neu transfectants of ALC would behave similarly to ALC transfected with a mHER-2/neu plasmid. We would thus anticipate that any results obtained on active immunization with hHER-2(9₄₃₅) or mHER-2(9₄₃₅) against ALC.A2.1.rHER should also be valid for ALC cells cotransfected with HLA-A2.1 and mHER-2/neu. The fast growth kinetics of ALC (inoculation of as few as 2 × 10⁴ cells create palpable tumors by days 22-26) render complete control of this tumor a difficult task. In spite of this, we observed complete cures in ∼30% of the mice inoculated with ALC tumor cells. Besides the percentages of long-term survivors, we interpret shifts in tumor outgrowth as indicators of successful immune-mediated intervention.

In this study, we show that vaccination with mHER-2(9₄₃₅) generates T cells with low avidity for this self-peptide resulting in weak immunity against ALC tumor cells expressing hHER-2/neu or rHER-2/neu. In contrast, the human variant of this peptide is a more potent immunogen acting as a better protective and therapeutic vaccine. hHER-2(9₄₃₅)-induced CTL from all HHD mice cross-recognized the mHER-2(9₄₃₅) self-peptide and T2 cells pulsed with mHER-2(9₄₃₅) were lysed more efficiently by HHDCD8-hHER₄₃₅ effectors compared with HHDCD8-mHER₄₃₅ effectors. In functional assays, we have found that both peptides bind with similar affinities to HLA-A2.1 molecules, an observation that is in agreement with SYFPEITHI [calculates binding scores 28 and 27 for mHER-2(9₄₃₅) and hHER-2(9₄₃₅), respectively]. These data suggest that immunization with the hHER-2(9₄₃₅) peptide plus GM-CSF generates CTL lines that recognize mHER-2(9₄₃₅) with higher avidity than their mHER-2(9₄₃₅) plus GM-CSF-induced counterparts. To this end, we showed that CD8⁺ CTL from HHD or C57BL/6 mice immunized with hHER-2(9₄₃₅) plus GM-CSF induce significantly higher levels of cytotoxicity against ALC.A2.1.rHER or ALC.rHER targets, respectively, compared with mHER-2(9₄₃₅)-induced CTL effectors. In agreement with the superior in vitro recognition of the transfected ALC tumors, the hHER-2(9₄₃₅)-specific CTL on adoptive transfer effectively controlled tumor growth in mice. Importantly, HHD mice which rejected HLA-A2.1⁺ HER-2/neu⁺ ALC tumors on vaccination with hHER-2(9₄₃₅) became resistant to a second challenge with the same tumor cells but not to wild-type ALC or HLA-A2.1 single transfectants of ALC lacking the expression of HER-2/neu showing the capacity of the vaccine to induce HER-2/neu peptide-specific memory T cells.

Based on transgenic mouse models, it is now clear that tolerance is capable of deleting or inactivating reactive high-avidity T cells against the transgene, thereby leading to self-tolerance (29). A typical example for this is tolerance induced to HER-2/neu in mice where the rat neu proto-oncogene is expressed in the mammary tissue under the control of the mouse mammary tumor virus promoter (30, 31). However, low-avidity self-specific T cells can be isolated from tolerant hosts (32–35) and there are reports that such cells can be activated, expanded, and involved in antitumor responses (32, 33). To this end, Lustgarten et al. (36) showed that multiple injections of dendritic cells pulsed with HER-2(9₃₆₉) or HER-2(9₇₇₃) in neuxA2.1 double transgenic mice stimulated low-avidity peptide-specific CD8⁺ CTL, which were capable of mediating antitumor activity in vivo. In agreement with this, we vaccinated our mice thrice with the HER-2/neu peptides to achieve tumor retardation. In other experimental models (37, 38), tolerance to self-melanocyte differentiation proteins tyrosinase and gp100 was successfully broken by immunization with dendritic cells pulsed with tyrosinase or gp100 peptide ligands containing conservative substitutions that were cross-reactive with the original antigens. Altered peptide ligands have been used to improve the reactivity of T cells specific for self/tumor antigens in mouse and human (18–21, 39–44). For melanocyte differentiation antigens, H-2D^d-restricted CD8⁺ T cells capable of recognizing B16 melanoma could only be elicited when the altered peptide was used (42). Recently, hgp100(9₂₅) that represents an altered peptide ligand form of mgp100(9₂₅) has been successfully used to induce in vivo tumor regression in combination with adoptively transferred hgp100(9₂₅)-specific T cells and IL-2 (43). The enhanced antitumor activity, induced on immunization with the hHER-2(9₄₃₅) vaccine in our model, is consistent with the generation of higher numbers of murine CTLs with increased avidity for mHER-2(9₄₃₅). Whether vaccination of mice with this hHER-2/neu peptide breaks tolerance to self-HER-2/neu will be definitively tested in HER-2/neu transgenic mice, which is currently under investigation in our laboratory.

An important issue for improvement of cancer immunotherapy is the identification of immunogenic tumor-associated peptide epitopes. HHD mice and the ALC.A2.1.HER tumor cell line constitute useful tools for rapid evaluation and optimization of such immunogenic epitopes in a context that would have relevance to therapy of HER-2/neu-overexpressing human cancers. In the present study, we used this preclinical model to show that repeated injections with mHER-2(9₄₃₅) and GM-CSF delayed outgrowth of ALC double transfectants expressing hHER-2/neu or rHER-2/neu. Furthermore, when the same protocol is applied but using the xenogeneic hHER-2(9₄₃₅), which contains a single amino acid substitution at position 4, as an altered peptide ligand, we could show a considerably enhanced antitumor effect. These data suggest that vaccination of cancer patients with similar, xenogeneic tumor peptides may be considerably more effective compared with self-tumor peptides.

Grant support: PENED 01ED55 and EPAN YB/3 (C.N. Baxevanis) and Regional Operational Program Attika no.20 MIS code 59605 G.R. (M. Papamichail).

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 with 18 U.S.C. Section 1734 solely to indicate this fact.