Evaluation of the HER2/neu-derived peptide GP2 for use in a peptide-based breast cancer vaccine trial

Authors

  • Elizabeth A. Mittendorf M.D.,

    1. Clinical Breast Care Project, Department of Surgery, Walter Reed Army Medical Center, Washington, DC
    2. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • Catherine E. Storrer B.S.,

    1. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • Rebecca J. Foley B.S.,

    1. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • Katie Harris B.S.,

    1. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • Yusuf Jama M.S.,

    1. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • Craig D. Shriver M.D.,

    1. Clinical Breast Care Project, Department of Surgery, Walter Reed Army Medical Center, Washington, DC
    Search for more papers by this author
  • Sathibalan Ponniah Ph.D.,

    1. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    Search for more papers by this author
  • George E. Peoples M.D.

    Corresponding author
    1. Clinical Breast Care Project, Department of Surgery, Walter Reed Army Medical Center, Washington, DC
    2. Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, Bethesda, Maryland
    • Clinical Breast Care Project, Immunology and Research Center, Henry M. Jackson Foundation, National Naval Medical Center, 8901 Wisconsin Avenue, Ci3CP 1RC Bldg. 139, Bethesda, MD 20889-5600===

    Search for more papers by this author
    • Fax: (301) 493-6840


  • This article is a US Government work and, as such, is in the public domain in the United States of America.

  • The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the views of the Department of the Army or the Department of Defense.

Abstract

BACKGROUND

E75 and GP2 are human leukocyte antigen (HLA)-A2-restricted immunogenic peptides derived from the HER2/neu protein. In a E75 peptide-based vaccine trial, preexisting immunity and epitope spreading to GP2 was detected. The purpose of this study was to further investigate GP2 for potential use in vaccination strategies. Importantly, a naturally occurring polymorphism (I → V at position 2, 2VGP2) associated with increased breast cancer risk was addressed.

METHODS

Prevaccination peripheral blood samples (PBMC) from HLA-A2 breast cancer patients and CD8+ T cells from HLA-A2 healthy donors were stimulated with autologous dendritic cells (DC) pulsed with GP2 and tested in standard cytotoxicity assays with HER2/neu+ tumor cells or GP2- or 2VGP2-loaded T2 targets. Additional cytotoxicity experiments used effectors stimulated with DC pulsed with E75, GP2, or the combination of E75+GP2.

RESULTS

GP2-stimulated prevaccination PBMC from 28 patients demonstrated killing of MCF-7, SKOV3-A2, and the HLA-A2 control target SKOV3 of 28.8 ± 3.7% (P<.01), 29.5 ± 4.0% (P<.01), and 16.9 ± 2.7%, respectively. When compared with E75, GP2-stimulated CD8+ T cells lysed HER2/neu+ targets at 43.8 ± 5.2% versus 44.2 ± 5.7% for E75 (P = .87). When combined, an additive effect was noted with 58.6 ± 5.4% lysis (P = .05). GP2-stimulated CD8+ T cells specifically recognized both GP2-loaded (19.6 ± 5.7%) and 2VGP2-loaded T2 targets (17.7 ± 5.2%).

CONCLUSIONS

GP2 is a clinically relevant HER2/neu-derived peptide with immunogenicity comparable to that of E75. Importantly, GP2-specific effectors recognize 2VGP2-expressing targets; therefore, a GP2 vaccine should be effective in patients carrying this polymorphism. GP2 may be most beneficial used in a multiepitope vaccine. Cancer 2006. Published 2006 by the American Cancer Society.

The identification of tumor-associated antigens (TAA) that can be recognized by immune effectors has generated increased interest in the development of cancer vaccines. Peptide-based vaccines attempt to utilize antigenic epitopes derived from TAA for the induction of peptide-specific cytotoxic T lymphocytes (CTL) that can recognize and lyse tumor cells expressing these immunogenic peptides complexed with MHC Class I molecules on their cell surface.1 One such example of a TAA is HER2/neu, which is overexpressed in a variety of epithelial malignancies including approximately 20% to 30% of breast cancers.2 HER2/neu is the source of several immunogenic peptides including E75 and GP2, which are human leukocyte antigen (HLA)-A2-restricted peptides that stimulate CTL to recognize and lyse HER2/neu-expressing cancer cells.3, 4 E75 is a 9 amino acid peptide (369-377: KIFGSLAFL) located in the protein's extracellular domain. E75 has high affinity for the HLA-A2 molecule and has been identified as the immunodominant CTL epitope of HER2/neu.3 GP2 is a 9 amino acid peptide (654-662: IISAVVGIL) from the transmembrane portion of the HER2/neu protein. It has relatively poor binding affinity to HLA-A2 and has been considered a subdominant epitope.4

Our group recently reported the use of the E75 peptide mixed with GM-CSF as a simple vaccine in a preventive strategy that has induced clonal expansion of E75-specific CD8+ T cells as detected by HLA-A2:Ig dimer molecules in node-positive (NP) breast cancer patients who were rendered disease-free through standard therapy. Documentation of clonal expansion correlated well with functional assays.5 Once proven safe, we expanded our clinical trial to include node-negative (NN) patients. Combining NN and NP patients, we have shown that the majority (98%) of patients generated clonal expansion of E75-specific CD8+ T cells after vaccination. Most NP patients (92%) also experienced intra-antigenic epitope spreading as evidenced by the clonal expansion of CD8+ T cells specific to GP2, not included in the administered vaccine, suggesting the immunogenicity of this subdominant epitope. Interestingly, even though patients with NN disease have a lower expected tumor antigen burden,6 85% of NN patients demonstrated epitope spreading to the GP2 peptide. However, the response in the NN patients was less robust than in NP patients, suggesting that patients with early stage disease may benefit from immunization with a multiepitope vaccine such as one administering both the E75 and GP2 epitopes.7

One theoretical concern regarding the use of GP2 in a peptide vaccine is its poor binding affinity for the HLA-A2 molecule. This raises the possibility that only a limited number of GP2-MHC Class I complexes are expressed on the cell surface, resulting in limited CTL induction in vivo.8–10 However, early in vitro studies evaluating GP2 demonstrate that, despite significantly lower binding affinity, this peptide has a similar capacity for CTL induction as E75, suggesting that it may be more immunogenic.11 Furthermore, several groups have investigated specific amino-acid substitutions introduced at the anchor residues of GP2 and found that increased binding of GP2 to the HLA-A2 molecule can be achieved.8, 9, 12

Even with improved binding affinity and adequate immunogenicity, GP2 may have theoretical limitations as a broadly applicable peptide vaccine. A naturally occurring polymorphism at codon 655 (isoleucine → valine, GP2 → 2VGP2) has been identified13 and estimated to occur in up to 25% of the population, depending on ethnicity.14, 15 Patients carrying this polymorphism have been found to have an elevated risk of developing breast cancer.14 Therefore, it is imperative to know if GP2-specific CTL can also recognize the 2VGP2 peptide before a GP2 peptide-based vaccine could be widely employed specifically for breast cancer.

In the present study we sought to confirm the presence of GP2-specific CTL in HER2/neu-expressing breast cancer patients with and without positive nodes in order to validate the potential of GP2 as a vaccine candidate. Additionally, we compared GP2 to E75 for immunogenicity, but more important, we investigated the combination of the 2 peptides as a potential multiepitope vaccine. Furthermore, in order to address the concern over the 2VGP2 polymorphism, we assessed the ability of vaccine-induced GP2-specific CTL to recognize the 2VGP2 peptide. These data taken together prove the clinical utility of the GP2 peptide, alone or in combination, as a viable peptide vaccine.

MATERIALS AND METHODS

Patient Characteristics and Clinical Trials

Two clinical trials evaluating a HER2/neu peptide (E75) vaccine are being conducted under an Investigational New Drug application (IND #9187) approved by the Food and Drug Administration (FDA). All patients were accrued to the studies through the Walter Reed Army Medical Center's (WRAMC) Comprehensive Breast Center after approval of the protocols by the institution's Department of Clinical Investigation. Before enrollment, patients were counseled and informed consent was signed. All patients had histologically confirmed breast cancer that expressed HER2/neu by standard immunohistochemistry (IHC). Initially, patients with NP disease were accrued. After documentation of vaccine safety in this population, a second protocol accepting patients with NN disease was begun. All patients had completed a standard course of surgery, chemotherapy, and radiation therapy, as required, before their enrollment. At the time of enrollment, patients were HLA typed to determine their HLA-A2 status because E75 binds this specific allele found in approximately 40% to 50% of the general population.16 HLA-A2+ patients were vaccinated and HLA-A2 patients were followed prospectively as matched controls for clinical recurrence. Before vaccination, patients were skin tested with a panel of recall antigens (Mantoux test = mumps, tetanus, candida). Patients were considered immunocompetent if they reacted (>5 mm) to ≥2 antigens.

Peptides

The E75 vaccine peptide (HER2/neu, 369-377, KIFGSLAFL) was produced commercially in good manufacturing practices (GMP) grade by Multiple Peptide Systems (MPS, San Diego, CA). The purity of the peptide was verified by high-performance liquid chromatography and mass spectrometry and the amino acid content was determined by amino acid analysis. The peptide was purified to >95%. Stability and general safety testing was carried out by the manufacturer. For the trials, formulation was performed with WRAMC Pharmacy oversight and sterility was assured by frequent testing by WRAMC Department of Microbiology per FDA guidelines. GP2 (HER2/neu, 654-662, IISAVVGIL) was similarly produced by MPS. The 2VGP2 peptide (IVSAVVGIL) was synthesized by the BIC Facility at the Uniformed Services University of the Health Sciences (Bethesda, MD).

T2 Cells and Tumor Cell Lines

The HLA-A2+ T-B lymphoblast hybrid cell line, T2, which is deficient in the transporter-associated protein (TAP), was obtained from the American Type Culture Collection (ATCC, Manassas, VA). The HER2/neu-expressing breast cancer cell line MCF-7 (HLA-A2+, HER2/neu+), and the ovarian cancer cell line SKOV3 (HLA-A2, HER2/neu+) were also obtained from ATCC. SKOV3 cells were cultured in RPMI-1640 supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gemini BioProducts, Woodland Hills, CA), 1% glutamine, and 1% penicillin-streptomycin (GIBCO-BRL, Gaithersburg, MD). MCF-7 cells were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS) and penicillin-streptomycin, glutamine, and insulin (Sigma Chemicals, St. Louis, MO). The HLA-A2-transfected variant of the SKOV3 parental line, SKOV3-A2 (HLA-A2+, HER2/neu+) was a kind gift from Dr. Constantin Ioannides and was cultured in culture medium in the presence of 0.2 mg/mL Geneticin.

Peripheral Blood Mononuclear Cell (PBMC) Isolation and Cultures

Blood was drawn from patients before beginning the vaccination series. Forty mL of peripheral blood was drawn into Vacutainer CPT tubes (Becton Dickinson, San Diego, CA) and centrifuged for the isolation of PBMC populations. PBMC were washed in Hanks balanced salt solution (HBSS) and resuspended in culture medium (CM) consisting of Isocove modified Dulbecco medium containing 10% human AB serum (Gemini) supplemented with 1× penicillin/L-glutamine/streptomycin, 1× sodium pyruvate, 1× nonessential amino acids, and 50 μM of 2-mercaptoethanol (Life Technologies, Gaithersburg, MD). Aliquots of freshly isolated PBMC were cryopreserved in 90% FCS and 10% dimethyl sulfoxide (DMSO) in liquid nitrogen for future testing. Leukopheresis blood products provided by healthy donors and commercially available from BRT Laboratories (Baltimore, MD) were used for the preparation of PBMC by density-gradient centrifugation using the lymphocyte separation medium, Lymphoprep (Accurate Chemical & Scientific, Westbury, NY).

In Vitro Stimulation of PBMC Cultures for Cytotoxicity Assay

The PBMC preparations resuspended in complete culture medium were used for the preparation of DC and purified CD8+ T cells and also as a source of lymphocytes for in vitro stimulation with the E75, GP2, and 2VGP2 peptides, as well as negative controls. Highly pure populations of CD8 T cells were isolated by a magnetic beads separation technique using CD4-, CD14-, and CD19-coated Dynal (Fort Lee, NJ) beads. Highly enriched populations of CD14 monocytes were prepared by magnetic bead depletion of PBMC that had been incubated with CD2-, CD19-, and CD8-coated beads from Dynal. The CD14 cell population was cultured in Macrophage Serum Free Medium (GIBCO-BRL) with 100 μg/mL of GM-CSF and 50 μg/mL of IL-4/mL (R&D Systems, Minneapolis, MN) to obtain monocyte-derived DC populations. Recombinant human TNF-α (R&D Systems) was added at 30 ng/mL on Day 3 to induce maturation. DC were then harvested at 6 days, incubated in the absence of peptide or with E75 or GP2 for 2 hours, washed, and added separately to purified preparations of CD8 T cells at a stimulator:effector ratio of 1:10. For experiments to test the combination of E75 and GP2 peptides, the CD8 T cells were stimulated with equal numbers of E75-pulsed and GP2-pulsed DC in the same well. For the stimulation of PBMC with E75 or GP2, the peptide was added directly to the cultures at 25 μg/mL in 48-well plates containing 2 × 106cells/mL. All of these cultures were set up in CM with 10 ng/mL of IL-7 and this was followed by the addition of 25 ng/mL of IL-2 on the second day. All cultures were maintained in a humidified incubator at 37°C in 5% CO2.

Chromium-51 Cytotoxicity Assay

Peptide-specific cytotoxicity in the CD8 T cell cultures stimulated with peptide-pulsed DC and in vitro-stimulated PBMC cultures were determined by standard 4-hour chromium release assay after 7-10 days in culture. Briefly, targets were labeled with 100-150 μCi of sodium chromate (Perkin-Elmer, Boston, MA) for 1.5 hours at 37°C then washed twice and plated at 2500 cells/well in 100 μL in 96-well U-bottom plates (Becton Dickinson). Effectors were added at an effector:target (E:T) ratio of 10:1-20:1 in 100 μL/well. After 4 hours of incubation, 100 μL of culture supernatant was collected and radionuclide release measured on a Microbeta Trilux counter (Perkin-Elmer). All determinants were done in triplicate. Results are expressed as percent-specific lysis as determined by: (experimental mean cpm-spontaneous mean cpm)/(maximum mean cpm-spontaneous mean cpm) × 100. The target cells used in the cytotoxicity assays were T2 cells loaded with GP2 or 2VGP2 peptide, MCF-7 cells (HLA-A2+, HER2/neu+), SKOV3 (HLA-A2, HER2/neu+), and SKOV3-A2 (HLA-A2+, HER2/neu+).

Statistical Analysis

Statistical analysis was performed using the SPSS program (Chicago, IL). Continuous data was compared using the paired, 2-tailed Student t-test. Probabilities <.05 were considered significant.

RESULTS

Cytolytic Activity of Peptide-Stimulated PBMC from Breast Cancer Patients

PBMC were obtained from 28 HLA-A2+ patients with HER2/neu-expressing breast cancer before initiation of the vaccination series. All patients had been rendered free of disease by conventional therapy including surgery, chemotherapy, and/or radiation therapy as dictated by their treating physician. PBMC samples from 10 NP and 18 NN patients were stimulated with the GP2 peptide, after which they were tested in cytotoxicity assays against the HER2/neu-expressing tumor targets MCF-7 (HLA-A2+, HER2/neu+), SKOV3 (HLA-A2, HER2/neu+), and SKOV3-A2 (HLA-A2+, HER2/neu+) at an E:T ratio of 20:1. In 23 of 28 patients, peptide-stimulated PBMC demonstrated specific cytotoxicity against HER2/neu-expressing cancer cells in an HLA-restricted fashion. As shown in Figure 1A, in NP patients the average cytotoxicity versus MCF-7 and SKOV3-A2 target cells was 37.8 ± 6.3% and 35.3 ± 6.8%, respectively, compared with 20.8 ± 5.2% against the HLA-A2 tumor cells SKOV3 (P = .001 vs. MCF-7; P = .02 vs. SKOV3-A2). As shown in Figure 1B for the NN patients, the average cytotoxicity versus MCF-7 and SKOV3-A2 target cells was 23.7 ± 4.2% and 26.3 ± 4.8%, respectively, compared with 14.7 ± 2.9% against the HLA-A2 SKOV3 (P = .0005 vs. MCF-7; P = .0002 vs. SKOV3-A2).

Figure 1.

Cytolytic activity of peptide-stimulated PBMC from breast cancer patients. PBMC from 28 HLA-A2+ breast cancer patients were initiated in culture and stimulated once with either (A,B) GP2 or (C,D) E75 HER2/neu peptides. The results of standard 4-hour chromium-release assays against HER2/neu-expressing targets [MCF-7 (HLA-A2+), SKOV3.A2 (HLA-A2+), and SKOV3 (HLA-A2)] at an E:T ratio of 20:1 are expressed in mean percent cytotoxicity ± SEM for (A,C) 10 NP and (B,D) 18 NN breast cancer patients.

These results are comparable to the results of cytotoxicity assays when prevaccination PBMC from the same 28 patients were stimulated with E75, the HER2/neu-derived immunodominant epitope that is currently being tested in our clinical trials. Figure 1C,D shows the results of cytotoxicity assays involving prevaccination PBMC from NP and NN patients, respectively, stimulated with E75 versus the HER2/neu-expressing tumor targets. In 24 of 28 patients, peptide-stimulated PBMC demonstrated specific cytotoxicity against HER2/neu-expressing cancer cells in an HLA-restricted fashion. In NP patients, the average cytotoxicity versus MCF-7 and SKOV3-A2 cells was 36.6 ± 6.1% and 31.2 ± 5.4%, respectively, compared with 20.1 ± 5.0% against SKOV3 (P = .0006 vs. MCF-7, P = .03 vs. SKOV3-A2). In NN patients, the average cytotoxicity versus MCF-7 and SKOV3-A2 cells was 27.1 ± 4.9% and 24.7 ± 4.9%, respectively, compared with 18.2 ± 3.7% against SKOV3, (P = .004 for MCF-7, P = .10 for SKOV3-A2).

In these experiments we attribute the difference between NP and NN patients to NP patients having a higher tumor burden, and therefore, more endogenously processed antigen and a greater number of prevaccination immune effectors recognizing HER2/neu-expressing tumor cells. Importantly, these experiments show evidence of the presence of GP2-specific precursor T cells. The percent of cytotoxicity is comparable in experiments involving GP2-stimulated PBMC versus E75-stimulated PBMC. In ongoing clinical trials, we have already demonstrated the ability to expand E75-specific CTL by vaccination with the E75-peptide as well as epitope spreading to the GP2 peptide. These results suggest that vaccination with the GP2 peptide should result in the expansion of GP2-specific CTL.

Detection of Cytolytic Activity against Tumor Targets in Peptide-Stimulated Cultures with E75, GP2, and E75+GP2

We next performed cytotoxicity experiments to further investigate the immunogenicity of GP2 compared with E75. Purified CD8+ T cells from 6 HLA-A2+ healthy donors were stimulated with autologous DC pulsed with either E75 or GP2. To investigate the potential efficacy of a multiepitope vaccine, we also performed cytotoxicity experiments using effectors stimulated with a combination of E75 and GP2. After stimulation with the designated peptide, these CD8+ T cells were used in chromium release assays versus HER2/neu-expressing MCF-7 tumor targets. As shown in Figure 2A, the average cytotoxicity for E75-stimulated effectors versus MCF-7 tumor targets was 44.2 ± 5.7%. The average cytotoxicity for GP2-stimulated effectors was not significantly different, 43.8 ± 5.2% (P = .87). These data further document that the immunogenicity of GP2 with respect to cytolytic activity is comparable to E75 despite the differences in binding affinities. Because GP2 is effective in stimulating an immune response against HER2/neu-expressing tumor cells, it may be useful for administration as a peptide-based cancer vaccine.

Figure 2.

Cytolytic activity of healthy donor CD8 T cells stimulated with individual or a combination of HER2/neu peptides. DC and CD8 T cells were purified from 6 HLA-A2+ healthy donor PBMC samples. CD8 T cells were stimulated once with GP2 or E75 peptide-loaded autologous DC and then utilized in standard 4-hour chromium-release assays against the HLA-A2+, HER2/neu+ breast cancer cell line, MCF-7, at an E:T ratio of 20:1. For the combination experiments, the CD8 T cells were stimulated with equal numbers of the GP2 and E75-loaded DC but at the same stimulator:effector ratio as the individual peptide experiments. (A) Results for all 6 samples expressed as mean percent cytotoxicity ± SEM. (B) Results of the individual experiments.

Interestingly, when CD8+ T cells were stimulated with DC that had been separately pulsed with both E75 and GP2, there was an increase in the average specific cytotoxicity compared with single peptide treatment. When the combination was used, the average specific cytotoxicity was 58.6 ± 5.4%, which was greater than the cytolytic activity using E75 (P = .04) or GP2 (P = .06) alone. Figure 2B details the results of individual experiments and, as shown, the combination worked better in 4 of 6 samples. These results suggest that administration of a multiepitope vaccine may be more efficacious in stimulating an immune response than administration of a single peptide vaccine.

Comparison of Cytolytic Activity of GP2 and 2VGP2-Specific CD8+ T cells

Having shown cytolytic activity of GP2-stimulated CD8+ T cells against HER2/neu-expressing tumor cells, we next sought to study the implications of the naturally occurring polymorphism 2VGP2 on a GP2-vaccine strategy. Initial cytotoxicity experiments were performed using CD8+ T cells purified from an additional 5 HLA-A2+ healthy donors. The CD8+ T cells were stimulated with either GP2 or 2VGP2 and used in cytotoxicity assays against SKOV3-A2 and SKOV3. Figure 3 depicts the specific cytotoxicity data from the GP2-stimulated CD8+ T cells and 2VGP2-stimulated CD8+ T cells after 1 round of peptide stimulation. Specific lysis was determined by subtracting the cytotoxicity against SKOV3 from that against SKOV3.A2. As seen in the figure, in 3 out of 5 experiments, both GP2 and 2VPG2-stimulated effector cells similarly lyse the HER2/neu-expressing tumor targets. These results suggest that GP2 and 2VGP2 have comparable immunogenicity with respect to cytolytic activity.

Figure 3.

Comparison of cytolytic activity of GP2 and 2VGP2-stimulated CD8 T cells. DC and CD8 T cells were purified from 5 additional HLA-A2+ healthy donor PBMC samples. CD8 T cells were stimulated once with GP2 or 2VGP2 peptide-loaded autologous DC and then utilized in standard 4-hour chromium-release assays against SKOV3.A2 (HLA-A2+, HER2/neu+) and SKOV3 (HLA-A2, HER2/neu+) cancer cell lines at an E:T ratio of 20:1. Results are expressed as specific lysis (SKOV3.A2 minus SKOV3) and are shown for each of the 5 samples.

After demonstrating comparable killing against HER2/neu-expressing tumor targets, we next investigated the cytolytic activity of GP2- and 2VGP2-stimulated CD8+ T cells against peptide-loaded T2 targets. Again, CD8+ T cells from 5 HLA-A2+ healthy donors were stimulated with either GP2 or 2VGP2. Cytotoxicity experiments were performed against GP2- or 2VGP2-loaded T2 targets. As shown in Figure 4A, GP2-stimulated effectors demonstrated killing of both GP2- and 2VGP2-loaded T2 targets, thereby showing complete cross-reactivity, which was not seen in the 2VGP2-stimulated cultures (Fig. 4B).

Figure 4.

GP2-stimulated CD8 T cells recognize targets expressing the 2VGP2 peptide. DC and CD8 T cells were purified from 5 HLA-A2+ healthy donor PBMC samples. CD8 T cells were stimulated once with (A) GP2 or (B) 2VGP2 peptide-loaded autologous DC and then utilized in standard 4-hour chromium-release assays against T2 cells loaded with either GP2 or 2VGP2 peptides at an E:T ratio of 20:1. Results are expressed as specific lysis (peptide-loaded T2 minus T2 alone) and are shown for each of the 5 samples.

These finding suggest that a GP2-peptide vaccine should be effective in breast cancer patients who may or may not carry the described isoleucine to valine polymorphism in the HER2/neu protein.

DISCUSSION

GP2, a peptide derived from the HER2/neu protein, has been shown to be immunogenic and capable of stimulating CTL.4 However, it has remained a less attractive candidate for vaccine development largely because it is a subdominant epitope with relatively poor HLA-A2 binding affinity, as well as being within the site known to contain a naturally occurring polymorphism. Acknowledging this, we performed the current study to assess the potential for utilizing GP2 in a simple vaccine aimed at preventing recurrent disease in patients treated for HER2/neu+ breast cancer. Specifically, we have shown that PBMC obtained from HER2/neu+ breast cancer patients contain the GP2-specific precursor CTL. In addition, the cytolytic activity of GP2-stimulated CD8+ T cells was equal to that obtained using E75-stimulated CD8+ T cells. This suggests that E75, which we have demonstrated in clinical trials to be effective as a peptide vaccine in stimulating the expansion of E75-specific CD8+ T cells, and GP2 have comparable immunogenicity in vitro. More important, the cytolytic activity achieved by CD8+ T cells stimulated with a combination of the 2 peptides was superior to either peptide alone, suggesting a potential benefit in using a multiepitope vaccine. Finally, we have shown that GP2-stimulated effectors can recognize and lyse 2VGP2-expressing targets, suggesting that a GP2 vaccine would be effective in patients carrying the 2VGP2 polymorphism.

GP2 was first described by Peoples et al.,4 who identified this HLA-A2-restricted, HER2/neu-derived 9 amino acid peptide (IISAVVGIL, p654-662) using tumor-associated lymphocytes isolated from patients with ovarian and breast cancer. Subsequent to this description, these investigators showed that GP2 is shared between several additional epithelial tumors, including nonsmall cell lung and pancreatic cancers,17–19 and that in vitro-generated allogeneic DC pulsed with the GP2 peptide could induce GP2-specific CTL capable of lysing human pancreatic cancer cells in vitro.20 Other investigators used autologous DC generated from peripheral blood monocytes pulsed with GP2 as antigen presenting cells for the induction of CTL that lysed colon and renal cell carcinoma cell lines expressing HER2/neu.21 Results from these in vitro experiments suggested that GP2 may be an attractive candidate for broadly applicable cancer vaccines.

Vaccinations using DC pulsed with TAA were initially shown to be effective in patients with B-cell lymphoma and malignant melanoma.22, 23 Brossart et al.11 were the first investigators to vaccinate patients with advanced breast and ovarian cancer with autologous DC pulsed with TAA. They vaccinated patients with metastatic disease with DC pulsed with the HER2/neu peptides, E75 or GP2. Postvaccination, they detected peptide-specific CTL in the peripheral blood of patients, thereby demonstrating the ability to induce antigen-specific cytotoxic T cells in vivo. Interestingly, the pattern of immunogenicity for E75 and GP2, as detected by intracellular IFN-γ staining, were comparable for the 2 peptides in their DC-based immunotherapy trial. A second Phase I/II study conducted by Dees et al.24 also demonstrated the feasibility of administering a vaccine consisting of DC pulsed with GP2 to patients with metastatic breast cancer. Despite the successful use of GP2 in these DC vaccines, this epitope has never been employed in a direct peptide vaccination strategy. In contrast, our group and others have been using the immunodominant epitope E75 in clinical trials as a simple exportable peptide-based vaccine.5, 25–27 E75 was chosen as the peptide for the first series of clinical trials by our group because of its high HLA-A2 affinity and proven immunogenicity.

One reason hypothesized for poor T-cell recognition of tumor-associated peptides is poor binding of those peptides to Class I MHC molecules. If a peptide dissociates from Class I MHC molecules too quickly, the cells presenting the peptide do not have a sufficient concentration of the specific peptide-MHC complex at the cell surface to be recognized by circulating T cells.9 Class I MHC molecules use a subset of amino acid sidechains within the peptide (termed “anchors”) to generate significant binding.28 Peptides that bind with high affinity to a given allotype typically have one of a few preferred amino acids at each anchor position (for example, leucine or methionine at position 2 and valine at position 9 for HLA-A2).29, 30 GP2 lacks dominant amino acids at both anchor sites. Kuhns et al.9 studied altered peptide ligands based on GP2 that replaced the isoleucine at position 2 with leucine or the leucine at position 9 with valine. These substitutions of dominant amino acids at the anchor residues did not significantly improve the binding affinity. The authors further studied GP2 binding to Class I MHC molecules by determining the crystallographic structure. They found that poor binding exists because the center of the GP2 peptide does not make stabilizing contacts with the binding cleft of the Class I MHC molecule.

Tanaka et al.8 investigated the effect on binding affinity of modifying GP2 with either single, double, or triple amino acid substitutions. They identified that double and triple amino acid modifications, especially 2L9V and 1F-based modifications (1F9V, 1F2L, 1F2L9V, 1F2L10V), markedly enhanced the peptide binding affinity to HLA-A2 compared with unmodified GP2. They then investigated the ability of the modified peptides to induce GP2-specific CTL. Using PBMC derived from three HLA-A2+ healthy donors stimulated with autologous DC pulsed with modified peptide, they demonstrated that CTL reactive against the parent GP2 epitope were induced most efficiently with 1F2L10V modified peptide. In experiments we performed comparing 1F2L10V and GP2 in cytotoxicity assays versus T2 and HER2/neu tumor targets, we also observed enhanced lysis after stimulation with this modified peptide, but were unable to demonstrate this response in a consistent manner (data not shown).

Despite the fact that GP2 has low-affinity binding to HLA-A2 based on standard assays, in experiments reported here and elsewhere,11 GP2 has immunogenicity that is comparable to that of E75. This suggests that the anchor residues of GP2 are adequate and the flexibility in the center of the peptide, while it prohibits strong binding to HLA-A2, may actually improve the peptide's immunogenicity. If the center of GP2 is not bound well to the HLA-A2 molecule, it is possible that it assumes multiple different conformations that can generate more molecular surfaces that can potentially be recognized by multiple T-cell receptors (TCR) on circulating T cells.9

Other investigators have recently reported on efforts to improve the affinity of an antigen for the TCR by altering the mid-portion of the peptide. Kawano et al.31 have altered the E75 antigen by adding methylene (-CH2) groups to the glycine molecule at position 4 (Gly4). It was found that the -CH2 extensions did not change the affinity for HLA-A2; however, they did cause increased binding affinity to the TCR. Specifically, the addition of 2 -CH2 groups at Gly4 resulted in optimal affinity. In turn, this led to expansion of CTL that were identified as having higher-than-average numbers of TCR as well as high functional avidity for their tumor targets. Similar studies have been initiated with the GP2 peptide.

Another consideration that must be addressed before the application of a GP2 vaccine is the naturally occurring polymorphism that encodes valine at codon 655 that has been described in the transmembrane coding region of HER2/neu.13 In a population-based, case-control study of Chinese women, this polymorphism was found to be associated with an increased risk of breast cancer, particularly in younger women.14 The frequency of the valine allele is variable among populations, with the prevalence of the polymorphism corresponding to the incidence of breast cancer in these different groups.15 A study by McKean-Cowdin et al.32 confirmed that women carrying the 2VGP2 polymorphism had an increased risk of developing breast cancer. Interestingly, these women were more likely to develop localized breast cancer, not regional or metastatic disease. Women with at least 1 copy of the valine variant were approximately one-half as likely to have high-stage as low-stage breast cancer, an effect that was present across ethnic groups. Another group identified that tumors from women who were homozygous for the Val allele were more likely to exhibit HER2/neu overexpression.33 This finding is particularly relevant as we investigate the impact of the polymorphism on a peptide-vaccination strategy targeting women with HER2/neu+ disease.

To assess the broad applicability of a GP2 vaccine, we have shown that the immunogenicity of GP2 and 2VGP2 is comparable, as determined in cytotoxicity assays versus HER2/neu-expressing tumor targets. This is likely because the polymorphism occurs at position 2 of the peptide, an anchor residue. T2 stabilization assays that we performed comparing GP2 with 2VGP2 did not show significant change in the stabilization of the HLA-peptide complex when comparing the 2 peptides (data not shown). Additionally, even with the 2VGP2 polymorphism, the center of the peptide is unchanged. This further supports the idea that the immunogenicity of GP2 and subsequently 2VGP2 is primarily due to the flexibility of the center portion of the peptide. Importantly, we have also shown that CTL stimulated with GP2 show complete cross-reactivity with 2VGP2-loaded T2 cells. This suggests that a GP2 peptide vaccine should be effective in breast cancer patients regardless of whether or not they carry the reported polymorphism.

In light of this preclinical work, we have recently begun enrolling patients into a clinical trial administering GP2 mixed with GM-CSF as a simple peptide vaccine. In order for a peptide vaccination strategy to be successful, we predict that ultimately patients will be administered a multiepitope vaccine. A recent clinical trial in melanoma patients administered a vaccine consisting of four gp100- and tyrosinase-derived peptides. Immune responses were detected against all four peptides, confirming their immunogenicity; however, there was substantial inter- and intrapatient variations in response to different peptides.34 Because of the variability of response to a given peptide as well as the heterogeneity of antigen expression in vivo, it is likely that vaccination with multiple different peptides will be required to provide an adequate immune response. Our preliminary results reported here looking at combination therapy with E75 and GP2 suggest an additive response. It is also possible that a multiepitope vaccine may be able to elicit a response against tumors that have lower levels of HER2/neu expression.

In summary, the present study confirms the presence of GP2-specific precursor CTL in women with HER2/neu+ breast cancer as well as the immunogenicity of the GP2 peptide and supports its further development as a peptide vaccine.

Ancillary