A modified HLA-A*0201-restricted CTL epitope from human oncoprotein (hPEBP4) induces more efficient antitumor responses

Abstract

We previously identified human phosphatidylethanolamine-binding protein 4 (hPEBP4) as an antiapoptotic protein with increased expression levels in breast, ovarian and prostate cancer cells, but low expression levels in normal tissues, which makes hPEBP4 an attractive target for immunotherapy. Here, we developed hPEBP4-derived immunogenic peptides for inducing antigen-specific cytotoxic T lymphocytes (CTLs) targeting breast cancer. A panel of hPEBP4-derived peptides predicted by peptide-MHC-binding algorithms was evaluated to characterize their HLA-A2.1 affinity and immunogenicity. We identified a novel immunogenic peptide, P40–48 (TLFCQGLEV), that was capable of eliciting specific CTL responses in HLA-A2.1/K^b transgenic mice, as well as in peripheral blood lymphocytes from breast cancer patients. Furthermore, amino-acid substitutions in the P40–48 sequence improved its immunogenicity against hPEBP4, a self-antigen, thus circumventing tolerance. We designed peptide analogs by preferred auxiliary HLA-A*0201 anchor residue replacement, which induced CTLs that were crossreactive to the native peptide. Several analogs were able to stably bind to HLA-A*0201 and elicit specific CTL responses better than the native sequence. Importantly, adoptive transfer of CTLs induced by vaccination with two analogs more effectively inhibited tumor growth than the native peptide. These data indicate that peptide analogs with high immunogenicity represent promising candidates for peptide-mediated therapeutic cancer vaccines.

Introduction

T-cell-based specific immunotherapy is considered to be one of the most promising strategies for tumor therapy.^1,2,3 CD8⁺ cytotoxic T cells (CTLs) are capable of recognizing epitope peptides derived from tumor-associated antigens (TAAs) presented on major histocompatibility complex (MHC) class I molecules for subsequent lysis of tumor cells. Effective cancer immunity in humans has been associated with the presence of CTLs directed against cancer antigens, a class of HLA-bound peptides derived from tumor-specific antigens or TAAs. As most TAAs are self-antigens, an antigen-specific CTL repertoire targeting tumor-associated self-antigens is uncommon in cancer patients as a consequence of thymic depletion and is poorly effective at mounting productive antitumor immune responses. It has been demonstrated that altering the amino-acid residues of peptides that interact with either the HLA molecule or T-cell receptor can increase their binding affinity to MHC class I molecules^4,5 and improve epitope immunogenicity.⁶ Moreover, effective induction of CTLs by peptides with enhanced immunogenicity has shown promising results in clinical studies.⁷

Human phosphatidylethanolamine-binding protein 4 (hPEBP4) was previously identified to be a novel antiapoptotic protein by our group.⁸ Its normal gene expression in healthy tissues is limited to low levels in the lung, liver, spinal cord, brain and bone marrow, whereas it has high expression levels in the testis, heart and thyroid. Interestingly, the hPEBP4 gene is overexpressed in several human cancers, such as breast, prostate and ovarian carcinoma. Examination of clinical specimens has shown that hPEBP4 is highly expressed in human breast carcinoma tissue and lowly expressed in normal breast tissue.^9,10,11 Thus, hPEBP4 might be an overexpressed TAA and may potentially serve as an attractive target for immunotherapy.

Here, we report the first identification of a 9-mer HLA-A2.1-restricted peptide derived from the hPEBP4 protein, designated P40–48 (TLFCQGLEV), that has the ability to induce peptide-specific, HLA-A2.1-restricted CTL responses in HLA-A2.1/K^b transgenic (Tg) mice in vivo, as well as in peripheral blood lymphocytes (PBLs) of HLA-matched breast cancer patients in vitro, by using dendritic cells (DCs) that were prepulsed with the peptides. We further performed amino-acid substitution in the sequence of this new CTL epitope and demonstrated that the HLA-A*0201-binding affinity and immunogenicity of at least two modified peptides could be greatly enhanced by amino-acid substitution at HLA-A*0201-binding anchor positions in native peptides. These data therefore provide a preclinical foundation for clinical trials of hPEBP4-directed vaccine immunotherapy for breast cancer or other tumors.

Materials and methods

Peptides

Peptides were synthesized by GL Biochem (Shanghai, China) using fluorenylmethoxycarbony l chemistry at >95% purity by reverse-phase high-performance liquid chromatography, as confirmed by mass spectrometry. Lyophilized peptides were dissolved in dimethyl sulfoxide, diluted in phosphate-buffered saline (pH 7.4) to a concentration of 10 mm and stored in aliquots at −80 °C as described previously.¹² The amino-acid sequences and predicted score for binding to HLA-A*0201 were generated by databases available online (nHLAPred, http://crdd.osdd.net/raghava//nhlapred/; ProPred-I, http://crdd.osdd.net/raghava//propred1/; SYFPEITHI, http://www.syfpeithi.de/; and PREDEP, http://margalit.huji.ac.il/Teppred/mhc-bind/).

Animals and cell lines

HLA-A2.1/K^b Tg mice, 6–8 weeks of age, were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Mice were bred and maintained under specific pathogen-free conditions. All research was performed in accordance with the Declaration of Helsinki and the Guide for Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health. All experimental protocols were approved by the Review Committee for the Use of Animal Subjects of the Second Military Medical University. Cell-surface HLA-A*0201 expression was assessed by flow cytometry using a fluorescein isothiocyanate (FITC)-labeled HLA-A2-specific monoclonal antibody BB7.2 (Serotec, Oxford, UK).

Human transporter associated with antigen processing (TAP)-deficient T2 cells (expressing HLA-A2.1 molecules), human breast adenocarcinoma MCF-7 (hPEBP4⁺/HLA-A2.1⁺), human Burkitt’s lymphoma Daudi (hPEBP4⁺/ HLA-A2.1⁻) and human lung carcinoma A549 (hPEBP4⁻/ HLA-A2.1⁺) cell lines were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). Human breast carcinoma MAD-MB-468 cells were kindly donated by Prof Jian Ding (Institute of Materia Medica Chinese Academy of Sciences, Shanghai, China).

Peptide-binding assay

The relative avidity of hPEBP4 peptides for HLA-A2.1 was measured using an MHC stabilization assay on T2 cells as described previously.^4,13 The SSp-1 peptide (RLNEVAKNL)¹² served as a positive control peptide, and OVA_257–264 (SINFKEL) was used as a negative control peptide. The fluorescence index (FI) was calculated as follows: FI=(mean FITC fluorescence with the given peptide−mean FITC fluorescence without peptide)/(mean FITC fluorescence without peptide). Peptides with a FI >1 were regarded as high-affinity epitopes.

Assessment of the peptide/HLA-A*0201 complex stability

This assay was performed as described previously.^15,16 Briefly, T2 cells were incubated overnight at 37 °C with 100 μm of each peptide in serum-free RPMI-1640 medium supplemented with 100 ng/ml β2-microglobulin.¹⁶ The cells were then washed four times to remove free peptides, incubated with brefeldin A (10 μg/ml; Sigma-Aldrich, USA) for 1 h to block cell-surface expression of newly synthesized HLA-A0201 molecules, and then washed and further incubated at 37 °C for 0, 2, 4 and 6 h. Subsequently, the cells were stained with monoclonal antibody BB7.2 to evaluate expression of the HLA-A0201 molecule. The mean fluorescence intensities (MFIs) measured at 0 h were set at 100%. The MFI at all other time points are expressed relative to the MFI at 0 h. For each time point, peptide-induced HLA-A2 expression was calculated as follows: [MFI (0 h)−MFI (2, 4 or 6 h)/MFI (0 h)] × 100%.

hPEBP4 RNA-interference assay

Transient knockdown of hPEBP4 using a small interfering RNA assay with a chemically synthesized small interfering RNA duplex and mutated control was performed as described previously.⁸

Generation of CTL in HLA-A2.1/K^b Tg mice

Preparation of bone marrow-derived DCs from Tg mice and vaccination of HLA-A2.1/K^b Tg mice (seven mice per group) with peptide-pulsed DCs was performed as described previously.^12,17,18,19 Immunization of Tg mice with Ad vector-transduced DCs was performed as described previously.²⁰ HLA-A2.1/K^b Tg mice were intraperitoneally immunized once with pAd-hPEBP4-transduced DCs (5 × 10⁵ cells per mouse). Control groups simultaneously received pAd-GFP-transduced and phosphate-buffered saline-treated DCs. On day 7 after the last immunization with peptide-pulsed DCs and on day 15 after immunization with pAd-hPEBP4-transduced DCs, total immune splenocytes from the same group of primed mice were collected and cultured at a density of 1 × 10⁷ cells per well in 6-well plates and stimulated with peptides (20 μm) for 7 days in vitro. Then, the bulk populations were functionally tested by enzyme-linked immunospot (ELISPOT), enzyme-linked immunosorbent assay (ELISA) and intracellular staining assays, as well as ⁵¹Cr-release cytotoxicity assays.

CTL induction in breast cancer patient-derived PBLs in vitro

Peripheral blood mononuclear cells from HLA-A2.1⁺ breast carcinoma patients (17 patients, all of whom provided informed consent; protocol approved by the Ethics Committee for the Application of Human Samples, Second Military Medical University) and human peripheral blood monocyte-derived DCs were generated as described previously.^21,22 Peptide-specific CTLs were generated as described previously¹² with minor modifications. After 7 days of coculturing with peptide-pulsed autologous DCs, lymphocytes were restimulated with peptide-pulsed autologous DCs in medium containing 10 ng/ml interleukinL-7 (IL-7) and then supplemented with 20 IU/ml rhIL-2 (Sigma-Aldrich) 72 h later. Lymphocytes were restimulated each week in the same manner. Media were changed every 3 days with half-fresh medium in the presence of recombinant human rhIL-2 (20 IU/ml) and expanded as necessary. On day 7 after the third stimulation, CD8⁺ T lymphocytes (6–7 × 10⁸) were enriched by positive selection using immunobeads (Miltenyi Biotec) following the procedure recommended by the manufacturer. Interferon-γ (IFN-γ) secretion of these CD8⁺ T cells was then examined by ELISPOT assays and ELISA, and the cells were tested for cytotoxicity using granzyme B ELISPOT and ⁵¹Cr-release assays.

ELISPOT assay

ELISPOT assays were performed using a commercially available kit (R&D Systems, Minneapolis, MN, USA) as described previously.¹² Splenocytes (1 × 10⁵ per well) from the immunized HLA-A2.1/K^b mice described above and purified CD8⁺ T cells (1 × 10⁵ per well, more than 95% pure) were used as effector cells. Peptide-pulsed T2 cells or tumor cells (1 × 10⁴ per well) were used as stimulator cells.

IFN-γ ELISA

The culture supernatants from the in vitro cytotoxicity assays were used to quantify the levels of IFN-γ using an ELISA Kit (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions. Peptide-pulsed T2 cells or tumor cells were used as stimulator cells. Splenocytes from immunized HLA-A2.1/K^b mice or purified CD8⁺ T cells were used as effector cells.

Intracellular detection of IFN-γ

Intracellular IFN-γ secretion was detected as described previously.²³ Briefly, effector T cells from immunized mice or breast cancer patients were restimulated in the presence of corresponding peptides or tumor cell lines (MCF-7, A549 and Daudi). After 48 h of stimulation, harvested cells were stained with antigen-presenting cell anti-CD3 and FITC anti-CD8 (BD Pharmingen, San Diego, CA, USA). After fixation and permeabilization, intracellular staining was performed using phycoerythrin anti-IFN-γ and a phycoerythrin isotype control (BD Pharmingen). Cells were fixed with 0.5% paraformaldehyde and analyzed by flow cytometry (FASCcan or FACSVantage SE; BD Biosciences). Data analysis was performed using CellQuest software (BD Biosciences).

Cell transfection

The hPEBP4¹ expression vectors were transfected into MAD-MB-468 using jetPEI (Polyplus, Illkirch-Graffenstaden, France) with pcDNA3.1/myc-His (−) B as a mock control. Transfected cells were selected in media containing 800 μg/ml of G418 and subsequently maintained using a 600 μg/ml dose of G418.

Cytotoxicity assay

The cytotoxic activities of peptide-induced CTLs were measured in a standard ⁵¹Cr-release assay as described previously.¹² Cellular cytotoxicity was calculated as follows: % specific lysis−[(mean experimental c.p.m.−mean spontaneous c.p.m.)/(mean maximum c.p.m.−mean spontaneous c.p.m.)] × 100%.

Adoptive immunotherapy in tumor-bearing nude mice

Splenocytes from each group of immunized HLA-A2.1/K^b mice were stimulated with 20 μm P40–48 for 7 days as described in the cytotoxicity assay. MAD-MB-468-hPEBP4 tumor cells (5 × 10⁶) were injected into the mammary fat pads of C57BL/6^nu/nu mice, 100% of which formed homogeneous tumors. Three days later, mice were intravenously injected with splenocytes (1 × 10⁸ cells per mouse) from each group of immunized HLA-A2.1/K^b mice that were stimulated with 20 μm P40–48 for 7 days as described in the cytotoxicity assay section. This adoptive transfer was performed twice at 1-week intervals, followed by intraperitoneal administration of 2000 U of hIL-2 every 2 days. Control mice received splenocytes from HLA-A2.1/K^b mice immunized with non-pulsed DCs or received only IL-2. Tumor size was measured in two perpendicular dimensions three times at weekly intervals. Mice were killed 80 days after tumor inoculation.

Statistical analysis

The statistical significance of the differences between means was assessed with the two-tailed Student’s t-test. Tumor sizes among groups were compared with the Mann–Whitney U-test. Kaplan–Meier curves were plotted for survival analyses. P<0.05 was considered statistically significant. SPSS manager software (v21; IBM Inc.) was used for all statistical tests.

Results

Screening of potential hPEBP4-derived peptides that bind to HLA-A*0201 molecules

To determine whether hPEBP4 has the potential to elicit an immunogenic response in the host, we screened the hPEBP4 amino-acid sequence using several well-established MHC-binding peptide prediction algorithms that assist in the identification of novel T-cell epitopes. We then synthesized 10 predicted nonameric peptides with the highest estimated half-life of dissociation from HLA-A*0201 and tested their affinity binding to HLA-A*0201 using the TAP-deficient T2 cell-peptide test (Table 1). Of the 20 candidate peptides, No. 2 (P12–20), No. 3 (P40–48) and No. 6 (P6–14) were high-affinity epitopes (FI=1.15, 1.59 and 1.04, respectively); all others were low-affinity epitopes (FI<1; Table 1). The positive control SSp-1 exhibited strong binding to HLA-A*0201 (FI=1.65),¹² whereas the negative control exhibited no binding (FI=0.03), as expected.

Table 1.

Binding affinity of hPEBP4-derived peptides to HLA-A*0201 molecules

Rank	Start position	Sequence	Score a	FI b,c
1	11	ALLLGLMMV	591.88	0.02±0.01
2	12	LLLGLMMVV	412.54	1.15±0.91
3	40	TLFCQGLEV	257.34	1.59±0.89
4	113	WLVTDIKGA	52.56	0.11±0.02
5	163	SLLPKENKT	27.52	—
6	6	RLVTAALLL	21.36	1.04±0.81
7	153	FVYLQEGKV	11.946	—
8	52	ELGNIGCKV	5.21	0.26±0.07
9	127	KIQGQELSA	2.39	0.43±0.12
10	4	TMRLVTAAL	1.17	—

Abbreviations: hPEBP4, human phosphatidylethanolamine-binding protein 4; MFI, mean fluorescence intensity.

^aPropred-I prediction software, available at http://crdd.osdd.net/raghava//propred1/.

^bPeptide binding was evaluated by the MFI. SSp-1 (RLNEVAKNL) and OVA_257–264 were used as positive controls and negative controls, respectively.

^cData are presented as the mean±s.e.m. of four independent experiments.

The hPEBP4-derived P40–48 peptide is immunogenic in HLA-A2.1/K^b Tg mice

To examine the in vivo immunogenicity of the peptides, syngeneic DCs pulsed with the high-affinity peptides were used to immunize HLA-A2.1/K^b Tg mice. In addition, to determine whether these peptides were naturally processed and presented by antigen-presenting cells, we also immunized HLA-A2.1/K^b Tg mice with syngeneic DCs infected with pAd-hPEBP4 containing the full-length hPEBP4 gene and encoding a humanized recombinant green fluorescent protein. After three rounds (peptide-pulsed DCs) and 15 days (pAd-hPEBP4-infected DCs) of immunization in vivo, splenocytes were restimulated in vitro with the related peptides for 7 days, and then analyzed for IFN-γ production and CTL cytotoxicity. Only P40–48 exhibited strong IFN-γ production (Figures 1a and b); P12–20 and P6–14 did not produce significant levels of IFN-γ in response to any stimuli (data not shown).

Figure 1.

Immunogenicity of the P40–48 peptide in HLA-A2.1/K^b Tg mice. HLA-A2.1/K^b Tg mice (seven mice per group) were immunized with phosphate-buffered saline (PBS) or P40–48-pulsed DCs, and pAd-GFP- or pAd-hPEBP4-transduced DCs as described in the Materials and methods. Splenocytes from immunized mice were restimulated in vitro with P40–48 for 7 days. SSp-1 was used as a nonspecific control peptide. (a) IFN-γ-positive SFCs/10⁶ splenocytes in ELISPOT assay. (b) IFN-γ-staining of CD8⁺ T cells (flow cytometry data are representative of three independent experiments). (c and d) Cytotoxicity measured by standard ⁵¹Cr-release assays at the indicated E:T ratio. (c) P40–48-pulsed T2 and (d) MCF-7 (hPEBP4⁺/HLA-A2.1⁺) cells were used as peptide-specific targets, while (c) SSp-1-pulsed T2 cells or T2 cells alone and (d) A549 (hPEBP4⁻/HLA-A2.1⁺) or Daudi (hPEBP4⁻/HLA-A2.1⁻) cells were used as control targets. Data are presented as the mean±s.e.m. of three independent experiments. **P<0.01; *P<0.05. CTL, cytotoxic T lymphocyte; DC, dendritic cell; ELISPOT, enzyme-linked immunospot; E:T, effector:target; hPEBP4, human phosphatidylethanolamine-binding protein 4; IFN, interferon; PE, phycoerythrin; SFC, spot-forming cell; Tg, transgenic.

In cytotoxicity assays, T2 cells pulsed with relevant hPEBP4-derived peptides were used to measure peptide-specific lysis of bulk CTLs, and T2 cells pulsed with the nonspecific peptide SSp-1 were used to measure nonspecific lysis. As shown in Figure 1c, only CTLs from pAd-hPEBP4- and P40–48-immunized mice could efficiently lyse P40–48-pulsed T2 cells (but not T2 cells alone or Spp-1-pulsed T2 cells). Cytotoxic effects of P12–20- and P6–14-specific CTLs against the related target cells were either less than those for P40–48 or near background levels (data not shown). Thus, P12–20 and P6–14 were excluded from further study. These results indicated that the CTL responses were peptide P40–4-specific.

Peptide P40–48 is a naturally processed peptide from hPEBP4

To address whether the immunogenic peptide P40–48 was naturally processed and presented by tumor cells, and may therefore serve as a tumor rejection antigen, we determined the capacity of peptide-specific CTLs to recognize HLA-A2.1⁺ tumor cells endogenously expressing the hPEBP4 protein. Previous work¹ has shown that the tumor cell lines MCF-7 and Daudi express hPEBP4, whereas A549 cells do not. Analysis of HLA-A2.1 expression by flow cytometry revealed that MCF-7 and A549 cells were HLA-A2.1⁺, and Daudi cells were HLA-A2.1⁻. As shown in Figure 1d, only CTLs directed against peptide P40–48 were able to specifically lyse MCF-7 cells expressing both hPEBP4 and HLA-A2.1 molecules. No lysis of Daudi cells (hPEBP4⁺/HLA-A2.1⁻) or A549 cells (hPEBP4⁻/HLA-A2.1⁺) was observed. This indicates that peptide P40–48 is not only immunogenic and naturally processed but also an HLA-A2.1-restricted CTL epitope in HLA-A2.1/K^b Tg mice.

P40–48-specific, HLA-A2.1-restricted CD8⁺ CTL responses can be induced in vitro from PBLs isolated from breast cancer patients

To investigate the ability of peptide P40–48 to induce tumor-specific CD8⁺ T-cell responses in vitro in HLA-A2.1⁺ breast cancer patients, specific CTLs were generated by in vitro sensitization of PBLs from HLA-A2.1⁺/hPEBP4⁺ breast cancer patients with autologous DCs prepulsed with P40–48. In ~ 73% (8/11) of the hPEBP4⁺/HLA-A2.1⁺ breast cancer patients tested, P40–48-specific CD8⁺ T cells were generated, eliciting marked IFN-γ secretion upon stimulation with P40–48-pulsed T2 or MCF-7 cells (hPEBP4⁺/HLA-A2.1⁺) (Figures 2a and b). Reduced or scant levels of IFN-γ were detected upon stimulation with T2 cells pulsed with a nonspecific SSp-1 peptide or A549 (hPEBP4⁻/HLA-A2.1⁺) or Daudi (hPEBP4⁺/HLA-A2.1⁻) tumor cells.

Figure 2.

Cytotoxicity of CD8⁺ CTLs raised by in vitro stimulation of PBLs of HLA-A*0201-positive breast carcinoma patients. CTLs were induced with autologous P40–48-pulsed DCs from PBLs of hPEBP4⁺/HLA-A2.1⁺ breast carcinoma patients. On day 7 after the third stimulation, CD8⁺ T cells were enriched by positive selection using magnetic immunobeads for analysis. (a) IFN-γ-positive SFCs/10⁶ CD8⁺ T cells detected by cytokine-specific ELISPOT. (b) IFN-γ staining of CD8⁺ T cells (flow cytometry data are representative of three independent experiments). (c) Cytotoxicity measured by ⁵¹Cr-release assays at the indicated E:T ratio. P40–48-pulsed T2 cells and MCF-7 cells (hPEBP4⁺/HLA-A2.1⁺) were used as peptide-specific targets, while SSp-1-pulsed T2 cells or T2 cells alone and A549 (hPEBP4⁻/HLA-A2.1⁺) or Daudi (hPEBP4⁻/HLA-A2.1⁻) cells were used as controls. Data are presented as the mean±s.e.m. of three independent experiments. **P<0.01; *P<0.05. CTL, cytotoxic T lymphocyte; DC, dendritic cell; ELISPOT, enzyme-linked immunospot; E:T, effector:target; hPEBP4, human phosphatidylethanolamine-binding protein 4; IFN, interferon; PBL, peripheral blood lymphocyte; PE, phycoerythrin; SFC, spot-forming cell; Tg, transgenic.

In ⁵¹Cr-release assays, the cytotoxicity of these activated bulk CTLs was evaluated against T2 cells loaded with either the peptide P40–48 or the nonspecific peptide SSp-1. As shown in Figure 2c, DCs loaded with peptide P40–48 were capable of inducing peptide-specific bulk CTLs. In addition, to investigate whether peptide P40–48 was naturally processed and presented, peptide-specific CTLs were evaluated for their ability to lyse target cells endogenously expressing hPEBP4. As observed in HLA-A2.1/K^b Tg mice, bulk CTLs against peptide P40–48 exhibited specific cytotoxicity to MCF-7 cells (hPEBP4⁺/HLA-A2.1⁺), but not Daudi (hPEBP4⁺/ HLA-A2.1⁻) or A549 cells (hPEBP4⁻/HLA-A2.1⁺) (Figure 2d).

To further confirm whether the immunogenic P40–48 peptide could be naturally processed and presented in tumor cells, we used MCF-7 cells (HLA-A2.1⁺, hPEBP4⁺) transiently transfected with an hPEBP4-specific small interfering RNA duplex⁸ and A549 cells (HLA-A2.1⁺, hPEBP4⁻) infected with pAd-hPEBP4 as target cells. As shown in Figure 3, the CTLs were capable of lysing hPEBP4-overexpressing A549 cells, but not the hPEBP4-silenced MCF-7 cells. These results further confirmed that hPEBP4-derived P40–48 could be naturally processed and presented in the context of HLA-A2.1 molecules by tumor cells.

Figure 3.

Changes in the levels of hPEBP4 affect the cytotoxic effects of P40–48-specific CTLs against cancer cells. hPEBP4-siRNA duplex and hPEBP4-expressing vectors were transfected into MCF-7 (MCF-7/si-hPEBP4) and A549 cells (A549+pAD- hPEBP4), respectively. Mock vectors served as controls. CD8⁺ CTLs from breast carcinoma patients were generated as described in Figure 2 and were tested for P40–48-specific lysis using a standard 4-h ⁵¹Cr-release assay. CTL-mediated specific lysis in response to hPEBP4-silenced MCF-7 cells (a) and hPEBP4-overexpressing A549 cells (b). Data are presented as the mean±s.e.m. of three independent experiments. *P<0.05. CTL, cytotoxic T lymphocyte; E:T, effector:target; hPEBP4, human phosphatidylethanolamine-binding protein 4; siRNA, small interfering RNA.

Modified peptide analogs exhibit high affinity to HLA-A2.1

Enhancement of the immunogenicity of a CTL epitope may be achieved by substitution of residues from the epitope that contact MHC class I molecules and/or residues that contact the T-cell receptor. Previous studies have indicated that residues at positions 1, 2, 3, 6, 7 and 9 may be replaced.^24,25,26 Residues at positions 2 and 9 are primary anchor motifs, and those at positions 1, 3, 6 and 7 are secondary anchor motifs. The native P40–48 peptide already contains preferred residues (leucine (L), valine (V)) at positions 2 and 9; we therefore used in silico analysis to determine further potential improvement in the ability of the peptide to bind to HLA-A0201 molecules. Replacement of the secondary anchor motif may increase the binding and affinity of the peptide to HLA-A*0201 molecules, including at position 1 by tyrosine (Y), phenylalanine (F), leucine (L), isoleucine (I) or valine (V); at position 3 by Y, F, L, I or tryptophan (W); at position 6 by Y, L, I, V or threonine (T); or at position 7 by proline (P), histidine (H), W, F or I. After these alterations, computer algorithm prediction showed that the binding affinity of modified peptides to HLA-A*0201 increased, up to 61 times that of the native peptide. Seventeen candidate nonameric peptides with the highest estimated binding affinities to HLA-A*0201 were selected and synthesized (Table 2).

Table 2.

HLA-A*0201 native peptides and synthetic analogs

Peptide	Sequence	Score a
P40–48	TLFCQGLEVb	257.34
1F3W6V7F	FLWCQVFEV	15791.56
1Y6L7F	YLFCQLFEV	15791.56
1Y6V7F	YLFCQVFEV	15791.56
1Y3W7F	YLWCQGFEV	15401.87
1Y7W	YLFCQGWEV	8878.51
1F7F	FLFCQGFEV	6865.90
1Y7F	YLFCQGFEV	6865.90
1Y3W6L	YLWCQLLEV	6107.64
1F3W6I	YLWCQILEV	6107.64
1Y6V	YLFCQVLEV	2722.68
7W	TLFCQGWEV	1930.07
7F	TLFCQGFEV	1492.59
1Y7Y	YLFCQGYEV	1183.78
1F	FLFCQGLEV	1183.78
1Y	YLFCQGLEV	1183.78
1L	LLFCQGLEV	437.48
7H	TLFCQGHEV	257.34

Modifications from the native sequence are indicated in bold.

Abbreviation: hPEBP4, human phosphatidylethanolamine-binding protein 4.

^aPropred-I prediction software available at http://crdd.osdd.net/raghava//propred1/.

^bNative peptide derived from hPEBP4.

High stability of the peptide analog/HLA-A2.1 molecule complex

The immunogenicity of MHC class I-restricted peptides requires the capacity to bind and stabilize MHC class I molecules on the cell surface. Therefore, we next sought to measure directly the interaction strength between the peptides and HLA-A2.1 using a conventional binding and stabilization assay with T2 cells. Of the 17 peptide analogs, 16 (except 7H) had stronger affinity than the native peptide P40–48 (Table 3).

Table 3.

Binding affinity of modified P40–48 peptides to HLA-A*0201 molecules

Peptide	Sequence	FI a,b
P40–48	TLFCQGLEV	0.48±0.22
1Y3W6L	YLWCQLLEV	1.48±0.68
1Y6V7F	YLFCQVFEV	1.39±0.82
1F3W6I	YLWCQILEV	1.37±0.85
1F3W6V7F	FLWCQVFEV	1.27±0.45
1F7F	FLFCQGFEV	1.09±0.62
1Y3W7F	YLWCQGFEV	1.00±0.32
1Y6V	YLFCQVLEV	0.97±0.35
1Y7W	YLFCQGWEV	0.96±0.42
1Y6L7F	YLFCQLFEV	0.94±0.62
1Y	YLFCQGLEV	0.92±0.8
1F	FLFCQGLEV	0.85±0.44
1L	LLFCQGLEV	0.84±0.62
1Y7Y	YLFCQGYEV	0.77±0.22
1Y7F	YLFCQGFEV	0.75±0.23
7F	TLFCQGFEV	0.71±0.58
7W	TLFCQGWEV	0.63±0.51
7H	TLFCQGHEV	0.18±0.21

Residues in bold represent modifications from the native sequence.

Abbreviations: FI, fluorescence index; MFI, mean fluorescence intensity.

^aPeptide binding was evaluated by the MFI.

^bData are presented as the mean±s.e.m. of six independent experiments.

A stable peptide-MHC complex could facilitate the formation of the synapses between T cells and antigen-presenting cells, and the stability of the peptide-MHC complex is a key factor for CTL activation.²² Thus, the stability of the peptide-MHC complex with wild-type or peptide analogs was investigated. The results (Table 4) showed that most peptide analogs (except 7W, 7H and 7F) had a higher stability with HLA-A2.1 molecules than the native peptide did over a 6-h observation period.

Table 4.

Stability of peptide-HLA-A*0201 complexes

Peptide	Time post-brefeldin A treatment (h)
	2	4	6
P40–48	79±7.8	72±3	55±5
1Y3W6L	95±3	88±6	79±3
1Y6V7F	93±4	85±5	77±7
1F3W6I	92±5	85±5	77±2
1F3W6V7F	90±5	81±7	73±6.5
1F7F	90±5	81±3	72±4
1Y6L7F	85±7	79±7	69±7
1F	84±4	75±6	69±9
1Y3W7F	84±8	76±6	67±7
1Y7W	80±7	75±5.5	66±8
1Y6V	82±8	77±7	64±4
1Y	84±8	74±5	63±6
1Y7Y	82±7	73±7	61±9
1Y7F	81±9	75±5	60±5
1L	80±3	71±5	59±9
7W	80±5	60±10	51±4
7F	70±5	59±9	51±6
7H	65±5	55±3	49±8

Results are expressed as the MFI, defined as 100% at 0 h; the MFI from other time points are relative to 0 h.

Data are presented as the mean±s.e.m. of six independent experiments.

Abbreviation: MFI, mean fluorescence intensity.

Peptide analogs can induce peptide-specific CD8⁺ CTLs in HLA-A2.1/K^b Tg mice in vivo

To determine whether the P40–48 analogs would be more efficient in inducing immunity in vivo, we induced peptide-specific CD8⁺ CTLs in HLA-A2.1/K^b Tg mice and then assayed them for IFN-γ production and CTL cytotoxicity. Once restimulated in vitro with the wild-type (WT) peptide, the splenocytes from 1Y3W6L-, 1F7F-, 1Y6V- and 1F-immunized Tg mice showed stronger IFN-γ production than the WT peptide (Figures 4a and b). However, bulk CTLs from other analog-immunized mice did not show significant IFN-γ production compared with the WT peptide.

Figure 4.

1Y3W6L, 1F7F, 1Y6V and 1F analogs induce more efficient hPEBP4-specific CTLs than the native P40–48 epitope in HLA-A2.1/K^b Tg mice. Splenocytes from HLA-A2.1/K^b Tg mice (seven mice per group) immunized with native P40–48- or analog-pulsed DCs were restimulated in vitro with P40–48 for 7 days. Ex vivo IFN-γ ELISA (a) and IFN-γ ELISPOT (b) of splenocytes. Data are presented as the mean±s.e.m. of three independent experiments. **P<0.01, *P<0.05. Cytotoxicity against P40–48-pulsed T2 cells examined by (c) granzyme B ELISPOT and (d) ⁵¹Cr-release assay. In (c), the number of granzyme B-positive SFCs from 10⁵ splenocytes is calculated. In (d), P40–48-pulsed T2, full-length hPEBP4-transfected MAD-MB-468 (MAD-MB-468-hPEBP4, HLA-A2.1⁺) and MCF-7 (hPEBP4⁺/HLA-A2.1⁺) cells were used as hPEBP4-specific HLA-A2.1-restricted target cells. SSp-1 peptide-pulsed T2, non-pulsed T2, Daudi (hPEBP4⁺, HLA-A2.1⁻) and MAD-MB-468 (hPEBP4⁻/HLA-A2.1⁺) cells were used as controls. The cytotoxicity of various CTLs against the corresponding target cells was tested at a 10:1 E:T ratio in the granzyme B ELISPOT assay and at a 50:1 E:T ratio in the ⁵¹Cr-release assay. Data are presented as the mean±s.e.m. of three independent experiments. **P<0.01, *P<0.05. All results. CTL, cytotoxic T lymphocyte; DC, dendritic cell; ELISA, enzyme-linked immunosorbent assay; ELISPOT, enzyme-linked immunospot; E:T, effector:target; hPEBP4, human phosphatidylethanolamine-binding protein 4; IFN, interferon; SFC, spot-forming cell; siRNA, small interfering RNA.

To address whether IFN-γ-producing CTL lines could lyse target cells pulsed with WT peptide or tumor cells expressing hPEBP4 protein, granzyme B ELISPOT and ⁵¹Cr cytotoxicity assays were performed. As shown in Figures 4c and d, 1Y3W6L-, 1F7F-, 1Y6V- and 1F-specific CTLs showed increased capacity to release granzyme B and to lyse native peptide-pulsed T2 cells. Interestingly, only 1Y3W6L and 1Y6V induced stronger CTL cytotoxicity against MCF-7- and hPEBP4-overexpressing MAD-MB-468 cells (MAD-MB-468-hPEBP4) (both HLA-A2.1⁺). CTLs induced with the remaining analogs showed either less effective or background-level cytotoxicity. Therefore, these other analogs were excluded from further studies. Overall, these results indicated that the 1Y3W6L and 1Y6V could induce stronger hPEBP4-specific CTL responses than the native peptide.

Peptide analogs induce enhanced peptide-specific CD8⁺ CTLs responses from breast cancer patients in vitro

To determine whether the immunogenicity of these substituted peptides was also enhanced in the induction of CTLs from breast cancer patients, PBLs from 17 HLA-A2.1⁺ breast cancer patients were stimulated with autologous DCs pulsed with either analog or native peptide, as described in the Materials and methods. Seven days after the third in vitro stimulation, CD8⁺ T cells from individual patients were tested for IFN-γ production and CTL cytotoxicity. In IFN-γ release assays, CTLs raised by the 1Y3W6L, 1F7F, 1Y6V, 1F and P40–48 peptides were challenged with native peptide-pulsed T2 cells or MAD-MB-468-hPEBP4 cells as stimulators. More than 35% (6/17 patients) of CTLs raised by 1Y3W6L and 65% (11/17 patients) of CTLs raised by 1Y6V were able to generate elevated levels of IFN-γ production against native peptide-pulsed T2 cells or MAD-MB-468-hPEBP4 cells, whereas 1F7F and 1F failed to elicit CTL responses to various stimulators (Figures 5a and b).

Figure 5.

1Y3W6L and 1Y6V analogs induced enhanced CD8⁺ T-cell responses in PBLs from HLA-A*0201-positive breast cancer patients. PBLs from HLA-A*0201-positive breast cancer patients were stimulated with 1Y3W6L, 1F7F, 1Y6V, 1F or native P40–48 peptides three times at weekly intervals as described in the Materials and methods. On day 7 after the third stimulation, CD8⁺ T cells were enriched by positive selection using magnetic immunobeads for analysis (a–d). The CTLs in PBLs were then evaluated for IFN-γ release (a and b), granzyme B release (c) and ⁵¹Cr-release (d) against T2 cells prepulsed with P40–48 or hPEBP4-expressing HLA-A*0201-positive tumor cells. The number of IFN-γ-positive and granzyme B-positive SFCs was determined from 10⁶ (b) or 10⁵ (c) CD8⁺ T lymphocytes, respectively. (d) P40–48-pulsed T2, MAD-MB-468-hPEBP4 (hPEBP4⁺/HLA-A2.1⁺) and MCF-7 (hPEBP4⁺/HLA-A2.1⁺) cells were used as the hPEBP4-specific HLA-A2.1-restricted target cells. SSp-1 peptide-pulsed T2, non-pulsed T2, Daudi (hPEBP4⁺/HLA-A2.1⁻) and MAD-MB-468 (hPEBP4⁻/HLA-A2.1⁺) cells were used as controls. The cytotoxicity of various CTLs against the corresponding target cells was tested at a 10:1 E:T ratio in the granzyme B ELISPOT assay and at a 50:1 E:T ratio in the ⁵¹Cr-release assay. Data are presented as the mean±s.e.m. of three independent experiments. **P<0.01, *P<0.05. CTL, cytotoxic T lymphocyte; DC, dendritic cell; ELISPOT, enzyme-linked immunospot; E:T, effector:target; hPEBP4, human phosphatidylethanolamine-binding protein 4; IFN, interferon; PBL, peripheral blood lymphocyte; SFC, spot-forming cell; siRNA, small interfering RNA.

We further evaluated CTL cytotoxicity using granzyme B ELISPOT and ⁵¹Cr-release assays. Compared with CTLs generated with native P40–48 peptide, CTLs generated in vitro in the presence of 1Y3W6L or 1Y6V were able to specifically lyse native peptide-loaded T2 cells, MCF-7 and MAD-MB-468-hPEBP4 cells (both hPEBP4⁺/HLA-A2.1⁺), but not T2 cells pulsed with a nonspecific peptide or without peptide, MAD-MB-468 cells (hPEBP4⁻/HLA-A2.1⁺) or Daudi cells (hPEBP4⁺/HLA-A2.1⁻). More importantly, T cells stimulated with 1Y3W6L or 1Y6V were able to recognize the native sequence, confirming the necessary heteroclitic response. In contrast, CTLs induced by 1F7F and 1F peptides generated no significant immune response against various targets (Figures 5c and d).

To confirm the increased immunogenicity of 1Y3W6L and 1Y6V analogs, the cytolytic activity of CTLs raised via the two analogs were assessed against T2 cells pulsed with different concentrations (100, 10 or 1 μm) of P40–48 peptide at a 25:1 effector:target (E:T) ratio and MAD-MB-468-hPEBP4 cells at various E:T ratios (12.5:1, 25:1 and 50:1). Remarkably, when recognizing T2 cells loaded with WT peptide at various concentrations, especially 10 or 1 μm, 1Y3W6L- or 1Y6V-induced CTLs showed stronger cytotoxicity compared with the native peptide. (Figure 6a). Regarding the E:T ratio, the ⁵¹Cr-release assay demonstrated that there was an obvious increase in the lysis of tumor cells for 1Y3W6L- or 1Y6V-induced CTLs, compared with WT (Figure 6b).

Figure 6.

1Y3W6L and 1Y6V analogs induced more effective CTLs that recognized the native peptide and hPEBP4-expressing HLA-A*0201-positive tumor cells. CD8⁺ T cells from different HLA-A*0201-positive breast cancer patients were restimulated in vitro after induction with 1Y3W6L, 1Y6V, or native P40–48 peptides, as described in the Materials and methods, and examined for cytotoxicity by ⁵¹Cr-release assays. (a) Cytotoxicity of P40–48, 1Y3W6L- and 1Y6V-induced CTLs against T2 cells prepulsed with various concentrations of native peptide (100, 10 or 1 μM) at a 25:1 E:T ratio. (b) Cytolytic activity of CTLs against MAD-MB-468-hPEBP4 cells at various E:T ratios (50:1, 25:1 and 12.5:1). (c) PBLs from HLA-A*0201-positive breast cancer patients were stimulated with autologous DCs loaded with various concentrations (100, 10 or 1 μM) of 1Y3W6L, 1Y6V or P40–48. Seven days after the third stimulation, purified CD8⁺ T cells from each individual were tested against T2 cells pulsed with 20 μm native peptide at a 25:1 E:T ratio. Data are presented as the mean±s.e.m. of three independent experiments. P<0.05. All results. CTL, cytotoxic T lymphocyte; DC, dendritic cell; E:T, effector:target; hPEBP4, human phosphatidylethanolamine-binding protein 4; IFN, interferon; PBL, peripheral blood lymphocyte.

We further determined the capacity of 1Y3W6L and 1Y6V analogs to induce peptide-specific CTLs by stimulating PBLs of HLA-A2.1⁺ breast cancer patients with autologous DCs loaded with various concentrations (100, 10 or 1 μm) of 1Y3W6L, 1Y6V, or control P40–48 peptides. Seven days after the third stimulation, these CTL activities against P40–48-pulsed T2 cells at a 25:1 E:T ratio were measured via ⁵¹Cr-release assays. As shown in Figure 6c, at the lower concentration (1 μm), P40–48 did not elicit any specific CTL response; in contrast, 1 μm of both 1Y3W6L and 1Y6V were surprisingly able to efficiently elicit CTL responses to the P40–48 peptide when cultures were stimulated with 1 μm of the corresponding peptide

Taken together, our results demonstrate that 1Y3W6L and 1Y6V can induce HLA-A2.1-restricted CTLs that recognize both P40–48-loaded T2 cells and hPEBP4-expressing tumor cells and produce higher IFN-γ levels and stronger cytotoxicity effects than native peptide.

Adoptive immunotherapy in C57BL/6^nu/nu mice bearing human breast carcinomas

To ascertain whether 1Y3W6L and 1Y6V peptides were capable of serving as effective vaccines against tumor growth in vivo, we established a model of adoptive transfer of HLA-A2.1/K^b Tg mouse-derived lymphocytes into C57BL/6^nu/nu mice bearing breast carcinomas. As shown in Figure 7a, tumor growth was greatly delayed in mice treated with 1Y3W6L- and 1Y6V-induced lymphocytes compared with native peptide-induced lymphocytes, whereas nonspecific peptide-induced CTLs did not prevent tumor growth. Furthermore, in 1Y3W6L and 1Y6V vaccination groups, 3/7 (43%) 1Y3W6L- and 4/7 (57%) 1Y6V-vaccinated mice exhibited extended, long-term survival (more than 80 days) following tumor inoculation (Figure 7b). However, only 2/7 (29%) mice treated with the P40–48 vaccine survived for more than 80 days. All mice in the control groups succumbed between days 26 and 40. Thus, these findings suggested that vaccination with 1Y3W6L or 1Y6V peptide induced effective in vivo antitumor responses that were superior to those induced by the native peptide.

Figure 7.

Adoptive immunotherapy of MAD-MB-468-hPEBP4 tumor-bearing nude mice. MAD-MB-468-hPEBP4 tumor cells (5 × 10⁶ cells/mouse) were injected into the mammary fat pads of female C57BL/6^nu/nu mice. At day 4 after tumor implantation, tumor-bearing mice received intravenous injections of splenocytes (1 × 10⁸ cells per mouse) from 1Y3W6L-, 1Y6V- or P40–48-immunized HLA-A2.1/K^b Tg mice, as described in the Materials and methods section. Control groups were treated with IL-2 alone or received no treatment. (a) Tumor growth was observed at regular intervals every 3 days and recorded as the mean tumor sizes (mm²). (b) Survival of mice after tumor inoculation (n=7 mice per group). hPEBP4, human phosphatidylethanolamine-binding protein 4; IL, interleukin; Tg, transgenic.

Discussion

In the present work, based on prediction and verification using computer algorithms and MHC peptide-binding assays, we identified an hPEBP4-derived HLA-A2.1 epitope, p40–48 (TLFCQGLEV), as an immunogenic, naturally processed and HLA-A2.1-restricted CTL epitope that can induce peptide-specific CTL responses in HLA-A2.1/k^b Tg mice and in culture with PBLs from HLA-A2.1⁺ breast cancer patients. In addition, amino-acid substitutions of the P40–48 sequence were able to increase its immunogenicity against hPEBP4. Peptide analogs 1Y3W6L and 1Y6V were shown to induce enhanced tumor control during adoptive immunotherapy in a murine tumor-bearing model.

In clinical trials of peptide-based immunotherapy, it has been demonstrated that tumor antigen-derived peptides can successfully induce antigen-specific CTLs against cancer.^27,28,29,30 We have previously reported^1,2,3,4 that hPEBP4 is selectively overexpressed in several types of tumor cells and promotes growth. These findings imply that hPEBP4 may be serve as a CTL-directed tumor antigen. In this study, we demonstrated that an hPEBP4-derived epitope (P40–48) with high affinity to HLA-A*0201 could induce antitumor CTL responses in HLA-A2.1/K^b Tg mice and PBLs of HLA-A*0201-matched patients with breast cancer.

Owing to the poor immunogenicity of CTL epitopes derived from tumor antigens, altering the amino-acid residues that interact with either HLA molecules or the T-cell receptor is one way to enhance immunogenicity.^24,25,26 Such a strategy has been demonstrated to significantly increase protective and therapeutic immunity against tumors in clinical trials.^31,32 Here, we engineered epitope-enhanced analogs of the native P40–48 peptide by replacing amino acids at positions 1, 3, 6 and 7, which are known to contain key residues. Interestingly, several of these peptide analogs exhibited enhanced binding to the HLA-A2.1 allele, as well as enhanced stability of the peptide-MHC complex. We therefore determined whether they could also elicit elevated antitumor immune responses compared with the native P40–48 peptide in vivo. In HLA-A2.1/K^b Tg mice, the 1Y3W6L and 1Y6V peptide analogs induced stronger CTL responses than the native peptide. These results are consistent with previous reports that improving the binding affinity of peptides to MHC class I can result in enhanced immunogenicity of epitopes.^{33,34,35,36,37}

Increasing the density of peptide-MHC complexes on antigen-presenting cells is important for enhancing in vivo immunogenicity. We demonstrated that the 1Y3W6L and 1Y6V peptides elicited increased immune responses against T2 cells pulsed with a low concentration of native peptide and against tumor cells endogenously expressing hPEBP4 at low E:T ratios. Thus, these CTLs may be superior to those raised with the native peptide in terms of efficiency and quantity. These findings suggest that the peptide analogs 1Y3W6L and 1Y6V have several advantages over the native peptide, including increased immunogenicity and enhanced therapeutic efficacy as part of an antitumor vaccine.

A higher-affinity binding to HLA molecules has been confirmed to significantly increase T-cell responses in vitro, but does not always result in a parallel clinical benefit.³⁸ Therefore, characterizing the in vivo immunogenicity of peptides that can function as immunotherapeutic agents is essential. Using a MAD-MB-468-hPEBP4 tumor-bearing nude mouse model, we demonstrated that CTLs raised from 1Y3W6L and 1Y6V peptides more efficiently controlled tumor growth and prolonged survival of tumor-bearing mice than the native peptide.

As a recently identified antiapoptotic molecule, hPEBP4 is preferentially expressed in several tumor cells, and its expression is further enhanced upon tumor necrosis factor-α treatment, which indicates that it is an overexpressed self-antigen.⁸ However, although overexpressed self-antigens involved in oncogenesis can represent useful targets for antitumor immunotherapy, they may also raise self-tolerance or autoimmunity against these antigens.^39,40 Mullins et al. ⁴¹ demonstrated that CTLs directed against melanocyte differentiation protein-derived epitopes (self-antigen) could be activated to elicit effective antitumor immunity in a preclinical model using Tg mice that expressed a recombinant MHC class I molecule. In the present study, we also induced CTLs against hPEBP4-expressing tumor cells from the PBLs of breast cancer patients, which suggested that the hPEBP4-specific CTL repertoire was not completely tolerized to the hPEBP4 protein. Furthermore, T lymphocytes with strong ‘self’-reactivity are often physically deleted during maturation in the thymus.⁴² However, T lymphocytes with low reactivity to autoantigens persist due to positive selection in the thymus, which requires a low level of autoreactivity.^42,43 To evaluate the risk of autoimmunity induction by overexpressed self-antigen-derived peptide vaccinations, several preclinical and clinical trials have demonstrated that peptides derived from overexpressed self-antigens were able to induce effective and specific CTL responses without any serious systemic adverse events.^44,45,46,47 However, studies by Overwijk and Restifo⁴⁸ and Overwijk et al. ⁴⁹ revealed that potent CD8⁺ T-cell responses induced with MDA in melanoma patients could interact with melanocytes in the skin, resulting in autoimmune depigmentation. Therefore, if hPEBP4-derived peptides are to serve as candidates for CTL-based immunotherapy against breast cancer or other hPEBP4-expressing carcinomas, their safety and tolerability need to be further assessed in basic and clinical studies.

In conclusion, we characterized P40–48 as a novel HLA-A*0201-restricted, immunogenic hPEBP4-derived CTL epitope that is naturally processed and presented by tumor cells. Additional replacement of secondary anchor residues within P40–48 resulted in peptides 1Y3W6L and 1Y6V, which had enhanced in vivo and in vitro immunogenicity compared with the native peptide. Taken together, these findings indicate that highly immunogenic P40–48 peptide analogs are potentially useful for the development of vaccines that are capable of effectively enhancing hPEBP4-specific, tumor-reactive CTL responses in hPEBP4-expressing tumor patients.⁵⁰

Conflict of interest

The authors declare no conflict of interest.