HLA-A2-restricted Cytotoxic T Lymphocyte Epitopes from Human Hepsin as Novel Targets for Prostate Cancer Immunotherapy


  • J. Guo,

    1. The Research Center of Stem Cell, Tissue and Organ Engineering, Kunming General Hospital of PLA, Kunming, China
    Search for more papers by this author
  • G. Li,

    1. Department of Hepatobiliary Surgery, The First People's Hospital of Liangshan Prefecture, Sichuan, China
    Search for more papers by this author
  • J. Tang,

    1. Department of Gastroenterology, the 105th hospital of Nanjing Military Command, Hefei, Anhui Province, China
    Search for more papers by this author
  • X.-B. Cao,

    1. The Research Center of Stem Cell, Tissue and Organ Engineering, Kunming General Hospital of PLA, Kunming, China
    Search for more papers by this author
  • Q.-Y. Zhou,

    1. Department of Urinary Surgery, Kunming General Hospital of PLA, Yunnan, China
    Search for more papers by this author
  • Z.-J. Fan,

    1. The Research Center of Stem Cell, Tissue and Organ Engineering, Kunming General Hospital of PLA, Kunming, China
    Search for more papers by this author
  • B. Zhu,

    Corresponding author
    1. Department of Gastroenterology, the 105th hospital of Nanjing Military Command, Hefei, Anhui Province, China
    • Correspondence to: B. Zhu, M.D., Department of Gastroenterology, the 105th hospital of Nanjing Military Command, Hefei 230031, Anhui Province, China. E-mail: binzhu105@yeah.net or X. -H. Pan, M.D., The Research Center of Stem Cell, Tissue and Organ Engineering, Kunming General Hospital of Chendu Military Command, Kunming 650032, Yunnan Province, China. E-mail: xinghuapan@yahoo.com.cn

    Search for more papers by this author
  • X.-H. Pan

    Corresponding author
    1. The Research Center of Stem Cell, Tissue and Organ Engineering, Kunming General Hospital of PLA, Kunming, China
    • Correspondence to: B. Zhu, M.D., Department of Gastroenterology, the 105th hospital of Nanjing Military Command, Hefei 230031, Anhui Province, China. E-mail: binzhu105@yeah.net or X. -H. Pan, M.D., The Research Center of Stem Cell, Tissue and Organ Engineering, Kunming General Hospital of Chendu Military Command, Kunming 650032, Yunnan Province, China. E-mail: xinghuapan@yahoo.com.cn

    Search for more papers by this author

  • Jun Guo, Guo Li and Jun Tang contributed equally to this work.


Hepsin is a type II transmembrane serine protease that is overexpressed in prostate cancer, and it is associated with prostate cancer cellular migration and invasion. Therefore, HPN is a biomarker for prostate cancer. CD8+ T cells play an important role in tumour immunity. This study predicted and identified HLA-A2-restricted cytotoxic T lymphocyte (CTL) epitopes in human hepsin protein. HLA-A2-restricted CTL epitopes were identified using the following four-step procedure: (1) a computer program generated predicted epitopes from the amino acid sequence of human hepsin; (2) an HLA-A2-binding assay detected the affinity of the predicted epitopes to the HLA-A2 molecule; (3) the primary T cell response against the predicted epitopes was stimulated in vitro; and (4) the induced CTLs towards different types of hepsin- or HLA-A2-expressing prostate cancer cells were detected. Five candidate peptides were identified. The effectors that were induced by human hepsin epitopes containing residues 229 to 237 (Hpn229; GLQLGVQAV), 268 to 276 (Hpn268; PLTEYIQPV) and 191 to 199 (Hpn199; SLLSGDWVL) effectively lysed LNCaP prostate cancer cells that were hepsin-positive and HLA-A2 matched. These peptide-specific CTLs did not lyse normal liver cells with low hepsin levels. Hpn229, Hpn268 and Hpn199 increased the frequency of IFN-γ-producing T cells compared with the negative peptide. These results suggest that the Hpn229, Hpn268 and Hpn199 epitopes are novel HLA-A2-restricted CTL epitopes that are capable of inducing hepsin-specific CTLs in vitro. Hpn229, Hpn268 and Hpn199 peptide-based vaccines may be useful for immunotherapy in patients with prostate cancer.


Prostate cancer (PC) is a major medical problem in the male population [1]. No curative treatment for metastatic prostate cancer is available despite advances in chemotherapy and androgen-deprivation therapy for patients with PC. Therefore, novel approaches for prostate cancer treatment are essential.

The hepsin (HPN) gene is prominently overexpressed in PC [2, 3]. HPN is a member of the hepsin/TMMPRSS/enteropeptidase subfamily within the type II transmembrane serine proteases [4]. The physiological function of hepsin is not clear, but HPN may play a role in PC cellular migration and invasion [5]. HPN cleaves several biological substrates, including blood coagulation factor VII [6], prohepatocyte growth factor [7], prourokinase-type plasminogen activator [8], epidermal growth factor receptor, the β3-subunit (LAMB3) of laminin332 (Ln332)[9] and promacrophage-stimulating protein (pro-MSP)[10]. HPN is a biomarker for PC [11]. Constitutive prostate-specific overexpression of HPN in transgenic mice produces a conspicuous disorganization of the prostate ECM [12]. Hepsin overexpression plays an important role in the promotion of prostate cancer metastasis and progression [12, 13]. HPN up-regulation in transgenic mice produces the transition of non-metastatic cancer into an aggressive carcinoma with metastases to bone, liver and lung [12]. HPN overexpression is generally noted in primary PC, which increases the rate of mortality [3]. Therefore, HPN is a potential therapeutic target for prostate cancer treatment.

A central tenet of cancer immunotherapy is generating tumour-associated antigen-specific cytotoxic T lymphocytes (CTLs) to lyse tumour cells. Antigen presentation is critical for the initiation of the adaptive immune response against tumours. Dendritic cells (DCs) are the most potent professional antigen-presenting cells with the strongest antigen-presenting capacity [14, 15]. CTLs are the chief mediators of tumour immunosurveillance via the recognition of the TAAs as cognate peptides that are bound to the MHC molecules on the surface of cancer cells. CTL epitope binding to MHCs, rather than integral proteins, induces CTL reactions. CTL epitopes are usually eight to eleven amino acids long. These epitopes possess two to three primary anchor residues that interact with MHC class I molecules and two to three amino acid residues that bind to the T cell receptor [16]. Numerous TAAs have been characterized, and the CTL epitopes from these TAAs have been identified. Several prostate cancer–associated antigens have been identified in recent years [17-19]. Different tissue-specific antigens have been employed for PC immunotherapy, which has produced promising clinical benefits in various patient populations [20, 21]. Nevertheless, clinical vaccination trials and promising clinical responses remain infrequent; therefore, tumour vaccination has not fulfilled its potential. These disappointing results are likely due to the ability of tumours to escape immune surveillance by the loss or downregulation of tumour-associated antigen expression. Therefore, the targeting of just one TAA for tumour immunotherapy may be insufficient to inhibit tumour growth, but the targeting of multiple TAAs for tumour immunotherapy may improve therapeutic efficacy. The identification of more prostate-associated antigens is important to reduce the risk of tumour recurrence.

This study identified HLA-A2.1-restricted CTL epitopes in human HPN with the ability to induce an HPN-specific antitumour immune response. First, bioinformatics methods predicted epitopes from human HPN. Second, HPN-specific CTLs were induced from HLA-A2.1-positive peripheral blood mononuclear cells (PBMCs) from six healthy donors and six prostate cancer patients with the five candidate peptides in vitro to screen for the CTL epitopes in human HPN antigen. The identification of HPN epitopes that induce an HPN-specific antitumour immune response will provide a foundation for immunotherapy for patients with PC, especially during advanced stages.

Materials and methods

Epitope prediction and synthesis

Computer analysis predicted the HLA-A2.1-restricted CTL epitopes from HPN. Briefly, Bioinformatics and Molecular Analysis Section (BIMAS) [website http://www-bimas.cit.nih.gov/molbio/hla_bind/], IMTECH [website: http://www.imtech.res.in/raghava/hla-pred/] and SYFPEITHI [website: http://www.syfpeithi.de/Scripts/MHCServer.dll/EpitopePrediction.htm] predicted the candidate HLA-A2.1-restricted CTL epitopes from the HPN antigen. Five nonapeptides that were derived from the human HPN amino acid sequence were synthesized by Shanghai C-Strong Co., Ltd, Shanghai, China. with a purity of ≥ 92% as determined by high-performance liquid chromatography and amino acid analysis. One nonapeptide from the HIV virus, HIVpol (476-484) (ILLEPVHGV), served as a positive control in HLA-A2.1 peptide-binding assays and a negative control for the induction of HPN-specific CTL reactions in vitro. The peptides were dissolved in DMSO (Sigma) and stored at −20 °C.

Cell lines

The human TAP-deficient T2 cell line synthesized HLA-A2.1. This cell line does not express normal HLA-A2.1 amounts on the cell surface due to the absence of a functional transporter that is associated with antigen processing heterodimers. The prostate cancer cell line, LNCaP, is HLA-A2.1+ and HPN+. The prostate cancer cell lines PC-3 and DU-145 are HLA-A2.1 and HPN. The colon carcinoma cell line SW480 is HLA-A2.1+ but HPN. The L-O2 cell line (Nanjing KeyGen Biotech Co. Ltd., Nanjing, China) is originating histologically from normal human liver tissue immortalized by stable transfection with the hTERT gene; the cell line has been used in previous studies [22-25]. L-O2 cell line expresses HPN at a low level, but it is HLA-A2.1 + . The BB7.2 cell line produces mAb against HLA-A2.1. All of these cell lines were purchased from the American Type Culture Collection (Manassas, VA, USA), except L-O2 cells. T2 cells, LNCaP cells, SW480 cells, DU-145 cells, L-O2 cells and BB7.2 cells were cultivated in RPMI-1640 medium (Life Technologies, Inc., Gaithersburg, MD; Invitrogen, Carlsbad, CA, USA) containing 10% foetal calf serum (FCS), penicillin (100 U/ml) and streptomycin (100 μg/ml). PC-3 cells were cultured in F-12K medium (Life Technologies, Inc.; Invitrogen) containing 10% FCS, penicillin (100 U/ml) and streptomycin (100 lg/ml). All of these cells were maintained at 37 °C in a humidified atmosphere containing 5% CO2.


The entire HLA-A2.1 encoding sequence was cloned into the vector pcDNA3.1 (were kindly provided by Prof. Wan Y from the Institute of Immunology, Third Military Medical University, Chongqing, PR China). The plasmid was grown in Escherichia coli (MAX Efficiency DH5-α Competent cells, as gift from Dr. Cai YG, from the Dept. of Gastroenterology, Southwest Hospital, Third Military Medical University, Chongqing, PR China) and isolated with an endotoxin-free preparation kit (Endo-free Plasmid Mini Kit I, Omega Bio-Tek Inc., Norcross, GA, USA). PC-3 and DU-145 cells were incubated with the mixture of plasmid and DOTAP (Roche, Mannheim, Germany) at a DNA/DOTAP ratio of 1:6 in serum-free medium for six hours, respectively. A total of 2 μg DNA was used per 6-cm-diameter culture dish. After incubation, the cells were washed and incubated with complete medium. After two-day incubation with complete medium, the selection medium (complete medium with G418 at the final concentration 600 μg/ml) was added into culture dish. Stably transfected clones were sorted by flow cytometry. The HLA-A2.1-transfected clone was designated PC-3/HLA-A2.1 and DU145/HLA-A2.1. The expression of HLA-A2.1 was identified by flow cytometry.

Construction of recombinant adenovirus encoding hepsin

The recombinant adenovirus vector encoding hepsin was constructed using the Adeno-XTM Expression System (Clontech, Palo Alto, CA, USA) according to manufacturer's instruction. The wild-type hepsin cDNA was purchased from Thermo Scientific Open Biosystems (Thermo Fisher Scientific Inc., Hanover Park, IL, USA); hepsin cDNA was cloned into the EcoR I site of shuttle vector pDC315 and sequenced to identify possible Taq polymerase-induced errors. The desired replication-deficient adenovirus containing the full-length cDNA of hepsin was generated by homologous recombination through cotransfection of plasmids pDC315-HPN and pBHGloXΔE1, 3Cre in HEK293 cells using DOTAP liposome reagent (Roche, Germany). The adenovirus containing the hepsin gene was amplified from cell lysates. The adenovirus containing the hepsin gene was purified by banding twice in CsCl density gradients as previously described [26]. Viral titres were determined using a BD Adeno-X™ rapid titer kit (Clontech). Aliquots of purified and dialysed viruses were stored at −70 °C for future use.

Peptide-binding assay with flow cytometry

The HLA stabilization assay is based on the stabilization of HLA class I molecules on the surface of T2 cells after loading different peptide concentrations. Flow cytometry detected the HLA-A2.1 binding affinity of synthesized peptides according to previously described methods [27, 28]. Briefly, T2 cells were cultured in 400 μl RPMI-1640 medium containing 0.1% FCS and 5 × 10−5 M β-mercaptoethanol at 5% CO2 and 37 °C for 16 h. The cells were seeded into a 96-well plate at 5 × 105 cells per well. The synthesized peptides were added to final concentrations of 10, 30 or 60 μm per well. Assays were performed in triplicate for each candidate peptide at each concentration in a 96-well plate. T2 cells were incubated with each peptide at 37 °C for 18 h and washed twice with serum-free RPMI-1640 medium. The HLA-A2.1 molecules on the surface of T2 cells were stained using an anti-HLA-A2.1 mAb (obtained from BB7.2 hybridoma cell culture supernatant) for 60 min at 4 °C. The T2 cells were washed twice with PBS and stained with fluorescein isothiocyanate-conjugated IgG antibodies for 30 min. The T2 cells were rinsed three times with PBS, and FACScan measured the fluorescence intensity (Becton Dickinson, San Jose, CA, USA). The relative binding affinity of the respective peptides was calculated using the mean fluorescence intensities (MFI). The fluorescence index (FI) was calculated as follows: FI = [MFI (with the given peptide) − MFI (without peptide)]/MFI (without peptide). The fluorescence index >1.5 was considered strong, 1.5–1.0 was considered intermediate, and <1.0 was considered low.

Immunohistochemical staining

LNCaP cells, HPN-PC-3/HLA-A2.1 cells, HPN-DU-145/HLA-A2.1 cells, L-O2 cells, SW480 cells and HPN-SW480 cells were grown on glass coverslips in six-well culture dishes and fixed with 4% paraformaldehyde (Sigma, St. Louis, MO, USA) for 10 min at room temperature (RT). The glass coverslips were then washed with buffer containing 0.1 m Tris (pH 7.5), 1.5 m NaCl and 1% bovine serum albumin and were permeabilized in the presence of 2% Triton X-100. Then, cells were then washed with PBS and subsequently incubated in PBS containing 10% normal goat serum for one hour at room temperature, followed by two-hour incubation with hepsin monoclonal antibody (10 μg/ml; R&D Systems, Inc. Clone#506230, Catalogue No. MAB47761). After incubation for 45 min at 4 °C and thorough washing in PBS for three times for 5 min, the glass coverslips were treated with secondary antibody. Finally, the cells were incubated for 15 min with an avidin–biotin enzyme reagent. Staining was developed by immersing slides in DAB/H2O2 solution. PBS was used as negative control in place of the primary antibody.

Western blot

Proteins in the cellular extracts were separated using sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) with an 8% polyacrylamide gel and transferred to a nitrocellulose membrane. The membrane was incubated with 5% non-fat milk in PBS followed by incubation with an anti-human-HPN mAb (Clone 506220, R&D Systems, Inc., Minneapolis, MN, USA) for 1 h at room temperature (24 °C). The membranes were washed and incubated with an alkaline phosphatase-conjugated goat anti-mouse IgG antibody (Amersham Biosciences, Buckinghamshire, England) for 1 h at 27 °C. Immunoreactive bands were detected by the ECL Western blot analysis system (Amersham Biosciences).

Generation of peptide-specific T cell

Peptide-specific T cells were generated by repetitive weekly in vitro stimulations with peptide-loaded DCs for one time and then stimulated by peptides at the concentration of 30 μm for 2 weeks. Dendritic cells from donor's PBMCs were generated according to previously described procedures [29]. This in vitro study was approved by the Ethics Committees of the Kunming General Hospital of PLA. All peripheral blood donors signed a written informed consent to participate in the study. Briefly, PBMCs from HLA-A2.1-positive individuals were seeded in culture flasks in RPMI-1640 medium containing streptomycin (100 μg/ml), penicillin (100 U/ml) and 10% FCS and incubated for 2 h. Non-adherent cells were collected and frozen in freeze medium (50% RPMI-1640, 40% FBS and 10% DMSO) for later use in CTL assays. The adherent cells were cultured for 6 days in the presence of 1000 U/ml rhGM-CSF (R&D Systems, Inc.) and 1500 U/ml rhIL-4 (R&D Systems, Inc.) and cultured for an additional day in the presence of 2000 U/ml tumour necrosis factor α (R&D Systems, Inc.) to induce final maturation. The DCs were morphologically and phenotypically characterized using flow cytometry. The autologous DCs were pulsed with 30 μm peptide for 4 h and irradiated with 20 Gy, which prevented all outgrowths in the control cultures. Autologous T cells were restimulated with peptide-pulsed DCs weekly to generate peptide-specific CTLs. RhIL-2 (50 U/ml, R&D Systems, Inc.) and 10 U/ml rhIL-7 (R&D Systems, Inc.) were added to the medium. The T cells were used as effectors in the cytotoxicity assay after three restimulations.

Cytotoxicity assay

CTL activity levels were evaluated 5 days after the last stimulation using a standard 4-h 51Cr release assay according to previously described methods [30, 31]. Briefly, target cells were labelled with 51Cr sodium chromate (100 μCi per 1 × 106 cells) for 2 h at 37 °C and washed three times. Target cells (104 per well) and effector cells were plated in a final volume of 200 μl in a 96-well round-bottomed plate at a 10:1, 20:1, 40:1 or 80:1 effector-to-target (E/T) ratios. Assays were performed in triplicate for each sample at each ratio. The supernatant (100 μl) of each well was harvested after a 4-h incubation at 37 °C in 5% CO2, and the radioactivity was measured using a gamma counter (Beijing Nuclear Instrument Factory, Beijing, China). The percentage of specific lysis was calculated as the percentage of specific 51Cr release according to the following formula:

display math
Enzyme-linked immunospot assay for IFN-γ

A 24-h ELISPOT assay determined the precursor frequencies of peptide-specific T lymphocytes according to a previously reported method [32]. Briefly, the CTLs that were generated by the above-mentioned methods were plated in triplicate at a final concentration of 1 × 105 cells per well in 96-well anti-INF-γ monoclonal antibody–precoated nitrocellulose plates (96-well coated microtiter plates, Dakewe Biotech Company Limited, Guangdong, China). Candidate peptide at a final concentration of 30 μm stimulated effector cells, phytohemagglutinin (PHA), was used as positive control. The plate was incubated in 5% CO2 at 37 °C for 24 h. The plate was processed using a biotin-labelled antihuman IFN-γ antibody and incubated with alkaline phosphatase-conjugated streptavidin. Freshly prepared developer was added, and the plates were incubated in the dark at 37 °C for 8 min (Quick Spot Mouse IFN-γ Precoated ELISPOT Kit, DAKEWE, Shenzheng, China). An ELISPOT reader (BioReader 4000 Pro-X, BIOSYS German) quantified spots.


All experiments were run in triplicate, and the results are given as the mean ± SD of triplicate determinations. Statistical analysis was performed using Student's t-test. Differences were considered statistically significant when the P value was <0.05. All statistical analyses were carried out with SPSS 11.5 software.


Sequence analysis of the predicted HLA-A2.1-binding peptides from HPN

Nonapeptides were predicted from the amino acid sequence of the human HPN transcript using BIMAS, IMTECH and SYFPEITHI software according to the peptide-binding motif of HLA-A2.1. The top five peptides using BIMAS, IMTECH and SYFPEITHI were selected as candidate epitope peptides. These peptides were derived from positions 20 to 28 (Hpn20; ALTAGTLLL), 229 to 237 (Hpn229; GLQLGVQAV), 191 to 199 (Hpn191; SLLSGDWVL), 179 to 187 (Hpn179; SLRYDGAHL) and 268 to 276 (Hpn268; PLTEYIQPV) of the HPN amino acid sequence.

MHC peptide–binding assay

The binding to HLA-A2.1 of the five candidate peptides was evaluated in in vitro T2 binding assays. An octapeptide that was derived from mouse heparanase [mHpa(519-526)(FSYGFFVI)] was used as negative control, and a nonapeptide that was derived from HIV [HIVpol(476-484)(ILLEPVHGV)] was used as a positive control. All of the predicted peptides upregulated HLA-A2.1 molecular expression and displayed high affinity for HLA-A21 (Fig. 1).

Figure 1.

HLA-A2.1 binding affinity of candidate peptides derived from human hepsin. A, Evaluation of peptide binding to HLA-A2.1 as assessed using MHC class I stabilization. The stabilization of HLA-A2.1 molecules on T2 cells was detected using flow cytometry and expressed as the mean fluorescence intensity (MFI). B, Fluorescence index (FI) of the binding of the hepsin-derived peptides to HLA-A2.1. NP, negative peptide control, H-2Kb restricted epitope mHpa(519-526)(FSYGFFVI) derived from heparanase was used as negative control. PP, positive peptide control, HLA-A2.1-restricted epitope HIVpol(476-484)(ILLEPVHGV) derived from human immunodeficiency virus, which served as a positive control in HLA-A2.1 peptide-binding assays. The fluorescence index (FI) was calculated as follows: FI = [MFI (with the given peptide) − MFI (without peptide)]/MFI (without peptide).

Expression of HPN and HLA-A2.1 in target cells

Western blotting assessed the expression of HPN protein in all cell lines. HPN protein was detected in LNCaP cancer cell lines, and the HPN protein was detected at low levels in L-O2 cells. The HPN protein was detected in HPN-SW480, HPN-DU-145/HLA-A2.1 and HPN-PC-3/HLA-A2.1 prostate cells that were transduced with the full-length cDNA of HPN. Immunohistochemistry confirmed the expression of HPN protein in the cytoplasm of these cells (Fig. 2A,B).

Figure 2.

The expression of hepsin and HLA-A2.1 in target cells. A, Expression of hepsin protein in different target cell lines as determined using immunohistochemistry. 1, LNCaP prostate cancer cells; 2, HPN-PC-3 prostate cancer cells; 3, HPN-DU-145 prostate cancer cells; 4, L-O2 normal liver cells; 5, HPN-SW480 colon carcinoma cells; 6, SW480 colon carcinoma cells. B, Expression of hepsin protein in different target cell lines detected using Western blot. C, Detection of HLA-A2.1 on the surface of various target cells using flow cytometry. Histograms are shown for 1, LNCaP prostate cancer cells; 2, SW480 colon carcinoma cells; 3, L-O2 normal liver cells; 4, DU-145 prostate cancer cells; 5, DU-145/HLA-A2.1 prostate cancer cells; 6, PC-3 prostate cancer cells; and 7, PC-3/HLA-A2.1 prostate cancer cells.

HLA-A2.1 expression on the surface of all target cell lines was detected using flow cytometry. The expression of HLA-A2.1 in LNCaP, SW480 and L-O2 cells was positive. HLA-A2.1 expression on DU-145 and PC-3 cells was negative. However, the expression of HLA-A2.1 on PC-3 and DU-145 cells results in 74.7% and 82.6% of the cells being positively stained relative to the isotype control after transfection with the HLA-A2.1 plasmid, respectively (Fig. 2C).

CTL cytotoxicity of cancer cells

The CTL cytotoxicity of cancer cells assay was performed for six times. All heathy donors and four of six prostate cancer patient donors respond to three epitopes. The cytotoxicity of stimulated PBMCs (effector cells) from healthy donors and patients with prostate cancer was determined using a standard 4-h 51Cr release assay. Three peptides, Hpn229, Hpn268 and Hpn191, elicited HPN-specific CTLs that lysed HPN- and HLA-A2.1-expressing LNCaP cells at an E/T ratio from 20:1 to 80:1 (Fig. 3A). The further study indicated that the cytotoxicity of effector cells from prostate cancer patients donors can lyse LNCaP prostate cancer cells, which expressing both HPN and HLA-A2.1 (Fig. 3B).

Figure 3.

Specific lysis of CTLs that were generated from healthy and prostate cancer peripheral blood by hepsin epitope-loaded dendritic cells against LNCaP prostate cancer cells. A, The CTLs from healthy donors' peripheral blood for screening the epitopes from hepsin could elicit potent antitumour immune response; B, The specific lysis of CTLs generated from prostate cancer patients' peripheral blood. *Statistically significant values at < 0.05 using a paired Student's t-test compared with the negative peptide control.

HPN-specific and HLA-A2.1 restriction analyses

PC-3 and DU-145 are HLA-A2.1- and hepsin-negative prostate cancer cell lines. PC-3 and DU145 cells were transfected with a eukaryotic vector containing the full-length cDNA of HLA-A using the DOTAP lipofection method according to the manufacturer's protocol. The G418 (600 μg/ml) selected clones were named PC-3/HLA-A2.1 and DU-145/HLA-A2.1. The PC-3, PC-3/HLA-A2.1, DU145 and DU-145/HLA-A2.1 cells resulting in 1.2%, 74.7%, 2.6% and 82.6% of the cells were labelled with anti-HLA-A2 monoclonal antibody, respectively (Fig. 2C). The cytotoxic assay demonstrated that the CTLs generated from Hpn229, Hpn268 and Hpn191 peptide-pulsed DCs lysed HPN-PC-3/HLA-A2.1 and HPN-DU-145/HLA-A2.1 cells (Fig. 4B,D). However, the previously mentioned induced effectors did not lyse HPN-PC-3 and DU-145/HLA-A2.1 cells at the highest E/T ratio (Fig. 4A,C). These results clearly demonstrated that CTLs are HPN-specific and restricted by HLA-A2 expression.

Figure 4.

HLA-A2.1 restriction of the CTLs that were generated from prostate cancer peripheral blood by different hepsin epitopes loading dendritic cells in vitro. CTLs generated from the HLA-A2-restricted nonapeptide HIVpol(476-484, ILLEPVHGV) derived from HIV served as an negative peptide control (NP). The cytotoxic activity of CTLs induced by hepsin Hpn229, Hpn268 and Hpn191 was determined against DU-145 (A), PC-3 (C), DU-145/HLA-A2.1 (B) and PC-3/HLA-A2.1 (D) cells in vitro at various E/T ratios using a 4-h 51Cr release assay. *Statistically significant values at < 0.05 using a paired Student's t-test compared with the negative peptide control.

The HLA-A2.1-positive and HPN-negative colon carcinoma cell line, SW480, was used to confirm that the effector cells lysed the target cells in an HPN-specific manner. SW480 cells were transduced with recombinant, replication-defective adenovirus (Ad-HPN) that encoded the full-length cDNA of human HPN at a multiplicity of infection (MOI) of 200. The cells were cultured for 2 days in RPMI-1640 media. Western blot analysis revealed HPN protein expression in HPN-SW480 cells, but HPN expression in control SW480 cells was negative (Fig. 2A,B). Effector cells that were induced by HPN-derived epitopes lysed HPN-SW480 cells, but no obvious lysis of SW480 cells at the highest E/T ratio was observed. These results indicated that the effectors lysed tumour cells in an HPN-specific manner (Fig. 5A,B).

Figure 5.

Hepsin-specific CTLs and side effects of CTLs generated from hepsin vaccines in prostate cancer patients' peripheral blood in vitro. CTLs generated from the HLA-A2.1-restricted nonapeptide HIVpol(476-484, ILLEPVHGV) derived from HIV served as an negative peptide control (NP). Specific lysis of CTLs generated from different hepsin-derived peptides against SW480 (A), HPN-SW480 colon carcinoma cells (B), L-O2 normal liver cells (C) and the inhibition of the induced cells by the anti-hepsin mAb (D). *Statistically significant values at < 0.05 using a paired Student's t-test compared with the corresponding control.

Inhibition of the recognition of effectors by anti-HPN monoclonal antibody

To determine whether the predicted peptide Hpn229-, Hpn268- and Hpn199-induced effectors recognized that the prostate cancer cells are hepsin specific, the mAbs against hepsin were used to block recognition by effectors. The results show that anti-hepsin antibody could significantly eliminate the cytotoxicity of the effectors against LNCaP cells, which indicates that the induced effectors lyse tumour cells in a hepsin-specific manner.

Hepsin-specific CTLs do not lyse normal liver cells

Hepsin is expressed in the normal liver and at lower levels in the kidneys, pancreas and testes [33]. Therefore, hepsin targeting for immunotherapy may have side effects in normal tissues. CTLs that were induced by hepsin-specific peptides were also used to lyse normal liver cells to investigate the effect of hepsin-specific CTLs on liver cells. The results demonstrated that hepsin-specific CTLs induced by peptides did not lyse normal liver cells in vitro (Fig. 5C).

ELISPOT test IFN-γ secretion

An ELISPOT assay measured IFN-γ-producing cells because CTLs produce this TH1 cytokine. The Hpn229, Hpn268 and Hpn191 peptides generated peptide-specific T cell responses because of their ability to induce increased frequencies in IFN-γ-producing T cells compared with negative controls (Fig. 6, < 0.01). No significant response was detected following stimulation with the other peptides. These results indicated that HPN peptide vaccines induced an HPN-specific CTL response.

Figure 6.

IFN-γ–producing cells were enumerated using an ELISPOT assay. A H-2Kb restricted epitope from mouse heparanase served as negative control. Phytohemagglutinin (PHA) served as a positive control. Column indicates the mean; bars, SE. *Statistically significant values at < 0.01 using a paired Student's t-test were compared with the NP group.


Hepsin is a type II transmembrane serine protease that is expressed in normal liver and at lower levels in the kidney, pancreas and testes. HPN-deficient mice grow normally and exhibit no defects in fertility, vision or liver function, which suggests that hepsin is not essential for embryonic development or the maintenance of vital adult functions [34, 35]. In contrast, hepsin overexpression is associated with a disorganization of the basement membrane and the promotion of metastases in a mouse transgenic model [12]. HPN neutralizing antibodies inhibit the invasion of prostate and ovarian tumour cells in vitro [33]. We determined the expression of hepsin in prostate cancer by immunohistochemistry, and our result showed that hepsin was over-expressed in 88.4% (61/69) of the samples of patients with prostate cancer. This result is similar to previously study [3]. Therefore, hepsin is a potential target for the immunizing treatment of prostate cancer.

Tumour-specific immunotherapy has become an attractive therapeutic approach for cancer in the past several years. Dendritic cell-based immunotherapy is a promising novel treatment modality because of its high specificity, strong immunogenicity and weak side effects. The results of immunotherapeutic cancer treatments have been promising in experimental models, but the overall success in clinical trials has been modest [36]. Therefore, the identification of more tumour-associated antigens and their CTL epitopes is important for the development of novel tumour immunotherapy strategies.

The spontaneous immune response to autologous tumours in patients with cancer has characterized tumour-associated antigens that can be targeted by TAA-specific immune responses based on the recognition of TAA by CTLs in an MHC-I-restricted manner. Several prostate cancer–associated antigens have emerged as targets for prostate cancer immunotherapy; clinical trials suggest a clinical benefit in prostate cancer patients treated with vaccines that target these prostate cancer–associated antigens [37, 38]. This study characterized three HLA-A2.1-restricted CTL epitopes, Hpn229, Hpn268 and Hpn191, which were derived from the amino acid sequence of HPN. Three epitopes were identified for pre-existing T cells that were specific for blood donors, and T cell clones were characterized specifically for these epitopes.

First, candidate epitopes were predicted from the human HPN antigen using computer algorithms. Computational methods are useful for identifying immunogenic T cell epitopes from defined antigens and pathogens [39, 40]. Five HLA-A2.1-restricted epitopes were predicted from human HPN using BIMAS, SYFPEITHI and IMTECH.

Second, a quantitative flow cytometric assay measured the surface induction of class I molecules on T2 cells to investigate the binding of the predicted epitopes to HLA-A2.1. This method is commonly used to measure the relative binding affinity of different peptides to the same class I allele [41]. Our five epitopes from HPN exhibited high affinity to HLA-A2.1 at peptide concentrations >30 μm.

Third, DCs that were generated from HLA-A2.1-positive donors were pulsed with one of the five peptides and used to stimulate T lymphocytes to investigate the CTL responses of the five epitopes. A standard 4-h 51Cr release assay determined the CTL induction of each epitope. Our results showed that Hpn229, Hpn268 and Hpn191 induced antitumour immunity to LNCaP, PC-3 and DU-145 prostate cancer cells, and SW480 colon carcinoma cells was in HPN-specific and HLA-A2.1-restricted manner. Therefore, the present study demonstrated that Hpn229(GLQLGVQAV), Hpn268(PLTEYIQPV) and Hpn191(SLLSGDWVL) are HLA-A2.1-restricted CTL epitopes that induced HPN-specific CTLs in vitro. HPN is highly expressed in prostate cancer cells, and these three HPN epitopes may contribute to the design of epitope-based vaccines for patients with prostate cancer. The standard 51Cr release CTL assay is semi-quantitative and measures the property of a population of cells, which may be polyclonal and multispecific. Recently, peptide–MHC tetramers have been widely used to detect CD8+ T cell responses to cancer vaccines and other immunotherapies. This technology allows researchers to look for correlations between the induced T cell responses and the clinical outcome [42].

Safety concerns limit the use of HPN epitopes for the development of prostate cancer vaccines in the near future. HPN is expressed in the normal liver and at lower levels in the kidney, pancreas and testes. Consequently, HPN-based cancer vaccine therapies require a thorough assessment of the potential side effects of autoimmunity to organs and HPN-positive cells. The effects of hepsin peptide-specific CTLs on normal liver cells were investigated. Hepsin peptide-specific CTLs could not lyse L-O2 normal liver cells, which suggested that the hepsin peptide vaccination did not markedly lyse normal liver cells. This result is similar to previous studies [29-32, 43]. HPN expression levels in normal tissue may be below the threshold for HPN-specific CTL recognition.

Finally, the epitope peptides may also promote the release of INF-γ in addition to the CTLs-induced HPN-specific lysis. These results demonstrated that the HPN epitopes Hpn229, Hpn268 and Hpn191 increased IFN-γ release, which suggested that HPN-specific peptides activated the CTLs. These lymphocytes displayed an HPN-specific cytotoxicity on prostate cancer cells and increased IFN-γ release, which enhanced the non-specific cytotoxicity on prostate cancer cells.


In conclusion, our results suggested that the Hpn229, Hpn268 and Hpn191 epitopes that were derived from human hepsin induced HLA-A2.1-restricted CD8+ CTLs, which may be lethal for prostate cancer cells that express both hepsin and HLA-A2.1 in vitro. Moreover, these epitope-induced HPN-specific CTLs did not lyse HPN-expressing normal liver cells. This study is the first report of human hepsin epitopes. Our results demonstrated that hepsin is a potential TAA target for prostate cancer immunotherapy. Further clinical trials with hepsin epitope vaccines are required to confirm these results.


This work was supported by grants from the National Nature Science Foundation of China (30800520) and the Nature Science Foundation of Nanjing Military Command (11MA041).

Authorship contributions

Guo J, Li G and Tang J designed and performed research; Cao XB and Zhou QY performed parts of immunological and RT-PCR studies and analysed the data; Fan ZJ performed part of the research and analysed data; Zhu B and Pan XH designed research, analysed data and wrote the paper.

Competing interests

The authors have declared that no competing interests exist.