Identification of new prostate stem cell antigen-derived peptides immunogenic in HLA-A2⁺ patients with hormone-refractory prostate cancer

Abstract

Purpose

Prostate cancer refractory to hormonal manipulation requires new treatment modalities. In the present study we attempted to identify prostate stem cell antigen (PSCA)-derived peptides immunogenic in HLA-A2⁺ prostate cancer patients in order to develop peptide-based immunotherapy against hormone-refractory prostate cancer (HRPC).

Methods

Eleven different PSCA-derived peptides, which were prepared based on the HLA-A2 binding motif, were examined to determine whether they would be recognized by cellular and humoral immune responses in 12 HLA-A2⁺ patients (11 with HRPC and 1 with non-HRPC).

Results

Among the PSCA-derived peptides, PSCA 7–15 and PSCA 21–30 peptides effectively induced HLA-A2-restricted peptide-specific and tumor-reactive cytotoxic T lymphocytes (CTLs) from peripheral blood mononuclear cells (PBMCs) of HLA-A2⁺ patients. The PSCA 21–30 peptide was capable of inducing peptide-specific CTLs in both cancer patients and healthy donors, whereas the PSCA 7–15 peptide was immunogenic in only cancer patients. Immunoglobulin G (IgG) reactive to the PSCA 21–30 peptide was detected in plasma of most patients and healthy donors, whereas IgG reactive to PSCA 7–15 was undetectable in all cases. These results indicate that the former peptide elicits both cellular and humoral immune responses in both patients and healthy donors, whereas the latter elicits only cellular responses in patients.

Conclusion

These two PSCA peptides should be considered for use in clinical trials of immunotherapy for HLA-A2⁺ HRPC patients.

Introduction

Prostate cancer is one of the most common cancers in older men [5]. Although hormone ablation therapy can temporarily palliate patients with advanced disease, the progression to hormone-refractory prostate cancer (HRPC) is inevitable in most cases [17]; therefore, the development of novel therapeutic modalities for the treatment of HRPC is necessary. One such therapy could be peptide-based immunotherapy, as recent advances in tumor immunology have enabled us to identify many genes encoding tumor antigens and their peptides that are recognized by cytotoxic T lymphocytes (CTLs) [1, 28]. Several prostate-specific antigens and antigen-derived peptides have also been reported, including prostate-specific antigen (PSA) [2, 3, 31] prostate-specific membrane antigen (PSMA) [10, 30], and prostatic acid phosphatase (PAP) [12, 25]. Some of these antigen-derived peptides have been used in the treatment of prostate cancer patients, but the clinical responses observed thus far have been unsatisfactory [19, 20, 21]; therefore, new antigens and peptides suitable for use in specific immunotherapy for patients with HRPC are needed.

The prostate stem cell antigen (PSCA) is a recently identified antigen expressed on the cell surface of prostate cancer cells [27]. This antigen is a glycosylphoshatidylinositol-anchored protein, and is overexpressed by both androgen-dependent and androgen-independent prostate cancers [27], suggesting that this antigen could be a good candidate for specific immunotherapy for patients with HRPC. In addition, immunotherapy targeting on PSCA could be promising for the treatment of patients with bone metastases, as the expression of PSCA in prostate cancer is higher in metastases than in primary tumors [6].

We report in this study new PSCA-derived peptides that can be recognized by both cellular and humoral immune responses in HLA-A2⁺ prostate cancer patients.

Materials and methods

Patients

Twelve HLA-A2⁺ prostate cancer patients (11 with HRPC and 1 with non-HRPC) and 5 HLA-A2⁺ healthy volunteers were enrolled in this study after informed consent was obtained. None of these participants were infected with human immunodeficiency virus (HIV). Twenty milliliters of peripheral blood was obtained, and peripheral blood mononuclear cells (PBMCs) were prepared by Ficoll-Conray density-gradient centrifugation. The expression of HLA-A2 molecules on PBMCs of cancer patients and healthy donors was first determined by flow cytometry, and HLA-A2 subtypes were determined by the sequence-specific oligonucleotide probe method.

Cell lines

T2 is an HLA-A*0201-expressing line. PC93 is an HLA-A2-negative prostate cancer cell line (HLA-A*6802⁺) that was established by K. Ohishi (Department of Urology, Kyoto University, Japan). To establish PC93 cells stably expressing HLA-A2 molecules, pCR3.1 vector (Invitrogen, Calif.), which was inserted with the HLA-A*0201 gene, was electroporated into PC93 cells and selected with neomycin (Gibco BRL, Grand Island, N.Y.) at a dose of 0.75 mg/ml. An established cell line was designated as PC93-A2. LNCap is a prostate carcinoma cell line. Both colo201 and colo302 are HLA-A2⁺ and HLA-A2⁻ colon carcinoma cell lines, respectively. All cell lines were maintained in RPMI-1640 medium (Gibco BRL) supplemented with 10% FCS.

RT-PCR

Total RNA was isolated from cancer cell lines using RNAzol B (Tel-Test, Friendswood, Texas). The cDNA was prepared using the SuperScript Preamplification System for First Strand cDNA Synthesis (Invitrogen, Calif.), and it was amplified using the following primers. The primer pair used for PSCA was as follows:

• Sense primer, 5’-GCAAGAAGAACATCACGTGC-3’

• Antisense primer, 5’-TAGGATGTGCCTCAGGAACC-3’.

The primer pair used for glyceraldehydes 3-phosphate dehydrogenase (GAPDH) was as follows:

• Sense primer, 5’-ACAACAGCCTCAAGATCATCAG-3’,

• Antisense primer, 5’-GGTCCACCACTGACACGTTG-3’.

The PCR was performed using Taq DNA polymerase in a DNA thermal cycler (iCycler, Bio-Rad Laboratories, Calif.) for 30 cycles (at 94°C for 1 min, 60°C for 2 min, and 72°C for 1 min).

Flow cytometric analysis

The expression of PSCA on tumor cell lines were examined using anti-PSCA mAb (1G8: mouse IgG1) [6], which was kindly provided by R.E. Reiter (Department of Urology, University of California). To examine the expression of HLA-A2 molecules on tumor cell lines, cells were stained by anti-HLA-A2 mAb, followed by FITC-conjugated goat anti-mouse IgG. The results were analyzed by the CELLQuest program (Becton Dickinson, Calif.).

Peptides

Eleven PSCA-derived peptides, which are listed in Table 1, were prepared based on the HLA-A*0201 binding motif. All peptides were of >90% purity and were purchased from Biologica (Nagoya, Japan). Influenza virus-derived (GILGFVFTL), Epstein-Barr virus (EBV)-derived (GLCTLVAML), and HIV-derived (SLYNTYATL) peptides with the HLA-A2 binding motif were used as a control. All peptides were dissolved with dimethyl-sulfoxide at a dose of 10 mg/ml.

Table 1.

Reactivity of prostate stem cell antigen (PSCA) peptide-stimulated peripheral blood mononuclear cells (PBMCs) from HLA-A2⁺ patients with prostate cancer. EBV Epstein-Bar virus. ND not determined

Peptides	Amino acid sequence	Scorea	Prostate cancer patients												Total
			1	2	3	4	5	6	7	8	9	10	11	12
			A0201	A0201	A0206	A0206	A0206	A0206	A0206	A0206	A0206	A0207	A0207	A0210
			IFN-γ production (pg/ml)b
PSCA4–13	VLLALLMAGL	309	0	415	4	5	19	629	20	86	0	25	26	19	2 of 12
PSCA5–13	LLALLMAGL	84	0	7	52	15	39	136	66	63	0	116	60	14	2 of 12
PSCA7–15	ALLMAGLAL	79	5	0	101	50	10	104	121	133	0	198	56	19	5 of 12
PSCA21–30	LLCYSCKAQV	118	4	124	0	302	89	1342	0	60	0	291	86	2	4 of 12
PSCA43–51	OLGEQCWTA	153	20	0	32	27	10	274	43	2	21	24	131	22	2 of 12
PCA70–79	CVDDSQDYYV	55	0	54	40	50	16	253	25	0	64	74	47	121	2 of 12
PSCA106–115	LLALLPALGL	36	0	5	35	34	34	143	116	27	61	34	147	55	3 of 12
PSCA108–117	ALLPALGLLL	79	7	20	46	40	39	28	15	0	50	11	59	46	0 of 12
PWA109–117	LLPALGLLL	36	10	8	63	40	1	58	14	23	36	57	92	93	0 of 12
PSCA14–22	ALQPGTALL	21	0	4	0	5	6	ND	ND	0	0	20	67	0	0 of 10
PSCA105–113	ALLALLPAL	25	0	0	24	29	5	ND	ND	0	17	50	79	17	0 of 10
EBV	GLCTLVAML	–	2	312	21	21	17	142	44	847	0	0	53	173	4 of 12
Flu	GILGFVFTL	–	0	0	554	22	292	236	395	34	26	105	57	134	5 of 12

^aScores represent the estimated half-time of dissociation of the PSCA peptides binding HLA-A2 molecules

^bThe PBMCs of HLA-A2⁺ prostate cancer patients were in vitro stimulated with the indicated PSCA peptides, as described in “Materials and methods”

On day 15, the cultured PBMCs were tested for their reactivity to T2 cells, which were pre-pulsed with a corresponding peptide in quadruplicate

Values represent the results of the best of the four wells, and background IFN-γ production in response to the HIV peptide was subtracted

Successful induction of peptide-specific CTLs was judged to be positive when more than 100 pg/ml IFN-γ was produced in response to corresponding peptide-pulsed T2 cells compared with HIV peptide-pulsed T2 cells. Positive responses are italicized

Assay for peptide-specific CTLs in PBMCs

The assay for the detection of peptide-specific CTLs was performed according to a previously reported method with several modifications [9]. In brief, PBMCs (1×10⁵ cells/well) were incubated with 10 μg/ml of each peptide in a U-bottom-type 96-well microculture plate (Nunc, Roskilde, Denmark) in 200 μl of culture medium. The culture medium consisted of 45% RPMI 1640, 45% AIM-V medium (Gibco BRL), 10% FCS, 50 U/ml of interleukin-2 (IL-2), and 0.1 mM MEM nonessential amino acid solution (Gibco BRL). Half of the culture medium was removed and replaced with the new medium containing a corresponding peptide (20 μg/ml) every 5 days. On the fifteenth day of culture, the harvested cells were tested for their ability to produce interferon (IFN)-γ in response to T2 cells (1×10⁴ cells/well), which were pre-loaded with either a corresponding peptide or the HIV peptide as a negative control. After an 18-h incubation, the supernatant was collected, and the level of IFN-γ was determined by enzyme-linked immunosorbent assay (ELISA; limit of sensitivity: 10 pg/ml).

A small number of PBMCs (1×10⁵/well) were applied in micro-well plates, and were separately stimulated with corresponding peptides in an assay for peptide-specific CTLs. Under these conditions, the reactivity varied considerably among individual wells, as reported previously [9]; therefore, each well was separately estimated to screen for the presence of peptide-specific CTL precursors.

In vitro culture for CTL assay

The PBMCs from patients (1×10⁵ cells/well) were incubated with 10 μg/ml of each peptide in a U-bottom-type 96-well microculture plate (Nunc, Roskilde, Denmark) at a volume of 200 μl of the culture medium containing 50 U/ml of IL-2. Half of the culture medium was removed and replaced with the new medium containing a corresponding peptide (20 μg/ml) and IL-2 (50 U/ml) every 5 days. On the fifteenth day of culture, half of the cultured cells were harvested, washed, and divided into three wells. The cultured cells in each well were cultured with or without PC93 or PC93-A2 cells (1×10⁴/well). After an 18-h culture, the supernatants were collected, and the levels of IFN-γ were determined by ELISA. Then, the cells in the wells producing IFN-γ in an HLA-A2-restricted manner were collected and further cultured in order to obtain a sufficiently large number of cells to carry out a CTL assay.

Cytotoxicity assay

Cultured cells were tested for cytotoxicity against both PC93 and PC93-A2 by a 6-h ⁵¹Cr-release assay.Two thousand ⁵¹Cr-labeled cells per well were cultured with effector cells in 96 round-well plates at the indicated effector/target ratios. The specific ⁵¹Cr-release was calculated according to the following formula:

Spontaneous release was determined by the supernatant of the sample incubated with no effector cells, and the total release was then determined by the supernatant of the sample incubated with 1% Triton X (Wako Pure Chemical Industries, Osaka, Japan). In some experiments, CD8⁺ T cells were positively isolated using CD8 Positive Isolation Kit (Dynal, Oslo, Norway) from the PSCA peptide-stimulated PBMCs. The positive percentage of CD8⁺ T cells was >97%. In other experiments of cytotoxicity assay, 10 μg/ml of either anti-HLA class I (W6/32: mouse IgG2a) or anti-HLA-DR (L243: mouse IgG2a) mAb was added into wells at the initiation of the culture.

Cold inhibition assay

The specificity of PSCA peptide-stimulated CTLs was confirmed by a cold inhibition assay. In brief, ⁵¹Cr-labelled target cells (2×10³ cells/well) were cultured with the CTLs (2×10⁴ cells/well) in 96 round-well plates with 4×10⁴ cold target cells. T2 cells, which were pre-pulsed with either the HIV peptide or the corresponding PSCA peptide, were used as cold target cells.

Detection of peptide-specific immunoglobulin G

Peptide-specific immunoglobulin G (IgG) levels in plasma were measured by ELISA, as previously reported [22, 23]. Briefly, a peptide (20 μg/well)-immobilized plate was blocked with Block Ace (Yukijirushi, Tokyo, Japan) and washed with 0.05% Tween 20-PBS, after which 100 μl /well of plasma sample diluted with 0.05% Tween20-Block Ace was added to the plate. After a 2-h incubation at 3°C, the plates were washed and further incubated for 2 h with a 1:1000-diluted rabbit anti-human IgG (γ-chain specific; Dako, Glostrup, Denmark). The plate was washed, and then 100 μl of 1:100-diluted goat anti-rabbit IgG-conjugated horseradish peroxidase (En Vision, Dako) was added to each well, and the plate was incubated at room temperature for 40 min. After the plate was washed again, 100 μl/well of tetramethyl benzidine substrate solution (KPL, Guildford, UK) was added, and the reaction was stopped by the addition of 1 M phosphoric acid. To estimate peptide-specific IgG levels, we compared the optical density (OD) values of each sample with those of serially diluted standard samples, and the values are shown as OD units/ml. To confirm the specificity of IgG to the PSCA 21–30 peptide, sample plasma was cultured with plates coated with either the PSCA 21–30 peptide or the control HIV peptide; thereafter, the levels of PSCA 21–30 peptide-specific IgG in the resultant supernatant were determined by ELISA.

Statistics

The statistical significance of the data was determined using a two-tailed Student’s t test. A p value of less than 0.05 was considered to be statistically significant.

Results

Induction of PSCA peptide-specific CTLs from prostate cancer patients

The PBMCs from 12 patients (11 with HRPC and 1, patient 2, with non-HRPC) were incubated with each of 11 kinds of PSCA-derived peptides or control peptides, and were examined for their IFN-γ production in response to corresponding peptide-pulsed T2 cells (Table 1). The assay was carried out in quadruplicate, and the results of the best well are shown. Successful induction of peptide-specific CTLs was judged to be positive when more than 100 pg/ml of IFN-γ was produced in response to the corresponding peptide-pulsed T2 cells compared with HIV peptide-pulsed T2 cells. The PSCA 14–22 and PSCA 105–113 peptides, which have been reported to be capable of inducing peptide-specific CTLs from HLA-A2⁺ cancer patients [4, 14], failed to induce peptide-specific CTLs with the employed culture system. The PSCA 108–117 and PSCA 109–117 peptides failed to induce peptide-specific CTL from any patients. Each of five PSCA peptides, including PSCA 4–13, PSAC 5–13, PSCA 43–51, PSCA 70–79, and PSCA 106–115 peptide, induced peptide-specific CTLs in 2 or 3 of 12 patients. The PSCA 7–15 and PSCA 21–30 peptides induced peptide-specific CTLs in 5 and 4 of 12 patients, respectively. Based on these findings, the PSCA 7–15 and PSCA 21–30 peptides were the focus in the subsequent experiments.

Induction of prostate cancer-reactive CTLs by the PSCA 7–15 and PSCA 21–30 peptides

It is important to determine whether or not peptide-induced CTLs show any reactivity against prostate cancer cells. The reactivity to either a parental PC93 cell line or its HLA-A2-transfectant, PC93-A2, was compared in order to test HLA-A2-restricted CTL responses to prostate cancer cells. RT-PCR revealed that PC93 was positive for the mRNA expression of PSCA, similar to LNCaP, which is widely used as a prostate cancer cell line (Fig. 1A). The expression of PSCA on tumor cell lines was examined using anti-PSCA mAb (Fig. 1B). In accordance with a previous report [6], LNCap cells were negative, although a part of them was non-specifically stained. PC93 was lowly but positively stained with anti-PSCA mAb. In contrast, both colo201 and colo320 tumor cell lines were negative. In addition, as shown in Fig. 1C, PC93 was negative for the expression of HLA-A2 molecules, whereas PC93-A2 was positive for HLA-A2 molecules on their cell surface.

Fig. 1A–C.

Expression of prostate stem cell antigen (PSCA) on PC93 cells and an HLA-A2-expressing transfectant cell line. A RT-PCR was carried out using cDNA from either LNCaP, PC93, or Epstein-Barr virus (EBV)-transformed B cells. Thirty PCR cycles were carried out. GAPDH glyceraldehyde 3-phosphate dehydrogenase. B Flow cytometric analysis was performed on LNCap, PC93, colo201, and colo320 cells. These cells were stained with anti-PSCA mAb, followed by FITC-conjugated anti-mouse IgG mAb. The dashed lines represent the staining without the first mAb. C Flow cytometric analysis was performed on PC93 and PC93-A2 cells. These cells were stained with anti-HLA-A2 mAb, followed by FITC-conjugated anti-mouse IgG mAb. The dashed lines represent the staining without the first mAb

The PBMCs from 4 HLA-A2⁺ patients were cultured in 10 wells in the presence of the PSCA 7–15 or PSCA 21–30 peptide to determine whether PSCA peptide-reactive CTLs from HLA-A2⁺ patients show reactivity against prostate cancer. After 15 days of culture, half of the cultured cells were used to test the reactivity against prostate cancer cells, and representative results are shown in Fig. 2A. Among 10 wells, the cultured cells in 3 or 4 wells produced a higher level of IFN-γ in response to PC93-A2 than to PC93. The cells in these positive wells, which are underlined in Fig. 2A, were further cultured with IL-2 for a cytotoxicity assay. The peptide-stimulated CTLs exhibited a higher level of cytotoxicity against PC93-A2 than against PC93 (Fig. 2B). These results indicate that PBMCs, which were in vitro stimulated with either the PSCA 7–15 or PSCA 21–30 peptide, can show HLA-A2-restricted cytotoxicity against prostate cancer cells.

Fig. 2A, B.

Induction of HLA-A2-restricted and prostate cancer-reactive cytotoxic T lymphocytes (CTLs) from peripheral blood mononuclear cells (PBMCs) of cancer patients. A The PBMCs from 4 HLA-A2⁺ prostate cancer patients were in vitro stimulated with the indicated PSCA peptides in 10 wells, as described in “Materials and methods.” On day 15, half of the cultured cells were separately harvested and cultured with or without PC93 or PC93-A2 cells for 18 h. The levels of IFN-γ in supernatants were then determined by ELISA. The cultured cells in wells, which are underlined at the well number, were further cultured with IL-2 alone for approximately 10 days for the CTL assay. B An assay of cytotoxicity carried out for 6-h was performed against PC93 and PC93-A2. Values represent the mean of triplicate assays. * p<0.05 was considered statistically significant

Induction of PSCA peptide-reactive CTLs from healthy donors

Some prostate-related antigens have the potential to induce peptide-specific CTLs even from healthy donors [7, 8]; therefore, we determined whether PSCA-derived peptides could induce peptide-specific CTLs from PBMCs of HLA-A2⁺ healthy donors (Table 2). The PSCA 21–30 peptide induced peptide-specific CTLs in PBMCs of 3 of 5 healthy donors. Each of the PSCA 4–13, PSCA 43–51, PSCA 70–79, PSCA 108–117, and PSCA 109–117 peptides induced peptide-specific CTLs in PBMCs of two healthy donors. The PSCA 106–115 peptide induced peptide-specific CTLs in 1 patient. The PSCA 5–13 and PSCA 7–15 peptides failed to induce peptide-specific CTLs in any patients. The inability of the PSCA 7–15 peptide of inducing peptide-specific CTLs from PBMCs of HLA-A2⁺ healthy donors was in sharp contrast to the results of peptide-specific CTL induction from PBMCs of cancer patients.

Table 2.

Reactivity of PSCA peptide-stimulated PBMCs from HLA-A2⁺ healthy donors

Peptides	Healthy donors
	1	2	3	4	5	Total
	A0201	A0202	A0203	A0206	A0207
	IFN-γ production (pg/ml)a
PSCA4–13	255	10	12	0	263	2 of 5
PSCA5–13	14	24	70	0	5	0 of 5
PSCA7–15	5	45	3	0	0	0 of 5
PSCA21–30	1016	106	250	0	0	3 of 5
PSCA43–51	129	11	5	9	568	2 of 5
PSCA76–79	143	20	5	16	530	2 of 5
PSCA106–115	50	20	43	2	455	1 of 5
PSCA108–117	5	33	106	121	0	2 of 5
PSCA109–117	24	14	2179	4	2006	2 of 5
Flu	272	62	33	179	741	3 of 5

^aThe PBMCs of HLA-A2⁺ healthy donors were in vitro stimulated with the indicated PSCA peptides, as described in “Materials and methods.” On day 15, the cultured PBMCs were tested for their reactivity to T2 cells, which were pre-pulsed with a corresponding peptide in quadruplicate

Values represent the results of the best of the four wells, and background IFN-γ production in response to the HIV peptide was subtracted

Successful induction of peptide-specific CTLs was judged to be positive when more than 100 pg/ml IFN-γ was produced in response to corresponding peptide-pulsed T2 cells compared with HIV peptide-pulsed T2 cells. Positive responses are italicized

Cytotoxicity of PSCA peptide-stimulated CD8⁺ T cells against prostate cancer cells

We further confirmed that PSCA peptide-specific CD8⁺ T cells could show cytotoxicity against prostate cancer cells. As shown in Fig. 3A, PSCA peptide-stimulated PBMCs, which were derived from healthy donors 4 and 6 and prostate cancer patients 3 and 9, produced higher levels of IFN-γ in response to corresponding peptide-pulsed T2 cells; thereafter, CD8⁺ T cells, which were positively purified from these PSCA-stimulated PBMCs, were examined for their cytotoxicity against four kinds of targets, including PSCA⁺ HLA-A2⁻ PC92, PSCA⁺ HLA-A2⁺ PC93-A2, PSCA⁻ HLA-A2⁺ colo201, and PSCA⁻ HLA-A2⁻ colo320. As a result, these CD8⁺ T cells showed a higher level of cytotoxicity against PC93-A2 cells compared with the other three target cells. This observation was conformed by experiments using anti-class-I or anti-class-II mAb and by cold inhibition assay. As shown in Fig. 3B, cytotoxicity of purified CD8⁺ T cells, which were stimulated with the corresponding PSCA peptide against PC93-A2 cells, were significantly inhibited by the addition of anti-class-I mAb, but not of anti-class-II mAb. In addition, their cytotoxicity against PC93-A2 cells were significantly suppressed when added with unlabelled corresponding PSCA peptide-pulsed T2 cells. These results indicate that the PSCA 7–15 and PSCA 21–30 peptides have the potential to generate HLA-A2-restricted CD8⁺ T cells cytotoxic to prostate cancer cells.

Fig. 3A, B.

Cytotoxicity of CD8⁺ T cells induced by in vitro stimulation with the PSCA peptides. A The PBMCs from 2 HLA-A2⁺ healthy donors (HD #4, HD #6) and 2 HLA-A2⁺ prostate cancer patients (Pt #3, Pt #9) were in vitro stimulated with the indicated PSCA peptides, as described in “Materials and methods.” On day 15, half of the cultured cells were harvested, pooled from 8 wells, and cultured with T2 cells which were pulsed with the HIV peptide (open symbols) or the indicated PSCA peptide (closed symbols) for 18 h. The levels of IFN-γ in supernatants were then determined by ELISA. Thereafter, these cells were examined for their cytotoxicity against four kinds of tumor cell lines. Values represent the mean of triplicate assays. * p<0.05 was considered statistically significant. B These PSCA peptide-stimulated CTLs from HD #6 and Pt #3 were examined for their cytotoxicity against PC93 and PC93-A2 cells with or without anti-class-I or anti-class-II mAb. Their cytotoxicity against PC93-A2 cells was also examined in the presence of unlabelled T2 cells, which were pre-pulsed with HIV or corresponding PSCA peptide. Values represent the mean of triplicate assays. * p<0.05 was considered statistically significant

IgG specific to the PSCA peptides

We previously reported that IgG reactive to CTL epitope peptides have been detected in patients with epithelial cancers and healthy donors [22, 23]. We therefore attempted to determine whether IgG reactive to PSCA-derived peptides can be detected in plasma of cancer patients and healthy donors (Table 3). IgG specific to the PSCA 21–30 peptide was detected in 9 of 12 cancer patients (75%). In patients 2, 5, and 10, IgG specific to the PSCA 43–51 or PSCA 70–79 peptide was detected. In patient 11, IgG reactive to the PSCA 43–51 or PSCA 108–117 peptide was detected. Representative results of cancer patients are shown in Fig. 3A. In healthy donors, IgG reactive to the PSCA 21–30 peptide was detected, but none of the other peptides were. Representative results for healthy donors are shown in Fig. 4B. The specificity of IgG against the PSCA 21–30 peptide was confirmed by an antibody absorption test. For example, it was found that the levels of anti-PSCA 21–30 peptide IgG in the plasma of patient 12 and healthy donor 1 were diminished by incubation with the PSCA 21–30 peptide-coated plate, but not with the HIV peptide-coated plate (Fig. 4C).

Table 3.

Immunoglobulin G reactive to the PSCA peptides in plasma of HLA-A2⁺ prostate cancer patients and healthy donors

Peptides	Prostate cancer patients													Healthy donor
	1	2	3	4	5	6	7	8	9	10	11	12	Total	1	2	3	4	5	Total
PSCA4–13	–	–	–	–	–	–	–	–	–	–	–	–	0 of 12	–	–	–	–	–	0 of 5
PSCA5–13	–	–	–	–	–	–	–	–	–	–	–	–	0 of 12	–	–	–	–	–	0 of 5
PSCA7–15	–	–	–	–	–	–	–	–	–	–	–	–	0 of 12	–	–	–	–	–	0 of 5
PSCA21–30	+	+	+	–	+	+	+	–	+	+	–	+	9 of 12	+	+	+	+	+	5 of 5
PSCA43–51	–	+	–	–	+	–	–	–	–	+	+	–	4 of 12	–	–	–	–	–	0 of 5
PSCA70–79	–	+	–	–	+	–	–	–	–	+	–	–	3 of 12	–	–	–	–	–	0 of 5
PSCA106–115	–	–	–	–	–	–	–	–	–	–	–	–	0 of 12	–	–	–	–	–	0 of 5
PSCA108–117	–	–	–	–	–	–	–	–	–	–	+	–	1 of 12	–	–	–	–	–	0 of 5
PSCA109–117	–	–	–	–	–	–	–	–	–	–	–	–	0 of 12	–	–	–	–	–	0 of 5

Immunoglobulin G reactive to a corresponding peptide was judged to be positive when a difference of optical density at a 1:100 diluted plasma was more than 0.05

Fig. 4A–C.

IgG reactive to the PSCA peptides in plasma from prostate cancer patients and healthy donors. A The positive results of IgG reactive to PSCA peptides are shown. These values are shown as optical density (OD) units/ml, and the HIV peptide was used as a negative control. B Representative results of three healthy donors are shown. C To confirm the specificity of IgG to the PSCA 21–30 peptide, 100 μl of the sample plasma from patient 12 and healthy donor 1 were cultured in plates which were pre-coated with either the PSCA 21–30 peptide or the HIV peptide. Thereafter, the levels of IgG reactive to the PSCA 21–30 peptide in the resultant samples were determined by ELISA

Recognition of glycine-substituted PSCA 21–30 peptide by IgG

Among 11 PSCA-derived peptides, both the PSCA 43–51 and PSCA 70–79 peptides contain one cysteine, and the PSCA 21–30 peptide has two cysteines. Cysteine can be easily oxidized by disulfide bondage, resulting in dimerization or cysteinylation [15, 18]; therefore, we attempted to exclude the possibility that preferential recognition of the PSCA 21–30 peptide by IgG is due to cysteine-based modification of the peptide (Fig. 5). Either or both cysteines at the third and sixth positions were replaced by glycine, and glycine-substituted peptides were found to be recognized by IgG of plasma from cancer patient 12 and healthy donor 4. Glycine-substitution did not inhibit recognition by IgG, but rather augmented recognition when the third position was substituted by glycine.

Fig. 5.

Recognition of glycine-substituted PSCA 21–30 peptides by IgG. Either or both of the cysteines in the PSCA 21–30 peptide were substituted by glycine. It was then determined whether these modified PSCA 21–30 peptides could be recognized by IgG of plasma from healthy donor 4 and cancer patient 12. The levels of peptide-reactive IgG were determined by ELISA

Discussion

Both the PSCA 14–22 and PSCA 105–113 peptides had been reported to be capable of inducing peptide-specific and prostate cancer-reactive CTLs from HLA-A2⁺ patients [4, 14]. We first screened 11 peptide candidates, including these two peptides, but both the PSCA 14–22 and PSCA 105–113 peptides failed to induce peptide-specific CTLs from PBMCs of HLA-A2⁺ cancer patients. On the contrary, our results identified both the PSCA 7–15 and PSCA 21–30 peptides as candidates for peptide-based immunotherapy for HLA-A2⁺ prostate cancer patients. The PSCA 7–15 peptide has not been found to be immunogenic in studies from the other laboratories [4, 14], but the PSCA 21-30 peptide had not previously been tested for its ability to induce peptide-specific CTLs. One explanation for these different results may simply be due to the different culture protocols for in vitro CTL induction. Another possibility may be the HLA-A2 subtypes differing between Caucasians and Japanese. Most Caucasians are HLA-A*0201-positive, but HLA-A2 subtypes vary considerably in Japanese, as shown in Tables 1 and 2. The PSCA-derived peptides were prepared based on the binding motif to HLA-A*0201 molecules [24], and both the T2 cells and PC93-A2 cells express HLA-A*0201 molecules; however, the latter possibility is less likely since CTLs reactive to PSCA-derived peptides were induced not only from HLA-A*0201⁺ patients, but also from those with other HLA-A2 subtypes, including HLA-A*0206 or HLA-A*0207 (Table 1). We have previously reported that epithelial tumor antigen-derived peptides, which were prepared based on the binding motif to HLA-A*0201 molecules, are also immunogenic in patients with several HLA-A2 subtypes [11, 13, 29]. These results suggest that both the PSCA 7–15 and PSCA 21–30 peptides are immunogenic in patients with HLA-A*0201, HLA-A*0206, and HLA-A*0207, which could cover most HLA-A2⁺ patients throughout the world.

We tested 11 PSCA peptides and identified that the PSCA 7–15 and 21–30 peptides have the potential to induce prostate cancer-reactive CTLs. As is the case with other prostate-associated antigens [7, 8, 16], CTLs reactive to PSCA-derived peptides could be induced from HLA-A2⁺ healthy donors. A similar phenomenon was observed with regard to melanocyte differentiation antigens [26]. The CTLs reactive to melanocyte differentiation antigens can be detectable in PBMCs of HLA-A2⁺ healthy donors. These findings may suggest that CTLs reactive to non-mutated self antigens circulate in the peripheral blood of both certain healthy donors and cancer patients, and that active stimulation of such CTL precursors could result in effective immunotherapy. On the other hand, the PSCA 7–15 peptide was immunogenic in prostate cancer patients but not in healthy donors. It may be that CTL precursors reacting to the PSCA 7–15 peptide might be efficiently primed only in HRPC patients. This finding suggests that PSCA 7–15 is a promising molecule with regard to the induction of prostate cancer-reactive CTLs in HRPC patients.

The PSCA 21–30 peptide was recognized by both cellular and humoral immune responses in patients and healthy donors. In particular, IgG specific to this peptide was detected in 75% of cancer patients and in all five healthy donors. We have previously reported that IgG reactive to CTL epitope peptides is often detected in cancer patients and healthy donors [22, 23]. We also reported that IgG reactive to the PSA 248–257 peptide was frequently detectable in HLA-A24⁺ healthy donors and prostate cancer patients [7]. The detection of IgG reactive to PSCA epitopes might reflect elicitation of immune responses toward the PSCA in both cancer patients and healthy donors.

It is important to determine whether plasma containing IgG reactive to the PSCA peptide could react positively to the LNCap cells and PC93 cells, and whether the LNCap cells and PC93 cells could absorb IgG reactive to the PSCA peptide. We tested these possibilities and found that no specific absorption of IgG reactive to the PSCA 21–30 peptide was observed when the plasma of reactive patients were incubated with these tumor cell lines, and that the heat-inactivated plasma containing IgG specific to the PSCA 21–30 peptide could not block cytotoxicity of PSCA 21–30 peptide-induced CTLs against the PC93-A2 cells (unpublished observation). Although we have no clear understanding of the roles peptide-specific IgG play in anti-tumor immune responses in cancer patients, it may be interesting to clarify the actions of IgG against CTL-peptides in cancer patients.

We examined the surface expression of PSCA on the LNCap, PC93, and two colon carcinoma cell lines. The PC93 was positive, but the LNCap was concluded to be negative for the surface expression of the PSCA, although a part of LNCap was non-specifically stained with anti-PSCA mAb. This result is consistent with a previous report that the LNCap cells are not to express PSCA on their cell surface [6]; however, our result of RT-PCR revealed that the LNCap cells expressed the mRNA expression of PSCA. Actually, we isolated a cDNA encoding PSCA gene by the PCR method from cDNA of the LNCap cells (unpublished observation). The LNCap cells might express the PSCA in their cytoplasm.

Conclusion

In conclusion, we identified new two PSCA-derived peptides immunogenic in HLA-A2⁺ prostate cancer patients. This information could extend the possibility of treating HLA-A2⁺ HRPC patients using peptide-based specific immunotherapy.