Correspondence to: J.-J. Lee, MD, PhD, Department of Hematology-oncology, Chonnam National University Hwasun Hospital, 322 Seoyangro, Hwasun, Jeollanamdo 519-763, Korea. E-mail: firstname.lastname@example.org
To induce a potent cytotoxic T lymphocyte (CTL) response in dendritic cell (DC)-based immunotherapy against prostate cancer, various tumour antigens should be loaded onto DCs. The aim of this study was to establish a method of immunotherapy for castration-resistant prostate cancer (CRPC) using prostate cancer–specific CTLs generated in vitro by DCs. Monocyte-derived DCs from patients with CRPC were induced to mature using a standard cytokine cocktail (in IL-1β, TNF-α, IL-6 and PGE2: standard DCs, sDCs) or using an α-type 1-polarized DC (αDC1) cocktail (in IL-1β, TNF-α, IFN-α, IFN-γ and polyinosinic:polycytidylic acid) and loaded with the UVB-irradiated CRPC cell line PC-3. Antigen-loaded DCs were evaluated by morphological and functional assays. The αDC1s significantly increased the expression of several molecules related to DC maturation, regardless of whether the αDC1s were loaded with tumour antigens or not, compared to sDCs. The αDC1s showed a higher production of interleukin-12 both during maturation and after subsequent stimulation with CD40L, which was not significantly affected by loading with tumour antigens, as compared to standard DCs (sDCs). Prostate cancer–specific CTLs against autologous CRPC cells were successfully induced by αDC1s loaded with dying PC-3 cells. Autologous αDC1s loaded with an allogeneic CRPC cell line can generate greater CRPC-specific CTL responses as compared to sDCs and may provide a novel source of DC-based vaccines that can be used for the development of immunotherapy in patients with CRPC.
In the modern era of localized prostate cancer treatment, rapid progress has led to a broad spectrum of therapeutic strategies for improving clinical outcomes. However, despite the improvement of therapeutic strategies, the disease still recurs in approximately 20–40% of patients and eventually develops into castration-resistant prostate cancer (CRPC) . Recently, the possible use of immunotherapy in prostate cancer has been supported by several findings. One such finding was that prostate cancer grows comparatively slowly , which could give the patient's immune system adequate time to mount an immune response, if appropriately stimulated. Another finding was that patients with prostate cancer develop multiple novel antibody specificities [3, 4], and newer laboratory techniques have led to the identification of several proteins that are relatively tissue specific to the prostate, which might be used as tumour antigens.
Dendritic cells (DCs) are the major antigen-presenting cells of the immune system. Because of their ability to stimulate T cells, DCs act as a link in antitumour immune responses between innate immunity and adaptive immunity . Thus, DC-based vaccines are the most attractive tool for cancer immunotherapy and have been used for more than 20 malignancies, most commonly melanoma , prostate cancer , colorectal carcinoma  and multiple myeloma . Previously, Jonuleit et al.  introduced ‘standard DCs (sDCs)’ induced by cytokine cocktail with tumour necrosis factor (TNF)-α/interleukin (IL)-1β/IL-6 and prostaglandin E2 (PGE2), and these sDCs have been used in several clinical studies [11-13]. However, the main disadvantage with sDC is the absence of IL-12p70 secretion , which is important for the induction of effective Th1 and cytotoxic T lymphocyte (CTL) responses that are assumed to be essential in connection with cancer vaccination therapy. In an attempt to increase the potency of DCs, alpha-type 1-polarized DCs (αDC1s) were developed through the use of cytokine combinations. The αDC1s are induced to mature by the addition of an αDC1-polarizing cytokine cocktail containing IL-1β, TNF-α, interferon (IFN)-α, IFN-γ and polyinosinic:polycytidylic acid [poly(I:C)]. Compared with sDCs, αDC1s generate strong, functional CTLs in several diseases [15, 16].
The ideal DC-based vaccination in prostate cancer would induce tumour-specific immunity to prostate cancer without causing clinically significant autoimmunity. One approach to enhancing the specificity of vaccines is to target them against protein antigens that are specifically expressed on prostate cancer cells. Several prostate-specific tumour antigens have been used in the treatment for late-stage prostate cancer, but the clinical responses observed thus far have been unsatisfactory . The use of whole tumour cells instead of a single protein antigen to load DCs may help to enhance prostate cancer–specific cytolytic effects with unfractionated tumour-derived antigens. However, it is not only impractical to obtain sufficient amounts of purified autologous tumour cells from patients with prostate cancer to be used as tumour antigens for this purpose, but also unsuitable for those with a lower tumour burden status. Therefore, to overcome this limitation, allogeneic tumour cells or established cancer cell lines from various tumours have been used as an alternative source of tumour-relevant antigens [18, 19].
In the present study, we investigated the feasibility of DC-based immunotherapy in patients with CRPC. To achieve this purpose, αDC1s were loaded with the UVB-irradiated CRPC cells as tumour antigens and were then assessed for their immunogenicity to elicit specific immune responses mediated by CTLs in vitro by means of IFN-γ enzyme-linked immunospot (ELISPOT) assays as compared to sDCs loaded with tumour antigens.
Materials and methods
Generation of αDC1s from patients with CRPC
Peripheral blood was collected from HLA-A2402-positive (HLA-A2402+) CRPC patients after obtaining informed consent according to the protocol approved by Institutional Review Board. CD14+ cells were isolated by positive selection using the magnetic activated cell sorter system (MACS; Miltenyi Biotec, Auburn, CA, USA). The purity of CD14+ cells was >90%. To generate immature DCs (iDCs), the CD14+ cells were cultured in IMDM (Gibco-BRL) with 10% heat-inactivated FBS (Hyclone) and 1% penicillin/streptomycin (Gibco-BRL) for 6 days in 24-well plates at 5 × 105 cells per well in the presence of 50 ng/ml granulocyte-macrophage colony stimulating factor (GM-CSF) (PEPROTECH, Rocky Hill, NJ, USA) and 20 ng/ml IL-4 (PEPROTECH). On day 6, the iDCs were matured either with conventional cytokine cocktail composed of IL-1ß (25 ng/ml, PEPROTECH), TNF-α (50 ng/ml, PEPROTECH), IL-6 (1000 units/ml, PEPROTECH) and PGE2 (106m/l, Sigma-Aldrich, St Louis, MO, USA) for sDCs  or with αDC1-polarizing cytokine cocktail composed of IL-1ß (25 ng/ml), TNF-α (50 ng/ml), IFN-α (3000 IU/ml, Intron-A-IFN-α-2b; Schering-Plough International, Kenilworth, NJ, USA), IFN-γ (1000 units/ml, Strathmann Biotech GmbH, Hannover, Germany) and poly(I:C) (20 μg/ml, Sigma-Aldrich) for αDC1s . The DCs were then loaded with the UVB-irradiated PC-3 (HLA-A2402+, CRPC cancer cell line) at a ratio of 2:1 at 2 h after the addition of the maturation cytokines. The matured DCs loaded with dying PC-3 tumour cells were harvested on day 8, washed and analysed by functional assays.
Preparation of the UVB-irradiated tumour cells as a source of tumour antigen
To load tumour cells onto DCs, PC-3 cells were irradiated with UVB (300 mJ/cm2) (International Light, Newburyport, MA, USA) followed by overnight culture in RPMI-1640 (Gibco-BRL) without FBS to induce apoptosis and were then thoroughly washed. The irradiated tumour cells were then loaded onto DCs 2 h after the addition of maturation cytokines. The irradiated dying cells were immediately confirmed by using Annexin-V and propidium iodide (PI).
Preparation of CRPC cells from patients
Tumour tissues obtained from patients with CRPC by palliative transurethral resection were minced and lysed for 2–4 h at 37 °C in AIM-V medium containing 0.4% collagenase type III. The mononuclear cells were separated by density gradient centrifugation with Ficoll-Hypaque (Lymphoprep) and cryopreserved until they were used as target cells in the cytotoxic assay.
Tumour antigen uptake and phenotypic expression of DCs
To measure the tumour antigen uptake by DCs, PC-3 cells were labelled with PKH67-GL-fluorescein isothiocyanate (FITC) (Sigma-Aldrich) before the UVB irradiation. After loading of tumour antigen onto DCs at a ratio of 1:2 on day 6, the αDC1s loaded with the dying PC-3 tumour cell line were stained with CD11c-phycoerythrin (PE) and analysed by flow cytometry for tumour antigen uptake (CD11c+/PKH67+). Flow cytometry was performed using a FACSAria cell sorter (Becton Dickinson, San Jose, CA, USA) after labelling of cells with CD86-PE, CD83-FITC and CCR7-FITC (PharMingen, San Diego, CA, USA) and the relevant isotype controls (mouse IgG1 and IgG2a, PharMingen). Cell debris was eliminated from the analysis by forward- and side-scatter gating, and the data were analysed with WinMDI version 2.9 software (Biology Software Net).
Cytokine analyses by enzyme-linked immunosorbent assay (ELISA)
The levels of IL-12p70 and IL-10 in the primary culture supernatants of the DCs were measured using the Quantikine Immunoassay Kits (R&D Systems, Minneapolis, MN, USA). Additionally, DCs harvested on day 8 were plated in 96-well plates at 2 × 104 cells/well and were stimulated to secrete IL-12 with CD40 ligand (CD40L)-transfected J558 cells (as an analogue of CD40L-expressing Th cells; a gift from Dr. P. Lane, University of Birmingham, UK) at a density of 5 × 104 cells/well. After 24 h, the supernatant was harvested and the production of IL-12p70 was determined by ELISA kits (R&D Systems).
Induction of CRPC-specific CTLs
Autologous CD3+ (purity > 90%) cells were positively isolated from the lymphocyte fraction after Percoll isolation using MACS (Miltenyi Biotec). T cells (1 × 106 cells) were sensitized by autologous αDC1s or sDCs (1 × 105 cells) loaded with dying PC-3 tumour cells. On day 3, rhuIL-2 (5 ng/ml, R&D Systems) and IL-7 (10 ng/ml, R&D Systems) were added. The CTLs were re-stimulated with the same DCs on day 10. On day 20, the number of antigen-specific T cells was analysed by IFN-γ enzyme-linked immunospot (ELISPOT) assay. PC-3, DU145, autologous CRPC cells from patients with CRPC, and K562 cells [chronic myeloid leukaemia cells with natural killer (NK) cell activity] were used as target cells. MHC class I- and MHC class II-restricted recognition of the prostate cancer–specific CTLs was analysed using MHC class I- and MHC class II-specific mAbs (clone W6/32 and clone CR3/43, respectively). The ELISPOT data were expressed as the mean number of spots (±SD) per 0.5–2 × 105 T cells. CTL alone was used as the control.
All statistical analyses were performed with the program spss 13.0 for Windows (SPSS Inc., Chicago, IL, USA). The Mann–Whitney U-test was performed to analyse the statistical significance of non-parametric differences between the groups. P <0.05 was considered statistically significant.
Preparation of dying PC-3 cells as tumour antigens and antigen uptake by αDC1s
Using Annexin-V/PI staining, about 66.9% of PC-3 cells were shown as dying cells with UVB irradiation (Fig. 1A). To measure tumour antigen uptake by αDC1s, αDC1s loaded with dying PKH67-FITC-labelled PC-3 cells were stained with CD11c conjugated with PE. As shown in Fig. 1B, αDC1s efficiently incorporated the dying PC-3 tumour antigen (36.3 ± 7.7%; n = 4) as measured by flow cytometry (CD11c+/PKH-67+). The efficacy of antigen uptake was similar between αDC1s and sDCs (not shown).
Characteristics of DCs generated from patients with CRPC
After the induction of iDCs, either sDCs or αDC1s were generated by the addition of the appropriate cytokine cocktail. The αDC1s showed typical morphology with large and branching structures that were aggregated among the cells (not shown). Phenotypically, αDC1s exhibited higher expression of the costimulatory molecule (CD86) compared to sDCs (P < 0.05). Both αDC1s and sDCs showed comparable expression of the maturation markers (CD83) and predictive marker of migratory ability (CCR7) (Fig. 2).
IL-12 production of DCs generated from patients with CRPC
One of the best ways of determining DC function is to examine cytokine secretion. IL-12p70 is an important cytokine for stimulating naive T cells for Th1 polarization to benefit cancer treatment, but IL-10 is the main inhibitory cytokine for cancer treatment. As shown in Fig. 3A, the αDC1s showed higher IL-12p70 level, which measured in primary culture supernatant collected at DC harvest (‘during maturation’), than sDCs (P < 0.05). Furthermore, the αDC1s showed even higher production of IL-12p70 after subsequent stimulation with CD40L-transfected J558 cells as compared to sDCs (P < 0.05, Fig.3B). This cytokine secretory capacity of αDC1s was not significantly suppressed by loading tumour antigen. In contrast, production of the inhibitory cytokine IL-10 by αDC1s was not significant (Fig. 3C).
Generation of potent CRPC-specific CTLs by autologous DCs loaded with dying tumour cells
We next determined the tumour-specific generation of CTLs by DCs. The secretion of IFN-γ by the CTLs was measured in three independent experiments using the ELISPOT assay. Consistent with the high ability to produce IL-12p70, primed CD3+ T cells generated by the αDC1s loaded with dying PC-3 prostate cancer cells showed a larger number of IFN-γ-producing cells against PC-3 prostate cancer cells (P < 0.05, Fig. 4A), DU 145 prostate cancer cells (P < 0.05, Fig. 4B) and autologous CRPC cells obtained from patients with CRPC (P < 0.05, Fig. 4C) as compared to sDCs. The MHC class I-restricted recognition of the CTL response was confirmed using a MHC class I-specific mAb. The NK-sensitive K562 cell line was used to exclude the possibility that NK cells contributed the CTL responses. No significant killing activity was observed against the K562 cell line.
For wide application of a DC vaccination in patients with cancer, easily available tumour antigens should be developed. Allogeneic cancer cells as universal tumour antigens could substitute to autologous tumour cells by having an easier culture of the tumour cells. In this study, we investigated the feasibility of cellular immunotherapy using autologous DCs loaded with allogeneic dying CRPC cells, which could generate potent CRPC-specific CTLs against autologous CRPC cells of patients. The αDC1s were successfully generated and significantly increased the expression of several costimulatory molecules without any difference by loading of tumour antigens. Furthermore, αDC1s showed a higher production of IL-12, without significant suppression by loading tumour antigens, in comparison with sDCs. In addition, potent CRPC-specific CTLs against autologous CRPC cells from patients were elicited by autologous αDC1s loaded with dying CRPC cells as compared to sDCs, which is a consistent finding of previous reports studied in other cancers [9, 16].
Among a vast array of novel therapies being developed for advanced prostate cancer, a number of immunotherapeutic approaches are being tested in clinics. DC-based therapies that utilize whole tumour cells  or tumour-related peptides [21, 22], proteins or DNA [23, 24], as the source of tumour-related antigens, offer the advantage of targeting many antigens that are overexpressed in prostate cancer, including PSA and prostatic acid phosphatase (PAP) [25, 26]. Vaccines targeting PSA and PAP have generated immense enthusiasm, given the recent evidence of clinical efficacy, including a survival advantage in the setting of advanced prostate cancer [27, 28]. However, these studies use monocyte-derived DCs maturated with TNFα, IL-1β, IL-6 and PGE2 for ‘standard DCs (sDCs)’, and recent data suggested that the current protocols used to generate mature monocyte-derived DCs may not result in optimal Th1 responses . Endogenous DCs may exhibit dysfunction due to the overall immunosuppression observed in patients with prostate cancer , and Koh et al.  reported that increased DC function may attribute to improved vaccination outcomes.
Aside from DC function, tumour antigens are important in the development of antigen-specific immunotherapy. Tumour antigens whose expression is necessary for tumour propagation offer a way to target tumour escape variants. The clinical success of previously mentioned studies using DC-tumour antigen peptides was limited. Also, whole tumour cell vaccines have multiple tumour antigens to stimulate immune system; however, these cells often lack costimulatory molecules .
Our study differs from the above in several important aspects. We used priming DCs with dying bodies from allogeneic cell lines. In addition to inducing higher numbers of tumour-specific T cells, the use of Th1-polarized αDC1s, with selectively increased ability to attract effector (rather than regulatory T cells) and to induce the IL-12-dependent effector (CTL) pathway of differentiation in CD8 T cells, may help to induce tumour-specific immunity of a more desirable pattern than either whole tumour cells or standard non-polarized, sDCs [33, 34]. Wieckowski et al.  reported that αDC1s generated from the blood of patients with prostate cancer were highly effective in inducing functional CD8 T cells. However, enrolled patients in their study were removed prostate previously and PSA only disease. Also, they did not show efficacy of CTL response to autologous tumour cells. Compared to their study, we tested the feasibility of employing αDC1-based vaccines for the treatment of chemo-naive CRPC patients with autologous tumour tissue. The GVAX, which is used whole tumour mixture (PC-3 cell), was successful in all measures in preclinical and early clinical trials, but ineffective as monotherapy in phase 3 trials in asymptomatic CRPC, and associated with increased cardiac deaths in combination with chemotherapy . Thus, in spite of the good immune data presented our study, further in vivo and preclinical studies are needed to the development of immunotherapy using αDC1s as a novel treatment for CRPC.
In clinical practice, allogeneic prostate cancer cells prepared from PC-3 cells might be an effective source of universal tumour antigen that could be used for loading onto αDC1s for the generation of prostate cancer–specific CTLs. This approach is similar to recent studies in other diseases showing that fusion cells generated by autologous DCs and allogeneic tumour cell lines could generate antigen-specific CTL responses [37, 38]. Vaccination with autologous DCs primed with allogeneic tumour cell lines has potential applicability to the field of antitumour immunotherapy; it provides a method that allows to overcome the difficulty of obtaining a sufficient number of tumour cells to be used as the source of tumour antigens. These universal cell lines would provide several advantages for clinical applications, such as no limiting factor with regard to the preparation of tumour cells, well-characterized features as a source of tumour-associated antigens, and no need for HLA typing .
In this study, we showed that autologous αDC1s loaded with dying PC-3 cells could prime T cells to generate the potent CRPC-specific CTLs, which can kill prostate cancer cells in patients with CRPC. More importantly, the PC-3-loaded αDC1s from the patients could induce CTL activity against not only autologous prostate cancer cells but also a relevant prostate cancer cells (DU-145). The inhibition of MHC class molecules by the use of anti-MHC class antibody prevented the development of CTL responses specific for autologous prostate cancer cells from patients. Thus, the use of an allogeneic tumour cell line as a source of tumour antigens could generate MHC class-restricted T cell responses against autologous prostate cancer cells.
In conclusion, our results showed that autologous αDC1s loaded with allogeneic dying CRPC cells could generate strong prostate cancer–specific CTLs against autologous CRPC cells. This may offer a highly feasible and effective method for DC-based cancer immunotherapy in patients with CRPC. Further studies will be necessary to enhance the in vivo antitumour effect of CRPC-specific CTLs generated by autologous dying CRPC cells against CRPC before clinical application.
This study was financially supported by Grant No. RTI05-01-01 from the Regional Technology Innovation Program of the Ministry of Commerce, Industry and Energy, Republic of Korea, by Grant No. A110167 of the Korean, Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea, by a grant (CRI 09082-1) Chonnam National University Hospital Research Institute of Clinical Medicine, Republic of Korea, and by grant no. 2011-0030034 from Leading Foreign Research Institute Recruitment Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (MEST), Republic of Korea.