Vaccination of prostatectomized prostate cancer patients in biochemical relapse, with autologous dendritic cells pulsed with recombinant human PSA

Abstract

This study was conducted in prostate cancer patients in biochemical relapse after radical prostatectomy, to assess the feasibility, safety, and immunogenicity of therapeutic vaccination with autologous dendritic cells (DCs) pulsed with human recombinant prostate-specific antigen (PSA) (Dendritophage-rPSA). Twenty-four patients with histologically proven prostate carcinoma and an isolated postoperative rise of serum PSA (>1 ng/ml to 10 ng/ml) after radical prostatectomy were included. The patients received nine administrations of PSA-loaded DCs by combined intravenous, subcutaneous, and intradermal routes over 21 weeks. Postbaseline blood tests were performed at months 1, 3, 6, 9, and 12 (PSA levels), at months 6 and 12 (circulating prostate cancer cells), at month 6 (anti-PSA IgG and IgM antibodies), and at up to eight time points before, during, and after immunization (PSA-specific T cells). Circulating prostate cancer cells detected in six patients at baseline were undetectable at 6 months and remained undetectable at 12 months. Eleven patients had a postbaseline transient PSA decrease on one to three occasions, predominantly occurring at month 1 (7 patients) or month 3 (2 patients). Maximum PSA decrease ranged from 6% to 39%. PSA decrease on at least one occasion was more frequent in patients with low Gleason score (p=0.016) at prostatectomy and with positive skin tests at study baseline (p=0.04). PSA-specific T cells were detected ex vivo by ELISpot for IFN-γ in 7 patients before vaccination and in 11 patients after vaccination. Of the latter 11 patients, 5 had detectable T cells both before and during the vaccination period, 4 only during the vaccination period, while 2 patients could for technical reasons not be assessed prevaccination. No induction of anti-PSA IgG or IgM antibodies was detected. There were no serious adverse events or otherwise severe toxicities observed during the trial. Immunization with Dendritophage-rPSA was feasible and safe in this cohort of patients. An immune response specific for PSA could be detected in some patients. A notable effect was the disappearance of circulating prostate cells in all patients who were RT-PCR positive before vaccination.

Introduction

Efforts in cellular immunotherapy have in recent times focused on antigen-presenting cells (APCs) that are capable of directing the immune system towards malignant cells. APCs process and present antigens on their major histocompatibility complex (MHC) molecules eliciting specific cytotoxic and helper T lymphocytes to generate cellular and humoral immune response. Major APCs are dendritic cells (DCs), macrophages, and, to a lesser extent, B lymphocytes [30]. DCs derive from CD34⁺ MHC class II negative precursors present in the bone marrow under the influence of cytokines and growth factors, e.g., IL-4, GM-CSF, and TNF-α. DCs are among the most potent APCs known and acquire this property during their maturation process [20]. DCs are mandatory for the activation of naïve T cells [4, 24] linked to the particularity of being capable of presenting exogenous antigens through both MHC class I and MHC class II pathways [12, 39]. This dual capacity explains their pivotal role in the initiation of T-cell responses and motivates an investigational use in immunotherapy.

A number of tumor-associated antigens (TAAs) have been identified [33]. They should in theory be recognized in vivo as non-self and trigger a major immune response leading to tumor destruction. However, they may escape the immune surveillance due to various mechanisms such as altered myeloid cell maturation [2, 3, 18] or induction of tolerance [5]. One attempt to overcome resistance mechanisms is to deliver the antigens to the DCs under controlled conditions in vitro, free from putative inhibitory factors. Immunotherapy based on the use of ex vivo generated DCs as natural adjuvants constitutes an attractive approach to cancer treatment. This approach is supported by the demonstration in experimental animal models that DCs, when pulsed ex vivo with tumor antigens and administered to tumor-bearing hosts, can elicit T-cell-mediated tumor destruction and memory response that protect against further tumor challenges [13, 23, 37]. In humans, in vitro DCs can be generated from CD14⁺ blood monocytes concentrated by apheresis, using a culture medium enriched with GM-CSF, IL-4, or IL-13 [11, 31, 32, 35]. Once pulsed ex vivo with TAAs, the DCs can then be reinfused into the patient. Early clinical trials [16, 25] indicate that DC-based therapy can lead to significant therapeutic benefit in cancer patients. Available data also indicate that cell therapy is not associated with any significant toxic side effects.

Radical prostatectomy provides excellent long-term local control of the primary tumor for many patients. Typically, after radical prostatectomy, serum PSA drops rapidly and should be undetectable within 3 or 4 weeks. However, 15% to 40% of the patients will present a rise of PSA level above the detection limit for residual cancer level within 5 years [26], and there is no clear consensus on the management of these patients [36]. The pretreatment PSA level is a dominant determinant of outcome whether treatment consists of radiation or surgery, but TNM staging and tumor differentiation (Gleason score) also have prognostic value for local or metastatic relapse [40]. In addition, the presence of circulating prostate cells at diagnosis appears to be a prognostic factor for later metastasis [17]. These observations prompted the initiation of a phase I/II study to assess the effects of therapeutic vaccination with autologous DCs pulsed with recombinant human PSA protein (Dendritophage-rPSA) in prostate cancer patients with no clinical manifestation of the disease following radical prostatectomy but with increasing serum PSA, suggestive of a subclinical local or metastatic relapse.

Methods

Patient population and study design

This was a phase II, nonrandomized, multicenter, open label study. All patients had an isolated biochemical relapse as reflected by serum PSA in the range of 1 to 10 ng/ml after radical prostatectomy for T1-T3, N0, M0 prostate adenocarcinoma. Gleason score and immunohistochemistry characterization had been determined from prostatectomy specimens. The increase in PSA had to be documented over three consecutive determinations, with the last PSA value ≥1 ng/ml. Patients with PSA values >10 ng/ml were not eligible, even if there was no evidence of clinical disease. Bone and pelvic-abdominal scans were performed before inclusion in the study to exclude the presence of any detectable but asymptomatic metastases. Patients previously treated with radiotherapy, antiandrogens, or chemotherapy were excluded. The patients received repeated injections of autologous DCs pulsed with rPSA over 5 months and subsequently were followed up at months 6, 9, and 12, in order to assess the safety of the treatment and to detect a possible late response. Patients were to be withdrawn from the study in case of unacceptable toxicity, cancer progression suggested by symptoms and confirmed by complementary investigations, or patient’s refusal. The study protocol was approved by the ethics committee (CCPPRB) of Hôpital la Pitié-Salpétrière, Paris (France), and signed informed consent was obtained from all patients. The study was conducted according to the International Conference on Harmonisation (ICH) Good Clinical Practices (GCP) and in compliance with the Declaration of Helsinki and subsequent amendments.

Study treatment

Dendritic cells were differentiated from monocytes with GM-CSF and IL-13 as previously described [11]. Briefly, about 1×10¹⁰ peripheral blood mononuclear cells (PBMCs) were collected before treatment and cultured for 7 days in the presence of 500 U/ml GM-CSF (Novartis, France) and 50 ng/ml recombinant human IL-13 (Sanofi-Synthélabo, France). IL-13 was renewed on day 4 of culture. On day 7, DCs were purified by elutriation and pulsed for 2 h with 20 μg/ml recombinant human PSA (rPSA, Novavax, Maryland, US). Excess rPSA was eliminated by washing. The cells were resuspended in a solution of 4% human serum albumin (LFB, France) and 10% DMSO (B. Braun, France), divided in cryopreservation bags, and then frozen. On a treatment day, a bag was thawed at 37°C, the cryopreservation solution was washed out, and cell count, viability, purity, sterility were reassessed. The DCs were CD11c⁺, CD40⁺, CD80⁺, CD86⁺, and HLA-DR⁺ (but CD83^-) as described elsewhere [11]. Mean purity of the DCs after thawing as assessed by microscopy, and trypan blue exclusion was 95.5% SD ±0.7%. Treatment consisted of nine series of injections of rPSA-loaded DCs at weekly intervals for the first three series, then fortnightly for the following three series and finally at 4-week intervals for the last three series. At each administration series, DCs were administered to six different sites. Four injections were given subcutaneously (both arms and both thighs), one intradermally (anterior side of forearm), and the remaining cells given intravenously. The number of DCs administered per visit ranged from 2.5×10⁷ to 1×10⁸.

Response evaluation

The patients treated had no measurable tumors. The antitumoral response to Dendritophage-rPSA vaccination was evaluated indirectly based on (1) the assessment of circulating tumor cells at baseline, and at 6 and 12 months after the start of the vaccination procedure, and (2) the change in PSA levels from baseline determined at months 1, 3, 6, 9, and 12. A PSA response was prospectively defined as an on-study 50% drop in peripheral blood PSA levels compared with baseline. Circulating prostate cancer cells were assessed from 5-ml blood samples by an RT-PCR assay [38]. The method used detects independently both PSMA and PSA mRNA transcripts using primers derived from PSMA and PSA cDNA sequences, respectively [21]. When applied to serial dilutions of LNCaP cells used as a prostate cancer cell model [15] in blood samples with known leukocyte count, this assay can detect as few as one prostate cancer cell per 10⁶ leukocytes [21] but does not distinguish between viable and nonviable cells. PSA was measured in sera from 5-ml blood samples by means of a time-resolved amplified cryptate emission (TRACE) assay (CIS Bio International, Gif-sur-Yvette, France) with a detection limit of 0.03 ng/ml [22]. PSA results were blinded to both the investigator and the patient during the first 6 months of the study period unless symptoms occurred.

Specific cellular immune response was assessed using an enzyme-linked immunospot (ELISpot) assay to detect PSA-specific, IFN-γ secreting, T cells [6, 14]. The assay was performed on fresh PBMCs (5×10⁵ cells/well) collected at weeks 2, 3, 5, 7, 9, 13, 17, and 21. PBMCs were incubated for 40 h with rPSA (30 μg/ml) or PSA-derived peptides (10 μg/ml). PSA peptides (PSA 141–150: FLTPKKLQCV; PSA 146–154: KLQCVDLHV; PSA 154–163: VISNDVCAQV) [1] were purchased from Cybergene, France, and were >80% pure. Tuberculin (Pasteur-Merieux, France) and a mix of influenza peptide 58–66: GILGFVFTL, and Epstein-Barr virus BMLFA1 peptide: GLCTLVAML (Cybergene, France) were used as positive controls to stimulate PBMC (2×10⁵ cells/well and 5×10⁵ cells/well, respectively). An ELISpot result was considered positive when the mean number of INF-γ secreting T cells in antigen-containing wells differed significantly from the mean number in control wells without antigen (p<0.05; Student’s t-test) and when the mean value exceeded 5 spots after subtraction of the mean mock control.

Skin delayed–type hypersensitivity tests (Multitest IMC, Pasteur Merieux, Lyon, France) were carried out at baseline and at 6 months. The skin test detects antigens from the following seven pathogens: Clostridium tetani, Corynebacterium diphtheriae, group C streptococci, Mycobacterium tuberculosis, Candida albicans, Trichophyton mentagrophytes, and Proteus mirabilis.

All patients were evaluated for safety during 7 months after the last administration. Toxicity of the treatment was assessed through clinical examination and laboratory tests, and graded according to the National Cancer Institute common toxicity criteria (NCI-CTC).

Data analysis

PSA level profile after baseline was characterized by (1) the number of times the value was below baseline level, (2) the time (in months) of first decrease below baseline level, and (3) the maximum decrease expressed as percentage of baseline level. For statistical inference, patients were categorized according to those presenting a PSA decrease below baseline on at least one occasion versus those presenting no decrease. Relationships between this definition of response and the following prognostic factors were investigated using the Fisher exact test: Gleason score at prostatectomy (patients with 5–6, 7, or 8–9 scores); PSA at initial diagnosis <20 ng/ml versus ≥20 ng/ml; time since prostatectomy <3 years versus ≥3 years; total number of DCs injected <3.5×10⁸ versus ≥3.5×10⁸; and number of positive skin-test antigens at baseline (patients with 0, 1–2, or 3–5 positive tests).

Results

Patient characteristics and dosing

Twenty-six patients were included in the study, of whom 24 actually received Dendritophage-rPSA administration. One patient withdrew before apheresis. Another patient underwent two aphereses but was never administered the study therapy since both preparations had to be rejected because of unsatisfactory DC viability. The characteristics of the disease of the 24 patients at initial diagnosis are summarized in Table 1. Mean age was 66.0 years (SD ±6.6 years). The majority of patients (19/24) were diagnosed with a stage T3, N0, M0 prostate adenocarcinoma and a Gleason score of 5–6 (n=3), a Gleason score of 7 (n=16), or a Gleason score of 8–9 (n=5). Time since initial diagnosis and time since radical prostatectomy ranged from 1 to 9 years (mean 3.7 and 3.4 years, respectively), and PSA values at baseline were all within the protocol specified 1–10 ng/ml range. At the time this study was designed, little information was available to guide our choice of injection route. Therefore, to increase chances of an effective vaccination, cells were administered by a combination of subcutaneous, intradermal, and intravenous routes. The mean count of DCs injected per administration ranged from 1.46×10⁷ to 7.22×10⁷. The total number of DCs injected over the trial ranged from 1.31×10⁸ to 6.50×10⁸ per patient. All 24 patients completed the 6-month treatment period. Seven patients were withdrawn posttherapy between month 6 and month 12. Reasons for discontinuation were local recurrence of disease or PSA progression treated with hormonal and/or radiation therapy. Seven additional patients were reported with disease progression but were followed up to the month-12 visit. All 24 patients were alive at the end of the study.

Table 1.

Characteristics of the disease at initial diagnosis and at baseline (n=24)

TNM stage at initial diagnosisa
T1	1
T2b	4
T3c	19
Gleason score at prostatectomy
Mean (SD)	7.1 (0.8)
Median (range)	7 (5–9)
PSA level at initial diagnosis (ng/ml)
Mean (SD)	36.9 (48.5)
Median (range)	21.7 (9–200)
Time since diagnosis (years)
Mean (SD)	3.7 (2.2)
Median (range)	2.9 (1.3–9.3)
Time since prostatectomy (years)
Mean (SD)	3.4 (2.0)
Median (range)	2.8 (1.0–8.6)
PSA level at baseline (ng/ml)
Mean (SD)	3.1 (1.9)
Median (range)	2.4 (1.0–7.6)

^aAll patients were N0, M0

^bIncluding two T2 and two T2C

^cIncluding eight T3, three T3A, four T3B, and four T3C

Safety

Four patients experienced a total of five adverse events considered as reasonably related to the study treatment: macular rash (n=3), asthenia (n=1), and halitosis (n=1). All these adverse events were of mild intensity (NCI-CTC grade 1). Two patients had transient elevation of total bilirubin from grade 0 at baseline to grade 2 during treatment.

Circulating prostate cancer cells

Six patients tested positive for the presence of circulating prostate cancer cells before DC vaccination but were negative after the last injection at month 6. All other patients were negative at both pretreatment and month 6. When the responding patients where subjected to a confirmatory analysis at month 12, the RT-PCR test remained negative for all six patients.

PSA levels

None of the treated patients had a 50% PSA decrease compared with baseline. However, decreases of 6–39% were observed on at least one instance in 11 patients. The decrease occurred as early as the 1st month after the start of vaccination in seven of these patients. It was delayed to month 3 in two patients and to month 5 and month 9 in the last two patients. All decreases in PSA from baseline value were transient and remained below baseline for up to three consecutive time points, or for a maximum of 6 months. The individual PSA profiles are shown in Fig. 1. A statistically significant relationship (p=0.016) was detected between the occurrence of postbaseline PSA decrease on at least one instance and Gleason score at prostatectomy, indicating that PSA decrease did not correlate with high (8–9) Gleason scores. Also, 90% of the patients with PSA decrease versus 46% without PSA decrease had at least one positive skin test at baseline (p=0.040). On the other hand, none of the other prognostic factors investigated showed a statistically significant relationship with the occurrence of postbaseline PSA decrease (Table 2).

Fig. 1a,b.

Time course of serum PSA in patients showing at least one postbaseline decrease. a Patients with first PSA decrease occurring at month 1. b Patients with first PSA decrease occurring at months 3, 5, or 9

Table 2.

Relationship between prognostic factors and PSA decrease after baseline

	Decrease in PSAa				p Valueb
	Yes (n=11)		No (n=13)
	n	%	n	%
Gleason score
5–6	3	27	0	0	0.016
7	8	73	8	62
8–9	0	0	5	38
Number of positive skin tests
0	1	10	7	54	0.040
1–2	7	70	3	23
3–5	2	20	3	23
Missing	1		0
PSA at initial diagnosis
<20 ng/ml	5	45	6	46	1.000
≥20 ng/ml	6	55	7	54
Time since prostatectomy
<3 years	4	36	9	69	0.217
≥3 years	7	64	4	31
Total number of injected cells
<3.5×108	8	73	6	46	0.240
≥3.5×108	3	27	7	54

^aOn at least one occasion

^bFisher exact test

Immunological responses

T-cell responses were monitored ex vivo in all 24 patients among whom 14 were HLA-A2 positive. ELISpot assays for IFN-γ were performed on PBMC collected immediately before the first vaccination, and after the 2nd, 3rd, 4th, 6th, 7th, 8th, and 9th vaccinations. PSA-specific T cells were detected in 13/24 patients. The results obtained in patients with positive responses are summarized in Table 3. All patients had detectable responses to recall antigens (i.e., tuberculin, flu- or EBV-derived peptides), indicating that they were immunocompetent (Table 3 and data not shown). Before vaccination, PSA-specific T cells were detected in seven patients including patients responding to PSA protein only (n=2), PSA protein and one PSA peptide (n=1), or to PSA peptides only (n=4). During the vaccination period, PSA-specific T cells could be detected in a total of 11 patients. Among these, four patients displayed an immune response to PSA protein or PSA-specific peptides postvaccination only, while five patients had detectable PSA-specific T-cell responses both before and after the first vaccination. Prestudy immune status could not be determined in two patients. In addition, two patients had a detectable immune response to PSA pretreatment that was undetectable at later time points. Antigen-specific T cells were detected in five of the six patients with circulating prostate cells at baseline, and among these five patients, three had preexisting T-cell response to the PSA protein. However, in two patients with both circulating prostate cells and PSA-specific T cells detected before vaccination, neither was detected postvaccination (Table 3). PSA-specific effector T cells were usually detected transiently during the vaccination period by direct ELISpot and the frequency of PSA-specific T cells were generally low (mean 15 SFCs/5×10⁵ PBMCs) except for three patients with strong responses to PSA peptides (>60 SFCs/5×10⁵ PBMCs).

Table 3.

Patients with positive PSA-specific T-cell tests. MD missing data; NA not applicable; patient nos. in bold refer to patients with circulating prostate cells at baseline

Patient No.	Time	T-cell response tests: positive/totala
		PSAproteinb	PSA peptidesc			Tuberculind	Flu and EBV peptidese
			1	2	3
0105	Before	0/1	1/1	1/1	1/1	1/1	1/1
	After	0/5	1/5	1/5	2/5	2/5	1/5
0117	Before	0/1	0/1	0/1	0/1	1/1	1/1
	After	0/8	0/8	0/7	1/8	8/8	8/8
0119	Before	1/1	1/1	0/1	0/1	1/1	1/1
	After	0/8	0/8	0/8	2/8	8/8	5/8
0120	Before	0/1	0/1	0/1	0/1	1/1	1/1
	After	1/8	0/8	0/8	0/8	6/8	8/8
0122	Before	0/1	0/1	0/1	1/1	1/1	1/1
	After	0/7	1/7	1/7	1/7	7/7	5/7
0125	Before	0/1	0/1	1/1	0/1	1/1	1/1
	After	0/8	0/8	1/8	0/8	8/8	7/7
0203	Before	MD	MD	MD	MD	MD	MD
	After	0/4	0/4	2/4	2/4	4/4	4/4
0207	Before	0/1	1/1	0/1	1/1	1/1	1/1
	After	0/2	1/2	1/2	0/2	2/2	2/2
0214	Before	0/1	0/1	0/1	0/1	1/1	1/1
	After	0/4	1/4	0/3	0/4	4/4	4/4
0602	Before	MD	MD	MD	MD	MD	MD
	After	0/4	0/4	1/4	1/4	3/4	1/4
0106	Before	1/1	NA	NA	NA	1/1	NA
	After	0/4				4/4
0126	Before	1/1	NA	NA	NA	1/1	NA
	After	0/8				8/8
0210	Before	0/1	NA	NA	NA	1/1	NA
	After	1/6				6/6

^aNumber of positive tests/total number of tests

^bTest responses for PSA specific T cells for both HLA-A2 positive and negative patients

^cTest responses for PSA specific T cells for HLA-A2 positive patients only

^dT-cell response to tuberculin: control of immunocompetence for both HLA-A2 positive and HLA-A2 negative patients

^eT-cell response to Flu + EBV peptides cocktail: control of immunocompetence for HLA-A2 positive patients

PSA-specific IgG and IgM antibodies were assayed in the plasma of treated patients and compared with control samples collected from healthy female subjects. The levels of PSA-binding IgG and IgM antibodies in plasma of vaccinated patients were of the same low magnitude as those observed in normal female plasma, indicating that the therapeutic vaccination with DCs did not induce an antibody response to PSA (data not shown).

Discussion

The reappearance of sustained detectable serum PSA levels after radical prostatectomy in prostate cancer patients, i.e., biochemical recurrence after PSA level had become undetectable, is a well-established sign of probable local or metastatic relapse of the disease [10, 28, 29]. The risk of disease progression is directly related to the T stage, the preoperative PSA levels, and prostatectomy Gleason score [7, 19]. Predictive models taking these parameters into account indicate that patients at high risk of relapse are those with disease stage greater than T2a, combined with high Gleason score (≥6) and high PSA value (>10 ng/ml) [9, 27]. According to these predictive criteria, most of the patients in the present study fell into high-risk categories for relapse. The best treatment but also the timing for intervention for these patients remains controversial. Second treatment options after radical prostatectomy include radiation therapy with or without androgen deprivation, or androgen deprivation alone. Distant relapse may be more likely in those patients with high-grade disease and very high pretreatment PSA levels. Such patients seem to be better candidates for experimental adjuvant therapy whereas those at risk of local relapse (with lower T stage, Gleason score, and PSA level) might be better candidates for radiation delivered to the prostatic bed [8]. These considerations and the promising results of a previous DC-based immunological approach of prostate cancer treatment [34] justified the assessment of Dendritophage-rPSA vaccine treatment of our study patients.

A first indication of an effect of Dendritophage-rPSA vaccination on prostate tumoral target is the disappearance at 6 months of circulating prostate cancer cells in all six patients who were positive at baseline. Positive RT-PCR detection of circulating cancer cells has been shown to be of prognostic value for metastasis, and in patients with hormone-refractory prostate cancer it is predictive of the duration of survival [17, 21, 38]. Although a causative role in the metastatic process is an attractive hypothesis, the possibility that these cells may simply have been shed by undetected metastases can not at present be excluded. In this context, it should be pointed out that three of the six patients with molecular clearance of circulating prostate cancer cells experienced a disease progression during the month 6–12 follow-up that necessitated therapeutic intervention. An effect of the Dendritophage-rPSA vaccination is also suggested by decreases in PSA below baseline (albeit only transiently) observed in 11 of the 24 treated patients. These decreases were principally observed within 1–3 months after the start of vaccination, which is consistent with a treatment-effect relationship. The results are consistent with those previously observed following treatment of hormone-resistant prostate cancer with DCs pulsed with PMSA peptides [34]. The effect on PSA is unlikely to have resulted from a humoral response since no induction of anti-PSA antibodies was detected in any patient. Further analysis demonstrated a statistically significant relationship between the occurrence of PSA decrease and Gleason score (p=0.016). In fact, patients with PSA decreases tended to have low or moderate Gleason scores (<7) at prostatectomy suggesting that such patients may be more likely to exhibit a positive response to Dendritophage-rPSA vaccination. However, in the absence of a control group, we can not rule out the possibility that patients with low Gleason scores could be naturally more prone to variations in PSA levels even in the absence of treatment. The other risk factors investigated failed to show any relationship with the occurrence of a PSA decrease. Again it must be emphasized that all but four patients had preoperative PSA values ≥10 ng/ml, a value regarded as corresponding to high-risk patients [9, 27] possibly more refractory to treatment.

Also of interest is the significant relationship between the occurrence of PSA decrease and the number of positive skin tests at baseline (p=0.040). This suggests that patients who are prone to cell-mediated immune reactions are also more likely to mount a PSA response. PSA-specific T cells could be detected on at least one occasion in a total of 13 patients (2 of whom were for technical reasons not assessed prevaccination; see Table 3). Seven patients had detectable PSA-specific T cells before the first vaccination indicating that a type 1 immune response to PSA might be induced spontaneously during the disease process. In two of these seven patients, PSA-specfic T cells were undetectable during the vaccination period, something for which we currently have no explanation. Four patients only displayed PSA-specific T cells after vaccination. It should be noted, however, that the detected T-cell responses were quite modest and appeared only transiently, which may reflect a certain level of self-tolerance to the soluble human PSA protein. It is also possible that the direct ex vivo approach we used (with no amplification) may not have been adapted to detect a low level response.

In this small pilot study, no statistically significant relationship could be found between PSA-specific T cells and either PSA decrease, circulating prostate cells, or skin tests. A possible way forward includes the use of more antigenic prostate-specific molecules, an optimized antigen-loading process, and an administration scheme carefully designed based on the current knowledge in the field of immunotherapy. These considerations will be taken into account in the preparation of further studies of Dendritophage vaccination.

Dendritophage-rPSA vaccination of prostate cancer patients in biochemical relapse following radical prostatectomy resulted in transient PSA decreases in 11 of the 24 patients and the disappearance of circulating prostate cells in all six patients in whom they were detected at baseline. Dendritophage-rPSA vaccination was well tolerated.