Thomsen-Friedenreich (TF) antigen as a target for prostate cancer vaccine: clinical trial results with TF cluster (c)-KLH plus QS21 conjugate vaccine in patients with biochemically relapsed prostate cancer

Abstract

The differential overexpression of self-antigens on tumor cells is a prime feature of malignant transformation. Thomsen-Friedenreich (TF), a core disaccharide of O-glycosylated complex glycoproteins, is one of many “self” antigens expressed on malignantly transformed cells that has served as a target for immune recognition and attack. Previously, we conducted clinical trials with a series of synthetic glycolipid, peptide and carbohydrate antigens conjugated to the immunological carrier keyhole limpet hemocyanin (KLH) mixed with the immunological saponin adjuvant, QS21. These trials resulted in the generation of high-titer IgM and IgG antibody responses specific for the individual antigens, and, in several cases, the capacity of those antibodies to mediate complement lysis. Four groups of five patients who had evidence of a biochemical relapse defined as rising prostate-specific antigens (PSAs) following primary therapy for prostate cancer with either prostatectomy or radiation were treated with escalating doses of 1, 3, 10 and 30 μg of synthetic TF in a clustered formation (c) which was conjugated to KLH and given with 100 μg of QS21. Patients received a total of five subcutaneous vaccines over 6 months and were monitored expectantly with scans every 3–4 months. Serum samples were obtained at weeks 1, 2, 3, 7, 9, 13, 19, 26, 50 and every 3 months. Antibody titers were monitored by ELISA and antibody binding to the cell surface of prostate cell lines was performed by flow cytometry. Complement-dependent cytotoxicity was performed on selected patients. Twenty evaluable patients were accrued to the study, of whom only one did not receive all six vaccinations. All patients developed maximum IgM and IgG antibody titers by week 9. The median IgM antibody titer by week 7 was 1/1,280 at 10 μg, 1/320 at 30 μg, 1/1,280 at 3 μg and 1/1,280 at 1 μg dose groups. The IgM titers from all groups remained greater than 1/320 by week 32 and beyond through week 50. We report here the results of a dose-escalating trial of a TF(c)-KLH conjugate vaccine in patients in the clinical state of a rising PSA in the absence of radiographic disease. For the first time, a synthetically made TF trimer or cluster (c) was made with three TF disaccharides attached to three sequential threonines on a peptide backbone. TF(c) doses of 1, 3, 10 and 30 μg were conjugated to KLH and administered with QS21. All doses induced high-titer IgM and IgG antibodies against TF. Unlike our findings in previous dose-escalating phase I trials, there did not appear to be increased antibody production with increasing doses of vaccine; higher titers of IgM and IgG antibodies developed at the lowest dose level (1 μg). An anti-tumor effect in the form of a change in post-treatment versus pretreatment logPSA slopes was also observed. The results justify the inclusion of TF(c) at a dose of 1 μg as a relevant antigenic target in a multivalent phase II vaccine trial in patients in the high-risk minimal disease state.

Introduction

Thomsen-Friedenreich (TF) antigen is a disaccharide oncofetal blood group-related antigen (Galβ1-3GalNAcα) normally O-linked to serines or threonines of mucins expressed on cells at the secretory borders of epithelial tissues [1–4]. It is a chemically well-defined carbohydrate antigen with a proven link to malignancy and plays a leading role in docking breast and prostate cancer cells onto endothelium by specifically interacting with endothelium-expressed beta-galactoside-binding protein, galactin-3 [5]. TF-bearing glycoproteins can mobilize galactin-3 to the surface of endothelial cells, thus priming them for interaction with metastatic cancer cells. The widespread expression of TF at the secretory borders of endothelial tissues does not induce immunological tolerance or autoimmunity once antibodies are formed, suggesting that at this site they are sequestered from the immune system [6]. Immunohistochemical stains using an IgM monoclonal antibody 49H.8 against TF showed that two of nine metastatic and 10 of 11 primary prostate cancers also showed >50% of cells reactive. Minimal staining was seen on prostate glandular epithelia [1–4].

Relatively few studies have been performed with TF antigen; so much of what we assume about TF is drawn by inference from two closely related antigens, Tn and sTn. Immunization with Tn and TF has been shown to protect mice from subsequent challenge with syngeneic cancer cell lines expressing these antigens [1]. Patients immunized with synthetic sTn-KLH and TF-KLH conjugate vaccines plus various adjuvants can generate high-titer IgM and IgG antibodies [7], but much of the reactivity was against antigenic epitopes present on the synthetic vaccine constructs which were not present on the naturally expressed mucins on tumor cells. Our studies with MoAb B72.3 resulted in the observation that there was preferential recognition of trimers or clusters of sTn rather than the single disaccharide molecules [8]. The availability of synthetic sTn clusters (c) enabled us to prove this hypothesis. In both direct tests and inhibition assays, B72.3 recognized sTn(c), and sera from mice immunized with sTn(c)-KLH reacted strongly with both natural mucins and tumor cells expressing sTn. Also, unlike our previous experience with unclustered sTn-KLH vaccines, all post-vaccination sera from patients vaccinated with sTn(c)-KLH reacted strongly with natural mucins and sTn-positive human tumor cells [9]. TF(c) consists of TF disaccharide molecules covalently attached to the chains of three serines or threonines, one disaccharide per amino acid residue to form clusters. TF(c)s are successfully synthesized by our group and conjugated to KLH, making it available as an immunogen for vaccination for the first time.

Our previous successful attempts at breaking immunological tolerance have used a series of synthetic mucin-related or glycolipid antigens conjugated to KLH and administered with the saponin immunological adjuvant, QS21 [10]. Men with microscopic disease recurrence as manifested by a rising biomarker, prostate-specific antigen (PSA), and no radiographic evidence of disease were eligible. This approach has been successful in inducing high-titer specific antibodies, suggesting that the immune system can recognize these “self” antigens. In many cases, stabilization of PSA slopes or diminution of logPSA slopes were observed with disease stabilization for close to 2 years [8]. Here we present our experience with a range of doses of TF(c)-KLH and suggest it is one of an increasing number of target antigens which may be used for immunologic strategies. The primary objectives of this phase I trial were: (1) to determine an optimal dose of TF(c)-KLH conjugate plus the immunological adjuvant QS21 that induces an antibody response to TF(c) and (2) to assess the safety of immunization with this conjugate vaccine. The secondary endpoints were to assess the post-immunization changes in PSA levels.

Materials and methods

TF(c)-KLH synthesis

Thomsen-Friedenreich in a clustered formation was synthesized on a threonine backbone by the glycal assembly method in the Laboratory of Bio-Organic Chemistry [11–13]. It was then covalently attached to KLH using a bifunctional cross-linker MBS [8] (Fig. 1). The epitope ratios were calculated by determining the amount of KLH using the Bio-Rad dye binding method as described in the manual and the amount of carbohydrate by the high-pH anion exchange chromatography pulsed amperometric detection (HPAEC-PAD). TF(c)-KLH ratios for 3, 10 and 30 μg were 466/1 and for the 1 μg dose 579/1, assuming a KLH molecular weight of 8.6×10⁶. TF(c)-KLH plus QS21 vaccine is used under an IND with the FDA held by MSKCC. Until vaccination, the TF(c)-KLH vaccine [equivalent to 1, 3, 10 μg of TF(c) in TF(c)-KLH] plus 100 μg of QS21 in 1 ml were prepared and stored at −70°C.

Fig. 1.

Synthetic structure of TF in the trimer (cluster formation) conjugated to KLH

Patient eligibility

Patients with biochemically relapsed prostate cancer were eligible for the study with the following restrictions: Karnofsky performance status >60%, white blood count ≥3,500/mm³, platelet count ≥100,000/mm³, bilirubin <2.0 mg /100 ml or serum glutamate-oxaloacetate transaminase <3.0 times the upper limit of normal, and serum creatinine ≤2.0 mg/100 ml or creatinine clearance ≥40 ml/min. Patients had to have recovered from the toxicity of any therapy and not have received chemotherapy or radiation therapy for at least 4 weeks prior to entry into the trial. No history of an active secondary malignancy within the prior 5 years, except for nonmelanoma skin cancer, was permitted. All patients gave informed consent. Allergy to seafood was a ground for exclusion, as was any history of rectal bleeding, save that of hemorrhoids, or a history of documented radiation-induced proctitis.

Patients were not enrolled if they had radiographically evident disease. To enter the trial, patients had to have had three consecutively rising PSA values determined at intervals of at least 2 weeks based on the following criteria: patients who had undergone prostatectomy had to have a minimal PSA entry value of 1.0 ng/ml with a 50% increase in range of values; patients who had undergone radiation therapy had to have minimal PSA entry of 2.0 ng/ml with a 50% increase in range of values. Patients with primary treatment who were on intermittent hormonal therapy were also permitted with no defined minimal PSA value, as long as they had non-castrate levels of testosterone (>50 ng/ml) and rising PSAs. Patients whose disease was resistant to hormonal therapy were excluded from the trial. Written informed consent was obtained from each patient.

Dose and immunization schedule

The immunization schedule was derived from our studies with other glycoprotein and carbohydrate conjugate vaccines in patients with colon [7], prostate [8], breast [14], and ovarian [15] cancers and melanoma [16]. Three sequential groups of five patients received TF(c)-KLH vaccine containing 3, 10, and 30 μg of TF(c) plus 100 μg of QS21 per vaccination corresponding to treatment groups 1, 2 and 3, respectively. The fourth group included five patients who, based on results of the previous dose levels, would go on to receive a lower (3 μg) or higher (100 μg) dose of the vaccine depending on whether antibody titers were higher with increasing doses. If the antibody response at the 3 μg dose was similar to that for the 10 μg and 30 μg doses, five patients were treated with 1-μg of TF(c)-KLH plus QS21. One hundred micrograms of QS21 was included in all vaccines. Five vaccinations were administered subcutaneously to random sites on the upper arms and upper legs during weeks 1, 2, 3, 7, and 19 with an additional booster vaccine at week 50.

Vaccine doses per vaccination (μg) for each cohort of 5 patients:

TF(c)-KLH + QS21 : 1 + 100 μg QS21

TF(c)-KLH + QS21 : 3 + 100 μg QS21

TF(c)-KLH + QS21 : 10 + 100 μg QS21

TF(c)-KLH + QS21 : 30 + 100 μg QS21

Serologic assays

Serum samples were obtained at weeks 1, 2, 3, 7, 9, 13, 19, 26, and 50 and then every 3 months. Patients were monitored expectantly upon completion of the vaccine trial on a monthly basis for the first 6 months, then at 3-month intervals thereafter, at which time routine blood samples were collected for both immunologic studies and biochemical markers such as PSA and acid phosphatase. IgM and IgG antibody titers were measured by ELISA as described previously, [17–19] and IgG subclasses were determined by using IgG subclass-specific antibodies (Zymed, San Francisco, CA, USA). Briefly, 96-well flat-bottomed plates (Nunc, Rochester, NY, USA) were precoated with TF(c)-HSA at 0.1 μg per well in carbonate buffer. Serially diluted sera in 3% human serum albumin (HSA) were added in triplicate along with sera from patients with known specific high-titer IgM and IgG antibodies or no antibodies, which served as positive and negative controls, respectively. Goat anti-human IgM or IgG conjugated to alkaline phosphatase (AP) (Southern Biotechnology Associates, Birmingham, AL, USA) was used to complete the assay. Plates were read at 10–15 min for IgG and 25–30 min for IgM on an ELISA plate reader (Bio-Rad model 550 Microplate Reader) at 405 nm. The titer was defined as the highest dilution yielding an optical density of ≥0.1.

Flow cytometric analysis

Fluorescence-activated cell sorting (FACS) was performed as previously described [16] to demonstrate antibody binding to the cell surface of the cell lines DU145, a prostate carcinoma, and MCF-7, a breast cancer cell line known to express TF [19]. Pre- and post-immunization sera were diluted twofold to fourfold from their maximum antibody titer determined by ELISA and incubated with DU145 cells or MCF-7 for 1 h on ice. The cells were washed, treated with goat anti-human IgM or IgG labeled with fluorescein isothiocyanate (FITC), and analyzed by flow cytometry (Becton Dickinson, San Jose, CA, USA) for percent positive cells as described previously [19]. Monoclonal antibody HB-T1 (DAKO, Denmark) served as a positive control.

Pre- and post-therapy evaluations

Interval safety assessments included a patient diary. Antitumor and restaging assessments were performed during week 13 and then at 3-month intervals. Patients were seen monthly for the first 6 months, then every 2–3 months thereafter. Blood samples for immune studies were drawn at the time of each visit to the clinic. If the patient did not demonstrate radiographic progression of disease, the patient received a fifth vaccination on week 19 and a sixth at week 50. The patients were restaged every 3 months following completion of the trial at week 26.

Biostatistical assessment of treatment effects—change in PSA slopes

Twenty patients were accrued to the trial, of whom all were evaluable. PSA measurements were collected serially before, during, and after vaccination. For this analysis, only those PSA measurements collected within 72 weeks prior to vaccination were used to calculate the prevaccine log PSA slopes. The treatment period was from week 0 to week 26. PSA measurements collected up to 50 weeks after completing the vaccination were used to calculate the postvaccine log PSA slopes. All PSA measurements collected during the initial 26-week vaccination period were excluded due to the observation that all PSAs continued to rise during the initial immunization period of 26 weeks. Percent change in log PSA slope was summarized. A 95% bootstrap-t confidence interval [20] was constructed for the percent difference in log PSA slopes between pre- and post-vaccine therapy.

Results

Patient profiles

Twenty evaluable patients (age range 52–77, median age 70.5 years) were accrued to the trial. Only one patient did not receive all six vaccinations due to sudden sepsis, resulting in death. Ten patients underwent localized radiation therapy (includes 3-D conformal and/or interstitial seed implants) to the prostate, eight underwent prostatectomy, with one patient undergoing treatment with intermittent hormonal therapy (defined as treatment either with the anti-androgen bicalutamide and the GnRH agonist leuprolide or with leuprolide alone) and one patient receiving chemotherapy prior to prostatectomy. Nine patients were taken off the study due to progression of disease, with one patient removed due to PSA anxiety and another due to death from endocarditis, which was unrelated to vaccine treatment.

Vaccine safety monitoring

Nineteen patients completed all six immunizations with the remaining patient receiving five of the six vaccinations. Of the 20 patients treated with TF(c)-KLH, 7 (35%) experienced grade I and 14 (70%) experienced transient grade II local reactivity, which included redness, tenderness and swelling at the injection site lasting 48–72 h. All 20 patients experienced immediate grade 1 pain during immunization. Five patients experienced grade 1 flu-like symptoms and fever lasting 24–48 h, with one patient having grade 2 fever and two patients showing grade 2 flu-like symptoms. Five (33%) patients complained of transient grade I arthralgias/myalgias. Six patients developed pruritic symptoms.

ELISA assessment

All patients developed maximum IgM and IgG antibody titers by week 9 (Fig. 2a, b). The median IgM antibody titers against TF(c)-HSA by ELISA prior to vaccination were 1:10 or less in all four treatment groups. Although antibody titers started to rise vigorously by week 3, peak titers developed between weeks 7 and 9. The median IgM antibody titer by week 7 was 1/1,280 at 10 μg, 1/320 at 30 μg, 1/1,280 at 3 μg, and 1/1,280 in the 1 μg dose group. The IgM titers from all groups remained greater than 1/320 by week 32 and beyond through week 50. Similarly, IgG antibodies (Fig. 2b) were maximal at weeks 7 through 9, with the 10 μg dose level showing no decline in titer through week 26, in contrast with the remaining three dose levels which declined markedly by week 19. IgG antibodies were slightly lower, with median peak titers of 1/1,280 at 10 μg, 1/640 at 30 μg, 1/720 at 3 μg and 1/320 at 1 μg dose levels, respectively. Antibodies were mainly of the IgG₁and IgG₃ subtype as determined by ELISA. All patients received a booster immunization by week 50. Antibody titers did not increase significantly with the sixth (booster) immunization. Overall, the IgM and IgG titers at the 1 μg and 3 μg doses were at least as high as at the higher doses. Every patient at the 1 μg dose level had peak IgM and IgG titers of at least 1/320 and 1/160, respectively.

Fig. 3.

a, b Representative changes in PSA slopes in two patients. Slopes did not change during the first 26 weeks of the trial; hence, only pre- and posttreatment slopes are shown

Fig. 2.

a, b Time course of median IgM and IgG antibody titers, respectively. Note higher sustained titer of IgM than of IgG

FACS reactivity of pre- and posttreatment sera with DU145 cell line

Reactivity of pre- and peak titer posttreatment sera was assayed by FACS against DU145 cells. Pretreatment sera were gated to reactivity of 10% positive cells and compared with posttreatment reactivity (Table 1). In patients vaccinated with TF(c)-KLH, the IgM antibodies in posttreatment sera of 6 of 15 patients screened at the 1, 10 and 30 μg dose levels demonstrated a two- to fivefold increase in percent positive cells. None of the 15 patients showed any change in IgG antibody reactivity of posttreatment sera compared with pretreatment sera against DU145.

Table 1.

FACS analysis of patients’ sera against prostate cancer cell line DU145

Dose level ( μg)	Patient#	Anti-human IgM (% Pos cells/ MFI*)
		Pre	Post
1	16	11/168	17/45
	17	11/58	18/42
	18	10/6	56/111
	19	11/21	48/36
	20	11/33	33/69
3	11	10/23	77/78
	12	10/28	20/25
	13	10/15	19/16
	14	10/64	3/25
	15	10/25	73/60
10	1	4/ 53	11/72
	2	10/149	19/260
	3	10/73	18/105
	4	10/60	15/67
	5	10/46	15/52
30	6	11/39	7/30
	7	11/22	19/26
	8	10/ 26	41/46
	9	10/29	11/34
	10	9/29	15/39

Higher reactivity noted against DU145 compared with the MCF-7 cell line

Pre-treatment sera from week 1; post-treatment sera from week 7. *MFI Mean fluorescent intensity.

Controls: Positive – MoAb HBT-1 + anti-mouse IgM (53%); Positive pt serum + anti-human IgM (77%); Positive pt sera + anti-human IgG (2%); Negative – anti-mouse IgM (1%); anti-mouse IgG (0.4%); FITC anti-human IgM (0.25%); FITC anti-human IgG (0.10%)

Complement lysis

Sera from ten patients (five patients from the 3 μg and five patients from the 1 μg dose level) were analyzed for the ability to mediate complement lysis using the MCF-7 cell line previously described [19]. One of five patients at the 1 μg dose showed significant increase in lysis between pre and postimmune samples, showing pretreatment lysis of 2.8% compared with posttreatment lysis of 27% at 100:1 dilution. No complement-mediated cytotoxicity was detected in any of the other sera studied.

Analysis of treatment effect

There appeared to be a small correlation between antibody titer and time to radiographic progression of disease in bone, with patients in the 30 μg dose level having lower antibody titers and progression of disease in bone, in two patients within 11 months of starting vaccine. One patient elected to stop treatment due to PSA anxiety and another developed recurrence of disease in the prostate bed. One patient in this group remained active and free of disease after 54 months. Within the first treatment cohort of 10 μg, two patients remained active and free of disease for 55 months, respectively, with the remaining three patients developing progression within a median time of 23 months. Similarly, within the 3 μg dose level, one patient remained without disease for 50 months with another patient having relapse in the prostate bed after 51 months. The remaining three patients had a median progression in bone at 13 months. Of those in the 1 μg dose level, three patients remained active at 36 months; one patient developed relapse in the prostate bed at 13 months; one patient elected to stop treatment due to PSA anxiety after 23 months.

Alteration in PSA

As seen in previous studies with globo H-KLH vaccine [19], and those involving treatment with MUC-1-KLH as well as Tn(c)-KLH and Tn(c)-PAM vaccines [10], patients developed slowly rising PSAs over the first 26 weeks of the trial. The demographics regarding the histologic grade of the tumor and initial level of tumor progression are seen in Table 2. Six (30%) patients had a pretreatment logPSA slope between 0 and 0.05, while 10 patients (50%) had a pretreatment logPSA slope between 0.05 and 0.0999. The four remaining patients had pretreatment logPSA slopes greater than 0.0999. Median pretreatment logPSA slope was 1.9. Median pretreatment doubling time was 11.1 months. Median pretreatment PSA was 5.8 ng/ml (range 1.3–33.7); median pretreatment logPSA was 1.9 ng/ml (range 0.8–3.5).

Table 2.

Demographics of patients enrolled in trial and responses by changes in log PSA slopes

	Number of patients (%)
Pretreatment:
Gleason grade <7	3 (16)
Gleason grade ≥7	16 (84)
Unassigned	1
Clinical stage:
T1a	1 (5)
T1c	3 (15)
T2	14 (70)
T3	2 (10)
Median doubling time	11.1 months
Median PSA	5.8 ng/ml (range:1.3–33.7)
Median logPSA	1.9 (range: 0.8–3.5)
Posttreatment:
Six months –
logPSA slope ↓ (≥25%)	7 (36)
logPSA slope ↑ (≥25%)	6 (32)
logPSA slope ↔	6 (32)
Twelve months –
log PSA slope ↓ (≥25%)	11 (55)
logPSA slope ↑ (≥25%)	6 (30)
logPSA slope ⇔	3 (15)
More than 12 months to current logPSA slope ↓ (≥25%)	13 (65)
Radiographic progression
No. progressed - (4 bone; 3 LN, 1 exp)	8
No. stable	12

By 6 months after vaccination, 6 (32%) patients’ logPSA slope had increased 25% more, while 7 (35%) patients’ logPSA slope had decreased at least 25%. The remaining 6 (32%) patients’ logPSA slopes were between these values. After 1 year, 11(55%) patients’ logPSA slopes had declined at least 25% while the number of patients whose logPSA slopes had increased 25% or more was down to 6 (30%). Three (15%) patients had logPSA slopes postvaccination somewhere in between. Using assessments up to the present, 13 (65%) patients had logPSA slopes which had decreased at least 25% from prevaccination, while 4(20%) patients had logPSA slopes which had increased at least 25% from prevaccination. The other three patients had logPSA postvaccine slopes between these values; however, after 6 months 19 (100%) had logPSA slopes which were continuing to rise. At 12 months post vaccination, 6 (30%) had decreasing logPSA slopes while 14 (70%) still had rising logPSA slopes. Eight patients’ disease progressed, while 12 patients remained progression-free or were censored. Median time to progression had not been reached. Median follow-up for those who did not progress or were censored was 30 months.

Discussion

Synthetic TF(c) conjugated to KLH and given with the saponin adjuvant, QS21, has been shown to elicit high-titer and durable IgM antibodies, albeit with moderate cell-surface reactivity. The TF molecule is interesting for several reasons: (1) it is widely expressed on prostate cancer cells, (2) its expression is associated with poor prognosis in several malignancies [1, 6], and (3) its interaction with galactin-3 during tumor cell docking into the endothelium may lead to the development of alternate immunologic strategies which may interfere or prevent micrometastatic disease [5]. TF and its precursor, Tn, are surprisingly simple mucin-bound carbohydrates with expression that appears to correlate with prognosis [6]. TF is one of the several self-antigens expressed on prostate cancer cells that are to be included in a larger multivalent trial which will investigate whether the combination of several synthetic antigens conjugated to KLH and given with QS21 can enhance the induction of antibodies and exert a more pronounced antitumor effect.

We present the results of a clinical trial using TF, which was synthesized in a cluster formation to simulate TF expression on cancer cell mucins. The impact of clustering was observed with Tn and sTn antigens. Clustering appears to be a means of more clearly mimicking the way these simple antigens are expressed on epithelial cancers but not normal tissues. Vaccines designed to induce an optimal antibody response have several components, each of which must be optimized. The first component is the antigen itself, which must closely resemble its expression on the target, in this case TF(c) expression on tumor mucins. While the TF antigen is a disaccharide covalently attached to serine or threonine, akin to the closely associated disaccharide sTn, MoAbs and sera selected for preferential reactivity with cancer cells (as opposed to normal tissues) react with clusters of three such disaccharide antigens.

The second component found to be necessary for an optimal antibody response is the antigen that is covalently conjugated to an immunogenic carrier protein. We have found KLH to be the optimal carrier for antibody induction. The conjugation must be achieved in a manner that does not interfere with the antigenic epitope itself [TF(c)] and that achieves as high an antigen/carrier ratio as possible [in this case 466 or more TF(c) molecules per KLH molecule]. This was achieved using the MBS hetero-bifunctional cross-linker which links the terminal cysteine group of the cluster backbone to amino groups on KLH. The final necessary component is the immunological adjuvant. In our experience, saponin adjuvants such as QS21 have been the most potent for augmenting the antibody response against conjugate vaccines. The 100 μg dose level of QS21 used here was found to be optimal, with higher doses resulting in excessive local and systemic toxicity and lower doses resulting in decreased immunogenicity.

The toxicity of the TF(c)-KLH plus QS21 vaccine utilized here was comparable to that observed in previous trials with a range of other antigen-KLH conjugates plus QS21. As in the previous trials, all patients experienced grade I or II local erythema and induration at the injection sites and some patients experienced grade I or II systemic flu-like symptoms lasting 1–4 days. No clinical signs of autoimmunity were observed, and there was no evidence of intradermal reactivity to the TF(c) skin tests at any time point.

The use of logPSA slopes as a measure of antitumor effect is under consideration as a means of assessing antitumor effect but remains to be validated [21]. Several other studies have demonstrated similar effects with PSA slopes using vaccines comprised of dendritic cells [22] or genetically altered prostate cancer cell lines given with systemic cytokines such as GM-CSF [23]. The clinical impact of stabilization or decline of posttreatment compared with pretreatment PSA log slopes and its relevance as an intermediate endpoint remains to be validated. The trial described here, assessed the safety of TF and its potential for inducing antibodies when used in a cluster formation. It did not address the relevance of a change in PSA in prognosticating the clinical course in this patient population and there are as yet no data to support a direct correlation between impact on PSA and a change in the biology of the tumor. It is important to identify patients who are at significant risk of radiographic recurrence as early as feasible and consider these patients for investigational approaches while they still have a low tumor burden [24]. Whether a multivalent vaccine will work better as a modality or in combination with hormonal therapy or even chemotherapeutic regimens in high-risk patients in early relapse with minimal residual disease is currently under consideration.

Abbreviations

TF

Thomsen-Friedenreich

TF(c)

TF in a clustered formation

sTF

Sialyted TF

PSA

Prostate-specific antigen

KLH

Keyhole limpet hemocyanin

HSA

Human serum albumin