Induction of cellular immune responses against carcinoembryonic antigen in patients with metastatic tumors after vaccination with altered peptide ligand-loaded dendritic cells

Abstract

Purpose: Dendritic cells (DCs) are characterized by their extraordinary capacity to induce T-cell responses, providing the opportunity of DC-based cancer vaccination protocols. In the present study, we conducted a phase I/II clinical trial to determine the capability of DCs differentiated from immunomagnetically isolated CD14+ monocytes and pulsed with a carcinoembyonic antigen-derived altered peptide (CEAalt) to induce specific CD8+ T cells in cancer patients. Experimental design: Nine patients with CEA-positive colorectal cancer (n=7) or lung cancer (n=2) were enrolled in this study. Autologous CD14+ monocytes were isolated by large-scale immunomagnetic separation and differentiated to mature DCs in sufficient numbers and at high purity. After incubation with the CEAalt peptide and keyhole limpet hemocyanin, DCs were administered to patients intravenously at dose levels of 1×10⁷ and 5×10⁷ cells. Patients received four immunizations every second week. Results: ELISPOT analysis revealed a vaccine-induced increase in the number of CEAalt peptide-specific Interferon (IFN)-gamma producing CD8+ T cells in five of nine patients and of CD8+ T lymphocytes recognizing the native CEA peptide in three of nine patients. In addition, CD8+ T lymphocytes derived from one patient exhibiting an immunological response after vaccination efficiently lysed peptide-loaded T2 cells and tumor cells. Immunization was well tolerated by all patients without severe signs of toxicity. Conclusion: Vaccination with CEAalt-pulsed DCs derived from immunomagnetically isolated CD14+ monocytes efficiently expand peptide-specific CD8+ T lymphocytes in vivo and may be a promising alternative for cancer immunotherapy.

Introduction

Dendritic cells (DCs) are professional antigen-presenting cells that play a pivotal role in the induction of immune responses [2, 3, 15]. Based on this unique ability DCs evolved as promising candidates for vaccination protocols in cancer therapy [10]. Animal models have demonstrated that DCs are capable of inducing protective and therapeutic antitumor responses [24, 27]. In addition, clinical trials have shown immune responses after DC vaccination in humans [28, 36]. DC-based vaccination approaches in man require a reproducible DC generation method that can be performed in conformity with good manufacturing practice (GMP) guidelines and that circumvents the need for multiple blood drawings to generate DCs [37].

There are several methods for the generation of clinical scale numbers of DCs. Monocyte-derived DCs are currently most widely used as they can be generated relatively simply without the need for cytokine stimulation of the donor. In addition, those monocyte-derived DCs have been shown to induce immunity in several clinical studies [8, 9, 36]. Most clinical protocols have used DCs derived from plastic-adherent monocytes for repetitive vaccinations. Peripheral blood mononuclear cells (PBMCs) obtained after leukapheresis or buffy coat separation can be used as starting fraction [4]. Plastic adherence leads to a relatively pure fraction of monocytes which can be differentiated into DCs by incubation with GM-CSF and IL-4 [32] or GM-CSF and IL-13 [5, 13, 35].

Besides the plastic-adherence method, immunomagnetic enrichment of CD14+ monocytes using antibodies linked to ferromagnetic beads can generate pure preparations of monocytes. A clinical scale immunomagnetic column (CliniMACS, Miltenyi Biotec, Bergisch Gladbach, Germany) has become available for the positive selection of hematopoietic stem cells (CD34, CD133). A new CD14+ microbead reagent enables the user to select large numbers of monocytes for subsequent clinical use. Several investigators have used the small-scale device to select monocytes for subsequent in vitro use of DC populations [6, 20]. Recently, the clinical-scale column has also been used to produce pure preparations of vital CD14+ monocytes [1, 16, 26, 29]. Monocytes can be differentiated in vitro into mature DCs as indicated by decreasing CD14+ expression, increasing CD83 expression and upregulation of the co-stimulatory molecules CD80 and CD86. The number of DCs generated from one leukapheresis product is sufficient for multiple vaccinations and obviates the need for repetitive blood drawings. In proliferation assays we could demonstrate the ability of fresh and cryopreserved DCs to stimulate allogeneic T cells in vitro [1].

One of the most interesting target antigens is the carcinoembryonic antigen (CEA) because it is overexpressed in the vast majority of gastrointestinal malignancies as well as carcinomas of the lung, breast and thyroid gland. However, the induction of immune responses against antigens expressed not only by tumor cells but also by benign tissues such as CEA is more difficult because rapid induction of tolerance against the so-called “self antigens” occurs [11, 17]. A possible solution to overcome this problem is the use of HLA-restricted peptides with modifications at certain positions to enhance the interaction with the T-cell receptor complex [18, 42]. One of those altered peptide ligands is the HLA-A*0201 restricted nonamer CEA610D (CEAalt). In this peptide aspartate replaces asparagine at position 610 resulting in an increased potency to induce cytotoxic T cells (CTLs) against CEA in vitro. This effect is presumably caused by altered interactions of CEAalt with the human T-cell receptor [31, 40]. In the present study, we performed a clinical phase I/II trial to test the capability of DCs derived from immunomagnetically isolated CD14+ monocytes to induce specific immunity against the HLA-A*0201 restricted peptide CEAalt and the peptide with the original binding-motif (CEAorg) after vaccination with CEAalt- and keyhole limpet hemocyanin (KLH)-loaded DCs.

Materials and methods

Study design

Patients included in this phase I/II trial underwent two autologous leukaphereses on day 1 and day 27 of the study. Immature autologous DCs for two subsequent vaccinations were generated from monocytes enriched by CD14+ selection. Fifty percent of the immature DCs underwent a 2-day maturation period in the presence of cytokines and KLH, whereas the other half was cryopreserved for subsequent use. On day 8, the peptide CEAalt was added to the DCs for 12 h. On the next day loaded DCs were harvested and administered to the patient intravenously at four occasions on two dose-levels on days 8, 22, 36 and 50. CD14 selection and DC generation were performed in a clean room.

Patients

The study had been approved by the local institutional review board and informed consent was obtained from all patients. Nine HLA-A*0201 positive patients with metastatic CEA positive tumors (two lung cancer, seven colorectal carcinoma) were included in the study. The median age was 60 years (range, 36–64). Only heavily pretreated patients with metastatic disease not treatable by further surgery, radiation or chemotherapy were included in the study. The study protocol required termination of all chemotherapy at least 4 weeks prior to the first DC application.

Immunomagnetic large-scale CD14+ selection

PBMCs were obtained from leukapheresis products as described elsewhere [1]. The CD14+ selection was performed according to the manufacturer’s instructions (Miltenyi Biotec, Bergisch Gladbach, Germany). Briefly, one centrifugation at 300 g was performed to decrease platelet contamination of the leukapheresis product. Subsequently, the PBMC suspension was incubated for 15 min at room temperature with one vial (7.5 ml) of the CD14 reagent (Miltenyi Biotec). After incubation, two washing steps with phosphate buffered saline (PBS) followed to eliminate nonbound antibody. The labeled PBMCs were then loaded on the CliniMACS column. To prevent cell clotting and obstruction of the column we used cold buffer solution for loading the device and interposed an additional 200 μm prefilter (Transfusionsgerat BB, MPL Lichtenberg GmbH, Lichtenberg, Germany) between cell bag and column. The product bag was put on ice to prevent adherence of monocytes. The selection took about 60 to 120 min depending on the initial cell counts.

DC generation

The selected CD14+ cells were incubated at 37°C in 5% CO₂ in culture flasks (185 qcm; Nalge Nunc, Naperville, USA) at a density of 2×10⁶/ml in X-Vivo 15 medium (Bio-Whittaker, Walkersville, USA) supplemented with IL-4 (1,000 IU/ml; R&D, Minneapolis, USA), GM-CSF (2,000 IU/ml; Novartis, Nuernberg, Germany) and 1% pretested AB serum (c-c pro, Neustadt/W., Germany). On day 6, 0.75×10⁸ cells at a concentration of 0.5×10⁶/ml were incubated with TNF-alpha (1,100 IU/ml; R&D, Minneapolis, USA), PGE₂ (1 μg/ml; Pharmacia, Puurs, Belgium), IL-1 beta (1,900 IU/ml; R&D, Minneapolis, USA), IL-6 (1,000 IU/ml; R&D, Minneapolis, USA) to induce maturation. At the same time KLH (50 μg/ml; Biosyn, Fellbach, Germany) was added. One half of the immature DCs were cryopreserved in autologous heat-inactivated plasma containing 10% dimethyl sulfoxide (DMSO)(Wak Chemie, Bad Soden, Germany) for the second and fourth vaccination, respectively. Trypan blue exclusion was performed in each DC preparation before clinical use. Additional quality controls included gram staining, a 48-h bacterial culture and regular screening for endotoxins using a standard Limulus Amebocyte Lysate test. A vitality >90% and a negative gram staining were release criteria for fresh preparations.

Peptide loading

Both the CEAorg peptide (YLSGANLNL; CEA_605-613) and the CEAalt (YLSGADLNL; CEA_610D) were synthesized and bought from Jerini Biotools GmbH (Berlin, Germany, http://www.jerini.de) with at least 95% HPLC-grade purity. All preparations were tested to be free of endotoxins. On day 8, the mature DCs were incubated with the CEAalt peptide at a concentration of 20 μg/ml for 12 h. Afterwards, DCs were harvested from the flasks and resuspended to a concentration of 1×10⁶/ml for clinical administration. For the second and the fourth infusion thawed immature DCs were incubated with the same cytokine cocktail and KLH for a 2-day period and were then also loaded with CEAalt before administration, as mentioned above.

Clinical application

Patients were immunized on four occasions every 2 weeks (days 8, 22, 36, 50). The first two doses on day 8 and 22 contained 1×10⁷ cells, on the third and fourth occasion (days 36 and 50) 5×10⁷ cells were injected intravenously. For the first and third vaccination fresh mature DCs were injected, whereas cryopreserved immature DCs were maturated over 2 days for the second and fourth application. For safety reasons blood pressure, heart rate, respiratory rate and oxygen saturation were monitored over a period of 4 h after vaccination. All patients received prophylactic anti-allergic medication (500 mg acetaminophen, 50 mg ranitidine, 4 mg dimetidin).

Clinical monitoring

Clinical and radiological staging was performed by the time of accrual and completion of vaccinations (day 57). Three-dimensional measurements of reference lesions were performed to assess the response according to the NCI criteria (http://www.nci.nih.gov).

FACS analyses

Expressions of CD14 (PE/FITC, Immunotech, Marseille, France), CD83 (FITC, Immunotech, Marseille, France), HLA-DR (PE, Becton Dickinson, San Jose, USA) CD80 (PE, Becton Dickinson, San Jose, USA) and CD86 (FITC, Pharmingen, San Diego, USA) were investigated by FACS analysis before and after CD14+ selection including positive and negative fraction. FACS analysis was done on a FACScan (Becton Dickinson, San Jose, USA). Viability was determined by PI staining on days 1, 6 and 8.

ELISPOT assay

To determine the frequency of CEAalt or CEAorg peptide-reactive CD8+ T cells in the blood of vaccinated tumor patients Interferon (IFN)-gammaELISPOT analysis was performed as previously described [19]. PBMCs were prepared before, after the second, and after the fourth vaccination. Wells of MultiScreen-HA plates (Millipore, Bedford, MA) were coated with 100 μl of mouse-anti-human IFN-gamma antibody (10 μg/ml; clone 1-D1K; Mabtech, Nacka, Sweden) in PBS, incubated overnight at 4°C, and washed with PBS. After blocking with 100 μl RPMI 1640 containing 10% human serum immunomagnetically (Miltenyi Biotec) isolated monocytes (5×10⁴/well) were pulsed separately with the CEAorg peptide, the CEAalt peptide or the HIV reverse transcriptase peptide ILKEPVHGV (Jerini Biotools) in a final volume of 100 μl/well at a concentration of 100 μg/ml. After 2 h, 1×10⁵ CD8+ T cells isolated by immunomagnetic purification (Miltenyi Biotec) were added to each well. Captured cytokine was detected by adding 100 μl of biotinylated anti-IFN-gammaantibody (2 μg/ml; clone 7-B6-1; Mabtech) per well for 2 h. After incubation with 100 μl of an avidin-biotin peroxidase complex (1/100; Vectastain Elite Kit; Vector Laboratories, Burlingame, CA) staining was performed with 3-amino-9-ethyl carbazole (Sigma-Aldrich, München, Germany) for 4 min and stopped by washing the plates under running tap water. Spots were counted using a stereomicroscope (Zeiss, Jena, Germany) at 40×magnification. Results were calculated by subtracting the mean background spot counts for HIV peptide from the mean spot counts for CEAorg or CEAalt peptide. The background reactivity of CD8+ T cells against the HIV peptide did not exceed six spots/1×10⁵ cells in all experiments.

Chromium release assay

To stimulate CEAalt-specific CTLs, monocyte-derived DCs were pulsed with the CEAalt peptide at a concentration of 20 μg/ml in serum-free RPMI 1640 medium for 4 h at 37°C. After washing, 2×10⁵ peptide-loaded DCs were cocultured with 2×10⁶ autologous PBMCs obtained pre-therapy and post-therapy in 2 ml complete RPMI 1640 medium supplemented with 25 U/ml IL-2 and 10 ng/ml IL-7 (both from Strathmann Biotech, Hanover, Germany) in a 24-well tissue culture plate (Greiner, Frickenhausen, Germany). Seven days later, cultures were washed and restimulated with peptide-loaded DCs. After 12 days of culture effector cells were analyzed for cytotoxic activity in a 4 h standard ⁵¹Cr-release assay [19]. Briefly, the HLA-A*0201-positive cell line T2 pulsed with CEAalt peptide or CEAorg peptide for 4 h at a concentration of 50 μg/ml, the HLA-A*0201-positive colorectal cancer cell line COLO 205 and the chronic myelogenous leukemia cell line K-562 (American Type Culture Collection, Manassas, VA) were labeled for 1 h at 37°C with 100 μCi ⁵¹Cr (sodium chromate; PerkinElmer Life Sciences, Rodgau Jügesheim, Germany). Labeled target cells were plated as triplicates in round bottomed 96-well plates at 3×10³ per well and were incubated with effector cells at different E:T ratios. Specificity of CD8+ T cell-mediated lysis was tested in the presence of the anti-HLA-A2 monoclonal antibody MA2.1 (American Type Culture Collection). Released ⁵¹Cr was determined in a beta-counter (PerkinElmer Life Sciences).

Results

CD14+ selection and DC generation

The median number of total nuclear cells collected with a 2–3 h leukapheresis was 1.51×10¹⁰ (range, 0.97% – 2.04%). The median percentage of CD14+ cells in the apheresis product was 26.6% (range, 16.1– 32.1%).

CD14+ selection was performed successfully in all cases leading to a CD14+ fraction with a median purity of 97.2% (range, 91.6–98.8%), the median viability of CD14+ cells after selection was 88.3% (range, 68.5–94.5%). After exposure to GM-CSF and IL-4 the cells developed typical dendritic morphology and on day 8 a high percentage of cells expressed the marker of mature DCs CD83 (median 86.3%; range, 61.8–93.9%) as well as high levels of the co-stimulatory molecules CD80 and CD86 (median 93.7%; range, 55.9–98.4%). As shown before, no significant difference in marker expression was detected between fresh DCs and those derived from thawed immature DCs [1].

Clinical course

Eight of nine included patients were vaccinated intravenously on four occasions every two weeks in an outpatient setting. The first two doses on day 8 and 22 contained 1×10⁷ cells, on the third and fourth occasion (days 36 and 50) 5×10⁷ cells were injected intravenously. There were no early signs of allergic reactions during the monitoring period, but after about 8 h patients developed grade 2–3 chills and fever. Only two patients developed fever on the first dose-level (1×10⁷ cells) but all patients did so at the higher dose (5×10⁷ cells). Fever was mostly accompanied by muscle and joint aches and nausea. No rash, dyspnoea or symptoms due to low blood pressure were observed. Fever usually resolved within 6–8 h. Because of obvious disease progression patient 9 was taken off the vaccination protocol after the second vaccination.

Re-staging after completion of four vaccinations showed progressive disease in seven patients. Patient 9 showed clinical signs of progression in terms of enlargement of palpable metastases after two vaccinations. Therefore, vaccination was stopped in this case. Patient 2 had stable disease for 4 months during the study period. Before entering the study an MRI scan revealed a tumor mass (4×2 cm) infiltrating the os sacrum causing neuralgia. In addition, two new small liver metastases had been diagnosed by a CT scan and ultrasound before the first vaccination. The size of all lesions and the clinical symptoms associated with the pre-sacral mass remained unchanged during 4 months after the last vaccination. The clinical course of this patient was associated with the detection of IFN-gamma- producing T-cells reactive against CEAalt and CEAorg peptides. Patient characteristics, the clinical course of each patient and ELISPOT results are provided in Table 1.

Table 1.

Characteristics of included patients: patient study number, age, gender, origin of CEA positive carcinoma (primary site), sites of metastasis, characteristics and number of lines of previous therapies (C chemotherapy, OP surgery, R radiation), number of vaccinations, immune response determined by ELISPOT assay for CEAalt and CEAorg and clinical response (PD progressive disease, SD stable disease)

Patient No.	Age	Sex	Primary site	Metastases	Treatment	No. of vaccinations	ELISPOT response	Clinical response
2	56	F	Rectum	Liver, bone, pelvis	3 C, 1 OP	4	+	SD
3	62	M	Rectum	Liver	1 C, 1 OP	4	+	PD
4	52	M	Rectum	Liver, lymph nodes	2 C, 1 R	4	+	PD
6	36	F	Rectum	Lung	1 C, 1 OP	4	+	PD
8	63	M	Rectum	Lung, liver	3 C, 1 OP, 1 R	4	−	PD
9	60	F	Rectum	Liver, chest wall, lung, lymph nodes, bone	1 C, 1 OP, 1 R	2	+	PD
7	64	M	Sigmoid	Liver	3 C, 1 OP	4	−	PD
1	47	F	Lung	Lung, brain, liver, bone	1 C, 2 R	4	−	PD
5	63	M	Lung	Pleura, lymph nodes	2 C, 1 OP, 1 R	4	−	PD

Vaccine-induced expansion of CEA peptide-reactive CD8+ T cells

By IFN-gammaELISPOT analysis, blood samples of nine tumor patients were evaluated for the presence of CEAalt or CEAorg peptide-reactive CD8+ T cells before, after the second, and after the fourth vaccination with CEAalt peptide-pulsed monocyte-derived DCs. As shown in Fig. 1, the frequency of CEAalt peptide-reactive CD8+ T cells secreting IFN-gamma was increased to detectable levels in five of nine tumor patients (patients 2–4, 6, 9) after vaccination. Whereas the highest frequency of CEAalt peptide-reactive CD8+ T cells in patient 2 was found after the second vaccination, the maximum number of specific CD8+ T cells in patient 3 was detected after the fourth vaccination. CD8+ T cells reactive against CEAorg peptide were increased in three of nine patients (patients 2, 3, 6) after treatment (Fig. 1). Patient 4 displayed pre-existing CEAorg-specific CD8+ T cells which were not expanded by vaccination. These results indicate that treatment with DCs derived from immunomagnetically isolated CD14+ monocytes induced an in vivo expansion of CD8+ T cells recognizing CEAalt peptide in five of nine vaccinated tumor patients.

Fig. 1.

Detection of CEA peptide-reactive CD8+ T cells by IFN-gamma ELISPOT analysis. Purified CD8+ T cells (1×10⁵ cells/well) were coincubated with monocytes (5×10⁴ cells/well) which were loaded separately with CEAorg peptide, CEAalt peptide or HIV peptide. The frequency of CEA peptide-reactive CD8+ T lymphocytes before vaccination (pre vac), after the second vaccination (post vac 2) and after the fouth vaccination (post vac 4) is demonstrated. Columns represent mean values of triplicate wells containing HIV peptide loaded monocytes substracted from mean values of triplicate wells containing CEAorg or CEAalt peptide-pulsed monocytes. Bars indicate SEM. n.d., not done; *, not detectable

Induction of CEA peptide-specific and tumor-reactive CTLs after vaccination

PBMCs obtained pre-vaccination and post-vaccination were analyzed for antigen-specific cytotoxic T-cell responses. Here, we evaluated the PBMCs of a patient who exhibited a T-cell reactivity as determined by ELISPOT assay (No. 3). Pre- and post-therapy PBMCs were stimulated twice with autologous CEAalt peptide-loaded DCs to generate CTLs which were analyzed in a chromium release assay. When peptide-loaded T2 cells were used as targets pre-therapy CTLs showed only a moderate reactivity against T2 cells loaded with the peptides CEAalt or CEAorg. In addition, only background lysis was observed against COLO 205 tumor cells. Interestingly, post-therapy CTLs efficiently lysed CEAalt peptide-loaded T2 cells as well as CEAorg peptide-pulsed target cells. Moreover, these CTLs displayed a marked cytotoxcity toward COLO 205 tumor cells. The specificity of CTL-mediated killing was further evidenced by the inhibition of COLO 205 cell lysis in the presence of the anti-HLA-A2 antibody MA2.1. To exclude natural killer cell-like activity K-562 cells were included as targets. Pre- as well as post-therapy CTLs showed only background lysis against K-562 cells. Results are depicted in Fig. 2. These data demonstrate that vaccination with CEAalt peptide-loaded monocyte-derived DCs can induce previously undetectable antigen-specific and tumor-reactive CTLs.

Fig. 2.

In vivo-generation of CEA peptide-specific cytotoxic effector T cells by monocyte-derived dendritic cells. PBMCs obtained pre-vaccination and post-vaccination from patient 3 were stimulated twice by CEAalt peptide-pulsed autologous DCs. a CTLs were tested against T2 cells loaded with CEAalt peptide, CEAorg peptide, an irrelevant peptide from HIV reverse transcriptase, and unpulsed T2 cells at an effector cell to target cell ratio of 20:1. Stimulated CTLs derived from PBMCs obtained pre-vaccination (b) and post-vaccination (c) were assayed against the tumor cell lines COLO 205 and K-562 at various effector cell (E) to target cell (T) ratios (3:1, 10:1, 30:1). d Inhibition of cytotoxic reactivity of activated CTLs derived from PBMCs obtained post-vaccination against COLO 205 tumor cells was tested in the presence of the monoclonal anti-HLA-A2 antibody MA2.1 at an E:T ratio of 30:1. All results represent the mean values of triplicate determinations, bars indicate SE

Discussion

In a previous phase I study we demonstrated the feasibility of generating sufficient numbers of CD14+ monocytes by the CliniMACS system and the safety of the administration of CD83+/HLA-DR++ DCs derived from these CD14+ cells via different routes. The ability of fresh and cryopreserved mature DCs to stimulate T cells in vitro could be shown in proliferation assays [1]. The use of mature DCs seems to be advantageous since the application of immature DCs carries the risk of inducing antigen-specific immune tolerance [7]. We carried out this study using the altered ligand peptide CEAalt to prove the functional activity of this DC preparation by inducing cellular immune responses in vivo. CEAalt interacts differently with the T cell receptor, leading to enhanced ZAP-phosphorylation [31, 40]. CTLs raised to this epitope have been shown to maintain their ability to lyse tumor cells expressing native CEA. CEAalt has been shown to induce immune responses in humans when administered with DCs selected from peripheral blood by density centrifugation after mobilizing patients with recombinant Flt3-ligand [11].

CD14+ selection by CliniMACS proved to be a very useful and reliable technique for the production of a pure preparaton of viable CD14+ cells. Compared to the phase I study the differentiation and maturation period was shortened from 11 days to 8 days without negative impact on DC quality. The cytokine combination used for DC maturation has been described previously by other investigators [37]. High numbers of mature DCs with typical morphology and marker expression sufficient for at least two vaccinations could be generated from all nine patients, demonstrating the broad applicability of this strategy even in unfavorable subjects carrying metastatic disease. Quality did not differ significantly between DCs produced from fresh or cryopreserved cells. The generation procedures were performed in accordance to GMP guidelines.

To evaluate the efficacy of T-cell based therapies, reliable methods for immunological monitoring are a prerequisite. The ELISPOT assay allows a sensitive quantification of antigen-specific T cells directly from peripheral blood samples. In this report, we show the induction of CD8+ T cells specific for CEAalt in five of nine patients vaccinated with DCs pulsed with CEAalt. T cells reactive against the CEAorg peptide could be induced in three of nine patients. Apart from patient No. 4, none of the patients did have IFN-gamma producing T cells specific for CEAalt or CEAorg before vaccination. These findings demonstrate that CEAalt loaded mature DCs derived from immunomagnetically isolated CD14+ monocytes represent effective T-cell stimulators of antigen-specific immune responses in vivo. The detection of specific CD8+ T cells in ELISPOT assays after vaccination is supported by the results of the chromium release assay for patient No. 3 showing specific lysis of CEA positive cells in vitro that can be inhibited by anti-HLA-A2 antibodies.

The possibility to induce T-cell responses against CEAorg with CEAalt had been shown by Fong et al. [11] in 2001. The authors used DCs isolated directly from the blood of Flt3-ligand treated patients. Immune responses against CEA using DCs generated from monocytes selected by plastic adherence could be shown by Itoh et al. [17] and Matsuda et al. [23] using the HLA-A24 restricted peptide CEA652. With this study we could show the capability of DCs generated from monocytes derived after CliniMACS immunomagnetic CD14+ selection to induce specific T-cell responses.

Like in previous tumor-vaccination trials KLH was added to induce CD4+ mediated help required for the induction of antigen-specific CTLs [25, 34]. Recent data underline the importance of CD4+ T cells for sustained anti-tumor immune responses [38]. The help provided by CD4+ cells is delivered by secretion of IL-2 and/or by CD40/CD40L interaction which can condition APCs for efficient CTL induction [30, 33].

Intravenous vaccinations were tolerated without severe immediate side effects. The intravenous route of DC application was chosen to allow a direct comparison of this study with the immune and clinical responses described by Fong et al. However, recent reports have shown that intradermal or subcutaneous application of DCs elicits more potent immune responses [7, 37]. Recently, encouraging immune responses against CEA have also been described for the intranodal application of peptide-loaded DCs [21]. All patients developed transient flu-like symptoms including fever a few hours after the injection of 5×10⁷ DCs. This is in concordance with other trials, whereas gastrointestinal symptoms like diarrhea which had been reported by other investigators were not seen in this trial [11, 23].

No objective tumor regressions were observed. The clinical course of the malignant disease showed progression in all but one patient. While patient 9 experienced tumor growth early during the vaccination period, seven patients were found to be progressive after all four vaccinations had been administered. Tumor manifestations of patient No. 2 did not change over 4 months. Fong et al. reported two patients with tumor regression, two with stable disease and one with mixed response out of 12 treated patients. Trials using different CEAalt-loaded DCs by Itoh et al. showed stable disease in two of 10 patients. Matsuda et al. [22] vaccinated 8 patients with CEAalt-loaded DCs and achieved stable disease in three patients, but no regression either. Although the majority of studies performed to date make use of a single tumor-associated epitope, the high rate of mutation in tumor cells allowing for loss of expression of a single antigen might explain the limited efficacy observed in these trials. It is likely that the use of multiple antigenic epitopes will induce a broader, longer lasting, and more effective tumor-specific immune response. Reasons for the missing tumor regressions might be the potential lack of HLA class I expression in about 20% of invasive colon carcinomas or the additional Fas ligand expression on tumor cells [12, 14]. Both factors were not specifically investigated in tumor biopsies in this study. The better response observed by Fong et al. can be explained either by the in vivo application of Flt3-ligand or the significantly higher dose of DCs which could be infused in their protocol. In addition, the lack of immunological response in four patients may also be related to an accelerated decay of the peptide-HLA2 complexes after loading of DCs in vitro. The stability and duration of these complexes have been shown to vary depending on the source of antigen-presenting cells and the peptide used [41]. Nevertheless, the current study as well as the other trials cited are further examples for the hypothesis that DC-based vaccination may not be able to induce clinical responses in patients with a high tumor burden due to the dominance of tolerogenic mechanisms and the impaired functional immune status of the patient having undergone several series of chemo- or radiotherapy [39].