A comparison of two types of dendritic cell as adjuvants for the induction of melanoma-specific T-cell responses in humans following intranodal injection



Dendritic cells (DCs) elicit potent anti-tumoral T-cell responses in vitro and in vivo. However, different types of DC have yet to be compared for their capacity to induce anti-tumor responses in vivo at different developmental stages. Herein, we correlated the efficiencies of different types of monocyte-derived DC as vaccines on the resulting anti-tumor immune responses in vivo. Immature and mature DCs were separately pulsed with a peptide derived from tyrosinase, MelanA/MART-1 or MAGE-1 and a recall antigen. Both DC populations were injected every 2 weeks in different lymph nodes of the same patient. Immune responses were monitored before, during and after vaccination. Mature DCs induced increased recall antigen-specific CD4+ T-cell responses in 7/8 patients, while immature DCs did so in only 3/8. Expansion of peptide-specific IFN-γ–producing CD8+ T cells was observed in 5/7 patients vaccinated with mature DCs but in only 1/7 using immature DCs. However, these functional data did not correlate with the tetramer staining. Herein, immature DCs also showed expansion of peptide-specific T cells. In 2/4 patients vaccinated with mature DCs, we observed induction of peptide-specific cytotoxic T cells, as monitored by chromium-release assays, whereas immature DCs failed to induce peptide-specific cytotoxic T cells in the same patients. Instead, FCS-cultured immature DCs induced FCS-specific IgE responses in 1 patient. Our data demonstrate that this novel vaccination protocol is an efficient approach to compare different immunization strategies within the same patient. Thus, our data define FCS-free cultured mature DCs as superior inducers of T-cell responses in melanoma patients. © 2001 Wiley-Liss, Inc.


APC, antigen-presenting cells; CTL, cytotoxic T cells; ELISPOT; enzyme-linked immunospot; FCS, fetal calf serum; HSA, human serum albumin; iDC/mDC, immature/mature dendritic cells; PPD (purified protein derivative), tuberculin; TT, tetanustoxoid.

Dendritic cells (DCs) are specialized for the initiation of T-cell immunity, including cytotoxic T cells (CTLs), which kill virus-infected as well as malignant cells.1–3 Immunization with antigen-pulsed or transfected DCs efficiently primes CD4+ and CD8+ T cells, resulting in protective immunity against infectious agents and tumors.4–7 Several strategies using antigen-bearing DCs as cancer vaccines are under investigation.8 DCs prime human CD8+ T-lymphocyte responses against tumor antigens in vitro.1, 2, 4In vivo, antigen-pulsed DCs rapidly generate broad T-cell immunity.9–11 However, the protocols for generation of human DCs vary considerably,10, 12–17 and little attention has been focused on the exact definition of the appropriate DC subtype for therapeutic application in humans.

It was our goal to directly compare mature DCs (mDCs) generated in serum-free medium with published protocols involving immature DCs (iDCs) generated in FCS-containing medium. We decided on a new strategy, which was to pulse different DC populations with different tumor antigens, which enabled us to compare 2 different vaccines within the same patient. We vaccinated advanced, stage IV melanoma patients simultaneously with iDCs and mDCs loaded with different melanoma-associated peptide antigens. Both DC populations were injected in different lymph nodes of the same patient. Intranodal rather than intracutaneous injection was used based on the fact that immunization with DCs or other adjuvants begins in the draining lymph node. Additionally, we wanted to circumvent differences in the migratory capacity15 of iDCs and mDCs in vivo. Our data demonstrate that FCS-free cultured mDCs act as adjuvants for CD8+ and CD4+ T cells, whereas iDCs cultured in the presence of FCS are more or less inactive. In vivo, DTH reactions were observed only after vaccination with mDCs. Thus, our data define FCS-free, cultured mDCs as superior inducers of T-cell responses in melanoma patients.



Eleven patients with stage IV cutaneous malignant melanoma (1988 AJCC/UICC pTNM staging system) and progressive disease despite chemo-/immunotherapy were included. Further inclusion criteria were as follows: HLA-A2+ (8 patients) or HLA-A1+ (3 patients, HLA serotyping done at the blood bank of the University of Mainz), expected survival ≥4 months, Karnofsky index ≥60%, age > 18 years and informed consent. Expression of the melanoma antigens MelanA/MART-1 and tyrosinase (clones A103/T311; Dako, Hamburg, Germany) was determined by immunostaining of paraffin material. Exclusion criteria were as follows: active CNS metastasis, severely impaired organ function, active auto-immune disease (except vitiligo), pregnancy or participation in any other clinical trial. Patients did not receive any concomitant chemo- or immunotherapy. Eight patients were fully evaluable.

Protocol and study design

The study was performed according to the standards of Good Clinical Practice (GCP). DC vaccinations were administered at 14-day intervals (Table I). Blinded analysis of patients, including clinical and laboratory staging, was performed at visits 1, 5 and 9 with a complete physical examination, staging procedures and standard laboratory tests (including tests for HIV-1/2, hepatitis B/C). Patients were monitored for 48 hr after intranodal DC injection. DTH tests were conducted together with the first, third and sixth vaccinations and evaluated after 1, 24 and 48 hr by measuring the size of the infiltrate and by photography. Side effects following vaccination were documented and graded according to the criteria of the National Cancer Institute.

Table I. Study Design
 Before vaccinationVaccination 1Vaccination 2Vaccination 3First evaluationVaccination 4Vaccination 5Vaccination 6Second evaluation
  1. Patients were vaccinated at 14-day intervals. Two weeks before the first vaccination, staging procedures were performed. Leukapheresis products were prepared from each patient before vaccination and after the third and sixth vaccinations. On the day of each vaccination, blood samples were taken and used for proliferation assays to recall antigens. To detect induced immune responses, DTH, ELISPOT and cytotoxicity assays were performed before vaccination and after the third and sixth vaccinations. At the same time point, staging procedures were repeated.

Clinical visit (day)1 (−14)2 (0)3 (+14)4 (+28)5 (+42)6 (+56)7 (+70)8 (+84)9 (+98)
Multitest MérieuxX
DTH to peptide-pulsed mDC/iDCXXX
Proliferation to recall antigensXXXXXXXX
Cytotoxicity assaysXXX

Generation of DC vaccines

DCs were generated from leukapheresis products, which were prepared from each patient 2 weeks before the first vaccination. PBMCs were isolated and stored frozen (2 × 108/tube) in human serum albumin (HSA; Behring/Centeon Pharma, Marburg, Germany) + 10% DMSO (Cryocerv; WAK-Chemie, Bad Soden, Germany). For each vaccination, frozen PBMCs were thawed and cultured in pre-warmed RPMI-1640 (BioWhittaker, Walkersville, MD) + 1% heat-inactivated autologous plasma for 30 to 40 min (5 × 107 PBMCs/dish, 100 × 20 mm; Corning/Costar, Bodenheim, Germany). Non-adherent cells were rinsed off and adherent cells cultured overnight in RPMI-1640 + 1% plasma. On day 1, medium was changed and cells were cultured in serum-free X-VIVO-15 (8 ml/dish, BioWhittaker) for the generation of mDCs or in RPMI-1640 + 10% FCS for the generation of iDCs. Both cultures contained 800 U/ml GM-CSF (Leukomax; Novartis, Basel, Switzerland) and 1,000 U/ml IL-4 (Strathmann, Hannover, Germany). At day 5, non-adherent cells were rinsed off, washed once in PBS and transferred to 6-well plates (Corning/Costar) at 7 × 105 cells/well. For generation of iDCs, precursors were cultured in the presence of GM-CSF + IL-4; for generation of mDCs, precursors were additionally stimulated on day 6 with 10 ng/ml IL-1β, 10 ng/ml TNF-α, 1,000 U/ml IL-6 (all from Strathmann) and 1 μg/ml PGE2 (Minprostin; Pharmacia-Upjohn, Uppsala, Sweden). Non-adherent iDCs and mDCs at day 7 were used for vaccination. Quality-control criteria for iDCs were as follows: CD1a++/CD83 non-adherent cells with few dendrites; intermediate expression of HLA-DR, CD40 and CD86; and weak T-cell stimulatory capacity in vitro. Quality-control criteria for mDCs were as follows: CD1a+/−/CD83++ non-adherent dendritic cells; >80% homogeneous high expression of HLA-DR, CD40, CD58, CD80 and CD86; stimulation of naive T cells at DC:T cell ratio <1:200; stable morphology; and T-cell stimulatory activity 48 hr after withdrawal of cytokines.

Loading of DCs with antigens

The recall antigens tetanustoxoid (TT) and tuberculin, purified protein derivative (PPD) (Chiron/Behring, Marburg, Germany), were added for the last 24 hr of culture at 10 μg/ml; peptides were added at 30 μM for the last 12 hr and again for the last hour of culture. The following melanoma-associated peptides, purchased from Clinalfa (Sissach, Switzerland), were used: MelanA/Mart-1 (EAAGIGILTV) and tyrosinase (YMDGTMSQV) for HLA-A2+ patients, MAGE-1 (EADPTGHSY) and tyrosinase (HKSDICTDEY) for HLA-A1+ patients and control peptides MAGE-3.A1 (EVDPIGHLY) and MAGE-3.A2 (FLWGPRALV).

Vaccination schedule

Intranodal vaccinations with mDCs and iDCs pulsed separately with 1 tumor peptide and 1 recall antigen were performed under ultrasound guidance at 14-day intervals (Fig. 1). mDCs and iDCs were injected in different lymph nodes of the same patient. Lymph nodes were distant from the location of primary melanoma. For the first 3 vaccinations, 8 × 106 mDCs and iDCs in 200 μl PBS were used for injection; for vaccinations 4–6, 12 × 106 mDCs and iDCs were injected.

Figure 1.

Vaccinations were performed at 14-day intervals. mDCs and iDCs were pulsed separately with 2 different tumor peptides (A and B) and 2 different recall antigens (A and B) and injected into different lymph nodes. DTH tests were performed before vaccination and after the third and sixth vaccinations. For each DTH, 4 × 105 mDCs and iDCs with or without peptide were applied intradermally.

In patients 03 and 04, we continued vaccinations at the end of the study as follows: 18 to 24 × 106 mDCs pulsed separately with tyrosinase-A2 and MelanA/MART-1 peptides were injected in monthly intervals.

Detection of tumor peptide-specific T-cell frequencies

To determine the frequency of tumor peptide-specific CD8+ T cells, an enzyme-linked immunospot (ELISPOT) assay was performed.18 CD8+ T cells were isolated immunomagnetically (MACS beads; Miltenyi Biotec, Bergisch-Gladbach, Germany) from PBMCs collected before vaccination, after the third vaccination and after the sixth vaccination. T cells (105) were used either directly or 6 days after in vitro restimulation with peptide-pulsed mDCs (104). IL-2 was added once at day 2 of culture (10 U/ml, Chiron/Behring). At day 6, T cells were harvested and cultured in 96-well nitrocellulose culture plates (Multiscreen; Millipore, Bedford, MA) coated with monoclonal antibody (MAb) for IFN-γ, IL-4 or IL-10 (anti-IFN-γ and anti-IL-10 were from Mabtech, Stockholm, Sweden and anti-IL-4 was from Pharmingen, San Diego, CA). HLA-A2+ or HLA-A1+ T2 cells (75 × 103/well; T2.A2 provided from Dr. P. Cresswell, Yale University, New Haven, CT, and T2.A1 provided from Dr. B. van den Eynde, Ludwig Institute for Cancer, Brussels, Belgium) were loaded with peptides (30 μM). After incubation for 20 hr (IFN-γ) or 48 hr (IL-4, IL-10), cytokine release of individual T cells was detected by a second-step antibody (Mabtech) followed by AEC staining (Vectastain ABC Kit; Linaris, Wertheim-Bettingen, Germany). Spots were counted by blinded observers and calculated as follows: spots in the presence of T2 cells pulsed with specific peptides minus spots in the presence of T2 cells pulsed with control peptides.

Detection of recall antigen-specific responses

Proliferation assays were performed before and 14 days after every vaccination, to detect the induction of immune responses to recall antigens. PBMCs from whole blood were cultured ± 10 μg/ml TT or PPD in X-VIVO-20 (BioWhittaker) in 96-well plates (1:2 dilutions from 2 × 105 to 1.5 × 103 PBMCs/well) for 4 days plus an additional 16 hr in the presence of 1 μCi [3H]-thymidine. PHA stimulation (2.4 μg/ml) was used as control.

Detection of antigen-specific CTL precursors

CD8+ T cells isolated by MACS beads from frozen PBMCs before and after vaccination were restimulated in vitro with peptide-pulsed syngeneic mDCs (30 μM/ml peptide for 1 hr, DC:T cell ratio 1:20) in X-VIVO-20 for 6 days. IL-2 (10 U/ml) was added once on day 2 of culture. At day 6, T cells were washed extensively in PBS and used in 4 hr 51Cr-release assays. As antigen-presenting cells (APCs), peptide-pulsed T2.A1 or T2.A2 cells were used (pulsed with 30 μM/ml, incubation for 1 hr) in different E:T ratios. Positive peptide-specific lysis is shown as specific lysis of peptide-pulsed T2 cells minus specific lysis of unpulsed T2 cells.

DTH reaction and analysis of humoral responses

DTH reactions to melanoma peptides were analyzed by intradermal injection of 4 × 105 peptide-pulsed mDCs or iDCs in 0.2 ml PBS before treatment and after 3 and 6 vaccinations. Unpulsed mDCs and iDCs served as controls. A Multitest Mérieux including TT and PPD was also performed before vaccination. For detection of antibody induction in patients with immediate-type inflammatory responses, the serum was analyzed for antibody induction against whole bovine proteins, bovine epithelia and BSA [FEIA-Uni-CAP-System, used according the instructions of the manufacturer (Pharmacia-Upjohn)].

Tetramer staining

Peptide–MHC tetrameric complexes were made as previously described.19, 20 Each tetramer was validated by staining against a CTL line or clone specific for HLA-A2 in association with the peptide of interest. Specificity was demonstrated by lack of staining of irrelevant CTLs. Each tetramer reagent was titered individually and used at the lowest concentration that still gave clearly discernable positive populations, generally 10 to 50 μg/ml, to minimize background staining. Only clustered populations by CD8 and tetramer staining (1 log above the negative population) were considered positive. Samples with tetramer-stained populations were repeated and considered positive only if confirmed. The limit of detection of each tetramer was determined by titrating known antigen-specific T cells into normal PBMCs. By collecting 106 or more events per sample to increase the number of potentially tetramer+ events and to observe clustering, we established a limit of detection of down to 0.01% CD8+ T cells. Generally, 1 to 2 × 106 PBMCs were incubated with anti-CD8-FITC, PE-conjugated tetramers and Cy5-PE–conjugated antibodies against CD4, CD14 and CD19 for 1 hr at room temperature. Cells were washed 3 times and analyzed by flow cytometry (FACScan; Becton Dickinson, San Jose, CA). FACS data were acquired through CELLQUEST, then analyzed using FlowJo (TreeStar, San Carlos, CA). Cells were progressively gated for lymphocytes by forward-scatter (FSC) and side-scatter (SSC); negatively gated for CD4, CD14 and CD19 staining; then plotted with CD8 vs. tetramer staining.


Clinical responses

A total of 11 patients with stage IV melanoma and progressive disease despite preceding chemo-/immunotherapy were vaccinated with mDCs and iDCs pulsed separately with 1 tumor peptide and 1 recall antigen (Table II). Three patients died after the second or third vaccination because of severe progression of melanoma; the remaining 8 patients received 3 to 6 vaccinations in 14-day intervals and were fully evaluable. Only patients 03 and 04 showed a positive reaction to PPD and TT before vaccination (Multitest Mérieux).

Table II. Clinical Characteristics of Stage IV Melanoma Patients before DC Vaccination
Patient/sex/age (years)HLA typeTumor type/size primary diagnosisMetastasesAntigen status of metastasesPrevious chemo-/immunotherapyVaccination
  • Eight of 11 patients (codes 01–08) were fully evaluable. All patients suffered from stage IV melanoma. Tumor types included superficial spreading melanoma (SSM), acrolentiginous melanoma (ALM) and nodular malignant melanoma (NMM). Tumor sizes ranged from 1.9 to 9.0 mm (level IV/V). Primary diagnosis of malignant melanoma was between 1996 and 1998 (except patient 03). Metastases of patients who could be tested showed positive immunostaining for MelanA/MART-1 and/or tyrosinase (ND, not detected;

  • 1

    weak staining;

  • 2

    strong staining).

  • All patients were progressive despite chemo-/immunotherapy.

01/male/52A2SSM, 7.0 mm (level V), 3/97Skin, lymph nodesMelanA/MART-1,1 tyrosinase2Melphalan, DNCB/DTIC, LeghamDC+ Tyr.A2+ TT iDC+MART-1+PPD
02/female/75A2SSM, 3.1 mm (level IV), 7/97Lymph nodes, liver, lung, skinNDIFN-α, taxoteremDC+ Tyr.A2+PPD iDC+ MART-1+ TT
03/female/31A2SSM, 1.9 mm (level IV), 8/85Lung, liverNDIFN-α/IL-2, LeghamDC+ Tyr.A2+PPD iDC+ MART-1+ TT
04/male/63A2SSM, 4/98Lymph nodes, liverMelanA/MART-1,1 tyrosinase1McClaymDC+MART-1+ TT iDC+Tyr.A2+PPD
05/male/59A2SSM, 3.2 mm (level IV), 6/98Lymph nodes, liver, spleen, bonesMelanA/MART-1,2 tyrosinase2IFN-α, McClaymDC+MART-1+ TT iDC+Tyr.A2+PPD
06/male/62A2NMM, 5.8 mm (level V), 1/97Lymph nodes, skinNDIFN-α, Kirkwood, IFN-α/IL-2mDC+MART-1+ TT iDC+Tyr.A2+PPD
07/female/35A1ALM, 2.3 mm (level IV), 10/98Lymph nodes, lung, liverTyrosinase2Kirkwood, IFN-α/IL-2, LeghamDC+ Tyr.A1+PPD iDC+ MAGE-1+ TT
08/male/62A1SSM, 9.0 mm (level IV), 8/96Skin, lymph nodesTyrosinase1IFN-α, taxoteremDC+MAGE-1+PPD iDC+ Tyr.A1+ TT

No major toxicity (>grade II) or side effects were observed. All patients transiently developed fever for 1 to 2 days following application of the DC vaccine (except patient 03). In patients with body temperature over 39°C (01, 02, 07), the dose of recall antigen was reduced to 1 μg/ml. In patients 02 and 04, we observed slight lymph node enlargement after the third and fifth vaccinations, respectively.

Results of the final evaluation are presented in Table III. Patient 03 experienced stable disease; 3 patients (01, 05, 07) progressed during vaccination. Patient 04 showed complete regression of 1 individual cervical lymph node; liver metastases appeared stable. The remaining patients showed partial regressions of individual s.c. and lymph node metastases (02, 08); some individual metastases in skin, liver and lymph nodes were stable (02, 06, 08). Strikingly, patients 02 and 08 showed massive progression after the end of the DC vaccination protocol. In patients 03 and 04, vaccinations with mDCs have continued, and both experience stable disease up to the present.

Table III. Clinical Response to DC Vaccination
PatientVaccinationMetastases at study entryClinical responseSurvival1
  • Five patients (01–04, 07) received 6 vaccinations and 3 patients (05, 06, 08) received 3 vaccinations at 14-day intervals. Clinical responses were measured 2 weeks after the third (01–08) and sixth (01–04, 07) vaccinations and disease was documented as stable, partial/complete regression of individual metastases or continuous progression.

  • 1

    Survival is listed as months since onset of stage IV until study entry + number of months since study entry. Patients 03 and 04 are still alive, while the other 6 patients are deceased.

  • 2

    Patients 02 and 08 showed massive progression after the end of the trial.

  • 3

    Patients 03 and 04 received further vaccinations in monthly intervals.

01mDC+ Tyr.A2+ TT iDC+MART-1+PPDLymph node, skinLymph nodeContinuous progression6 + 9
022mDC+ Tyr.A2+PPD iDC+ MART-1+ TTLymph node, skinLiver, lung, lymph nodePartial regression of s.c. metastases, partial regression of inguinal and retroperitoneal lymph nodes, stable liver metastases6 + 10
033mDC+ Tyr.A2+PPD iDC+ MART-1+ TTLiver, lungStable disease50 + 15
043mDC+MART-1+ TT iDC+Tyr.A2+PPDLymph node, enoralLiverComplete regression of an individual cervical lymph node, stable liver metastases5 + 15
05mDC+MART-1+ TT iDC+Tyr.A2+PPDLymph nodeLiver, spleen, bonesContinuous progression4 + 2
06mDC+MART-1+ TT iDC+Tyr.A2+PPDLymph node, skinLymph nodeStable s.c. metastases2 + 5
07mDC+ Tyr.A1+PPD iDC+ MAGE-1+ TTLymph nodeLung, liver, lymph nodeContinuous progression4 + 5
082mDC+MAGE-1+PPD iDC+ Tyr.A1+ TTLymph node, skinLymph nodePartial regression of s.c. metastases, inguinal lymph nodes stable3 + 5

mDCs strongly enhance CD4+ T-cell reactivity to recall antigens

Recall antigen-specific T-cell responses were analyzed before treatment and after the third and sixth vaccinations. In all patients, there were little or no proliferative responses to recall antigens prior to vaccination. Upon vaccination, 7/8 patients showed increasing T-cell responses to the recall antigen pulsed on mDCs (Fig. 2, responses of patient 05 were too weak to evaluate). In contrast, iDCs induced a more heterogeneous profile of T-cell proliferation in the same patients. In 5/8 patients, no efficient T-cell response was detectable: in 3 of these patients (01, 03, 08), iDCs induced no or only little proliferation in response to the relevant recall antigen. In patient 04, we detected weak and transient T-cell responses. Only 3/8 patients (02, 06, 07) showed T-cell responses comparable to vaccination with mDCs. Thus, enhanced CD4+ T-cell reactivity to recall antigens was observed in 7/8 patients treated with serum-free mDCs but in only 3/8 patients treated with FCS-cultured iDCs.

Figure 2.

Recall antigen-specific immune responses were detected in proliferation assays. PBMCs were cultured in the presence of 10 μg/ml TT or PPD or without antigen. Values (PBMCs with recall antigen minus PBMCs without recall antigen) for each patient are shown before vaccination and after the third and sixth vaccinations. ND, not done.

Expansion of tumor peptide-specific IFN-γ–producing CD8+ T cells requires mDCs

To determine the frequencies of individual antigen-specific, cytokine-producing T cells and to monitor the kinetics of T-cell responses, we used the ELISPOT technique. Due to limitations of material of patient 06, only 7/8 patients could be analyzed. As shown in Figure 3, vaccination of melanoma patients with mDCs resulted in expansion of peptide-specific, IFN-γ–producing CD8+ T cells. In patient 03, we were able to detect T-cell responses directly ex vivo. Serum-free mDCs induced increasing de novo peptide-specific T-cell responses against tyrosinase peptide, whereas vaccination with iDCs showed no effect in this patient. The ex vivo response in patient 05 was too weak to evaluate, and there was not enough material left to do restimulation assays. In the remaining 5 patients, T-cell responses were analyzed after restimulation as ex vivo responses were negligible. In patient 02, serum-free mDCs were able to induce de novo peptide-specific T-cell responses, whereas in 3/7 patients (01, 04, 08) a significant number of CD8+ T cells already showed antigen-specific IFN-γ production before vaccination. After the third vaccination, increasing frequencies of specific T cells were detectable only for peptides loaded on mDCs. This effect was enhanced after further vaccinations (in patients 03, 04) and was independent of an individual peptide since it was observed for peptides of tyrosinase and MelanA/MART-1 in HLA-A2+ patients and MAGE-1 and tyrosinase peptides in HLA-A1+ patients. IFN-γ release was reduced to 50% by anti-MHC class I antibodies (not shown), indicating the specificity of the T-cell response. In contrast, specific T cells induced with iDCs for vaccination showed only little expansion (in patients 01, 02, 04, 07) or were not detectable (in patients 03, 05). Only patient 08 showed strong expansion of IFN-γ–producing CD8+ T cells when using tyrosinase peptide-pulsed iDCs. IL-4– or IL-10–producing CD8+ T cells were not detectable in any of the patients (not shown). Our data demonstrate that serum-free mDCs induced de novo generation of peptide-specific T cells in 2/7 patients (patient 03 ex vivo, patient 02 after restimulation) and expansion of pre-existing peptide-specific CD8+ T cells in the remaining patients. In contrast, only 1 patient (08) showed comparable peptide-specific T-cell responses following vaccination with FCS-cultured iDCs.

Figure 3.

To analyze the frequencies of peptide-specific CD8+ T cells induced by vaccination, we used the ELISPOT technique. T cells were isolated from PBMCs collected before vaccination and after the third and sixth vaccinations and were used either directly ex vivo (105 T cells/well, patients 03 and 05) or after restimulation when no ex vivo responses were detectable (104 T cells/well, patients 01, 02, 04, 07 and 08), as described. T cells were co-cultured with HLA-A2+ or HLA-A1+ T2 cells in the presence of different tumor peptides (MART-1, MelanA/MART-1; Tyr.A2, tyrosinase), as described. T2 cells labeled with an irrelevant peptide (e.g., MAGE-3.A2 or MAGE-3.A1) or unpulsed T2 cells served as controls. Background for T2 cells alone was <8 spots (<40 spots after restimulation) and for T2 cells pulsed with irrelevant peptide <10 spots (<80 spots after restimulation). Numbers of individual IFN-γ spots (T2 + relevant peptide minus T2 + MAGE-3 peptides) are shown for each patient.

mDC-induced increased frequency of antigen-specific CTLs

In 4/8 patients (01, 06–08), enough material was available to assess the induction of peptide-specific cytotoxicity after in vitro restimulation of CD8+ T cells with peptide-pulsed DCs. In 2/4 patients (01, 06), vaccination with peptide-pulsed mDCs resulted in a detectably increased cytotoxic activity of specific CTLs (Fig. 4); 2 patients (07, 08) showed only little or no response with very low levels of lysis following vaccination. Using FCS-cultured iDCs for vaccination, 3/4 patients showed no lysis of peptide-pulsed target cells. Interestingly, patient 06 showed detectable cytotoxicity against the MelanA/MART-1 peptide before vaccination but decreased cytotoxicity after vaccination using MART-1–pulsed iDCs.

Figure 4.

CD8+ T cells were isolated from frozen PBMCs before and after vaccination (patients 01, 06–08), pre-cultured as described and analyzed in a 4 hr cytotoxicity assay. 51Cr-labeled, peptide-pulsed T2.A1 or T2.A2 cells were used as targets. T2 cells loaded with irrelevant peptides or without peptides served as controls. Specific lysis was calculated as follows: 100% × [(experimental release − spontaneous release)/(maximal release − spontaneous release)] − non-specific lysis without peptide]. E:T ratios of 1:20 are shown.


DTH reactions were performed in 5/8 patients by intradermal injection of peptide-pulsed and unpulsed serum-free mDCs or FCS-cultured iDCs. All 5 patients were negative before vaccination (not shown). Typical DTH reactions (>10 mm) after the third vaccination were detectable against peptide-pulsed mDCs 24 and 48 hr after injection in 3/5 patients [02, 03, 07; Fig. 5 (patient 03), Table IV]. In contrast, strong local reactions (>30 mm) were detectable in 3/5 patients 1 hr after application of iDCs regardless of whether peptide-pulsed or unpulsed iDCs were injected, indicating that this induration was not antigen-specific. This immediate-type inflammatory reaction disappeared within the first 24 to 48 hr. As these observations suggested induction of FCS-specific immune responses by iDCs, we tested the humoral and cellular anti-FCS responses after vaccination. Using FCS-containing medium in comparison to serum-free X-VIVO-20, no FCS-specific proliferation could be observed (not shown). In contrast, patient 03 with strong immediate-type inflammatory response to FCS-cultured iDCs developed a specific IgE response for bovine epithelia (0.88 IU/ml, RAST class 2.0) and whole bovine protein (1.23 IU/ml, RAST class 2.1).

Figure 5.

DTH tests of tumor peptides were performed by injecting peptide-pulsed or unpulsed mDCs or iDCs intradermally. PBS injection served as control. Infiltration at the injection site was measured and documented in millimeters: 1, mDCs; 2, mDCs + tyrosinase; 3, iDCs; 4, iDCs + MelanA/MART-1; 5, PBS. DTH reactions of patient 03 after the third vaccination are shown. Strong local reactions (>30 mm) were detectable after application of iDCs with and without the specific peptide even 1 hr after injection. Strong DTH reaction (>10 mm) not before 20 hr after injection was detectable only with mDCs + peptide. Similar local reactions were also observed in patients 02 and 07 (not shown).

Table IV. Positive DTH Reactions After Intranodal Application of Peptide-Pulsed mDC in 3/5 Patients
PatientTime after application (hr)Size of local reaction (mm)
  1. DTH tests were performed in 5 patients by i.d. injection of peptide-pulsed and unpulsed mDC or iDC before vaccination and after the third and sixth vaccinations. Injection of PBS served as control. At different time points (1, 24 and 48 hr) after application, local reactions were measured and documented in millimeters. DTH reactions after the third vaccination are shown.


Expansion of peptide-specific T cells in vivo

To monitor the total number of peptide-specific T cells in the peripheral blood before and after vaccination directly ex vivo, we used the MHC-peptide tetramer technique for a limited number of patients [Table V (patients 02–04), Fig. 6 (patient 04)] with appropriate numbers of frozen PBMCs. CD8+ T cells were stained with a set of 3 tetramers (A2/MART27–35, A2/Tyr368–376 and A2/gp100209–218 as irrelevant control peptide) before vaccination, after the third vaccination, after the sixth vaccination and partially after the ninth vaccination. Interestingly, all 3 patients already had endogenous peptide-specific T-cell populations even pre-vaccination (03 and 04 for MART-1, 02 for gp-100). Whereas the MART-1-specific T-cell population of patient 04 remained fairly stable through vaccinations 1–6 using MART-1–pulsed mDCs, we could induce significant expansion of tyrosinase peptide-specific CD8+ T cells injecting mature tyrosinase-pulsed DCs for vaccinations 6 to 9. We also detected expansion of peptide-specific T cells in all 3 patients using iDCs for vaccination. Nevertheless, there was no correlation between the expansion of tetramer+ CD8+ T cells and their function in vitro since iDCs failed to induce IFN-γ–producing cytotoxic T cells in all 3 patients. Furthermore, patient 02 showed a 3-fold expansion of gp100-specific T cells, though this peptide was not part of our vaccine. These results demonstrate that the expansion of peptide-specific T cells in tumor patients detected by tetramer staining is not necessarily associated with enhanced anti-tumor activity of the expanded T cells. We suggest that additional assays on the function of induced T cells are necessary for the immunomonitoring of vaccinated patients.

Table V. Tetramer Analysis Sample
PatientVaccinationCollectionCD8+ tet+ as % of CD8+
  1. Tetramer analyses of patients 02, 03 and 04 were done before and after vaccination using PBMCs directly ex vivo. CD8+ T cells were stained with specific tetramers (gp100 tetramers served as controls) and analyzed by flow cytometry. Tetramer-positive T cells are listed as % of CD8+ T cells.

04mDC + MART-1Post-vaccination 30.63<0.01<0.01
 Vac. 1–6iDC + Tyr.A2Post-vaccination 60.81<0.010.024
 Vac. 6–9mDC + MART-1 + Tyr.A2Post-vaccination 90.76<0.010.041
03mDC + Tyr.A2Post-vaccination 30.044<0.010.03
 Vac. 1–6iDC + MART-1Post-vaccination 60.055<0.010.07
 Vac. 6–9mDC + MART-1 + Tyr.A2Post-vaccination 90.41<0.010.13
02mDC + Tyr.A2Post-vaccination 30.160.50<0.01
 Vac. 1–7iDC + MART-1Post-vaccination 70.150.65<0.01
Figure 6.

Tetramer analysis was performed before vaccination, after the third vaccination, after the sixth vaccination and partially after the ninth vaccination using PBMCs directly ex vivo. CD8+ T cells were stained with anti-CD8-FITC and a set of 3 specific melanoma tetramers (A2/MART27–35, A2/Tyr368–376 and A2/gp100209–218 as irrelevant control) and analyzed by flow cytometry. Only clustered populations by CD8 and tetramer staining 1 log above the negative population were considered positive. Results of the tetramer staining of patient 04 are shown.


Different sources of antigen-loaded DCs have been used for vaccination of cancer patients.10, 11, 21–23 To directly compare the immunostimulatory properties of different DC vaccines in vivo, we established a new immunization strategy by vaccinating melanoma patients simultaneously with different DC populations: immature and mature, peptide-pulsed DCs. iDCs, cultured in FCS-containing medium in the presence of GM-CSF and IL-4, were compared to mDCs, generated in FCS-free medium and by additional stimulation with a cocktail of pro-inflammatory cytokines and prostaglandin. These 2 DC populations were characterized by distinct T-cell stimulatory functions in vitro and in vivo.1, 2, 8, 10, 14, 15, 17, 24 Both populations have been used in clinical trials for the vaccination of melanoma patients, with somewhat conflicting results.10, 22 We simultaneously injected iDCs and mDCs to directly compare the efficiency of the 2 DC populations as tumor vaccines in vivo in the same patient. The populations were loaded separately with distinct tumor-derived peptides and recall antigens to distinguish between the T-cell responses induced by iDCs and mDCs. To consider varying T-cell precursor frequencies to individual antigen epitopes,25 antigens were exchanged between iDCs and mDCs from patient to patient.

Injection of both DC vaccines induced only minor side effects, e.g., grade I fever and transient mild lymph node enlargement after injection. In contrast, induction of consistent recall antigen-specific CD4+ T-cell responses and tumor peptide-specific CD8+ T-cell responses were detectable only after injection of antigen-loaded mDCs. Only 3/8 patients vaccinated with the combination of tumor antigens associated with iDCs developed antigen-specific CD4+ T-cell responses against recall antigens. Furthermore, only injection of peptide-pulsed mDCs induced expansion of peptide-specific CD8+ T cells in 7/8 patients with cytotoxic activity in 2/4 patients tested. The different capacities of the 2 vaccines were consistent for all tumor peptides and recall antigens used for vaccination and therefore cannot be simply explained by varying T-cell precursor frequencies for an individual antigen. Our results are in agreement with several animal experiments and clinical trials using primarily mDCs for vaccination.4–6, 8–10, 26–28

Nestle et al.22 used FCS-cultured iDCs for the treatment of melanoma patients and detected tumor responses in 5/16 patients. In contrast to our study, they used tumor lysates and cocktails of melanoma peptides for loading of iDCs; tumor load was low in these patients, and prior chemotherapy was administered only to 4/16 patients.22 In our study, all patients had progressive disease with high tumor load after chemotherapy and before DC vaccination.

The IgE responses specific for bovine proteins are in agreement with the results of other investigators using FCS-cultured DCs.29 We were not able to distinguish between strong immunostimulatory effects induced by FCS-cultured iDCs themselves and potential peptide-specific reactions. Nevertheless, iDCs cultured with or without FCS induce only the expansion of antigen-non-specific CD8+ T cells in vitro, whereas mDCs induce expansion of tumor peptide-specific CTLs.28 Furthermore, terminally differentiated DCs are necessary for the induction of Th1 responses in vitro, whereas iDCs rather induce regulatory Th cells.30

Tetramer analysis of a limited number of patients revealed that mDCs and iDCs were able to expand peptide-specific CD8+ T cells. However, the results of tetramer staining did not correlate completely with the results of functional analyses. This may be due to the possibility that peptide-specific T cells elicited by iDCs may not be functionally active. Therefore, we suggest that the parallel use of different assays is necessary for appropriate immunomonitoring of vaccinated patients.

We found complete regression of individual metastases in 4/8 patients and stable disease in patients 03 and 04. Furthermore, 2 patients who were stable while being vaccinated showed massive progression after the termination of DC vaccination. Nevertheless, no complete regression was detectable in these advanced stage IV patients. Possibly, the large tumor mass and immune escape mechanisms are responsible for the lack of more striking anti-tumor responses.

We found that mDCs are able to expand melanoma peptide-specific CD8+ effector cells in fresh blood, indicating that stage IV melanoma patients are not fully tolerant. Probably, the use of mDCs in the treatment of stage II/III melanoma patients with lower tumor mass will significantly increase the prospects of successful DC vaccinations. Identifying the best maturation state, including resistance toward inhibitory factors produced by tumors,3 dose and life span of DCs, will significantly improve the successful use of DCs as natural adjuvants. In addition, we present a novel approach of comparing 2 immunization strategies in the same patient, which is of wide interest for research on immunotherapy and vaccination.


We thank Drs. E. von Stebut, E. Schmitt and R.M. Steinman for critical reading of the manuscript and helpful discussions. We also thank Mr. C. Britten and Dr. W. Herr for help with the ELISPOT.