Increased regulatory T-cell frequencies in patients with advanced melanoma correlate with a generally impaired T-cell responsiveness and are restored after dendritic cell-based vaccination

Authors


Andrea Tüttenberg, Department of Dermatology, Johannes Gutenberg-University, Langenbeckstr 1, D-55101 Mainz, Germany, Tel.: +49-6131-172957, Fax: +49-6131-17473541, e-mail: tuettenberg@hautklinik.klinik.uni-mainz.de

Abstract

Please cite this paper as: Increased regulatory T-cell frequencies in patients with advanced melanoma correlate with a generally impaired T-cell responsiveness and are restored after dendritic cell-based vaccination. Experimental Dermatology 2010; 19: e213–e221.

Abstract:  Naturally occurring CD4+ CD25+ regulatory T-cell (Treg) activity is assumed to facilitate tumor development and progression. To elucidate the possible role of Tregs in the course of melanoma progression, we analysed the frequency of Tregs in the peripheral blood of patients at melanoma stages I–IV and in patients at melanoma stage IV that underwent dendritic cell (DC)-based immunotherapy. Using CD25, Foxp3, CD127 and HLA-DR as Treg associated markers, we observed increased Treg frequencies in patients at the late melanoma stage (stage IV) when compared to healthy donors. Accumulation of Tregs in patients with progressed melanoma correlated with a general reduction of T-cell responsiveness to the recall antigens tetanus toxoid and tuberculin-GT. However, DC-based immunotherapy not only restored antigen-specific immunity, but also decreased the frequency of Tregs in peripheral blood of patients with melanoma. These findings indicate that tumor progression in patients with melanoma result in general immunosuppression that is associated with Treg expansion in the periphery and can be overcome by DC-based vaccination.

Introduction

Despite the observation that melanomas are immunogenic tumors and thus often induce an immunological reaction, immunotherapy in patients with melanoma rarely succeeds (1), suggesting a counter-regulatory immune modulation (2). The mechanisms contributing to tumor-induced immunological escape include down-regulation of MHC class-I antigens on tumor cells, lack of costimulatory function, secretion of immunosuppressive factors and induction of tolerance (3,4). Many strategies to enhance antigen-specific immunity in patients with melanoma have been explored in clinical studies (5–8). We recently completed two dendritic cell (DC)-based vaccination trials in different collectives of patients with melanoma (9,10). Whereas patients with a progressive melanoma developed only low and transient T-cell responses to melanoma antigens after vaccination with mature DC (9), patients with a minimal residual disease (stage II) showed high frequencies of de novo-induced antigen-specific T cells and long-lasting immune responses even after completion of vaccination (10). As immunosuppression seems to be an important way by which tumors succeed in immune escape, regulatory T cells (Tregs) – key players in the natural tolerance network – may contribute to the therapeutic resistance of progressive melanoma tolerance (2,3,11).

Tregs comprise 5–10% of the CD4+ T cells in peripheral blood and constitutively express high levels of CD25, the alpha chain of the IL-2 receptor (12–14). They are crucial for the maintenance of peripheral immune tolerance and the prevention of autoimmunity (12). Although the precise cellular and molecular mechanisms of Treg-induced immunosuppression are not entirely understood, both cell contact-dependent and -independent mechanisms have been described (12). Whereas Treg deficiency is associated with several autoimmune diseases and type-I allergy, high levels of Tregs are assumed to facilitate development and progression of cancer (15,16). In a variety of malignancies, elevated Treg frequencies have been observed in the tumor environment as well as in the peripheral blood (17–21). Likewise, patients with melanoma appear to display increased Treg frequencies in peripheral blood, local tumor environment and also in lymph nodes containing metastases (22,23). Other groups reported increased frequencies of circulating Tregs in patients undergoing IL-2 therapy (24,25). However, none of these analyses characterized Tregs by novel Treg marker combinations including CD25, Foxp3, CD127 and HLA-DR (26–29). In contrast, here we defined Treg frequencies using these recently identified markers to elucidate the possible role of Tregs in melanoma progression. CD25, Foxp3 and CD127 revealed increased Treg ratios in the peripheral blood of patients with melanoma when compared to healthy volunteers. In addition, Treg frequencies increased with the progression of disease. Most important, accumulation in patients with progressed melanoma correlated with a general reduction of T-cell responsiveness to recall antigens such as tetanus toxoid (TT) and tuberculin-GT (PPD). Surprisingly, DC-based vaccination of patients with advanced melanoma induced a decrease in Treg frequencies that correlated with partially recovered T-cell responses in these patients.

Our results show a clear association between the stage of melanoma progression and Treg frequencies in the peripheral blood that correlate with immunologic responses after DC-based cancer therapy. Together, these observations suggest a tumor-induced systemic suppression of T-cell reactivity supported by increased Treg activity in patients with progressive melanoma and may, at least in part, explain the poor success of DC-based immunotherapies in these patients. Nevertheless, DC-based vaccination helped to decrease Treg-mediated suppression in patients with melanoma.

Material and methods

Patients and healthy volunteers

Blood samples for isolation of peripheral blood mononuclear cells (PBMC) were collected from patients with malignant melanoma in different disease stages [stage I–IV, 2001 AJCC/UICC pTNM staging system (30)]. All patients had histological confirmed melanoma. Tumor types included superficial spreading melanoma (SSM), acrolentiginous melanoma (ALM) and nodular malignant melanoma (NMM). Patients did not receive any concomitant immunotherapy, radiotherapy or chemotherapy at the time point of analysis. In all patients, the period between the last systemic therapy and analysis of Treg frequencies was at least 3 months (Table 1). PBMC from healthy volunteers were obtained from buffy coats and served as controls. Informed written consent was obtained from all patients and healthy volunteers in accordance with local ethical committee approval (protocol 837.029.05 (4687) approved by the local ethics committee, Landesaerztekammer Rheinland-Pfalz).

Table 1.   Clinical characteristics of melanoma patients
StageGenderAgeSizeMetastasesTherapy
  1. Patients suffered from melanoma stages I–IV (AJCC classification 2003). Tumor types included superficial spreading melanoma (SSM), acrolentiginous melanoma (ALM) and nodular malignant melanoma (NMM). Tumor sizes ranged from 0.5 to 35.0 mm (LI-V). Location of tumor included face/neck, back and upper/lower limb. Two patients in the stage IV group had a melanoma of unknown origin. Primary diagnosis of malignant melanoma was between 1980 and 2008. Patients did not receive any concomitant immunotherapy, radiotherapy or chemotherapy at the time point of analysis. In all patients, the time period between the last systemic therapy and analysis of Treg frequencies was at least 3 months.

I–IIFemale n = 5
Male n = 8
54.3 years
(range: 37–74 years)
0.5–4 mmnonew/o therapy n = 10; after intermediate dose IFN-alpha therapy n = 3
IIIFemale n = 4
Male n = 5
55.6 years
(range: 38–71 years)
0.6–7.8 mmlymph nodesw/o therapy n = 6; after high dose IFN-alpha therapy n = 3
IVFemale n = 7
Male n = 11
54.8 years
(range: 32–83 years)
1.3–35 mmlymph nodes, liver, lung, brain, skin, bonesw/o therapy n = 9; after chemo n = 7; after immunotherapy n = 2

Preparation of peripheral blood mononuclear cells

PBMC were isolated from blood samples (patients with melanoma) or buffy coats (healthy individuals) by Ficoll density gradient centrifugation (PAA Laboratories GmbH, Pasching, Austria and Biochrome AG, Berlin Germany) as previously described (10).

Flow cytometric analysis

All antibodies were direct fluorochrome conjugates as follows: anti-CD4-FITC, anti-CD4-APC, anti-CD25-PE-Cy5, anti-CD127 (IL-7Rα)-PE and anti-HLA-DR-FITC (BD Bioscience, San Jose, CA, USA), anti-Foxp3-PE, anti-Foxp3-APC (clone PCH101; eBioscience, San Diego, CA, USA). For staining, PBMC were isolated from whole blood or buffy coats as previously described and surface staining was performed for 30 min at 4°C with optimal dilution of mAbs. Intracellular analysis of Foxp3 expression was performed using the Foxp3 staining buffer set from eBioscience according to the manufacturer’s instructions. Cells were analysed using FACSCalibur and CELLQuest software or LSRII and FACSDiva software (all Becton Dickinson, San Jose, CA, USA).

To analyse Treg frequencies, cells expressing defined marker combinations were quantified as a percentage of CD4+ cells after gating on lymphocytes in FSC = Forward Scatter/SSC = Side Scatter.

Sorting and functional testing of CD4+ T-cell subsets

CD4+ T cells were enriched from human PBMC via Magnetic Cell Separation cell sorting by magnetic labelling with CD4 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany) and stained, using anti-CD4-APC, anti-CD25-PE-Cy5, anti-CD127 (IL-7Rα)-PE and anti-HLA-DR-FITC mAb (BD Bioscience). T cells were sorted for CD4+ CD25++, CD4+ CD25+ CD127low and CD4+ CD25+ HLA-DR+ populations using FACSAria (Becton Dickinson). Sorted cell populations had a purity of 78–82% (CD4+ CD25++), 91–93% (CD4+ CD25+ CD127low) and 84–94% (CD4+ CD25+ HLA-DR+). For functional testing, sorted cell populations were cocultured with isolated CD4+ T cells as responder cells in 96-well plates in X-VIVO-15. A total of 105 responder cells were present in each well with decreasing numbers of sorted cells added at 1:1 ratio (105:105 sorted cells),1:1/2 (5 × 104 sorted cells), 1:1/4 (2,5 × 104 sorted cells) in comparison to 105 sorted cells alone. In this experiment 3 × 105 irradiated PBMC and 0.5 μg/ml anti-CD3 mAb served as stimulation. After 3 days of culture and additional 16 h in the presence of 37kBq/well [3H]-thymidine, proliferation was measured by [3H]-thymidine incorporation. After subtracting background activity determined from irradiated PBMC, proliferation was displayed normalized to responder T-cell proliferation.

Analysis of recall antigen-specific T-cell proliferation

Proliferation assays were performed at the indicated time points to detect T-cell responses to recall antigens. A total of 2 × 105 PBMC were cultured in the presence of 10 μg/ml TT or PPD in X-VIVO-15 in 96-well plates for 3–4 days plus additional 16 h in the presence of 37kBq/well [3H]-thymidine. PHA stimulation (2.4 μg/ml) served as a positive control. Background proliferation of unstimulated cells was subtracted.

Production of the DC-based vaccine and immunization schedule

DCs were generated from leukapheresis products prepared from each patient 2 weeks before the first vaccination as previously described (10). Briefly, for each vaccination, frozen PBMC were thawed and cultured in serum-free X-VIVO-15 (Cambrex, Verviers, Belgium) + 1.5% heat-inactivated autologous plasma containing 800 U/ml GM-CSF (Leukomax; Novartis, Basel, Switzerland) and 1000 U/ml IL-4 (Strathmann Biotec, Hamburg, Germany). At day 5, non-adherent cells were harvested and transferred to fresh culture plates. Cells were additionally stimulated on day 6 with 10 ng/ml IL-1β, 10 ng/ml TNF-α, 1000 U/ml IL-6 (all from Strathmann) and 1 μg/ml PGE2 (Pharmacia-Upjohn, Uppsala, Sweden). Mature DCs at day 7 were used for vaccination. The recall antigens TT or PPD (Chiron/Behring, Marburg, Germany), were added for the last 24 h of culture at 1 μg/ml.

Intradermal injections of 6 × 106 DC pulsed with a recall antigen (stage II: patients a–c: TT, patients d,e: PPD; stage IV: 01, 04, 05, 06: TT, patients 02, 03, 07, 08: PPD) were performed six times at 14-day intervals. Injections were given distant from the location of primary melanoma. For vaccination, 6 × 106 peptide-pulsed DC in 200 μl PBS were used.

DTH-reactions

Delayed type hypersensitivity (DTH) reactions to recall antigens were performed by intradermal injection of 4 × 105 DC either pulsed with TT or PPD in 200 μl PBS. Unpulsed DC served as control. DTH-reactions were measured 48 h after the application: the diameter of the palpable skin induration was measured in two dimensions and the mean diameter was given as cm infiltrate.

Statistical analysis

The data are expressed as mean and standard deviation (SD) for percentages. Differences between the groups were assessed using the non-paired two-sided Student’s t-test. P values <0.05 were considered statistically significant: *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001.

Results

Enhanced frequencies of T cells with regulatory T-cell phenotype in the peripheral blood of progressing patients with melanoma

Several molecules, including CD25, Foxp3, CD127 and HLA-DR are associated with the phenotype of human Tregs. Because of the fact that a unique lineage-specific marker is still elusive, we used multiple marker combinations to investigate Treg frequencies in peripheral blood from healthy volunteers. CD4+ T lymphocytes were gated and characterized for the Treg marker combinations CD4+ CD25++, CD4+ CD25+ CD127low, CD4+ CD25+ HLA-DR+, CD4+ Foxp3+, CD4+ Foxp3+ CD25+, CD4+ Foxp3+ CD127low and CD4+ Foxp3+ HLA-DR+ as indicated in Fig. 1a. Using these settings, we analysed samples from 15 to 28 healthy individuals. Frequencies of identified T-cell populations (CD4+ CD25++ 3.6 ± 0.9%, CD4+ CD25+ CD127low 5.8 ± 1.4%, CD4+ CD25+ HLA-DR+ 1.4 ± 0.6%, CD4+ Foxp3+ 4.9 ± 1.7%, CD4+ Foxp3+ CD25+ 4.9 ± 2.3%, CD4+ Foxp3+ CD127low 6.6 ± 1.5% and CD4+ Foxp3+ HLA-DR+ 1.9 ± 0.8% in CD4 +  T cells) resembled previously published results (28, 29, 26, 31).

Figure 1.

 Characterization of T-cell populations by different marker combinations associated with regulatory T cells in healthy volunteers. T-cell populations expressing Treg-associated markers were analysed in peripheral blood mononuclear cells (PBMC) of healthy volunteers. (a) PBMC from healthy volunteers were analysed by flow cytometry. CD4+ T cells were gated as CD4+ lymphocytes and characterized for the Treg marker combinations CD4+ CD25++, CD4+ CD25+ CD127low, CD4+ CD25+ HLA-DR+, CD4+ Foxp3+, CD4+ Foxp3+ CD25+, CD4+ Foxp3+ CD127low and CD4+ Foxp3+ HLA-DR+. Density plots show one typical experiment. Frequencies of T-cell populations are indicated as means of 15–28 individuals. (b) T-cell populations characterized by surface marker combinations were sorted based on the expression of CD4+ CD25++, CD4+ CD25+ CD127low and CD4+ CD25+ HLA-DR+ and tested for their suppressive capacity in coculture with CD4+ responder T cells stimulated by anti-CD3 mAb and irradiated PBMC. Proliferation was measured by 3H-TdR incorporation and displayed normalized to responder T-cell proliferation after subtracting the background activity of the irradiated PBMC. Data show mean values of three separate experiments.

As previously reported (26,28,29), we could show that T-cell populations sorted based on the expression of the surface marker combinations CD4+ CD25++, CD4+ CD25+ CD127low and CD4+ CD25+ HLA-DR+ (Fig. S1) were anergic after activation (Fig. 1b). Moreover, they suppressed T-cell proliferation in a dose-dependent manner (suppression by CD4+ CD25++ T cells: 74 ± 5% ratio 1:1, 45 ± 8% ratio 1/2:1, 38 ± 3% ratio 1/4:1; suppression by CD4+ CD25+ CD127low T cells: 84 ± 5% ratio 1:1, 70 ± 8% ratio 1/2:1, 54 ± 14% ratio 1/4:1; suppression by CD4+ CD25+ HLA-DR+ T cells: 86 ± 3% ratio 1:1, 59 ± 2% ratio 1/2:1, 35 ± 9% ratio 1/4:1; Fig. 1b). T-cell populations targeted by the intracellular transcription factor Foxp3 could not be tested functionally. However, as they largely comprise the same cells as the functionally tested T-cell populations, in particular as the CD25+ CD127low T-cell population (Fig. S2), Foxp3 expressing CD4+ T cells could be assumed to show comparable functional properties.

Applying the above tested staining patterns, we subsequently analysed peripheral blood of patients with melanoma in different stages of disease (Table 1) and thus were able to compare the frequencies of Treg populations expressing Treg-associated markers in healthy individuals and patients with melanoma (Fig. 2). Whereas the marker combination CD4+ CD25+ CD127low and CD4+ Foxp3+ HLA-DR+ only indicated a slight trend of Treg increase in the course of melanoma progression, CD4+ CD25++, CD4+ Foxp3+, CD4+ Foxp3+ CD25+ and CD4+ Foxp3+ CD127low clearly revealed significantly higher Treg frequencies in patients with melanoma at late stage (stage IV) when compared to healthy donors. Additionally, for CD4+ Foxp3+, CD4+ Foxp3+ CD25+ and CD4+ Foxp3+ CD127low, a significant increase in Treg ratios was displayed during the course of tumor progression. In contrast, the CD4+ CD25+ HLA-DR+ subset showed no significant differences within the four groups. Taken together, using distinct marker combinations, our results show a clear association between the stage of melanoma progression and the Treg frequency in the peripheral blood.

Figure 2.

 Expression of different markers associated with regulatory T cells in healthy volunteers and during the course of melanoma progression. Peripheral blood mononuclear cells from healthy volunteers (H) and melanoma patients (M I–IV) were analysed by flow cytometry. CD4+ T cells were gated as CD4+ lymphocytes and characterized for different Treg marker combinations. The percentages of CD4+ CD25++, CD4+ CD25+ CD127low, CD4+ CD25+ HLA-DR+, CD4+ Foxp3+, CD4+ Foxp3+ CD25+, CD4+ Foxp3+ CD127low and CD4+ Foxp3+ HLA-DR+ (H n = 15–28, M I+II n = 4–12, M III n = 3–9, M IV n = 11–18) subsets in CD4+ T cells were determined. In scatter plots, each data point corresponds to one individual, horizontal bars represent mean values.

Reduced recall antigen-specific immune responses in stage IV compared to patients at melanoma stage II

Several clinical studies, using immunotherapeutic approaches to treat melanoma, have failed in patients at melanoma stage IV, implying an overall immune suppression associated to increased Treg frequencies. Therefore, we correlated antigen-specific immune responses in patients with melanoma in vivo (DTH-reactivity) and in vitro (antigen-specific proliferation) with the obtained Treg frequencies. Specific DTH-reactivity to different recall antigens was determined in patients at melanoma stage II and stage IV and documented as mean diameter of the palpable skin induration. Whereas the reaction to both tested recall antigens (TT and PPD) was only minimal in patients at melanoma stage IV (infiltrate < 1 cm), patients at melanoma stage II showed a positive DTH response to TT as well as to PPD (Fig. 3a). To analyse these findings in more detail, in vitro proliferation assays were performed. We compared the TT-stimulated and PPD-stimulated proliferation of PBMC from healthy individuals and tumor-free patients with melanoma (stage II) as well as from patients with progressive disease (stage IV). The results shown in Fig. 3b demonstrate a significantly reduced antigen-specific T-cell response in patients at melanoma stage IV when compared to stage II (PPD) and healthy volunteers (TT and PPD). Interestingly, tumor-free melanoma patients with minimal residual disease also showed significantly reduced T-cell responses towards PPD when compared to healthy volunteers. These findings support the idea of a tumor-induced systemic suppression of T-cell reactivity through increased Treg ratios and activities in patients with progressive melanoma.

Figure 3.

 Stage-dependent T-cell responsiveness in patients with melanoma. (a) Delayed type hypersensitivity (DTH)-tests were performed by intradermal injection of dendritic cells either pulsed with tetanus toxoid (TT) or tuberculin-GT (PPD). DTH-reactions were measured 48 h after application and documented as mean diameter of the palpable skin induration. (b) Peripheral blood mononuclear cells from healthy volunteers (H n = 22) and patients with melanoma (M II n = 13, M IV n = 7) were stimulated in vitro with the recall antigens TT and PPD. Antigen-specific proliferation was measured by 3H-TdR incorporation.

Partially restored recall antigen-specific T-cell responses in patients with melanoma after vaccination with antigen-pulsed mature DC

Recall antigens are widely used in DC-based cancer immunization studies as control antigens to analyse CD4+ T-cell activity that provides crucial ‘help’ for effective CD8+ T-cell responses (9,10,32). During two recent clinical vaccination trials in patients at melanoma stage II and stage IV, we collected PBMC before treatment, and after the third and sixth vaccination. Recall antigen-specific reactivity against TT or PPD was analysed by proliferation assays. We were able to induce antigen-specific proliferation in the majority of vaccinated patients (Fig. 4). Whereas most of the patients at stage II showed increasing and long lasting T-cell reactivity over the period of vaccination (Fig. 4, upper panel), in six of eight patients at melanoma stage IV antigen-specific reactivity only exhibited a transient character and declined or even entirely vanished after several rounds of vaccination (Fig. 4, lower panel). In addition, T-cell reactivity on average was lower in patients at melanoma late stage than in patients at melanoma stage II.

Figure 4.

 Partially restored antigen-specific immune responses in patients with melanoma after dendritic cell (DC)-based immunotherapy. Vaccinations with recall antigen-pulsed mature DC were performed six times at 14-day interval in patients at melanoma stage II and stage IV. Before vaccination and 2 weeks after the third and the sixth vaccination. Peripheral blood mononuclear cells were collected and cultured in the presence of the recall antigens tetanus toxoid or tuberculin-GT. Antigen-specific proliferation was measured by 3H-TdR incorporation and corrected for background proliferation.

DC-based vaccination of patients with advanced melanoma leads to a decrease in regulatory T cells

Analyses of T-cell responses suggested a suppressed cellular immunity in patients with advanced melanoma when compared to patients with stage II melanoma and healthy volunteers. Nevertheless, after DC vaccination, patients at stage IV showed at least temporarily increased antigen-specific responses to recall antigens (Fig. 4, lower panel). To investigate whether this restored T-cell reactivity is associated with an alteration in Treg occurrence, we determined Treg frequencies in the peripheral blood of these patients. PBMC, as far as available, were analysed before, during and after vaccination, using different marker combinations as described previously. Indeed, we observed a clear reduction of Treg ratios after DC-based immunotherapy based on the marker combinations CD4+ CD25+ CD127low, CD4+ Foxp3+, CD4+ Foxp3+ CD25+ , CD4+ Foxp3+ CD127low and CD4+ Foxp3+ HLA-DR+. Concerning CD4+ CD25++ and CD4+ CD25+ HLA-DR+, only one patient showed a reduction in Treg frequency. However, this individual showed a significant decrease for all marker combinations, demonstrating the interindividual differences characteristic of clinical studies in the human system (Fig. 5). Taken together, our data demonstrate that DC-based vaccination is able to restore T-cell immunity even in patients at the melanoma late stage, and this effect may be because of a reduction of the level of Tregs.

Figure 5.

 Regulatory T-cell marker expression in the course of a dendritic cell (DC)-based immunotherapy. Peripheral blood mononuclear cells from patients at melanoma stage IV (n = 5) were analysed by flow cytometry before, after the third and after the sixth DC-vaccination, as far as available. The percentages of CD4+ CD25++, CD4+ CD25+ CD127low, CD4+ CD25+ HLA-DR+, CD4+ Foxp3+, CD4+ Foxp3+ CD25+, and CD4+ Foxp3+ HLA-DR+ subsets in CD4+ T cells were determined in the course of vaccination.

Discussion

Tumor-induced immunosuppression is a phenomenon that in part explains the poor clinical effectiveness of DC-based immunotherapies (33). It is an increasingly well-recognized paradigm that Tregs play an important role in suppression of antitumor T-cell immunity (34). Several studies in animal models have shown that depletion of Tregs or inhibition of Treg function in vivo significantly improves tumor clearance and survival (35,36). In humans, there is evidence for accumulation of Tregs not only in tumor tissue but also in the periphery, suggesting that tumor-induced immunosuppression may be a Treg mediated and systemic effect (23,37,38).

Unlike in rodents (39), a unique Treg-specific marker is still lacking in humans. We therefore used a panel of markers associated with a Treg phenotype to analyse Treg frequencies from different melanoma stages and healthy volunteers. We enumerated subsets of CD4+ T cells characterized by the expression of CD4+ CD25++, CD4+ CD25+ CD127low, CD4+ CD25+ HLA-DR+, CD4+ Foxp3+, CD4+ Foxp3+ CD25+, CD4+ Foxp3+ CD127low and CD4+ Foxp3+ HLA-DR+. CD25, Foxp3 and CD127 indicated enhanced Treg frequencies in the peripheral blood of patients at melanoma stage IV when compared to healthy volunteers. Furthermore, particularly regarding Foxp3 expression, an increase of Treg frequencies was observed in the course of melanoma progression. Moreover, in association with DC-mediated vaccination, a decrease in CD4+ Foxp3+, CD4+ Foxp3+ CD25+, CD4+ Foxp3+ HLA-DR+ and CD4+ CD25+ CD127low Tregs in peripheral blood of all monitored patients at melanoma stage IV was demonstrated. As expected, the use of CD25 alone was not sufficient to characterize Tregs, because CD25 is also expressed at high levels by activated CD4+ T helper cells. During the course of vaccination, CD25 expression either remained stable or increased – a possible indication of a vaccination induced T-cell activation. The combination of different Treg associated markers, especially in combination with Foxp3, allowed a more reliable quantification of Tregs in patients with melanoma and healthy individuals. Nevertheless, Foxp3 is discussed controversially as a key marker for Tregs, as it is also expressed by CD4+ T helper cells upon T-cell receptor stimulation without induction of regulatory function (40–45). Other findings, however, indicate that there might still be a difference between Foxp3 expression in regulatory T cells and activated T helper cells, for example in respect of permanence and expression level (46). Applying new methods for Treg quantification like quantitative DNA Methylation(47), though, could lead to more consolidated findings in the future.

Expression of MHC class II determinants (HLA-DR) defines a functionally distinct subpopulation of human Tregs that typically comprises 20–30% of the CD4+ CD25+ Treg population (29). Although the functional role of MHC class II expression on these cells is unclear, it was found to strictly correlate with contact-dependent suppressive activity. Moreover, MHC class II expression is upregulated on Tregs upon TCR-mediated stimulation and thus may identify a population of recently activated Tregs. Surprisingly, we did not find the frequency of CD4+ CD25+ HLA-DR+ Tregs to be significantly altered during tumor progression or vaccination. This observation can be interpreted in at least two ways: MHC class II expression may relate to a state of activation that is only transiently achieved and is therefore only observed on a limited number of cells at a given time. The other possibility is that MHC class II expression requires stimuli that are not provided by a growing tumor or by DC vaccination such as self-peptides and counterligands of the MHC class II molecule. Regardless of the underlying reasons, MHC class II expression does not seem to be affected on Tregs by melanoma growth.

The detection of a substantial increase in circulating Tregs in the peripheral blood of patients with melanoma is in line with previous studies in patients with cancer which suggest that Tregs play a pivotal role in a complex mechanism that suppresses cell-mediated immunity in patients with cancer (23,34,37,38). Nevertheless, this is the first study comparing Treg frequencies, characterized by the described panel of marker combinations, in patients with cancer not only to healthy individuals but also during the course of cancer progression and after immunotherapy with DC-based vaccination.

Our findings partially are in contrast to those from a study that showed an increase or expansion of Tregs after DC-based vaccination (48,49). However, that study only used staining combinations including CD25 (CD4+ CD25+ and CD4+ CD25+ Foxp3+) for Treg identification. As CD25 is known to be an activation marker in T helper cells, their observed increase in CD25+ T cells could be ascribed to T-cell activation because of DC vaccination, rather than Treg expansion. This interpretation is supported by reported discrepancies in Treg expansion at the cellular and mRNA levels concerning Foxp3 (49).

The present study found that immunosuppression increased during the progression of melanoma and was associated with enhanced Treg activity in these patients. There was a close correlation between stage progression and the increase of Treg frequencies in the peripheral blood, leading to a general suppression of T-cell-mediated immune responses in patients with advanced melanoma. DC-based immunotherapy is able to restore T-cell immunity to recall antigens and to induce T-cell responses against tumor antigens associated with decreased Treg ratios in the peripheral blood. However, the effectiveness of immunotherapy is dependent on the stage of disease. In progressing patients at stage IV, the stimulatory effects of the vaccine were transient, and long-lasting immune responses were only induced in tumor-free patients at stage II, detectable more than 1 year after the last vaccination (9,10,33). This discrepancy shows that Treg play a significant role in the general immunosuppression of patients with tumor, however, not an exclusive one. Other factors besides Treg seem to have an important impact on the overall antigen-specific T-cell responsiveness of patients at melanoma late stage during the course of vaccination and should be investigated in future studies.

We are aware that the correlation of increased Treg frequencies with reduced T-cell reactivity in patients with melanoma gives only indirect evidence for the role of Tregs in the general immune suppression of patients with cancer. Because of limited Treg numbers, we have not provided any functional data concerning the Treg activity in those patients with melanoma. The small volume of available blood samples combined with the relatively low Treg frequency in peripheral blood made such experiments unfeasible. However, the fact that DC vaccination in patients with melanoma reduced Treg frequencies together with the observation that impaired T-cell activity was restored provides strong evidence for a Treg-associated tumor-induced immunosuppression in progressive melanoma.

The data presented in this study show an association between the stage of melanoma progression, the frequency of cells with a Treg phenotype in the peripheral blood, the overall T-cell reactivity and the immunological efficacy of a DC-based vaccination trial. Thus, strategies aimed at inhibiting the function of Tregs could contribute to improve the efficiency of DC-based vaccination. Moreover, combining DC-based vaccination with adoptive immunotherapy, as expansion of TIL devoid of Treg amplification (50), could open up new perspectives and should be evaluated in future clinical trials.

Acknowledgements

We thank Kerstin Steinbrink, Mario Hubo and Stephen F. Marino for critically reading the manuscript. This work was supported by grant SFB 432 B11 to H.J. and Transregio52 to A.T. and H.J. Any commercial sponsorship or commercial affiliations are disclosed.

Ancillary