Introduction
The role played by T cell mediated immunity in the control of tumour growth has been established over recent years. As a result, most immunization strategies adopted in clinical trials of cancer treatments have aimed at enhancing tumour antigen (TA)-specific cellular immunity. The induction and expansion of TA-specific T cells requires optimal antigen presentation and T-cell co-stimulation. Dendritic cells (DCs) are specialized antigen-presenting cells with a remarkable ability to stimulate naïve T lymphocytes and generate memory T lymphocytes [
1]. However, objective response rates to vaccine or DC trials in cancer remain low [
2]. Differentiation and maturation of DCs are important to their protective activity against tumour development [
3]. Exposure to necrotic tumor cells can induce maturation of immunostimulatory DCs [
4] but the involved mechanisms are still unresolved [
5].
Cytotoxic T lymphocytes (CTLs) directed against tumour cells can be amplified
in vitro with the use of DCs pulsed with treated tumour cells (DC-Tu) [
6]. When assays were done with cells from healthy donors, DC-Tu stimulation repeatedly increased the final cytolytic activity of T cells more than twofold. However, we observed that a similar procedure applied to cells from cancer patients enhanced the final cytotoxic activity against autologous tumour only in half of the assays [
6]. We noticed in these experiments that the final yield and phenotype of blood-derived myeloid immature DCs was heterogeneous in cancer patients [
6]. These findings could be related to a relationship between immune suppression instilled during tumour development, as previously described by Kusmartsev and Gabrilovich [
7], and increased production of immature myeloid cells in patients with advanced cancers [
8].
Our aim in the present study was to detail the differences in characteristics of DCs between patients with cancer and healthy donors. We investigated blood cells from patients with breast, colorectal, or renal carcinoma and compared them, using the same assays, with cells from healthy donors. DCs were obtained from peripheral blood [
9] and matured using various cocktails combining proinflammatory cytokines and danger or co-stimulating signals that are known for their ability to induce a T-helper-1 phenotype [
3,
10]. Tumour cells were from the M74 melanoma cell line in all of the assays. Treatment of tumour cells was done for induction of late apoptosis (postapoptotic necrotic tumour cells) [
11]. Necrotic cells were chosen for DC pulsing, in accordance with previous reports [
5,
12,
13] and preliminary experiments by our group that demonstrated that processing and cross-presentation of TA led to specific CTL responses in DCs pulsed under these conditions (Gervais A, unpublished data).
The ultimate mechanisms by which DC deficiency is established are not understood. The tumour microenvironment is rich in growth factors and molecules that are able to modulate the immune response of the host. Polyamines, which are conducive to proliferation and metastatic invasion, are synthesized in large amounts by tumour cells [
14]. A therapeutic strategy combining inhibition of all cellular and exogenous sources of polyamines has been evaluated in several murine tumour models, with positive findings [
15]. However, the role played by polyamines in immune processes is poorly understood [
16,
17]. Nevertheless, our group showed that polyamine deprivation can prevent the development of
in vivo tumour-induced immunosuppression [
18]. In the present study, the hypothesis that putrescine is involved in immunodeficiency was tested by investigating the effects of putrescine on functional activity of DCs from donors. This treatment was able to mimic the abnormalities observed in DCs from patients with breast cancer.
Materials and methods
Patients
Thirty-four patients with histologically confirmed cancer were enroled in the study. Seventeen (age 47–76 years) had breast cancer: 13 had infiltrating ductal carcinoma (grade I-III; Elston Ellis grading); one was invasive lobular carcinoma (grade III); one was mixed ductal-lobular carcinoma (grade I); and two were in situ ductal carcinomas (low and high grade). Seven (age 33–86 years) had colorectal cancer (stage 2–4 adenocarcinoma) and 10 (age 42–78 years) had renal cell carcinoma (Fuhrman grade III clear cell carcinoma). Patients were newly diagnosed and peripheral blood samples were collected at the time of initial surgery, with no prior therapy. The study was approved by the regional ethics committee (Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale de Rennes, Rennes, France). Eleven healthy volunteers (HLA-A2) served as control individuals (Etablissement Français du Sang, Rennes, France).
Isolation of cells from peripheral blood
Cells were centrifuged by applying a density gradient (UNISEP®; Novamed, Jerusalem, Israel). Mononuclear cells (MNCs) were frozen in human serum albumin and 10% dimethyl sulphoxide until use for DC and lymphocyte preparation.
Tumour cell treatment
The HLA-A2 MelanA-Mart1 expressing M74 melanoma cell line was used for both antigen DC pulsing and as a target for evaluation of specific cytotoxic activity. This cell line and the K562 natural killer (NK) cell-sensitive erythroleukaemia cell line were maintained in RPMI 1640 medium (Eurobio, Les Ulis, France) containing 10% foetal calf serum (Gibco Invitrogen Life Technologies, Cergy Pontoise, France), 1 mmol/l L-glutamine, 50 μg/ml streptomycin and 50 IU/ml penicillin (ICN Biomedicals, Aurora, OH, USA).
M74 cells were used for TA DC pulsing following necrosis-inducing treatment. The method was adapted from that presented by Lennon and coworkers [
11]. Briefly, cells were treated with hydrogen peroxide 10 μmol/l (Sigma-Aldrich, Saint Quentin-Fallavier, France) for 3 consecutive days. Supernatant cells were collected each day, pooled and kept at 4°C. Collected cells (M74 per) were used for DC pulsing (DC-Tu
per).
Treated tumour cells were examined for degree of apoptosis and secondary necrosis using a standard fluorescence-activated cell sorting assay (Annexin V-FITC detection kit; Immunotech, Marseille, France), which detects binding of annexin V (A) and propidium iodide inclusion/exclusion. With 10 μmol/l peroxide, 16% of the collected cells were apoptotic (annexin V positive/propidium iodide negative) and 50% were in a state of postapoptotic necrosis (annexin V positive/propidium iodide positive).
Culture of dendritic cells
DCs were prepared from MNCs (from patients or healthy donors) in accordance with the method described by Sallusto and Lanzavecchia [
9]. Briefly, 10 × 10
6 MNCs were seeded in 5 ml serum-free X-Vivo 10 medium (Biowhittaker, Walkersville, MD, USA) in a 25 cm
2 culture flask (Cellstar
®; Greiner Labortechnik, Frickenhausen, Germany). Nonadherent cells were collected after 2 hours for lymphocyte culture. The remaining adherent cells were cultured in DC medium: serum free X-Vivo 10 medium supplemented with 10% AB serum (EFS de Rennes, Rennes, France), 10 μg/ml steptomycin and 100 IU/ml penicillin. Granulocyte–macrophage colony-stimulating factor (GM-CSF) 1000 IU/ml (Leucomax 400™; Novartis/Shering Plough, Huningue, France, Switzerland) and 400 IU/ml IL-4 (Promokine; Promocell, Heidelberg, Germany) were added on days 0, 2 and 5 of culture. After 7 days nonadherent cells (immature DCs) were collected and added to peroxide-treated M74 cells (ratio 1:10) for antigen processing. After 18 hours of contact, supernatant cells (DC-Tu
per) were added to autologous lymphocytes (DC:lymphocyte ratio 1:100) for lymphocyte stimulation. DCs were phenotypically characterized before lymphocyte stimulation (day 8).
For maturation assays, immature DCs were seeded in DC medium (density 106 cells/ml) in 24-well plates (Falcon®; Becton Dickinson, Franklin Lakes, NJ, USA) and maturating agents were immediately added. Three different maturation cocktails were evaluated. The first (cocktail A) was a combination of tumour necrosis factor-α (25 ng/ml; Pharmingen, San Diego, CA, USA), lipopolysaccharide (10 μg/ml; Sigma-Aldrich) and CD40L (0.4 μg/ml; Alexis Biochemicals, QBiogene, Illkirch, France). The second (cocktail B) was a combination of IL-1β (10 ng/ml), IL-6 (1000 U/ml), tumour necrosis factor-α (10 ng/ml; R&D systems, Lille, France) and prostaglandin E2 (1 μg/ml; Sigma-Aldrich). The third (cocktail C) was a combination of bacterial extracts (1 μg/ml; Ribomunyl®; Pierre Fabre Médicament, Boulogne, France) and IFN-γ (1000 U/ml; Imukin®, Boehringer Ingelheim, Reims, France). After 18 hours of contact, supernatant cells (matured DCs) were collected and phenotypically characterized (day 8).
Phagocytic activity of dendritic cells
Phagocytic activity was evaluated using FITC-labelled opsonized bacteria (Escherichia coli; Phago Test®; OrpeGen Pharma, Heidelberg, Germany). Immature DCs from patients or donors and opsonized bacteria were co-cultured for 2 hours at 37°C and their internalization was evaluated by flow cytometry, in accordance with the manufacturer's recommendations. Controls were run at +4°C.
IL-10 and IL-12 production
Measurements of IL-10 and IL-12 were done in supernatants of immature DCs and mature DCs after 72 hours in culture at 37°C. Assays were done using enzyme-linked immunosorbent assay methods according to the manufacturer's instructions (Ready-set-Go®; eBioscience, San Diego, CA, USA) and were performed in duplicate.
Dendritic cell treatments
Immature DCs were treated for 18 hours with 10 mmol/l putrescine (1,4-diaminobutane dihydrochloride; Sigma-Aldrich) before phenotype or functional activity were measured. Because it was reported that all-
trans retinoic acid (ATRA) can reduce the number of immature myeloid cells and favour their differentiation into DCs [
8], putrescine-treated immature DCs were added with ATRA (Sigma-Aldrich); 1 μmol/l ATRA was added each day for 5 days, in accordance with the procedure described by Almand and coworkers [
8]. Phenotype and functional analyses were conducted after 5 days with or without ATRA treatment.
Lymphocyte culture
Lymphocytes were cultured from MNCs in lymphocyte medium: RPMI 1640 containing 10% AB serum, 1 mmol/l L-glutamine, 2% pyruvate, 1% nonessential amino acids (Bioproducts, Gagny, France), 100 μg/ml streptomycin, 100 IU/ml penicillin and 150 IU/ml IL-2 (Proleukin®; Chiron, Suresnes, France). After 8 days in culture (density 106 cells/ml), lymphocytes were stimulated with DC-Tuper. Number, phenotypic and functional characteristics of lymphocytes were evaluated 7 days after DC-Tuper stimulation. Viability was evaluated using the trypan blue exclusion test. Controls were performed with nonstimulated lymphocytes.
Flow cytometry analysis
Cells (10
5) were suspended in phosphate-buffered saline supplemented with 0.5% bovine serum albumin and labelled for characterization of lymphocyte or DC phenotype by incubation at 4°C for 30 min with the following PE-, FITC-, or PC5-conjugated antibodies and corresponding isotypes: anti-CD3 (clone UCTH1), anti-CD4 (13B8.2), anti-CD8 (B9 11), anti-CD25 (B1.49.9), anti-CD40 (mAb 89), anti-CD56 (NKH-1), anti-αβ-TCR (BMA 031), anti-γδ-TCR (immu 510), anti-CD80 (MAB 104), anti-CD83 (HB15A) and anti-CD152 (CTLA-4; after saponin permeabilization) from Immunotech (Marseille, France); anti-CD11c (S-HCL-3), anti-HLA-DR (L243) and Lin1 cocktail (anti-CD3, anti-CD14, anti-CD16, anti-CD19, anti-CD20 and anti-CD56) from Becton Dickinson/Pharmingen (CA, USA); and CD86 (BU63) from Immunotech (Oxford, UK). Cells were washed and suspended in 250 μl phosphate-buffered saline added with 0.3% formol. CD4
+CD25
+CTLA4
+ was considered to be the T regulatory cell phenotype, in accordance with the findings of Jonuleit and coworkers [
19]. Data analysis was performed using a FACScan flow cytometer (Becton Dickinson).
Cytotoxicity assays
T-cell mediated cytotoxicity was tested in triplicate with a standard 51Cr release assay. The assays were conducted in U-bottomed microtitre plates. Depending on the assays, target cells were M74 tumour cell line or K562 cells pulsed with 51Cr ([51Cr]sodium chromate; Amersham Life Sciences, Buckinghamshire, England) for 1 hour.
A total of 5000 target cells/well were mixed with effector cells (ratio of effector to target cells 50:1) and incubated for 4 hours. Chromium release was assessed in culture supernatants using a γ-counter (Topcount, Packard Instrument, Rungis, France). Specific release was calculated as follows: ([mean experimental counts/min - mean spontaneous counts/min]/ [mean maximum counts/min - mean spontaneous counts/min]) × 100.
IFN-γ production
Responder cells were evaluated for their production of IFN-γ in response to contact with antigenic cells. Analyses were performed 8 days after stimulation with DC-Tuper. Briefly 2 × 105 M74 cells were seeded in 24-well plates for 12 hours. The supernatant was discarded before adding 105 lymphocytes in a final volume of 500 μl of medium without IL-2. The plates were then incubated at 37°C for 72 hours and IFN-γ was measured in supernatant using enzyme-linked immunosorbent assay methods, in accordance with the manufacturer's instructions (Ready-set-Go®; eBioscience). Duplicate wells were run for each assay.
Statistical analysis
Each assay was repeated with at least three different donors or patients. The nonparametric Mann–Whitney rank test was used for statistical analysis.
Discussion
In recent years several groups have described defective immune function in tumour-bearing animals [
18,
20,
21] and in cancer patients [
8,
22,
23]. Of note, it was reported that factors produced by tumour cells could influence differentiation of DCs from CD34
+ progenitors, and that low concentrations of IL-4 could reverse the inhibitory effect of cancer cell conditioned medium, at least in terms of phenotype and some functional differentiation of DCs [
22]. We show here that, even in the presence of IL-4, differences in differentiation of circulating monocytes into DCs persisted in cancer patients as compared with healthy donors. Using the classic procedure of blood monocyte derived DC culture (in the presence of IL-4 and GM-CSF), the
ex vivo yield of DCs was found to be significantly reduced in patients with cancer, particularly in those with breast cancer. Furthermore, the phenotype of collected cells using this procedure was different in patients with breast cancer. Expression of MHC class II (HLA-DR
+Lin
- cells) was found to be lower and the percentage of HLA-DR
-Lin
- to be higher than in donors. In contrast, these subpopulations were not significantly modified in patients with colon or renal cell carcinoma.
Whatever combination of maturating agents was used, significantly lower expressions of mature DC markers were observed in patients with breast cancer. Maturation induced by Ribomunyl®/Imukin® resulted in lower expressions of CD80 and CD86 in patients than in donors, but, interestingly, it also resulted in greater production of IL-12.
Other groups have reported that, in breast cancer patients, monocyte-derived DCs have substantially lower level of expression of HLA-DR than do DCs isolated from control donors, leading to a reduced ability to stimulate allogenic and Flu-specific T-cell responses [
8]. We confirm here that DCs from such patients not only exhibit low expression of MHC class II but they also have reduced ability to cross-prime exogenous antigens. Stimulation of CTLs by pulsed DCs was less efficient in patients than in donors. In a similar procedure for lymphocyte stimulation, using the same antigen preparation (peroxide-treated tumour cells) and tumour target (M74 cell line), we repeatedly observed defective stimulation when DCs were from patients with breast cancer. In general, the natural cytolytic activity of lymphocytes against the M74 or NK target cell line was found to be lower in patients than in donors. Unlike donors, patients were not selected for their expression of HLA-A2 class I molecules. This could represent an advantage in terms of CTL activation, but the opposite was observed. Cytolytic activity was enhanced by up to 40% when DCs were from donors but only up to 10% when they were from patients.
IFN-γ production after DC-Tu stimulation was repeatedly found to be enhanced in donors. In contrast, nonspecific lysis of the NK-sensitive K562 cell line was the same after DC-Tu
per stimulation, clearly indicating that in donors DC-Tu
per stimulation induced TA-specific T cells. Induction of TA-specific T cells did not occur in all cancer patients. Blockade of DCs at an early stage in differentiation could be responsible for this inconsistency between patients. For example, in patient S137 a high percentage of HLA-DR
-Lin
- cells was observed in DCs, and concomitantly IFN-γ lymphocyte production was reduced twofold after DC-Tu stimulation, indicating that DCs were not only nonimmunogenic but were actually tolerogenic in this patient. However, the percentage of regulatory T cells was not changed after DC-Tu stimulation (<0.1% in S137). Correlation could not be demonstrated in this study between clinical grade of disease and HLA-DR
-Lin
- DC phenotype. Defective function and poor ability of immature DCs to mature in some patients could represent an additional reason why DC cell therapy in cancer patients has, contrary to expectations, not yet yielded significant clinical responses [
2].
Defective DC function can be mimicked by adding putrescine to the culture medium of DCs from healthy donors. The percentage of cells with HLA-DR
-Lin
- phenotype was found to be enhanced after putrescine treatment. In addition, expansion and final cytolytic activity of lymphocytes was reduced following the DC-Tu
per stimulation procedure, leading us to conclude that adding putrescine to the microenvironment of antigen-presenting cells blocks their ability to cross-prime exogenous antigens efficiently, indicating a reduction in their immunogenic function. It was reported by other authors that spermine, another polyamine, is responsible for severe inhibition in proinflammatoty cytokine synthesis when added to cultures of human peripheral blood MNCs stimulated with lipopolysaccharide [
24].
Our group previously established that polyamine deprivation leads to significant reduction in tumour growth in murine experimental models. Consistent with that effect, an enhancement in CD8
+ T lymphocytes was observed in the spleens of the animals [
18]. With similar experimental tumours we observed that combining polyamine deprivation with cyclophosphamide, which is known to downregulate regulatory T cells [
25], enhances macrophage tumouricidal activity, indicating that the two treatments have synergistic effects [
26].
In addition, breast cancer tissues are characterized by high polyamine levels. In a study including 174 patients with invasive breast cancer [
27], a correlation was established between enhancement of putrescine and spermidine levels and tumour aggressiveness. Taken together, these observations led to the conclusion that putrescine release by tumour cells may be involved in the defective DC function observed in breast cancer patients. Interestingly, we showed in the present study that
in vitro treatment of DCs with ATRA could reverse the putrescine-induced deficiency in DC function. ATRA and retinoic derivatives are known to influence DC differentiation, favouring a T-helper-1 response [
28]. Further investigations are needed to detail the mechanism underlying the reversal in putrescine-induced deficiency in DC function. Nevertheless, use of ATRA treatment to initiate TA-specific CTL expansion in cancer patients could be of particular interest.
Competing interests
The author(s) declare that they have no competing interests.
Authors’ contributions
AG carried out the DC-Tu preparation, lymphocyte stimulation procedures and measurements of functional activities, and participated in writing the manuscript. JL selected the patients with breast cancer and took biological samples. FB-T participated in the design and coordination of the study. FB carried out anatomo-pathological examinations. TL participated in drawing blood samples in breast patients. LS selected the patients with colonic carcinoma and took biological samples. J-JP selected the patients with renal cell carcinoma and took biological samples. NG carried out cytometric analyses. VC-Q conceived the study, participated in its design and coordination, and wrote the manuscript.