It has been reported that regulatory T cells (T regs) play important roles in immunological self-tolerance,1–4 and are functionally immune-suppressive subsets of T cells. T regs are identified as a small population of CD4+ cells that constitutively express CD25 (IL-2 receptor α chain) on their surface, and several other markers, including CD45RO, glucocorticoid-induced tumor-necrosis factor receptor-related protein (GITR), or cytotoxic T lymphocyte-associated antigen 4 (CTLA-4), are also known to be expressed on T regs.3–5 Recently, it has been reported that Foxp3, forkhead/winged helix transcription factor, is the most reliable marker of T regs6, 7; therefore, it is possible to define T regs more strictly as CD4+ CD25+Foxp3+ cells.
In mice, it is known that auto-immune diseases such as ulcerative colitis or Crohn's disease occur due to the depletion of T regs.6, 8 Also in humans, immune dysregulation polyendocrinopathy enteropathy X-linked syndrome (IPEX) is an auto-immune disease developed from a deficiency of T regs.8–11 These observations indicate that T regs play important roles in immunological homeostasis. Although the mechanisms of suppression by T regs are still unclear, it has been reported that T regs can inhibit the function of effector T cells directly by cell to cell contact or indirectly via the secretion of immune-suppressive cytokines.12–14
Recently, in murine models, many studies have shown that the depletion or inhibition of T regs can enhance anti-tumor immunity.15, 16 Moreover, in humans, it has been reported that the population of T regs in tumor-infiltrating lymphocytes (TILs) is significantly higher than that in normal tissues in several malignancies.17–21 We have reported that the frequency of T regs among TILs, lymphocytes derived from tumor-draining regional lymph nodes (LNLs), and peripheral blood lymphocytes (PBLs) was higher in gastric cancer and esophageal cancer patients in comparison to their normal counterparts.22 In addition, patients with a higher proportion of T regs showed poorer survival rates in comparison to those with a lower proportion.23, 24 Interestingly, after patients received curative resection for gastric cancer, the increased proportion of T regs was significantly restored to levels comparable with normal healthy donors.24 These results strongly suggest that tumor-related factors induce and expand the accumulation of T regs in cancer-bearing hosts. In fact, Curiel et al. reported that CCL22 chemokines derived from a tumor induce the migration of T regs through CCR4, which is a chemokine receptor for CCL22, and impair anti-tumor immunity in ovarian cancer.25 Also, it has been reported that T regs were able to migrate in response to CCL22 and CCL17 through CCR4.26 There is, however, still limited information describing the mechanisms behind the accumulation of T regs within the tumor microenvironment.
In the present study, we investigated the frequency of CD4+CD25+Foxp3+ T regs in TILs, LNLs and PBLs of patients with gastric cancer, and evaluated the relationship between the presence of CCL17- or CCL22-producing CD14+ cells and the infiltration of T regs within gastric cancer.
Material and methods
Forty-five patients with gastric cancer, who were operated on in the University of Yamanashi Hospital from May 2006 to March 2007, were enrolled in the present study. They were 67.1 ± 10.6 years old (mean ± SD); 30 patients were men and 15 were women. The characteristics of the study subjects are summarized in Table I. Patients with gastric cancer were divided into 2 groups: those with early disease (n = 29) corresponding to Stage I according to the TNM classification for gastric cancer (UICC), and those with advanced disease corresponding to Stages II, III and IV (n = 16). Regional lymph nodes in the stomach of patients with gastric cancer were classified into N1 regional lymph nodes adjacent to the gastric tumor and N2 regional lymph nodes marginally distant from the tumor, according to the Japanese Classification of Gastric Carcinoma.27 N1 and N2 regional lymph nodes and mesenteric lymph nodes used as a control, which were not regional lymph nodes of the stomach, were collected during surgery.
None of the patients received radiotherapy, chemotherapy or other medical interventions before the study. This study was approved by the ethics committee of the University of Yamanashi, and written informed consent was obtained from all patients.
PBLs were isolated with Ficoll-Paque Plus (Amersham Biosciences, Uppsala, Sweden) density gradient solution. Tumor tissue, normal gastric mucosa and regional lymph nodes from gastric cancer were isolated during surgery and homogenized by mechanical mincing with X-VIVO15 medium (CAMBREX, East Rutherford, NJ). Then, cell mixtures were passed through a cell strainer (Becton Dickinson, Franklin Lakes, NJ) and suspended as a single-cell suspension with X-VIVO15. Subsequently, the single-cell suspension was purified by centrifugation with Ficoll-Paque Plus.
CD4+CD25+ T-cell separation by magnetic beads
CD4+CD25+ T cells were isolated from PBLs or LNLs by CD4-negative selection followed by CD25-positive selection using a Dynal® CD4+CD25+ T-reg Kit according to the manufacturer's protocol (Invitrogen, Carlsbad, CA).
Flow cytometric analysis
To analyze the prevalence of T-regs, CD4+CD25+Foxp3+ cells were evaluated using the Human Regulatory T-cell staining Kit (eBioscience, San Diego, CA) according to the manufacturer's protocol. Briefly, single-cell suspension was stained using a cocktail of anti-CD4-FITC mAb and anti-CD25-APC mAb, then, after permeabilization, samples were blocked by normal rat serum, and stained using anti-Foxp3-PE mAb or PE conjugated rat IgG2a used as an isotype control. Then, the number of Foxp3-positive cells in comparison to the isotype control (PE conjugated rat IgG2a) on the gating of CD4+ cells was evaluated, and the frequency of Foxp3+ T regs was expressed as a percentage of total CD4+ cells.
In some experiments, single-cell suspension derived from tumor and normal gastric mucosa was stained using anti-CD14-FITC mAb (DAKO, Glostrup, Denmark) and anti-CCL17/TARC-PE mAb or anti-CCL22/MDC-PE mAb (R&D Systems, Minneapolis, MN) with an intracellular staining technique. FITC-conjugated mouse IgG2a (DAKO) or PE-conjugated mouse IgG1 and IgG2b (DAKO) were used as an isotype control. CCL17- or CCL22-positive cells on the gating of CD14+ cells were counted and the frequency of CCL17- or CCL22-positive cells was expressed as a percentage of total CD14+ cells.
Double- or triple-color flow cytometry was performed using FACSCalibur (BD Biosciences, San Jose, CA). Cells were analyzed using CellQuest Pro software.
Foxp3 staining in gastric cancer tissue sections was conducted using the avidin-biotin-peroxidase complex method on paraffin-embedded 4-μm-thick sections. Briefly, each paraffin section was deparaffinized, followed by antigen retrieval with Epitope Retrieval Solution (10 mmol citrate buffer (pH 6.0), Dakocytomation, Glostrup, Denmark) in a preheated water bath (98°C, 40 min), and endogenous peroxidase was blocked using Chemmate Peroxidase Blocking Solution (DAKO). Then, biotinylated anti-human Foxp3 antibody (diluted by PBS, 1:20, eBioscience) was applied for 40 min at room temperature. Thereafter, the sections were incubated with streptavidin conjugated horseradish peroxidase (DAKO) for 10 min, followed by development with 3,3′-diaminobenzidine (DAKO) for 5 min, and counter-staining with hematoxylin. Negative control staining was performed with Negative Control Reagent (Dakocytomation) instead of the specific primary antibody.
In vitro migration assay
CD4+CD25+ cells and CD4+CD25− cells were isolated from regional lymph nodes or PBLs of gastric cancer patients using magnetic beads. Migration of T cells was assessed by a 96-well chemotaxis chamber assay (Neuro Probe, Gaithersburg, MD). Briefly, the bottom chamber was filled with X-VIVO15 with recombinant human CCL17 or CCL22 chemokines (0, 0.1 and 1 ng/ml) (PeproTech, Rocky Hill, NJ) in the combination with neutralizing anti-CCL17 mAb or anti-CCL22 mAb (2 μg/ml, R&D Systems). Then, a 5-μm pore polyvinylpyrrolidine-free polycarbonate filter (Neuro Probe) was placed over the plate, and 5 × 104 cells in 65 μl X-VIVO15 were placed in the upper chamber in triplicate. After incubation at 37°C for 2 hr, cells that had migrated into the bottom chamber were quantified by MTT methods. MTT reagents (5 mg/ml, Sigma-Aldrich, St. Louis, MO) were added to each chamber and incubated for 4 hr. Then, DMSO (Sigma-Aldrich) was added to dissolve the crystals and samples were incubated overnight. The plate was read on a spectrophotometer at 570 nm, and cell numbers were calculated from the standard curve of a known cell number.
Differences between values were determined using the nonpaired Student's t-test. Correlation between values was evaluated using nonparametric Spearman's rank correlation. Significance was considered at p < 0.05.
Increased frequency of Foxp3+ T regs in TILs of patients with gastric cancer
To analyze the prevalence of T regs, CD4+CD25+Foxp3+ cells were evaluated by flow cytometry (Fig. 1) and expressed as a percentage of total CD4+ cells. In all gastric cancer cases, the frequency of T regs in TILs was significantly higher than that in normal gastric mucosa in the same patient (mean ± SD; 12.4% ± 7.5% vs. 4.1% ± 5.3%, respectively, p < 0.01) (Fig. 2a). In advanced disease, the frequency of T regs in TILs was significantly higher than in normal mucosa (13.1% ± 7.4% vs. 3.4% ± 3.7%, respectively, p < 0.01). Furthermore, the frequency of T regs in TILs in early disease was also significantly increased compared to in normal mucosa (12.1% ± 7.6% vs. 4.5% ± 6.0%, respectively, p < 0.01). Of note, the degree of infiltrating T regs in early disease occurred to the same extent as in advanced disease, indicating that the accumulation of T regs had already occurred in a relatively early stage of the disease.
Representative Foxp3 staining by immunohistochemistry demonstrated that the accumulation of Foxp3+ T regs could be observed within the tumor (Fig. 3a). Also, an increased amount of Foxp3+ T regs was seen in a peritumor region, which comprised small, early gastric cancer (Figs. 3b and 3c).
Increased frequency of Foxp3+ T regs in PBLs of patients with disease progression
The frequency of Foxp3+ T regs in PBLs of all patients was not significantly higher than in normal healthy donors (mean ± SD;3.0% ± 1.9% vs. 2.1% ± 0.7%, respectively, p = 0.14) (Fig. 2b); however, the frequency of Foxp3+ T regs in PBLs was significantly increased in advanced disease as compared to in early disease (3.9% ± 2.7% vs. 2.5% ± 1.3%, respectively, p < 0.05) or in healthy donors (3.9% ± 2.7% vs. 2.1% ± 0.7%, respectively, p < 0.05). Furthermore, as shown in Table II, the frequency of Foxp3+ T regs in PBLs in cases of lymph node metastasis was significantly increased than without lymph node metastasis (3.8 ± 2.6 vs. 2.5 ± 1.2, respectively, p < 0.05). Thus, it is likely that the prevalence of T regs in PBLs increased according to disease progression. We also analyzed the correlation of T reg prevalence between TILs and PBLs in the same patient, but no significant correlation was observed (data not shown).
Table II. Prevalence of Foxp3+ T regs relation to lymph node metastasis or histological classification
Each value indicates the frequency of Foxp3+ T regs as a percentage of total CD4+ cells (mean ± SD). PBL; peripheral blood lymphocytes, LNL; lymph node lymphocytes, TIL; tumor infiltrating lymphocytes, N1; regional lymph nodes adjacent to a gastric tumor, N2; regional lymph nodes marginally distant from a gastric tumor according to the Japanese Classification of Gastric Carcinoma (26).
Increased frequency of Foxp3+ T regs in regional lymph nodes adjacent to the tumor
Regional lymph nodes in the stomach of patients with gastric cancer were classified into N1 regional lymph nodes adjacent to the gastric tumor and N2 regional lymph nodes marginally distant from the tumor. Summarized data from all individuals indicated that the percentage of Foxp3+ T regs of N1 regional lymph nodes was significantly higher than in control mesenteric lymph nodes (mean ± SD; 7.1% ± 6.0% vs. 2.9% ± 3.3%, respectively, p < 0.01), as shown in Figure 2c. In advanced disease, the percentage of Foxp3+ T regs of N2 lymph nodes (5.2% ± 4.3%) as well as N1 lymph nodes (8.1% ± 5.4%) was significantly higher than in control mesenteric lymph nodes (2.7% ± 2.9%); however, in early disease, there were no significant differences in the prevalence of T regs among N1, N2 and the control.
These observations indicated that tumor-draining lymph nodes with advanced disease exhibited an increased prevalence of Foxp3+ T regs in comparison to control mesenteric lymph nodes. Furthermore, a more extended area (N2) of regional lymph nodes as well as lymph nodes (N1) adjacent to the tumors was involved in the increased prevalence of Foxp3+ T regs according to disease progression.
Correlation between the levels of CCL17 or CCL22 and frequency of Foxp3+ T regs in the tumor
Single cells derived from tumor tissues and normal gastric mucosa were evaluated for CD14-FITC/CCL17-PE-positive cells or CD14-FITC/CCL22-PE-positive cells by flow cytometry. Representative flow data of CD14/CCL17 or CD14/CCL22 staining are demonstrated (Figs. 4a and 4b). CCL17- or CCL22-positive cells on the gating of CD14+ cells were counted and the frequency of CCL17- or CCL22-positive cells was expressed as a percentage of total CD14+ cells. As a result, the frequency of CCL17- or CCL22-positive cells among CD14+ cells in the tumor was significantly higher than in normal tissues (mean ± SD; 30.6% ± 15.3% vs. 11.3% ± 6.6%, n = 6, p < 0.05; 31.3% ± 14.9% vs. 14.4% ± 7.9%, n = 17, p < 0.01, respectively) (Figs. 5a and 5c). Furthermore, we found a significant correlation between the frequency of CCL17- or CCL22-positive cells and Foxp3+ T regs in TILs on Spearman's rank correlation analysis (n = 6, p < 0.05; n = 17, p < 0.02, respectively) (Figs. 5b and 5d).
T reg migration induced by CCL17 or CCL22 chemokines
To evaluate the migration of T regs induced by CCL17 or CCL22, an in vitro migration assay was performed using CD4+CD25+ and CD4+CD25− cells purified from regional lymph nodes of patients with gastric cancer (n = 9). Since Foxp3 is the most reliable marker of T regs,6, 7 Foxp3+ cells should be purified as T regs for migration assays; however, Foxp3 staining was performed by intracellular staining with permeabilization, and it is impossible to subject sorted Foxp3+ cells to a migration assay. We confirmed that CD4+CD25+ cells separated by magnetic beads expressed 90.7% ± 2.3% (mean ± SD) Foxp3-positive cells on flow cytometric analysis in 10 independent experiments (data not shown). Thus, CD4+CD25+ cells selected by magnetic beads were used for the migration assay as T regs.
Representative data of the migration assay from 9 independent experiments are shown in Figure 6. The migration of T cells induced by CCL17 or CCL22 was significantly observed in CD4+CD25+ cells in a dose-dependent manner and, furthermore, migration was inhibited by neutralizing anti-human CCL17 or CCL22 mAb. On the other hand, the migration of CD4+CD25− cells was not significantly induced by CCL17 or CCL22.
Moreover, to clarify whether T regs migration induced by chemokines is specific for tumor-associated T regs, we performed migration assays using PBLs from gastric cancer patients (n = 3) and healthy donors (n = 3). As a result, there were not significant differences in the extent of the number of migrated cells between T regs derived from PBLs of gastric cancer patients and healthy donors (Fig. 7).
The present study shows important and novel findings relevant to the distribution of T regs in gastric cancer. First, the frequency of Foxp3+ T regs in TILs was significantly higher than that in normal tissues in gastric cancer. Of note, the accumulation of Foxp3+ T regs in TILs was seen in early gastric cancer to the same degree as that in advanced gastric cancer. Second, the frequency of Foxp3+ T regs in lymph nodes adjacent to the tumor was significantly higher than that in those marginally distant from the tumor or control lymph node. Third, the degree of CCL17- or CCL22-positive cells among CD14+ cells was significantly correlated with the frequency of Foxp3+ T regs in the tumor, and the in vitro migration assay revealed that CCL17 or CCL22 could induce the migration of T regs derived from gastric cancer.
Previously, T regs that have functionally suppressive properties for other effector T cells were characterized as a CD4+CD25high population among CD4+ T cells12; however, it was difficult to discriminate T regs clearly from conventional effector T cells that express CD25, and the gating of the CD4+CD25+ population might contaminate effector or memory T cells with no suppressive capacity.28–31 Recently, it has been reported that their suppressive property was due to the regulatory master gene, Foxp3, and since then, Foxp3 has been recognized as a reliable marker of T regs.28 Hence, in the present study, we analyzed the prevalence of Foxp3+ T regs, as more strictly defined T regs, in patients with gastric cancer.
Through the results, we confirmed that the frequency of T regs in TILs was significantly higher than that in normal tissues in patients with gastric cancer, in line with previous reports on several malignancies.17–21 These observations were also confirmed by immunohistochemical analysis in the present study.
Of note, we demonstrated that T regs are already present in small localized tumors at an early stage of gastric cancer, in line with a few reports indicating that the accumulation of T regs occurred in an early cancer stage or carcinoma in situ,32, 33 suggesting that they might be related to disease progression. Furthermore, it is likely that tumor-related factors mediate T regs trafficking to the tumor microenvironment in an early stage of cancer. In fact, Curiel et al. presumed that T regs were induced to migrate by CCL22, a macrophage-derived chemokine, toward the tumor microenvironment in ovarian cancer.25 Moreover, it has been shown that T regs strongly express CCR4, a chemokine receptor for CCL17 or CCL22, on their surface as compared to effector T cells in leukemia studies26, 34; however, in solid tumors, there are few reports evaluating CCL17 or CCL22 expression within tumors.25, 35 Therefore, we investigated whether the proportion of CCL17- or CCL22-positive cells within tumors was increased in comparison to normal tissues. As a result, the frequency of CCL17- or CCL22-positive cells among CD14-positive cells within tumors was significantly higher than in normal gastric mucosa. Moreover, a positive correlation was noted between the frequency of Foxp3+ T regs and the degree of CCL17- or CCL22-positive cell infiltration. These results strongly suggest that CCL17 or CCL22 within the tumor microenvironment is related to the accumulation of Foxp3+ T regs in gastric cancer. In addition, an in vitro migration assay in the present study further confirmed that T regs derived from gastric cancer had higher affinity for CCL17 or CCL22 than effector T cells, and CCL17 or CCL22 could induce the migration of T regs. Although the mechanisms leading to the upregulation of CCL17 or CCL22 within gastric cancer have not been clarified, it is possible that the tumor microenvironment induces macrophages to secrete CCL17 or CCL22. Furthermore, it has been reported that both CCL17 and CCL22 were also produced by dendritic cells (DCs).36, 37 These observations suggested that inflammatory reactions, including the production of CCL17 or CCL22, were induced by tumor-infiltrating macrophages and DCs, leading to the migration of T regs.
On the other hand, some reports indicated that activated effector T cells were converted into Foxp3-positive T cells, capable of suppressing autologous effector T cells.29–31 It is possible that tumor-related factors induce transient Foxp3+ T cells from CD4+CD25intFoxp3− effector T cells; however, their suppressive function is thought to be temporary, not intrinsic and unstable.29–31 Thus, it is likely that naturally occurring Foxp3+ T regs in peripheral sites faintly perceive tumor-related signals such as CCL17 or CCL22, migrate to the tumor, and create a favorable environment for tumor growth.
Since T regs have an inhibitory effect on the surrounding effector T cells, the elimination of T regs might be an effective therapeutic approach against cancer. With respect to T regs migration to tumors, our data revealed that treatment with anti-human CCR4 mAb might be an immunotherapeutic option for solid tumors. Indeed, Ishida and Ueda reported that CCR4 could be a novel target for immunotherapy against cancer, for example, CCR4+ T-cell leukemia/lymphoma.26
In conclusion, CCL17 and CCL22 chemokines within the microenvironment of gastric cancer are related to the high frequency of Foxp3+ T regs in TILs, with such an observation occurring in the early stage of gastric cancer.