CD4+CD25+ regulatory T cells in patients with gastrointestinal malignancies

Possible involvement of regulatory T cells in disease progression

Authors


Abstract

BACKGROUND

Active suppression by CD4+CD25+ regulatory T cells plays an important role in the down-regulation of the response of T cells to foreign and self antigens. Experimental tumor models in mice revealed that regulatory T cells inhibit antitumor immune responses. The purpose of the current study was to demonstrate the possible involvement of CD4+CD25+ regulatory T cells in immune system impairment in patients with gastrointestinal malignancies.

METHODS

The phenotypes of lymphocytes, particularly those of CD4+CD25+ T cells, were analyzed in peripheral blood in 149 patients with gastrointestinal malignancies and in ascites in 7 patients with peritoneal dissemination. In addition, cytokine production after in vitro stimulation was examined in CD4+CD25+ and CD4+CD25 T cells isolated from patients with malignant disease.

RESULTS

Compared with healthy volunteers, patients with gastrointestinal malignancies had a higher proportion of CD4+CD25+ T cells in peripheral blood, due to the presence of a drastically smaller number of CD4+CD25 T cells. Among patients with gastric carcinoma, those with higher percentages of CD4+CD25+ T cells had a poorer prognosis than did those with lower percentages. CD4+CD25+ T cells also were present in greater proportions in ascites from patients who had advanced-stage disease with peritoneal dissemination. Isolated CD4+CD25+ T cells from patients with malignant disease produced interleukin (IL)-4 and IL-10 but not IL-2 or interferon-γ; these cells also inhibited cytokine production by CD4+CD25 T cells after in vitro stimulation.

CONCLUSIONS

The relative increase in CD4+CD25+ regulatory T cells may be related to immunosuppression and tumor progression in patients with gastrointestinal malignancies. This finding suggests that the use of immunomodulatory therapy to treat patients with gastrointestinal malignancies may be an effective strategy. Cancer 2003;98:1089–99. © 2003 American Cancer Society.

DOI 10.1002/cncr.11618

There is considerable evidence of the critical role of the immune system in controlling the growth of malignant cells. Impairment of tumor immunity, which leads to immunologic toleration of malignant cells, is believed to contribute to the development and progression of malignancies; however, the precise mechanisms by which the immune system is modulated in patients with malignant disease remain poorly understood.

Regulatory T cells play an active role in the maintenance of the immune system's tolerance of both foreign and self antigens. Recently, great progress has been made in understanding the ontogeny, functioning, and mechanisms of action of regulatory T cells, and several subclasses of these cells have been identified.1, 2 Some subclasses of regulatory T cells produce immunosuppressive cytokines, such as interleukin (IL)-10 and transforming growth factor (TGF)-β, and exert their suppressive influence via the effects of these cytokines.3, 4 Other regulatory T cells suppress the activation and/or proliferation of other T cells without producing cytokines, via a mechanism that requires direct cell-cell contact.1, 2

In rodent models, a unique line of CD4+ T cells, which continuously expresses the α chain of the IL-2 receptor (CD25) and possesses a remarkable suppressive capacity both in vivo and in vitro, has been identified and characterized as a regulatory T cell line.1, 2 Depletion of CD4+CD25+ T cells in mice leads to spontaneous development of various organ-specific autoimmune diseases, which can be prevented by reconstitution using this same T cell subset. In addition, it has been reported that elimination or reduction of CD4+CD25+ regulatory T cells can induce effective tumor immunity in otherwise nonresponding mice by activating tumor-specific cytotoxic T lymphocytes (CTL) and nonspecific lymphokine-activated killer (LAK)/natural killer (NK) cells.5–9 The beneficial effect on tumor immunity associated with the removal of CD4+CD25+ regulatory T cells suggests that these cells play a critical role in immunologic toleration of tumor cells.

Recent studies have demonstrated that CD4+CD25+ T cells exhibiting regulatory/suppressive properties are naturally present in humans.10–16 Identification and characterization of human CD4+CD25+ regulatory T cells allow for the monitoring of these cells in patients with various diseases and have important implications regarding understanding and treating autoimmunity, graft rejection, and malignant disease in humans. Nonetheless, to our knowledge, there is scant information regarding the significance of this cell population in individuals with malignant disease. The results of the current study provide evidence that in peripheral blood and ascites, human CD4+CD25+ regulatory T cells are present in greater proportions in patients with gastrointestinal (GI) malignancies compared with healthy individuals. These findings suggest that CD4+CD25+ regulatory T cells play a critical role in immunosuppression and tumor progression in patients with malignant disease.

MATERIALS AND METHODS

Patients

The current study involved 149 patients with GI malignancies who were admitted to the Department of Surgery, Tazuke-Kofukai Kitano Hospital (Osaka, Japan), between December 2001 and June 2002; 10 healthy volunteers also were included in the study. Malignancies represented in the study population included gastric carcinoma (n = 55), colorectal carcinoma (n = 48), esophageal carcinoma (n = 17), liver carcinoma (n = 16), and pancreatic carcinoma (n = 13) (Table 1). Tumor stages were determined using the TNM classification system.17 Six patients with recurrent gastric carcinoma and one with recurrent pancreatic carcinoma developed ascites due to peritoneal dissemination. The 10 healthy volunteers included 5 men and 5 women (age range, 23–78 years). Informed consent was obtained from all healthy volunteers and all patients according to institutional review board–approved protocols. Heparinized blood samples were collected from all study participants. Ascites were obtained at paracentesis from 7 patients who had recurrent disease with peritoneal dissemination; samples from peritoneal lavage with 50 mL saline were obtained at the time of surgery from 5 patients with Stage I gastric carcinoma and were used as controls. Mononuclear cells were prepared from peripheral blood samples and ascites by centrifugation over Ficoll-Paque gradients (Pharmacia, Piscataway, NJ).

Table 1. Patient Characteristics
 HealthyDisease site
StomachColon/rectumPancreasEsophagusLiver
  • a

    Mean ± standard deviation.

Gender      
 Male5353391613
 Female52015413
Agea (yrs)40.1 ± 16.765.7 ± 10.265.4 ± 11.364.9 ± 13.265.4 ± 9.666.1 ± 7.0
Disease stage      
 I 196011
 II 210033
 III 49231
 IV 1331043
Recurrent disease 1720168
Total105548131716

Antibodies, Flow Cytometric Analysis, and Purification of CD4+CD25+ and CD4+CD25 T Cells

The following monoclonal antibodies (MoAbs) were used in the current study: fluorescein isothiocyanate (FITC)-conjugated anti-CD3 (SK7), FITC- or R-phycoerythrin (R-PE)-conjugated anti-CD4 (SK3), FITC-conjugated anti-CD8 (SK1), and R-PE-conjugated anti-CD25 (2A3) (Becton Dickinson, San Jose, CA); as well as FITC-conjugated anti-CD45RO (UCHL-1), FITC-conjugated anti-CD69 (FN50), and CyChrome-conjugated anti-CD4 (RPA-T4) (Becton Dickinson–Pharmingen, San Diego, CA). For flow cytometry, single-cell suspensions (concentration, 1 × 106 cells/ml) were stained in phosphate-buffered saline–2% fetal calf serum (FCS) at saturating concentrations according to standard procedures. Phenotypes and proportions of each subpopulation were analyzed with a FACSCalibur flow cytometer using CellQuest software (Becton Dickinson). CD4+CD25+ and CD4+CD25 T cells were isolated from peripheral blood mononuclear cells (PBMC) by sorting with the FACSCalibur system after staining with anti-CD4 and anti-CD25 MoAbs. The purity of the isolated CD4+CD25+ and CD4+CD25 T cells was greater than 90% in each of 3 separate experiments.

Measurement of Cytokine Levels

CD4+CD25+ and/or CD4+CD25 T cells in 96-well round-bottom plates (1 × 104 cells/well) were incubated at 37 °C in RPMI-1640 medium supplemented with 10% FCS (Life Technologies, Grand Island, NY) in the presence of 1 ng/mL phorbol 12-myristate-13-acetate (PMA) and 500 ng/mL A23187 (Sigma, St. Louis, MO). After a 48-hour incubation period, supernatants were collected and assayed for IL-2, IL-4, IL-10, and interferon (IFN)-γ production using an enzyme-linked immunosorbent assay (ELISA) kit (OptEIA; Becton Dickinson Pharmingen). Ascites from patients who had GI malignancies with peritoneal dissemination also were assayed for IL-2, IL-4, IL-10, and IFN-γ using the ELISA kit. Results were calculated as the mean values obtained from duplicate wells.

Statistical Analysis

Results are presented as mean values with associated standard deviations. Statistical analysis was performed using the Student t test. The Kaplan–Meier method was used to generate survival curves. The impact of the percentage of CD4+CD25+ T cells on survival was evaluated by comparing survival curves using the log-rank test. P values less than 0.05 were considered significant.

RESULTS

Recent studies have shown that CD4+CD25+ T cells exhibiting regulatory/suppressive properties are naturally present in humans.10–16 To determine whether CD4+CD25+ regulatory T cells, which have been reported to be closely associated with immunosuppression and tumor progression in mice,5–9 are affected in patients with GI malignancies, we compared the profiles of cell surface molecules on peripheral blood lymphocytes (PBL) in patients with malignancies and in healthy volunteers. Figure 1A shows representative profiles of CD4 and CD25 expression on PBL from a patient with malignant disease and from a healthy donor; the relative prevalence of the CD25+ subset in CD4+ T cells was enhanced in a patient with recurrent gastric carcinoma compared with a healthy donor. To further characterize CD4+CD25+ T cells in patients with malignancies, we sorted CD4+CD25 and CD4+CD25+ T cells from PBMC of patients with malignant disease and examined the expression of two other cell surface activation markers, CD45RO and CD69, in these sorted cells. Most CD4+CD25+ T cells (73%) exhibited a high level of CD45RO expression, whereas a smaller percentage of CD4+CD25 T cells (37%) were positive for CD45RO (Fig. 1B). In contrast, neither CD4+CD25+ nor CD4+CD25 T cells expressed CD69. These findings suggest that the CD4+CD25+ T cell population in patients with malignancies is composed of highly differentiated T cells that have not been recently activated but that are primed; these features reportedly are characteristic of human regulatory CD4+CD25+ T cells.10–16

Figure 1.

Characterization of CD4+CD25+ T cells in peripheral blood from a patient with malignant disease. (A) The expression of CD4 and CD25 in peripheral blood lymphocytes from a healthy donor and a patient with recurrent gastric carcinoma. Peripheral blood mononuclear cells (PBMC) from a healthy donor and from a patient with recurrent gastric carcinoma were double-stained with fluorescein isothiocyanate (FITC)-anti-CD4 and phycoerythrin-anti-CD25 monoclonal antibodies (MoAbs) and assayed using flow cytometry. The percentage of each type of cell is indicated by the number on the right-hand side of each box within a given graph. (B) Expression of activation markers CD45RO and CD69 in CD4+CD25 and CD4+CD25+ T cells from a patient with malignant disease. CD4+CD25 and CD4+CD25+ T cells in PBMC from a patient with malignant disease were sorted and stained with FITC-anti-CD45RO or FITC-anti-CD69 MoAbs. Percentages of cells within the relevant gates are indicated.

Relative Prevalence of the CD25+ Subset among CD4+ T Cells

Figure 2A shows the relative prevalence of the CD25+ subset in CD4+ T cells found in peripheral blood from healthy volunteers and from patients with GI malignancies (gastric, colorectal, pancreatic, esophageal, and liver). In healthy volunteers, CD25+ cells account for 26.5 ± 4.7% of all CD4+ T cells (range, 18.5–34.5%; n = 10), compared with 43.7 ± 12.7% (range, 24.3–86.8%; n = 55) in patients with gastric carcinoma, 39.7 ± 11.3% (range, 18.3–71.2%; n = 48) in patients with colorectal carcinoma, 44.0 ± 9.0% (range, 32.5–60.0%; n = 13) in patients with pancreatic carcinoma, 47.0 ± 9.8% (range, 32.0–65.3%; n = 17) in patients with esophageal carcinoma, and 42.1 ± 14.7% (range, 20.7–70.6%; n = 16) in patients with liver carcinoma. Compared with healthy volunteers, patients with GI malignancies of any type had significantly higher proportions of the CD25+ subset in their CD4+ T cells (P = 4 × 10−9, 9 × 10−7, 9 × 10−6, 1 × 10−7, and 9 × 10−4 for gastric carcinoma, colorectal carcinoma, pancreatic carcinoma, esophageal carcinoma, and liver carcinoma, respectively). To determine whether the increase in the relative prevalence of CD4+CD25+ T cells was related to disease stage, we examined the prevalence by disease stage of the CD25+ subset among CD4+ T cells in patients with gastric or colorectal carcinoma. The proportion of CD25+ T cells in patients with recurrent gastric carcinoma was significantly higher than in patients with primary gastric carcinoma (P = 0.0003 vs. Stages I–III; P = 0.0005 vs. Stage IV), whereas there were no significant differences in the prevalence of CD25+ T cells between patients with earlier (Stage I–III) and more advanced (Stage IV) stages of primary gastric carcinoma (Fig. 2B). Among patients with colorectal carcinoma, no significant differences by disease stage were observed in the prevalence of CD25+ T cells (Figure 2C). Next, we examined the correlation between the prevalence of the CD25+ subset among CD4+ T cells and disease prognosis in patients with gastric or colorectal carcinoma. Among patients with gastric carcinoma, those with higher percentages of CD25+ T cells had significantly poorer prognoses compared with those with lower percentages of CD25+ T cells (P = 0.007) (Fig. 3). In contrast, we did not observe a correlation between the prevalence of the CD25+ subset among CD4+ T cells and prognosis in patients with colorectal carcinoma (data not shown).

Figure 2.

Increased prevalence of the CD25+ subset among peripheral CD4+ T cells from patients with gastrointestinal malignancies. Percentages of CD25+ cells among CD4+ T cells in peripheral blood from (A) healthy donors (n = 10) and patients with gastrointestinal malignancies, including malignancies of the stomach (n = 55), colon/rectum (n = 48), pancreas (n = 13), esophagus (n = 17), and liver (n = 16); (B) healthy donors (n = 10) and patients with Stage I–III (n = 25), Stage IV (n = 13), or recurrent (n = 17) gastric carcinoma; and (C) healthy donors (n = 10) and patients with Stage I (n = 6), Stage II (n = 10), Stage III (n = 9), Stage IV (n = 3), or recurrent (n = 20) colorectal carcinoma. Percentages of CD25+ cells were determined using flow cytometry (see Fig. 1A). Open circles and bars represent values for individual patients and average values for patient groups, respectively. *: P < 0.01; **: P < 0.05.

Figure 3.

Survival of patients with gastric carcinoma, stratified by percentage of CD25+ cells among peripheral CD4+ T cells. Survival curves were generated using the Kaplan–Meier method for patients with gastric carcinoma whose percentages of CD25+ cells among CD4+ T cells were less than 40% (n = 23), between 40% and 50% (n = 17), or greater than 50% (n = 14). Statistical analysis was performed using the log-rank test (P = 0.007).

Absolute Counts of CD4+CD25+ T Cells in Patients with Gastrointestinal Malignancies

To further analyze the phenotypic changes that are characteristic of PBL from patients with malignant disease, we examined absolute counts of cell subpopulations in lymphocytes from patients with GI malignancies. Absolute counts of PBL were significantly decreased in patients with GI malignancies (except for those with Stage I–III gastric carcinoma) compared with healthy volunteers (Table 2). In particular, a reduction in the absolute number of CD3+ T cells was evident and significant in all types of GI malignancies. In contrast, absolute counts of other lymphocyte subsets, including B cells and NK cells, were not significantly changed (data not shown). These findings indicate that the phenotypic changes in PBL from patients with GI malignancies mainly depend on reduction of the number of CD3+ T cells. To explore the way in which each T cell subset was altered in patients with GI malignancies, we calculated absolute counts of each subset by determining the total number of lymphocytes and then multiplying by the percentage of each subset as found by flow cytometric analysis. The decrease in CD4+ T cells was more prominent and significant than the decrease in CD8+ T cells in all types of GI malignancies (Table 2). In particular, out of all the T cell subsets, the decrease in CD4+CD25 cell counts was the most evident one among patients with GI malignancies (in all stages). In contrast, absolute counts of CD4+CD25+ T cells were unchanged or slightly decreased among patients with advanced-stage disease. These findings suggest that in patients with GI malignancies, increases in the relative prevalence of the CD25+ subset among CD4+ T cells are not due to increases in the absolute numbers of CD4+CD25+ T cells, but rather to a selective reduction in the number of CD4+CD25 T cells.

Table 2. Absolute Counts of T Cell Subsets from Healthy Volunteers and Patients with Gastrointestinal Malignancies
Disease siteNo. of patientsMean cells/μL ± SD (P valuea)
PBLCD3+CD4+CD8+CD4+CD25CD4+CD25+
  • SD: standard deviation.

  • a

    Versus healthy volunteers.

None (healthy)102053 ± 4861229 ± 265862 ± 220443 ± 214590 ± 137209 ± 43
Stomach55      
 Stage I–III251725 ± 358 (0.07) 879 ± 334 (0.004)645 ± 254 (0.02)351 ± 136 (0.23)368 ± 143 (0.0005)247 ± 116 (0.16)
 Stage IV131371 ± 583 (0.006) 708 ± 486 (0.004)499 ± 284 (0.002)305 ± 194 (0.12)290 ± 200 (0.0004)166 ± 96 (0.17)
 Recurrent171154 ± 526 (0.0002) 332 ± 225 (8 × 10−8)248 ± 187 (1 × 10−6)102 ± 77 (0.0006)111 ± 92 (1 × 10−7)126 ± 97 (0.006)
Colon/rectum48      
 Stage I–III251639 ± 590 (0.04) 704 ± 402 (0.0001)495 ± 267 (0.0004)288 ± 176 (0.07)303 ± 198 (5 × 10−5)180 ± 88 (0.21)
 Stage IV or recurrent231395 ± 467 (0.002) 636 ± 295 (2 × 10−5)448 ± 210 (0.0001)272 ± 136 (0.04)268 ± 132 (1 × 10−5)176 ± 89 (0.21)
Pancreas13      
 Stage I–III21326 ± 156 (0.001) 697 ± 75 (0.001)595 ± 144 (0.02)347 ± 40 (0.22)210 ± 54 (0.002)192 ± 9 (0.30)
 Stage IV or recurrent11 962 ± 402 (9 × 10−5) 497 ± 388 (8 × 10−5)281 ± 270 (0.0003)170 ± 134 (0.004)226 ± 198 (0.0002)154 ± 108 (0.15)
Esophagus17      
 Stage I–III71445 ± 338 (0.007) 834 ± 292 (0.02)545 ± 264 (0.02)311 ± 140 (0.14)302 ± 104 (0.0002)297 ± 182 (0.25)
 Stage IV or recurrent101070 ± 566 (0.001) 341 ± 205 (2 × 10−7)203 ± 131 (8 × 10−8)251 ± 230 (0.07) 98 ± 64 (2 × 10−7) 87 ± 62 (0.0002)
Liver16      
 Stage I–III5 987 ± 368 (8 × 10−5) 557 ± 258 (0.002)377 ± 170 (7 × 10−5)252 ± 103 (0.04)229 ± 144 (0.002)135 ± 70 (0.08)
 Stage IV or recurrent11 951 ± 259 (3 × 10−5) 467 ± 231 (2 × 10−6)329 ± 209 (9 × 10−5)174 ± 103 (0.004)177 ± 114 (8 × 10−7)136 ± 74 (0.02)

CD4+CD25+ T Cells in Malignant Ascites from Patients with Gastrointestinal Malignancies

Some patients with gastric or pancreatic carcinoma who had relatively high percentages of CD4+CD25+ T cells developed ascites due to peritoneal dissemination (Table 3). We subsequently examined the phenotypes of the cells in ascites from patients who had gastric or pancreatic carcinoma with peritoneal dissemination. Figure 4A shows a representative profile of CD4 and CD25 expression in cells from malignant ascites. Significant percentages of the CD25+ subset among CD4+ T cells (30.3–75.9%) were detected in ascites from all patients who had malignant disease with peritoneal dissemination (Table 3), whereas CD4+ T cells were not present in peritoneal lavage samples from patients with early-stage gastric carcinoma who were tested as controls (Fig. 4A). Figure 4B shows the phenotypic alteration of cells in ascites, as determined by staining cells with anti-CD4, anti-CD8, and anti-CD25 MoAbs, from a patient who had recurrent gastric carcinoma with peritoneal dissemination and was treated with OK-432. Before treatment, CD4+CD25+ T cells were the dominant subpopulation among CD3+ T cells in ascites. However, after the presence of ascites was decreased dramatically by repeated intraperitoneal injection of 10 klinische einheit (clinical units) of OK-432 (3 injections), a substantial increase in the CD4+CD25 and CD8+ T cell subsets was observed, along with little change in the percentage of CD4+CD25+ T cells. Because the volume of ascites was decreased, the absolute number of CD4+CD25+ T cells was greatly reduced after treatment. This finding suggests that CD4+CD25+ T cells may be associated with the development of malignant ascites in patients with advanced-stage disease.

Table 3. CD4+CD25+ T Cell Subset Population and Cytokine Production in Malignant Ascites
Patient no.Age (yrs)GenderDisease sitePBLAscitesa
CD4+CD25+ (%)bCD4+CD25+ (%)bIL-2IFN-γIL-4IL-10
  • PBL: peripheral blood lymphocytes; IL: interleukin; IFN: interferon; F: female; M: male; ND: not detectable.

  • a

    Cytokine levels in pg/mL.

  • b

    Percentage of CD25+ cells among CD4+ T cells.

157FStomach86.875.910163518
269MStomach67.331.3NDNDND53
364MPancreas51.131.6NDNDND1197
476MStomach58.638.5ND13ND221
558MStomach41.130.3NDNDND76
657MStomach51.743.7NDNDND19
741FStomach42.131.4NDNDND594
Figure 4.

Increased prevalence of CD4+CD25+ T cells in ascites from patients who had gastrointestinal malignancies with peritoneal dissemination. The percentage of each type of cell is indicated by the number on the right-hand side of each box within a given graph. (A) Expression of CD4 and CD25 in cells found in ascites from a patient who had recurrent gastric carcinoma with peritoneal dissemination. Cells found in samples from peritoneal lavage with saline at the time of surgery in a patient with Stage I gastric carcinoma were used as controls. All cells were double-stained with fluorescein isothiocyanate (FITC)-anti-CD4 and phycoerythrin (PE)-anti-CD25 monoclonal antibodies (MoAbs) and assayed using flow cytometry. (B) Expression of CD4, CD8, and CD25; before and after treatment with repeated intraperitoneal injections of OK-432; in cells found in ascites from a patient who had recurrent gastric carcinoma with peritoneal dissemination. All cells were double-stained with FITC-anti-CD8 and PE-anti-CD4 MoAbs or with FITC-anti-CD4 and PE-anti-CD25 MoAbs and then assayed using flow cytometry.

In Vitro Modulation of Cytokine Production by CD4+CD25+ T Cells

To investigate the function of CD4+CD25+ T cells in patients with GI malignancies, we sorted CD4+CD25+ and CD4+CD25 T cells in PBMC from patients with malignant disease and cultured the sorted cells in vitro in the presence of 1 ng/mL PMA and 500 ng/mL A23187 for 48 hours; we then analyzed the cells for cytokine production using ELISA. Sorted CD4+CD25 T cells predominantly secreted IL-2 and IFN-γ, and there was little production of IL-4 and IL-10; these findings were reminiscent of a Th1-like phenotype after in vitro stimulation. In contrast, sorted CD4+CD25+ T cells produced greater quantities of Th2 cytokines, IL-4, and IL-10, and smaller quantities of Th1 cytokines, IL-2, and IFN-γ (Fig. 5). In addition, the production of IL-2 and IFN-γ by CD4+CD25 T cells after in vitro stimulation was inhibited significantly by coculturing with CD4+CD25+ T cells in a 1:1 ratio (Fig. 6). These results indicate that CD4+CD25+ T cells from patients with malignant disease can produce suppressive cytokines and inhibit the cytokine production of CD4+CD25 helper T cells; this inhibitory activity is characteristic of CD4+CD25+ regulatory T cells.10–16 The detection of high levels of IL-10 in ascites from patients with peritoneal dissemination (Table 3) was consistent with the cytokine profiles observed in CD4+CD25+ T cells after in vitro stimulation.

Figure 5.

Cytokine production after in vitro stimulation by CD4+CD25+ and CD4+CD25 T cells from patients with gastrointestinal malignancies. CD4+CD25+ and CD4+CD25 T cells (1 × 104 cells/well), which were isolated by sorting peripheral blood mononuclear cells, were stimulated in vitro by incubation with 1 ng/mL phorbol 12-myristate-13-acetate and 500 ng/mL A23187 for 48 hours at 37 °C. Subsequently, supernatants were collected and assayed for production of interleukin (IL)-2, interferon (IFN)-γ, IL-4, and IL-10. Means (circles) and standard deviations (error bars) from three separate experiments are shown. *: P < 0.01; **: P < 0.05.

Figure 6.

CD4+CD25+ T cells inhibit cytokine production by CD4+CD25 T cells in patients with gastrointestinal malignancies. Sorted CD4+CD25 or CD4+CD25+ T cells (5 × 103 cells/well), or a coculture of both cell types in a 1:1 ratio (total, 1 × 104 cells/well), were stimulated in vitro by incubating with 1 ng/mL phorbol 12-myristate-13-acetate and 500 ng/mL A23187 for 48 hours at 37 °C. Supernatants were collected and assayed for production of interleukin (IL)-2 and interferon (IFN)-γ. Results are representative of three independent experiments.

DISCUSSION

In recent years, much attention has been given to factors that regulate immune responses to malignant tumor cells. Ineffective immune responses to malignant cells contribute to the establishment and progressive growth of tumors. There are many proposed explanations for why tumor cells do not stimulate immune responses and/or are able to evade an immune attack in some cases, but two of these explanations are dominant: 1) the failure of the host immune system to recognize tumor antigens (ignorance) and 2) the failure of tumor-specific T cells to proliferate and function at the levels required for eradicating tumors (immunosuppression).

There is increasing evidence that most tumor-associated antigens are self antigens18; this evidence suggests that immunosuppression in tumor-bearing hosts may be associated with regulatory CD4+CD25+ T cells, which play a key role in the maintenance of immunologic self-tolerance via the inhibition of activated T cells. In animal models, reduction or elimination of CD4+CD25+ regulatory T cells can induce effective tumor immunity in otherwise nonresponding mice by bringing about the activation of tumor-specific CTL and nonspecific LAK/NK cells.5–9 The current study has revealed that the relative prevalence of CD4+CD25+ T cells with phenotypes that are characteristic of regulatory T cells is significantly increased in peripheral blood and ascites from patients with GI malignancies. Recently, Woo et al.19, 20 reported that tumor-infiltrating lymphocytes (TIL) from patients with early-stage nonsmall cell lung carcinoma and from patients with late-stage ovarian carcinoma contained increased proportions of CD4+CD25+ T cells, which secreted an immunosuppressive cytokine, TGF-β. In addition, Wolf et al.21 demonstrated that the proportion of CD4+CD25+ T cells was increased in malignant tumors originating in various organs. Therefore, the immune dysfunction observed in patients with malignant disease may be explained, at least in part, by the increase in the proportion of regulatory T cells in peripheral blood as well as in the local tumor environment.

It is noteworthy that the increase in the proportion of CD4+CD25+ regulatory T cells was correlated with clinical stage in patients with gastric carcinoma. Compared with patients with primary gastric carcinoma, patients with recurrent gastric carcinoma exhibited a significantly higher relative prevalence of the CD25+ T cell subset. In addition, although not shown, our preliminary observations indicate that the percentage of CD4+CD25+ regulatory T cells in peripheral blood increases gradually in parallel with tumor progression in patients with gastric carcinoma. Therefore, it appears likely that CD4+CD25+ regulatory T cells are directly involved in the mechanisms that allow tumor progression to occur. Indeed, among patients with gastric carcinoma, those with a higher prevalence of the CD25+ subset in CD4+ T cells exhibited significantly poorer prognosis compared with those with a lower prevalence of CD25+ T cells. These findings suggest that the relative prevalence of CD4+CD25+ regulatory T cells may be a determinant for predicting the prognosis of patients with gastric carcinoma. However, unlike in patients with gastric carcinoma, there were no significant differences by clinical stage in the prevalence of CD4+CD25+ T cells among patients with colorectal carcinoma. In addition, no significant correlation was observed between the prevalence of CD25+ T cells and prognosis in patients with colorectal carcinoma. Although further investigation is necessary, it is possible that the effects of progressing tumors on the host's immune system, especially on CD4+ T cells, differ in a way that depends on the type of malignancy.

The relative prevalence of the CD25+ subset among peripheral CD4+ T cells was significantly higher in patients with GI malignancies compared with healthy volunteers, whereas the absolute counts of CD4+CD25+ T cells were unchanged or decreased slightly among patients with advanced-stage disease. Because the immune system is regulated by a complex network composed of various types of cells, the relative abundance of CD4+CD25+ T cells in patients with malignant disease should have a major impact on those patients' immune responses to tumor cells.

The relative increase in the proportion of CD4+CD25+ T cells in patients with malignant disease was the result of a selective reduction in the number of CD4+CD25 T cells, possibly due to differences between CD4+CD25+ and CD4+CD25 T cells in their sensitivities to clonal deletion or apoptosis. It has been reported that in mice, regulatory CD4+CD25+ T cells are resistant to clonal deletion induced by viral superantigen in vivo22 and to Fas-dependent apoptosis in vitro.23 Furthermore, it is possible that in humans with malignant disease, some factors, such as tumor-derived antigens or molecules, can induce apoptosis selectively in the CD4+CD25 T cell subset but not in the CD4+CD25+ subset. For example, the phenotypes detected in patients with malignancies may be the result of repetitive stimulation of lymphocytes by tumor-derived antigens, which may cause activation-induced cell death selectively in the CD4+CD25 T cell subset. In fact, it has been demonstrated that for patients with melanoma, repetitive in vitro stimulation of TIL with peptide-pulsed or tumor lysate–pulsed autologous antigen-presenting cells (APC) or with autologous tumor cells leads to the outgrowth of regulatory CD4+ T cells that exhibit the characteristic phenotype of Th2 cells.24 It also has been postulated that the continuous presence of self antigens activates tissue-specific regulatory T cells via immature dendritic cells (DC),25 which have been reported to have an increased prevalence in the tumor site,26 as well as in peripheral blood,27 in patients with malignant disease. In the presence of progressing tumors, tumor-associated antigens could be constitutively presented by immature DC to regulatory T cells in tissue and to tumor-reactive CTL, thus tipping the balance in favor of toleration of tumor cells by activating the former and suppressing the latter. This hypothesis is supported by the recent findings that CD4+CD25+ regulatory T cells are present in high numbers in tumors isolated from humans19, 20 and mice.28 In addition, as has been shown in mouse models,29, 30 impairment of the homeostatic proliferation of CD4+CD25 T cells, which is mediated by CD4+CD25+ T cells, may accelerate the decline in the number of CD4+CD25 T cells in patients with malignant disease.

The role of immunoregulatory cytokines in immunosuppression mediated by CD4+ regulatory T cells remains an open question. While some claim that these cells do not produce cytokines,12, 15, 16 others have demonstrated that they can produce IL-10,11, 13, 14 IL-4,11 TGF-β,13 and a small amount of IFN-γ.13 Several studies have reported that the Tr1 and Th3 subsets of regulatory T cells suppress immune responses via production of immunosuppressive cytokines, such as IL-10 and/or TGF-β.3, 4 In contrast, it has been reported that CD4+CD25+ regulatory T cells suppress the activation and proliferation of other CD4+ and CD8+ T cells in an antigen-nonspecific manner, via a mechanism that requires cell-cell contact and is independent of the production of immunosuppressive cytokines.1, 2

In the current study, isolated CD4+CD25+ T cells from patients with malignancies produced Th2 cytokines, IL-4 and IL-10, after in vitro stimulation. In addition, IL-10 consistently was produced in ascites from patients with advanced-stage disease and peritoneal dissemination, who had increased CD4+CD25+ T cell populations in ascites. IL-10 is known to have antiinflammatory and suppressive effects on most hematopoietic cells, including T cells, NK cells, macrophages, and DC.31 IL-10 directly inhibits cytokine production by T cells and is involved in the induction of peripheral tolerance via effects on T cell–mediated responses. IL-10 also indirectly suppresses T cell responses by acting as a potent inhibitor of the antigen-presenting capacity of APC, including DC and macrophages. Therefore, in vivo, the mechanism involving the cytokines produced by CD4+CD25+ T cells may interact synergistically with the cell-cell contact–dependent mechanism via the immunosuppressive effects of the mechanisms on various types of immune cells. In agreement with our results, several reports have demonstrated that patients with malignant disease develop Th2-dominant status, in which the proportion of CD4+ cells producing Th2 cytokines, such as IL-4, IL-6, and IL-10, increases in peripheral blood32, 33; furthermore, a negative correlation between IL-10 production and prognosis has been described in patients with gastric or colorectal carcinoma.34

Because of the demonstrated inhibitory effects of CD4+CD25+ T cells on CD4+CD25 T cells, it seems clear that CD4+CD25+ T cells from patients with malignant disease have regulatory/suppressive functions. However, the CD4+CD25+ T cell subset that we defined as being regulatory may be a heterogeneous population, because CD25 is an imperfect marker that is expressed in every activated T cell. Recently, Foxp3 has been identified as an important transcription factor that is expressed predominantly in CD4+CD25+ T cells and converts naïve T cells into CD4+CD25+ regulatory T cells.35 A more detailed evaluation of molecules, such as Foxp3, that are specific for regulatory T cells in patients with malignant disease may clarify the role of regulatory T cells in tumor immunity and make it possible to turn regulatory T cells on and off at will for the purposes of managing tumor immunity.

Recently, vaccination therapy using specific tumor-associated antigens has been tested extensively in patients with highly advanced disease. Clinical effectiveness has been limited, although most patients with malignant disease possess precursors of CTL that are specific for the antigens used in vaccination. Because most tumor-associated antigens are self antigens,18 increases in the relative prevalence of CD4+CD25+ regulatory T cells may be responsible for the poor clinical response of these patients to repeated vaccination therapy.

The findings of the current study may help make current immunotherapies more effective or enable the development of a novel immunotherapy for malignant disease. For example, as has been demonstrated in animal models,5–9 depletion or inhibition of CD4+CD25+ T cells by administration of anti-CD25 or anti-CTLA-4 to tumor-bearing hosts for a limited period may lead to enhanced tumor immunity, although the development of autoimmune diseases also is possible. Furthermore, removal of CD4+CD25+ T cells from PBL or TIL before in vitro stimulation with IL-2 may lead to the production of more potent cytotoxic cells, including CTL and LAK/NK cells. Such elimination of CD4+CD25+ regulatory T cells could be used as a novel and effective strategy for enhancing tumor immunity if combined with current attempts to augment the immunogenicity of tumor cells using, for instance, cytokine gene transduction in tumor cells or vaccination of patients with tumor antigens and/or peptides or with antigen-pulsed DC. In fact, a recent report demonstrated that adoptive transfer of autologous, tumor-reactive T cells after nonmyeloablative lymphodepleting conditioning, which eliminates regulatory T cells in the host, yields good clinical responses and results regarding the regression of metastatic tumors in patients with melanoma.36

Ancillary