CD8⁺CD28⁻ T Regulatory Lymphocytes Inhibiting T Cell Proliferative and Cytotoxic Functions Infiltrate Human Cancers

Patients

Biological samples were collected from a series of 42 cancer patients affected with colorectal (15), gastric (4), lung (3), pancreatic (2), kidney (2), ovarian (2), head-neck (2), thyroid (1), breast (1), esophagus (1), prostate (1), neuroendocrine (1) cancers, Hodgkin lymphomas (3), non-Hodgkin lymphoma (1), melanoma (1), seminoma (1), or sarcoma (1). Fresh surgical specimens from primitive tumor lesions and/or satellite lymph nodes, as well as peripheral blood samples, were obtained from single patients with respect for their clinical condition and diagnostic/therapeutic needs. To avoid the risk of impairing the diagnostic value of surgically excised material, only small (≈5-mm diameter size) tissue specimens were provided to us for immunological examinations. Tissue slices were collected from each bioptic sample to undergo pathologic examination. Clinical features (stage, survival after surgery) and available biologic material from each patient are specified in Table I⇓. The study was approved by the local ethical committee and all patients included in the study provided their informed written consent.

Purification of CD8⁺CD28⁺, CD8⁺CD28⁻, and CD4⁺CD25⁺ T lymphocytes

Lymphocytes from surgical specimens were purified filtering minced tissues using a sterile cell strainer (Falcon; BD Biosciences) and running the collected cells on Ficoll gradient. PBMC were purified by centrifugation on Ficoll gradient. CD8⁺CD28⁺, CD8⁺CD28⁻, and CD4⁺CD25⁺ T cell subsets were isolated from the different cell preparations by sequential cycles of cell sorting on magnetic beads (Miltenyi Biotec) following the manufacturer’s instructions. MultiSort CD8 MicroBeads and microbeads conjugated with the anti-CD28 9.3 mAb (8) were used for purification of CD8⁺CD28⁺ (by positive selection) and CD8⁺CD28⁻ (by negative selection) T lymphocytes; the CD4⁺CD25⁺ Regulatory T cell Isolation kit was used for purification of CD4⁺CD25⁺ T cells. The purity of sorted cells resulting was ≥95% as demonstrated by flow cytometric analysis.

Abs used for cytofluorimetric analyses

The mAbs used for cytofluorimetric analyses were the following: allophycocyanin-conjugated anti-CD3 and anti-CD25, PerCP-cyanin 5-conjugated anti-CD8 and anti-CD4, FITC-conjugated anti-CD8 and anti-CD4, PE- or FITC-conjugated anti-CD28, PE-conjugated anti-CD25, allophycocyanin-cyanin 7-conjugated anti-CD3 (all obtained from BD Biosciences), FITC-conjugated anti-CCR4 and unconjugated anti-CCR2 (R&D Systems), PE-conjugated anti-IL-10 (R&D Systems), FITC-conjugated anti-Foxp3 mAb (eBioscience). Polyclonal goat anti-mouse Ig FITC-conjugated F(ab)₂ (DakoCytomation) was used as secondary Ab in indirect immunofluorescence analyses. Fluorochrome-conjugated isotype-matched Abs were used as controls (BD Biosciences).

Immunofluorescence analyses of cell suspensions

Cell expression of membrane Ags was analyzed by immunofluorescence incubating tested lymphocytes (1 × 10⁵ cells in 100 μl of PBS) with specific mAbs at 4°C for 30 min in the dark. The intracellular expression of IL-10 by CD8⁺CD28⁻ Treg was analyzed as follows. The cells (resuspended in culture medium at the concentration of 1 × 10⁶/ml) were stimulated with PMA (50 ng/ml) and ionomycin (2 μg/ml) for 36 h at 37°C. Brefeldin (10 μg/ml) was added to the cells 18 h before the end of incubation. After washings, the cells were stained with a FITC-conjugated anti-CD28 mAb before fixing and permeabilizing the lymphocytes with the Cytofix/Cytoperm kit (BD Biosciences) following the manufacturer’s instructions. The cells were then incubated with the PE-conjugated anti-IL-10 23738 mAb for 15 min at room temperature; isotype-matched Abs were used as controls. To analyze Foxp3 expression by CD4⁺CD25⁺ T lymphocytes, the cells were initially stained with the PerCP-cyanin 5-conjugated anti-CD4 and the allophycocyanin-conjugated anti-CD25 and allophycocyanin-cyanine 7-conjugated anti-CD3 mAbs. Then, they were fixed and permeabilized using the FoxP3 Staining Buffer Set (eBioscience) following the manufacturer’s instructions, and incubated with the FITC-conjugated anti-Foxp3 mAb (eBioscience) for 30 min at 4°C in the dark; isotype-matched Abs were used as controls. After staining procedures, the cells were analyzed by a FACSCanto flow cytometer (BD Biosciences).

Immunofluorescence analyses of tissue sections

Slices (5-μm thick) from frozen tissue specimens were rehydrated in PBS and incubated with normal serum (5% v/v) from the host species of the labeled secondary Abs. After washings with PBS, slices were incubated at room temperature for 45 min with anti-human mAbs specific to the following Ags: CD4, CD8, CD28 (BD Biosciences), CD25 (Biomeda), IL-10 (R&D Systems). After washings, specimens were incubated with FITC-conjugated secondary Ab (monovalent Fab; Jackson ImmunoResearch Laboratories) at room temperature for 30 min. In the case of double labeling immunofluorescence, two different secondary Abs (monovalent Fab; Jackson ImmunoResearch Laboratories) conjugated with different probes (FITC or PE, respectively) were used. Cover was slipped using glycerol-PBS and sections were examined by an epifluorescent microscope (DMLB; Leica).

In vitro generation of CD8⁺CD28⁻ Treg from the peripheral blood of healthy subjects

Purified CD8⁺CD28⁻ T lymphocytes (1 × 10⁵ cells/well), resuspended in culture medium, constituted by RPMI 1640 culture medium (Invitrogen Life Technologies) added with 10% FCS (Invitrogen Life Technologies), 2% glutamine (Invitrogen Life Technologies), and 100 U/ml penicillin-0.1 mg/ml streptomycin (Sigma-Aldrich), were incubated either with 20 U/ml IL-2 (Chiron) and 40 ng/ml IL-10 (R&D Systems), or with 20 U/ml IL-2 and 10 ng/ml GM-CSF (Roche) in the presence of autologous monocytes (2.5 × 10⁴ cells/well), in 96-well flat-bottom plates (Corning Costar) at 37°C for 7 days. In some experiments, IL-10 or GM-CSF were substituted by 100 μl of supernatant from BJAB (9) or MEL-P (a melanoma cell line) (10, 11) cells preincubated (for 1 h at 37°C) or not with neutralizing anti-IL-10 23738 or anti-human GM-CSF mAb (10 μg/ml; R&D Systems), respectively. In other experiments, CD8⁺CD28⁻ T lymphocytes were incubated in the presence of exogenous IL-2 (20 U/ml) and MEL-P supernatant collected from cell cultures performed following MEL-P cell preincubation with the oligonucleotide 5′-CACTCTCACCACATC-3′ (antisense of the GM-CSF genetic sequence; GENSET). To favor oligonucleotide penetration into cells, DOTAP Liposomal Transfection Reagent (Boehringer Mannheim) was used according with the manufacturer’s instructions.

Proliferation suppression assay

PBMC pulsed with the anti-CD3 UCTH-1 mAb (5 μg/ml) were cultured for 5 days in a 96-well flat-bottom plate (1 × 10⁵ cells/well) in the presence (or not) of CD8⁺CD28⁺, CD8⁺CD28⁻, or CD4⁺CD25⁺ T lymphocytes (1 × 10⁵ cells/well). The experiments with CD8⁺CD28⁻ T cells were performed in transwell plates (Corning Costar) seeding anti-CD3- stimulated PBMC and regulatory lymphocytes at the opposite sides of the pored membrane to avoid cell-to-cell direct contact. Twelve hours before harvesting, 0,5 μCi [³H]thymidine were added to each well. The radioactivity of cells from individual wells was measured by a beta counter (Hidex Oy). These experiments were also replicated in the presence of the neutralizing anti-IL-10 23738 mAb (10 μg/ml) or of an unrelated, isotype-matched mAb.

Analysis of cytotoxic activity of tumor-infiltrating CD8⁺CD28⁻ T lymphocytes

A proliferation suppression assay was performed culturing in transwell plates anti-CD3-stimulated PBMC in the presence or not of tumor-infiltrating CD8⁺CD28⁻ T lymphocytes. After 5 days of culture, the anti-CD3-stimulated PBMC were collected and the percentage of dead cells was measured by immunofluorescence analysis. To identify apoptotic cells, the stimulated PBMC were stained with allophycocyanin conjugated anti-CD3 mAb and FITC annexin V (Bender MedSystem) for 30 min at 4°C. Then, 2.5 μg of propidium iodide were added to samples to identify necrotic cells. After staining procedures, cells were analyzed by a FACSCanto flow cytometer (BD Biosciences).

Generation of a p540-specific CTL line

PBMC from one HLA-A2⁺ cancer patient (1.5 × 10⁶ cells/ml) were incubated in 24-well plates in culture medium containing human AB serum and 10 μg/ml p540 peptide. After 2 days, 12 U/ml IL-2 were added to each well. Five days later, cells were collected, resuspended at 5 × 10⁵ cells/ml, and seeded in the presence of autologous, irradiated (5000 rad) PBMC pulsed with the p540 peptide at a responder:stimulator cell ratio of 1:5; on the day after, 12 U/ml IL-2 and 30 U/ml IL-7 were added to the wells. After 1–2 wk, cells were collected, resuspended at 5 × 10⁵ cells/ml in medium containing 30 U/ml IL-7 and plated in the presence of autologous, irradiated APCs prepared as follows. Autologous, irradiated PBMC were resuspended at 4 × 10⁶ cells/ml in medium containing both 5 μg/ml β₂-microglobulin and 10 μg/ml p540 peptide, and distributed in 24-well plates at 1 ml/well at 37°C. After 2 h, nonadherent cells were removed and adherent cells were used as APCs. The latter cycle of stimulation was repeated weekly for five times before analysis of CTL cytotoxicity.

Cytotoxicity suppression assay

Pelleted TAP-deficient T2 cells (12) were labeled using 150 μl (1 μCi/μl) of Na₂⁵¹CrO₄ for 90 min at 37°C. During the labeling, T2 cells were also incubated (or not) with 10 μg of p540 peptide. After washings, 5 × 10³/well target cells were incubated with a p540 telomerase peptide-specific CTL line at a CTL:target cell ratio of 40:1 in the presence or not of CD8⁺CD28⁻ T lymphocytes. The experiments were performed in flat-bottom 96-well transwell plates, thus separating the suppressor lymphocytes from the cytotoxic and target cells through the pored membrane. The experiments were also performed in the presence of the neutralizing anti-IL-10 23738 mAb (10 μg/ml) or of an unrelated, isotype-matched mAb. Target cells incubated with medium alone or with Triton X-100 diluted 1/100 were used to calculate spontaneous and maximum ⁵¹Cr release, respectively. After 6 h of incubation, supernatants were collected and the radioactivity was detected by gamma counter (Wallac). The percentage of cytotoxicity was calculated as follows: percentage of lysis = sample cpm − spontaneous release cpm/maximum release cpm − spontaneous release cpm × 100.

Cytokine and chemokine measurements

CD8⁺CD28⁻ T lymphocytes (1 × 10⁵ cells/well) purified from tumor specimens were cultured in 200 μl of culture medium for 5 days. Supernatants were, then, collected and the concentrations of IL-10, IL-4, TGFβ, TNF-α, and IFN-γ were determined by commercially available ELISA kits (Human EliPairs IFN-γ, Human EliPairs IL–4, Human EliPairs TNF-α, Human EliPairs Il-10, Human TGFβ ELISA kit; Diaclone).

The concentrations of IL-10, GM-CSF, and CCL22 in culture supernatants from BJAB, MEL-P, and LK (a prostate cancer primary cell line) (13) cells were measured by the CBA Human Soluble Protein Flex Set System (BD Biosciences) as follows. A Flex Set capture bead is a single bead population with a distinct fluorescence intensity and is coated with a capture Ab specific for a soluble protein. The bead population is resolvable in the FL3 and FL4 channels of a flow cytometer. Each bead population is given an α-numeric position designation indicating its position relative to other beads in the CBA Flex Set system (BD Biosciences). In a CBA Flex Set assay (BD Biosciences), the capture bead, the PE-conjugated detection reagent, and the standards or samples are incubated together to form sandwich complexes. On this basis, our analyses were performed as follows. Supernatants (50 μl) or CBA Human Soluble Protein Flex Set Standard dilutions (50 μl; BD Biosciences) were incubated with 50 μl of mixed capture beads and 50 μl of the appropriate PE-conjugated anti-cytokine Ab detector at room temperature for 3 h in the dark. After incubation, samples were washed once (5 min at 200 g) with washing buffer to remove the excess of detector Abs. The samples were then resuspended in 300 μl of washing buffer. Immediately afterward, data acquisition was performed by a FACSCanto flow cytometer (BD Biosciences), using a FACSDIVA software program following the manufacturer’s instructions. The sample results were generated in a graphical and tabular format using the FCAP Array Software.

Statistical analyses

Statistically significant correlation between variables was searched by Spearman test for nonparametric data. Statistically significant differences between mean percent concentrations of different T cell subsets, as well as between mean percent suppressor activities of CD8⁺CD28⁻ T cells from patients with <12 and >12 mo survival were analyzed by Mann-Whitney U test for nonparametric values. Calculations were performed using the GraphPad Prism version 4.00 software.

CD8⁺CD28⁻ Treg are present in TILs

TILs were purified from tumor specimens of primitive cancer lesions in a series of 23 patients affected with colorectal (13), gastric (3), ovarian (2), kidney (1), esophagus (1), thyroid (1), breast (1) cancers or non-Hodgkin lymphoma (1). Phenotypic and/or functional analyses were performed in relation to the limited number of recovered cells. The mean percentage of CD8⁺ T cells (41 ± 11%) was significantly higher than that of CD4⁺ T cells (33.6 ± 16%) (p < 0.03) (Fig. 1⇓A). CD8⁺CD28⁻ and CD4⁺CD25⁺ T lymphocytes were detected in all tumors with mean percentages of 35 ± 12% and 3 ± 2%, respectively (p < 0.001). The mean percentage of CD8⁺CD28⁻ T cells was also significantly higher than that of CD8⁺CD28⁺ T cells (6% ± 5) (p < 0.001). Superimposable results were obtained considering absolute numbers (instead of percentages) of TILs purified from tissue specimens of comparable size (data not shown). The prevalence of CD8⁺CD28⁻ on CD4⁺CD25⁺ T cells in primitive lesions of human cancer was also evident at immunofluorescence analysis of tissue sections.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Phenotypic and functional analyses of TILs. A, Percent concentrations of CD8⁺, CD8⁺CD28⁺, CD8⁺CD28⁻, CD4⁺, CD4⁺CD25⁺ T cell subsets in the total TIL population from patients 3, 6, 7, 10–14, 16, 18–22, 25–32. B, Suppressor activity, tested in a T cell proliferation assay, exerted by tumor-infiltrating CD8⁺CD28⁻ T cells from patients 3, 6, 7, 10–14, 16, 18–22, 25–32, 42 or CD8⁺CD28⁺ T cells from patients 3, 7, 10, 11, 12, 14, 16, 18, 20, 22, 25, 26, 28–30, 32, 42. The experiment was performed incubating tumor-infiltrating CD8⁺CD28⁻ or CD8⁺CD28⁺ T cells purified from single patients with autologous, anti-CD3 mAb-stimulated PBMC. Allogeneic, anti-CD3 mAb stimulated PBMC were used as proliferating cells only for the case of patient 32 due to the unavailability of autologous PBMC. Data are expressed as percentage of inhibition of anti-CD3 mAb-induced T cell proliferation measured by cpm originating from [³H]thymidine incorporation. Analytical data (cpm ± SD) are presented in Table II⇓. Analysis of the existence of a statistically significance difference between cpm of proliferation assays performed in the presence of tumor-infiltrating CD8⁺CD28⁻ T cells and cpm of assays performed in the presence of both CD8⁺CD28⁻ T cells and an anti-IL-10 mAb was performed by Wilcoxon t test for matched, nonparametric data. C, Cytotoxic activity of a p540 telomerase peptide-specific CTL line against T2 cells pulsed with the p540 peptide in the presence or not of CD8⁺CD28⁻ T cells from primitive tumor masses of patients 29 and 30. The cultures containing the CD8⁺CD28⁻ T cells were performed in transwell plates to physically separate these lymphocytes from cytotoxic and target cells. a, CTL plus nonpulsed T2 cells. b, CTL plus p540 peptide-pulsed T2 cells. c, As b plus CD8⁺CD28⁻ T lymphocytes. d, As c plus an anti-IL-10 mAb (10 μg/ml). e, As c plus an unrelated isotype-matched mAb (10 μg/ml). D, Foxp3 expression by tumor-infiltrating CD4⁺CD25⁺ T cells from patient 21. a, Flow cytometric characterization of CD4⁺CD25^+bright and CD4⁺CD25^+dim T cell subpopulations within total CD4⁺CD25⁺ tumor-infiltrating T cells. b, Intracellular staining of CD4⁺CD25^+bright T cells by an isotypic control of the anti-Foxp3 mAb. c, Intracellular staining of CD4⁺CD25^+bright T cells by the anti-Foxp3 mAb. Foxp3 expression in CD4⁺CD25^+dim T cell subset was 14% (data not shown).

To analyze their function, tumor-infiltrating CD8⁺CD28⁻ T cells were sorted from TILs and tested in suppression assays. Fig. 1⇑B shows that tumor-infiltrating CD8⁺CD28⁻ T cells (but not CD8⁺CD28⁺ T cells) exerted IL-10-dependent suppressor function on T cell proliferation (see Table II⇓ for analytical data). This effect was not due to cytotoxic activity because cell death rates (including both necrotic and apoptotic cells) in anti-CD3-stimulated PBMC were 15–20% and 30–50% when cultures were performed in the presence or not of tumor-infiltrating CD8⁺CD28⁻ T cells, respectively (data not shown). Similarly, they inhibited, in an IL-10-dependent fashion, the cytotoxic activity of a p540 telomerase peptide-specific CTL line (Fig. 1⇑C shows the results of representative experiments performed with cells from patients 29 and 30).

From specimens of patients 18–21 and 25, enough cells were recovered to perform a suppression assay with purified CD4⁺CD25⁺ T cells: the suppressor activities were 65, 26, 43, 54, 56%, respectively, thus confirming that tumor-infiltrating CD4⁺CD25⁺ T cells in our series also belonged to the regulatory compartment of immune system (Table III⇓). Accordingly, a relevant proportion of these cells expressed Foxp3 (Fig. 1⇑D).

To characterize their secretion pattern, CD8⁺CD28⁻ T lymphocytes from the primitive tumor mass of patients 21–23 and 26–33 were cultured for 5 days in the absence of stimulation and the concentration of IL-10, IL-4, TGFβ, IFN-γ, and TNF-α were measured in the supernatants (Table IV⇓). Interestingly, a detectable amount of both effector (IFN-γ, TNF-α) and potentially regulatory (IL-10, IL-4, TGFβ) cytokines were detected in all cases likely suggesting a mixed composition of the tumor-infiltrating CD8⁺CD28⁻ T cell subset.

CD8⁺CD28⁻ Treg are also present in metastatic satellite lymph nodes

Phenotype and function of CD8⁺CD28⁻ and CD4⁺CD25⁺ T lymphocytes were also studied in lymphocytes purified from a series of metastatic or metastasis-free satellite lymph nodes. Phenotypic analyses on cells from metastatic lymph nodes of patients 33–41 showed no significantly higher mean percentages of CD4⁺ (47 ± 21%) than of CD8⁺ (30 ± 25%) T cells, as well as comparable mean concentrations of CD4⁺CD25⁺ (3 ± 1%) and CD8⁺CD28⁻ (6 ± 5%) T lymphocytes (Fig. 2⇓A). Functional analyses performed on CD8⁺CD28⁻ T cells sorted from metastatic lymph nodes of patients 1–12, 15–19, 21, 22, 24–26, 32, and 41 showed IL-10-dependent suppressor activity in 22 of 24 (91%) cases. On the contrary, CD8⁺CD28⁻ T cells purified from metastasis-free satellite lymph nodes of patients 9, 11, 13, 21, 23, and 26 did not exert relevant suppressor activity (Fig. 2⇓B and Table II⇑). Similar findings were observed when the function of CD4⁺CD25⁺ T cells, purified from metastatic lymph nodes of patients 11, 17, 18, 21, 25, 26 or from metastasis-free lymph nodes of patients 9, 11, 13, 21, 23, and 26 was analyzed (Fig. 2⇓C and Table III⇑). Accordingly, CD4⁺CD25⁺ T cells from metastatic but not from metastasis-free lymph nodes showed relevant Foxp3 expression (Fig. 2⇓E). In the cases of patients 33–40, the limited number of available cells obliged us to perform the suppressor assay directly with the total lymphocyte population (without any round of sorting and selection). According to previous results, cells from metastatic, but not from metastasis-free lymph nodes, exerted a remarkable IL-10-dependent suppressor activity (Fig. 2⇓D).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Phenotypic and functional analyses of lymphocytes purified from metastatic or metastasis-free lymph nodes. A, Percent concentrations of CD8⁺, CD8⁺CD28⁺, CD8⁺CD28⁻, CD4⁺, CD4⁺CD25⁺ T cell subsets in metastatic lymph nodes of patients 33–40. B, Suppressor activity, tested in a T cell proliferation assay, exerted by CD8⁺CD28⁻ T cells purified from metastatic lymph nodes of patients 1–12, 15–19, 21, 22, 24–26, 32 or from metastasis-free lymph nodes of patients 9, 11, 13, 21, 23, 26. Data are expressed as percent inhibition of anti-CD3 mAb-induced T cell proliferation measured by cpm originating from [³H]thymidine incorporation. Analytical data (cpm ± SD) are presented in Table II⇑. Analysis of the existence of a statistically significance difference between cpm of proliferation assays performed in the presence of CD8⁺CD28⁻ T cells and cpm of assays performed in the presence of both CD8⁺CD28⁻ T cells and an anti-IL-10 mAb was performed by Wilcoxon t test for matched, nonparametric data. C, Suppressor activity, tested in a T cell proliferation assay, exerted by CD4⁺CD25⁺ T cells purified from metastatic lymph nodes of patients 11, 17, 18, 21, 25, 26 or from metastasis-free lymph nodes of patients 9, 11, 13, 21, 23, 26. Analytical data (cpm ± SD) are presented in Table III⇑. D, Suppressor activity, tested in a T cell proliferation assay, exerted by total lymphocytes purified from metastatic lymph nodes of patients 33–41. Analytical data (cpm ± SD) are presented in Table III⇑. Analysis of the existence of a statistically significance difference between cpm of proliferation assays performed in the presence of total lymphocytes from metastatic lymph nodes, and cpm of assays performed in the presence of both total lymphocytes from metastatic lymph nodes and an anti-IL-10 mAb, was performed by Wilcoxon t test for matched, nonparametric data. E, Foxp3 expression by CD4⁺CD25⁺ T lymphocytes from a metastatic (upper row) and a metastasis-free (lower row) lymph node from patient 21 (representative of 6 tested patients). The analyses performed on CD4⁺CD25^+bright (a–d, gray histograms) T cells are shown: b–e refer to cell staining with isotypic controls, while c and f refer to the anti-Foxp3 mAb cell staining. The analyses performed on CD4⁺CD25^+dim T cells detected Foxp3 expression in 8% of cells from the metastatic lymph node and in 1% of cells from the metastasis-free lymph node (data not shown).

CD8⁺CD28⁻ Treg in the peripheral blood of cancer patients

To verify whether CD8⁺CD28⁻ Treg were uniquely localized at the sites of tumor development or recirculated in the periphery, CD8⁺CD28⁻ T cells were purified from the peripheral blood of patients 1–32 and 42. Relevant suppressor activity was detected in 30 of 33 cases (90%) (Fig. 3⇓ and Table V⇓). This is remarkable because in preliminary experiments performed on >50 healthy subjects, we never observed suppression activity analyzing circulating CD8⁺CD28⁻ T cells. Functional analysis of CD4⁺CD25⁺ T cells purified from the peripheral blood of cancer patients showed relevant suppressor activity in 8 of 11 cases (72%) (Fig. 3⇓ and Table V⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Suppressor activity exerted by CD8⁺CD28⁻ (▵) or by CD4⁺CD25⁺ (□) T lymphocytes purified from the peripheral blood of cancer patients. CD8⁺CD28⁻ T cells were extracted from blood of patients 1–32, 42; CD4⁺CD25⁺ T lymphocytes were from patients 11, 17, 18–21, 23, 25–28. Data are expressed as percentage of inhibition of anti-CD3 mAb-induced T cell proliferation measured by cpm originating from [³H]thymidine incorporation. Analytical data (cpm ± SD) are presented in Table V⇓. Analysis of the existence of a statistically significance difference between cpm of proliferation assays performed in the presence of peripheral blood CD8⁺CD28⁻ T cells and cpm of assays performed in the presence of both peripheral blood CD8⁺CD28⁻ T cells and an anti-IL-10 mAb was performed by Wilcoxon t test for matched, nonparametric data.

Tumor cells may induce both generation and recruitment of CD8⁺CD28⁻ Treg

The above findings suggest a direct relationship between the presence of tumor cells and tissue infiltration by Treg. To verify whether Treg might be generated at the tumor site, a panel of tumor cell lines was screened for the capacity to secrete IL-10 and/or GM-CSF (the cytokines involved in the generation of CD8⁺CD28⁻ Treg) (14, 15, 16). BJAB, LK, and MEL-P cells were selected because of their secretion of IL-10 (BJAB cells) or GM-CSF (LK and MEL-P) (Fig. 4⇓A). The supernatant from BJAB cells, incubated (in the presence of exogenous IL-2) with CD8⁺CD28⁻ T cells from the peripheral blood of a healthy donor, induced generation of CD8⁺CD28⁻ Treg (Fig. 4⇓B, upper panel). The same occurred with the supernatant from LK or MEL-P cells (in the presence of exogenous IL-2 and autologous monocytes) (Fig. 4⇓B, medium and lower panels, respectively). The generation of CD8⁺CD28⁻ Treg was strictly dependent on cytokine secretion by tumor cells because it was inhibited by an anti-IL-10 mAb (supernatant from BJAB cells), or by an anti-GM-CSF mAb (supernatants from LK and MEL-P cells). Accordingly, when MEL-P cells were incubated with oligonucleotides antisense of the GM-CSF gene sequence, their supernatant did not contain a detectable amount of GM-CSF (data not shown) and lost the capacity to induce generation of CD8⁺CD28⁻ Treg (Fig. 4⇓C). In all these sets of experiments, when CD8⁺CD28⁺ T cells were incubated under culture conditions able to induce generation of Treg from CD8⁺CD28⁻ T cell precursors (e.g., in the presence of either IL-2 and IL-10, or IL-2 and GM-CSF, or tumor cell supernatants), they never acquired suppressor activity, as demonstrated by their incapability to inhibit proliferative T cell responses (data not shown). Suppressor activity of CD8⁺CD28⁻ Treg generated by incubation with tumor cell supernatant was IL-10 dependent as demonstrated by counteraction of their function by an anti-IL-10 mAb (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Analysis of the capacity of tumor-secreted IL-10 or GM-CSF to induce generation of CD8⁺CD28⁻ Treg. A, Detection of IL-10 (□) or GM-CSF (▪) in the culture supernatants of a Burkitt lymphoma cell line (BJAB), a primary prostate cancer cell line (LK), and a primary melanoma cell line (MEL-P). B, Suppression assay of T cell proliferation performed with CD8⁺CD28⁻ Treg generated in vitro in the presence of culture supernatant from BJAB cells (upper row), LK cells (medium row), or MEL-P cells (lower row). The cultures containing the CD8⁺CD28⁻ T cells were performed in transwell plates to physically separate these lymphocytes from proliferating and stimulating cells. Neg. Contr., Irradiated macrophages plus nonstimulated PBMC; anti-CD3 PBMC, irradiated macrophages plus anti-CD3 mAb-stimulated PBMC; anti-CD3 PBMC plus CD8⁺CD28⁻, as anti-CD3 PBMC plus unstimulated CD8⁺CD28⁻ T cells; anti-CD3 PBMC ⁺ act.CD8⁺CD28⁻, as anti-CD3 PBMC plus CD8⁺CD28⁻ T cells prestimulated with IL-2 alone (20 U/ml for 7 days); anti-CD3 PBMC plus CD8⁺ Treg, as anti-CD3 PBMC plus CD8⁺ Treg generated with rIL-2 and rIL-10; anti-CD3 plus snCD8 plus Treg, as anti-CD3 PBMC plus CD8⁺ Treg generated with rIL-2 and the tumor cell line supernatant (in the case of supernatants from GM-CSF-secreting cell lines, autologous monocytes were also added to the culture for generation of CD8⁺ Treg lymphocytes); anti-CD3 PBMC plus snCD8 plus Treg plus antag, as anti-CD3 PBMC plus CD8⁺ Treg generated with rIL-2 and the tumor cell line supernatant preincubated with an anti-IL-10 (20 μg/ml) (in the case of BJAB cell supernatant) or an anti-GM-CSF mAb (20 μg/ml) (in the case of supernatants from LK and MEL-P cells) for 1 h at 37°C; anti-CD3 PBMC plus snCD8 plus Treg plus contr, as anti-CD3 PBMC plus SN-CD8 plus Treg plus antag but the anti-IL-10 or the anti-GM-CSF mAbs were replaced by unrelated, isotype-matched mAbs. Triplicate cultures were performed and the results are expressed as mean cpm ± SD. Percent suppression activities are indicated in parentheses. C, Suppression assay of T cell proliferation performed with CD8⁺CD28⁻ Treg generated in vitro in the presence of culture supernatant from MEL-P cells incubated or not with a GM-CSF antisense oligonucleotide. The cultures containing the CD8⁺CD28⁻ T cells were performed in transwell plates to physically separate these lymphocytes from proliferating and stimulating cells. a, Irradiated macrophages plus nonstimulated PBMC; b, irradiated macrophages plus anti-CD3 mAb stimulated PBMC; c, as b plus CD8⁺ Treg generated with rIL-2 and rGM-CSF in the presence of autologous monocytes; d, as c using the MEL-P supernatant instead of rGM-CSF; e, as d using MEL-P supernatant preincubated with an anti-GM-CSF mAb (20 μg/ml); f, as d using supernatant from MEL-P cells preincubated with a GM-CSF antisense oligonucleotide. D, Immunofluorescence analysis on a tissue section from primitive tumor mass of patient 18 performed with an anti-IL-10 mAb. Magnification, ×250.

To verify whether Treg might be prone to chemokine attraction by the tumor, the expression of CCR2 (receptor for CCL2) and CCR4 (receptor for CCL22) on tumor-infiltrating CD8⁺CD28⁻ and CD4⁺CD25⁺ T lymphocytes was analyzed. Cells positive for both chemokine receptors were detected at variable frequencies in the two regulatory T cell subsets (Fig. 5⇓). This is of importance because CCL2 and CCL22 are the chemokines most frequently secreted by cancers (17, 18). Accordingly, CCL2 was detected in the supernatants of LK and MEL-P primary cell lines (369 and 197 pg/ml, respectively).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Expression of CCR2 and CCR4 chemokine receptors by CD4⁺CD25⁺ and CD8⁺CD28⁻ T lymphocytes from metastatic lymph nodes. Analyses were performed on cells from patient 35 (upper row) and from patient 38 (lower row). Cell double staining using an allophycocyanin-conjugated anti-CD3 mAb and PerCP-cyanin 5-conjugated anti-CD4 (upper row, a) or anti-CD8 (lower row, a) mAbs was performed to identify CD4⁺ or CD8⁺ T cell subpopulations, respectively; within each of these T cell subsets, CD4⁺CD25⁺ (upper row, b) or CD8⁺CD28⁻ (lower row, b) T cells were detected using PE-conjugated anti-CD25 or anti-CD28 mAbs, respectively; the expression on these two T cell subpopulations of CCR4 (upper and lower rows, c) was analyzed using a FITC-conjugated anti-CCR4 mAb, while CCR2 expression was studied using an unconjugated anti-CCR2 mAb followed by staining with a FITC-conjugated goat anti-mouse secondary Ab (upper and lower rows, d). Allophycocyanin-, PerCP-cyanin 5-, PE-, or FITC-conjugated unrelated, isotype-matched mAbs were used as negative controls (data not shown). The percentages of positive cell subpopulations are indicated.

Relationship between CD8⁺CD28⁻ Treg function and clinical stage or survival

A statistically significant direct correlation was found between survival and the absolute number of CD8⁺ (p = 0.04, Spearman r = 0.57) as well as that of CD8⁺CD28⁻ (p = 0.03, Spearman r = 0.61) TILs. Interestingly, the percent suppressor function of CD8⁺CD28⁻ T cells deriving from primitive tumor mass was found in statistically direct correlation with clinical stage (p = 0.01, Spearman r = 0.50) and in statistically inverse correlation with patient survival (p = 0.0001, Spearman r = 0.92) (Fig. 6⇓). Moreover, the mean suppressor activity of CD8⁺CD28⁻ T lymphocytes purified from the primitive tumor mass of patients with <12 mo survival from surgery (59 ± 18%) was significantly higher than that of patients with >12 mo survival (31 ± 5%) (p = 0.001). Similar findings were observed comparing the mean suppressor activities of CD8⁺CD28⁻ T lymphocytes purified from metastatic lymph nodes of patients with <12 (46% ± 19) or >12 (30% ± 10) month survival (p = 0.01), as well as of CD8⁺CD28⁻ T cells from peripheral blood of patients with <12 (44 ± 24%) or >12 (27 ± 8%) month survival (p < 0.01).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Analysis of the statistical correlation between the percent suppressor activity of tumor-infiltrating CD8⁺CD28⁻ Treg and (A) disease stage or (B) patient survival.