TGF-β1 is a potent immunoregulatory cytokine. However, its impact on the generation and effector function of Ag-specific human effector memory CD8 T cells had not been evaluated. Using Ag-specific CD8 T cells derived from melanoma patients immunized with the gp100 melanoma Ag, we demonstrate that the addition of TGF-β1 to the initial Ag activation cultures attenuated the gain of effector function by Ag-specific memory CD8 T cells while the phenotypic changes associated with activation and differentiation into effector memory were comparable to control cultures. These activated memory CD8 T cells consistently expressed lower mRNA levels for T-bet, suggesting a mechanism for TGF-β1-mediated suppression of gain of effector function in memory T cells. Moreover, TGF-β1 induced a modest expression of CCR7 on Ag-activated memory CD8 T cells. TGF-β1 also suppressed cytokine secretion by Ag-specific effector memory CD8 T cells, as well as melanoma-reactive tumor-infiltrating lymphocytes and CD8 T cell clones. These results demonstrate that TGF-β1 suppresses not only the acquisition but also expression of effector function on human memory CD8 T cells and tumor-infiltrating lymphocytes reactive against melanoma, suggesting that TGF-β1-mediated suppression can hinder the therapeutic benefits of vaccination, as well as immunotherapy in cancer patients.

Transforming growth factor β1 is a pleiotropic cytokine with a potent immunoregulatory function. TGF-β1 knockout mice develop multifocal infiltration of inflammatory cells in multiple organs with increased numbers of proliferating cells within lymphoid organs leading to early death (reviewed in Refs. 1, 2, 3). Moreover, disruption of the TGF-β-signaling pathway exclusively in T cells was also sufficient to cause T cell-mediated autoimmune disease in mice carrying this transgene (4, 5).

In humans, TGF-β suppresses proliferation of T cells in vitro through inhibition of IL-2 production (6, 7). In addition, TGF-β1 can also have an inhibitory effect on the differentiation of murine naive CD4 (8, 9) and CD8 T cells (10) independent of its antiproliferative effect because addition of IL-2 could not reverse this inhibition (8). More recently, Flavell and colleagues demonstrated that TGF-β1 inhibits the differentiation of murine naive CD4 T cells by suppressing the expression of T-bet (11) and GATA-3 (12), the key transcriptional activators for development of CD4 Th1 (13) and Th2 (14) cells, respectively. However, the impact of TGF-β1 on the activation and differentiation of effector and memory CD8 T cells is unknown.

Because of its profound immunoregulatory function, TGF-β1 is implicated to play a pivotal role in regulating antitumor immune responses. TGF-β1 has been secreted by many tumors, including melanoma (15, 16, 17), and mutations in TGF-β-signaling components can abrogate the growth inhibitory effect of this cytokine on tumor cells (18, 19). In mice, tumor-derived TGF-β can suppress the cytolytic function of tumor-reactive CD8 T cells both in vitro and in vivo (20) and can attenuate the efficacy of in vivo tumor vaccines (21). Abrogation of the TGF-β-signaling pathway by ectopic expression of dominant negative TGF-β receptor type II in human or murine CD8 T cells increased EBV-specific CTL responses in vitro (22) and enhanced the eradication of B16 melanoma in vivo (23), respectively. Collectively, the evidence from these studies suggests that TGF-β1 may play a crucial role in facilitating the immunosuppression of tumor-specific responses and highlights the significance of establishing the potential impact of TGF-β1 in antitumor immune responses in humans.

Although the inhibitory effect of TGF-β1 on the differentiation of naive into effector T cells has been established, the impact of this cytokine on the activation of Ag-specific human memory CD8 T cells into functionally active effector memory populations has not been investigated previously. In this study, we have evaluated the impact of recombinant human TGF-β1 on the activation of tumor Ag-specific memory CD8 T cells derived from melanoma patients vaccinated with gp100, a melanoma Ag (24). Our results show that the addition of TGF-β1 to initial Ag activation cultures attenuated the gain of effector function by human memory CD8 T cells, although induction of activation and differentiation markers were unaffected. Furthermore, our data demonstrate that TGF-β1 can suppress the secretion of inflammatory cytokines by fully activated Ag-specific effector memory CD8 T cells, tumor-reactive human tumor-infiltrating lymphocytes (TIL),2 and CD8 T cell clones. These results demonstrate, for the first time, the immunosuppressive impact of TGF-β1 on the effector function of memory CD8 T cells, as well as fully activated effector CD8 T cells reactive to melanoma tumor Ags in humans.

HLA-A*0201-positive melanoma patients with completely resected disease who had completed three or four courses of immunization with s.c. injections of 1 mg of the modified gp100:209-217 (210M) peptide in IFA were selected for this study (24). PBMC were obtained by leukapheresis before and 3 wk following completion of the immunization course, isolated by Ficoll gradient separation, and cryopreserved until analyses. The protocol was approved by the Institutional Review Board of the National Cancer Institute.

CD8 clones were generated by limiting dilution from melanomas that were removed surgically from patients treated at the Surgery Branch, National Cancer Institute (25). These clones were highly reactive to either gp100 or MART-1 Ags expressed by HLA-A2+ melanoma tumor cell lines. TIL were also generated from melanoma tumors by culturing in high-dose IL-2 (6000 IU/ml) as previously described (26) for patient treatment. The cryopreserved TIL were thawed, washed once in complete medium without exogenous IL-2, and cultured with melanoma tumor cells in the absence of any exogenous cytokines.

The native peptide gp100:209-217 (ITDQVPFSV) (designated as g209) and the modified peptide gp100:209-217 (210M) (IMDQVPFSV) (designated as g209-2M) were derived from the gp100 melanoma Ag and were prepared according to good manufacturing practice using solid phase synthesis by Multiple Peptide Systems. The irrelevant peptide gp100:280-288 (YLEPGPVTA) (designated as g280) was used as a negative control for Ag specificity responses. The g209, g209-2M, and g280 peptides used for in vitro cultures were provided by the National Cancer Institute Cancer Therapy Evaluation Program.

Cryopreserved PBMC from immunized patients were thawed into complete medium consisting of IMDM (Biofluids) supplemented with 10% heat-inactivated human AB serum (Gemini Bio-Products), 100 U/ml penicillin, 100 μg/ml streptomycin (Biofluids), 50 μg/ml gentamicin (Biofluids), 10 mM HEPES buffer (Biofluids), 2 mM l-glutamine (Biofluids), and 1.25 μg/ml amphotericin B (Biofluids). On day 0, PBMC (1.5 × 106 cells/ml) were activated with 1 μM g209-2M peptide with or without recombinant TGF-β1. On day 1, IL-2 was added at a final concentration of 300 IU/ml. On day 6, memory cells were analyzed for the ability to express effector function. For generating fully activated effector memory CD8 T cells, PBMC were activated as described above in the absence of exogenous TGF-β1, and an additional dose of IL-2 was added following splitting the cultures with fresh media on day 6. Confluent cultures beyond 9 days were split with fresh media and IL-2 as needed. Activated T cells, designated as effector memory T cells, were analyzed 9–15 days after initiation of the cultures.

In some experiments, CD8 T cells were negatively enriched using Miltenyi CD8+ T Cell Isolation kit II (Miltenyi Biotec), according to the manufacturer’s instructions. Briefly, cells harvested from different activation cultures were labeled with a mixture of biotin-conjugated Abs against CD4, CD14, CD16, CD19, CD36, CD56, CD123, TCRγδ, and glycophorin A, followed by incubation with anti-biotin microbeads. CD8+ T cells were enriched by depletion of magnetically labeled non-CD8 cells and used for either 24-h Ag restimulation cocultures or lysed for total RNA extraction.

For Ag restimulation of activated memory CD8 T cells, CD8 clones, and TIL, 105 cells were cocultured with 105 peptide-pulsed T2 cells or melanoma cell lines for 18–24 h in 96-well round-bottom plates. MEL-526 and MEL-624 (HLA-A*0201+, designated as A2+) and MEL-888 and MEL-938 (HLA-A*0201, designated as A2) are melanoma cell lines that were derived in the Surgery Branch, National Cancer Institute, as described previously (27). Culture plates were spun at 2000 rpm for 10 min at room temperature, and cell-free supernatants were assayed for IFN-γ, GM-CSF, and TNF-α using human IFN-γ ELISA kits and SearchLight Proteomic Arrays by Pierce. Percent cytokine suppression was calculated as shown below: percent suppression = (((stimulation − background) − (stimulation + TGF-β1 − background))/(stimulation − background)) × 100.

PE-labeled g209 (ITDQVPFSV) and HIV gag (SLYNTVATL) peptide-HLA-A*0201 tetramers were purchased from Beckman Coulter. Purified anti-TGF-β (clone 1D11 mAb) was provided generously by Genzyme and used at 10 μg/ml. The following mAb specific for human surface Ags were purchased from BD Pharmingen: FITC-conjugated anti-CD45RA, CD45RO, CD62L, CD25, CD69, and allophycocyanin-conjugated anti-CD8. FITC-conjugated CCR7 Ab was purchased from R&D Systems. Recombinant human TGF-β1 was purchased from R&D Systems, reconstituted, and stored at −20°C as suggested by the manufacturer. Recombinant human IL-2 was kindly supplied by Chiron.

Cells were resuspended in staining buffer (PBS containing 3% FBS) and blocked with mouse Ig (Caltag Laboratories) for 15–30 min at room temperature. Cells were stained with PE-conjugated g209 tetramer, allophycocyanin-conjugated anti-CD8 mAb, and FITC-conjugated mAb or isotype control Abs and negative tetramer for 30 min on ice or at 4°C in the dark. Cells were washed subsequently in staining buffer twice and briefly stained with propidium iodide (PI) for nonviable cell exclusion before analyzing in a FACSCalibur (BD Biosciences).

To determine gene expression, total RNA was extracted from CD8+ enriched cell lysates using the RNeasy Mini kit (Qiagen), according to the manufacturer’s instructions. cDNA was synthesized using oligo(dT) primers and ThermoScript RT as reverse transcriptase (Invitrogen Life Technologies). Expression of T-bet and CD3ε was measured using Assays-On-Demand primers and probes (Applied Biosystems). PCR thermal cycling conditions were 50°C for 2 min, 95°C for 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min in a total volume of 25 μl/reaction. Data were collected using ABI Prism 7700 Sequence Detection System Software. All samples were run in duplicates, and the mean values were used for quantification. T-bet expression was normalized by dividing the relative quantity of T-bet cDNA for each sample by the relative quantity of CD3ε cDNA.

To examine the impact of TGF-β1 on the generation of activated memory CD8 T cells, we used memory cells derived from melanoma patients who had undergone multiple immunization courses with the modified melanoma tumor Ag, g209-2M (24). We had demonstrated previously that these immunized patients had an increased frequency of memory T cells reactive to the native g209 epitope as indicated by phenotypic and functional analyses (24, 28, 29). After immunization, PBMC from melanoma patients were cultured with g209-2M peptide and titrated amounts of TGF-β1 (0.01–10 ng/ml) on day 0, followed by addition of IL-2 (300 IU/ml) to all cultures on day 1 and phenotypic and functional analyses on day 6.

To determine the impact of TGF-β1 on the activation of memory CD8 T cells, we examined changes in the surface expression of activation markers, which are indicative of differentiation into an effector memory phenotype. The tumor Ag-specific memory CD8 T cells before in vitro activation (no stimulation) and following Ag activation in the absence (stimulation) or presence of TGF-β1 (stimulation + TGF-β1) were visualized with g209 tetramer (Fig. 1). The histograms were gated on live cells (PI), CD8+, and g209 tetramer+ cells. Because activated lymphocytes are larger in size, we compared the forward scatter (FSC) on tetramer+ CD8 T cells. As expected, tetramer+ cells in the no stimulation group had lower FSC with mean fluorescence intensity (MFI) of 325 compared with tetramer+ cells from activated cultures (Fig. 1). Following Ag activation, the FSC of tetramer+ cells in both activation cultures were increased with comparable amounts (Fig. 1; stimulation, MFI 491; stimulation + TGF-β1, MFI 456), suggesting that the tetramer+ populations in both cultures were activated. CD25 expression was up-regulated on memory CD8 T cells from both activation cultures; however, tetramer+ cells from TGF-β cultures (stimulation + TGF-β1) had overall lower MFI (156) compared with control cultures (234). In addition, induction of other activation markers, such as CD69 and HLA-DR, were enhanced equally in both TGF-β1 and control cultures (Fig. 1 and data not shown, respectively). Therefore, TGF-β1 does not inhibit the induction of activation markers on Ag-specific CD8 memory T cells upon Ag activation.

FIGURE 1.

The phenotype of tumor Ag human memory CD8 T cells following Ag activation with or without TGF-β1. PBMC from immunized melanoma patients were activated with the immunizing Ag, g209-2M peptide, in the absence (stimulation (Stim)) or presence of 1 ng/ml TGF-β1 (Stim + TGF-β1). All cultures received IL-2 (300 IU/ml) the following day and were analyzed 6 days later. Cells were stained with native g209 tetramer, anti-CD8, and different activation markers. For control (no stimulation (No Stim)), PBMC corresponding to the starting population were thawed and rested overnight in media and stained the next day with the other groups. The histograms were gated on PI, CD8+, and g209 tetramer+ cells.

FIGURE 1.

The phenotype of tumor Ag human memory CD8 T cells following Ag activation with or without TGF-β1. PBMC from immunized melanoma patients were activated with the immunizing Ag, g209-2M peptide, in the absence (stimulation (Stim)) or presence of 1 ng/ml TGF-β1 (Stim + TGF-β1). All cultures received IL-2 (300 IU/ml) the following day and were analyzed 6 days later. Cells were stained with native g209 tetramer, anti-CD8, and different activation markers. For control (no stimulation (No Stim)), PBMC corresponding to the starting population were thawed and rested overnight in media and stained the next day with the other groups. The histograms were gated on PI, CD8+, and g209 tetramer+ cells.

Close modal

We also examined surface expression of CD45RA and CD45RO that are expressed inversely on activated memory T cells (30). Before in vitro activation, CD45RA was expressed heterogeneously on tetramer+ memory CD8 T cells (Fig. 1, no stimulation). Following Ag activation, expression of CD45RA was reduced in both stimulation and stimulation + TGF-β1 cultures at similar levels. Conversely, CD45RO expression was expressed at low levels before activation (no stimulation); its expression was enhanced equally in both stimulation and stimulation + TGF-β1 cultures upon Ag activation (Fig. 1). Thus, TGF-β1 did not interfere with alteration in surface expression of CD45RA and CD45RO, typically associated with T cell activation.

In addition to activation and differentiation markers, we sought to examine the impact of TGF-β1 on expression of lymph node homing receptors, CD62L and CCR7. Before activation, tetramer+ memory CD8 T cells were CD62L (Fig. 1, no stimulation); expression of CD62L was minimally induced in both activation cultures regardless of addition of TGF-β1 (Fig. 1). In contrast, expression of CCR7 was significantly altered depending on whether TGF-β1 was added to the in vitro activation cultures. Before activation, tetramer+ memory cells were all CCR7 with MFI 6 (Fig. 1, no stimulation); however, following Ag activation and in the presence of TGF-β1, CCR7 was up-regulated in a predominant population of tetramer+ cells with MFI 52 (Fig. 1, stimulation + TGF-β1), while its expression was slightly increased (MFI 25) yet significantly lower in the absence of exogenous TGF-β1 (Fig. 1, stimulation). Thus, addition of TGF-β1 to activation cultures can result in a higher induction of CCR7, whereas expression of CD62L is not altered.

Table I represents a summary of phenotypic changes from three patients (patient 1, whose phenotype is presented in Fig. 1, is included in this table for reference). The percentages for each surface marker represent the frequency of tetramer+ CD8 T cells expressing that particular marker within the entire Ag-specific population; MFI of the entire tetramer+ population is also shown for FSC, CD25, and CCR7 to evaluate the level of intensity of the phenotypic markers. As shown in Table I, the frequency of g209-tetramer+ cells in CD8 T cell population increases in Ag-activated cultures, even in the presence of TGF-β1 for all patients analyzed. Thus, addition of TGF-β1 to the initial activation cultures had no detectable impact on the expansion of Ag-specific CD8 memory T cells. Overall, the phenotypic changes were reproducible and indicate that Ag activation of memory CD8 T cells in the presence of TGF-β1 leads to the generation of activated memory CD8 T cells that phenotypically exhibit similar activation profile as control cultures yet with higher expression of CCR7.

Table I.

Summary of phenotypic changes in tetramer+ CD8 memory T cells

Tet+a %FSC (MFI)CD25 % (MFI)bCD69 %CD45RA %CD45RO %CCR7 % (MFI)CD62L %
Patient 1         
 No stimc 1.6 (325) 0 (5) 0.8 88.5 67.8 0.1 (6) 0.2 
 Stim 4.2 (491) 98.2 (234) 77.3 51.5 99.7 14.8 (25) 3.5 
 Stim + TGF-β1 5.4 (456) 85.9 (156) 72.6 39.9 99.1 79.1 (52) 
Patient 2         
 No stim 1.6 (317) 0.3 (5) 5.7 99.4 39 1.5 (13) 1.4 
 Stim 2.4 (464) 98.1 (426) 56.3 48.6 99.7 8.9 (30) 3.1 
 Stim + TGF-β1 3.8 (416) 54.1 (106) 49.7 55.7 99.6 26.4 (38) 0.7 
Patient 3         
 No stim 0.5 (306) 0.8 (6) 51.3 97.6 91.6 6.0 (12) 2.9 
 Stim 2.1 (466) 100 (1352) 42.4 20.3 99.5 77.1 (179) 2.8 
 Stim + TGF-β1 3.2 (477) 99.9 (1347) 48.6 20.4 99.8 99.1 (373) 2.3 
Tet+a %FSC (MFI)CD25 % (MFI)bCD69 %CD45RA %CD45RO %CCR7 % (MFI)CD62L %
Patient 1         
 No stimc 1.6 (325) 0 (5) 0.8 88.5 67.8 0.1 (6) 0.2 
 Stim 4.2 (491) 98.2 (234) 77.3 51.5 99.7 14.8 (25) 3.5 
 Stim + TGF-β1 5.4 (456) 85.9 (156) 72.6 39.9 99.1 79.1 (52) 
Patient 2         
 No stim 1.6 (317) 0.3 (5) 5.7 99.4 39 1.5 (13) 1.4 
 Stim 2.4 (464) 98.1 (426) 56.3 48.6 99.7 8.9 (30) 3.1 
 Stim + TGF-β1 3.8 (416) 54.1 (106) 49.7 55.7 99.6 26.4 (38) 0.7 
Patient 3         
 No stim 0.5 (306) 0.8 (6) 51.3 97.6 91.6 6.0 (12) 2.9 
 Stim 2.1 (466) 100 (1352) 42.4 20.3 99.5 77.1 (179) 2.8 
 Stim + TGF-β1 3.2 (477) 99.9 (1347) 48.6 20.4 99.8 99.1 (373) 2.3 
a

Percent tetramer+ cells in CD8+ population.

b

Percent tetramer+ cells expressing surface marker; MFI is for the entire tetramer+ population.

c

Stim, stimulation.

Given that Ag-activated memory CD8 T cells in the presence of TGF-β1 displayed activation markers associated with fully activated CD8 T cells, we asked whether their functional response to recall Ag were also similar. Varying amounts of TGF-β1 were added only to the initial activation cultures (day 0), and the acquisition of effector function by memory CD8 T cells was assessed 6 days later. Activated memory CD8 T cells were cocultured with T2 cells pulsed with titrated amounts of native g209 peptide for 24 h, and recall response was assessed by quantifying the level of IFN-γ release in cell-free supernatants. T2 cells pulsed with irrelevant peptide (g280) were used as a control. We used effector cytokine secretion as a functional readout for effector function of CD8 T cells because IFN-γ secretion, in our experience, has correlated with cytolytic ability of effector cells (28) and is used continuously for selection of tumor-reactive TIL for patient treatments (25, 31). As shown in Fig. 2,A, memory CD8 T cells activated with the Ag in the absence of exogenous TGF-β1 (stimulation) released large quantities of IFN-γ when restimulated with peptide-pulsed T2 cells, whereas the addition of TGF-β1 to the initial activation cultures resulted in significantly lower amounts of IFN-γ release in a dose-dependent manner. This reduced ability to secrete IFN-γ was independent of Ag concentration (Fig. 2,A) and affinity of the antigenic peptide (g209-2M; data not shown). The IFN-γ release was Ag specific because cells from both activation cultures released undetectable amounts of IFN-γ in response to restimulation with the irrelevant peptide (Fig. 2 A, g280).

FIGURE 2.

Functional and molecular analyses of tumor Ag-reactive memory CD8 T cells. PBMC were activated with g209-2M with titrated amounts of TGF-β1 for 6 days. Activated cells were washed in media and cocultured with T2 cells pulsed with native g209 peptide or irrelevant peptide (g280) as control. A, IFN-γ release was quantified from 24 h supernatants using ELISA. B, Percent IFN-γ suppression with SE from four independent experiments is calculated from cocultures stimulated with T2 cells pulsed with 1 μM g209 peptide (g209; 1 μM). C, To quantify the relative T-bet expression in activated memory CD8 T cell cultures, memory CD8 T cells from five patients were activated with g209-2M (1 μM) in absence or presence of TGF-β1 (1 ng/ml) for 6 days as described earlier. RNA was extracted from cultures enriched for CD8 T cells (see Materials and Methods). T-bet expression is normalized for CD3ε.

FIGURE 2.

Functional and molecular analyses of tumor Ag-reactive memory CD8 T cells. PBMC were activated with g209-2M with titrated amounts of TGF-β1 for 6 days. Activated cells were washed in media and cocultured with T2 cells pulsed with native g209 peptide or irrelevant peptide (g280) as control. A, IFN-γ release was quantified from 24 h supernatants using ELISA. B, Percent IFN-γ suppression with SE from four independent experiments is calculated from cocultures stimulated with T2 cells pulsed with 1 μM g209 peptide (g209; 1 μM). C, To quantify the relative T-bet expression in activated memory CD8 T cell cultures, memory CD8 T cells from five patients were activated with g209-2M (1 μM) in absence or presence of TGF-β1 (1 ng/ml) for 6 days as described earlier. RNA was extracted from cultures enriched for CD8 T cells (see Materials and Methods). T-bet expression is normalized for CD3ε.

Close modal

The percent suppression of IFN-γ release in response to restimulation with T2 cells pulsed with 1 μM peptide was calculated from four independent experiments (Fig. 2,B). Although small suppression is detectable at a very low concentration of TGF-β1 (0.01 ng/ml), it was variable among the patients tested. However, there was a steady increase in the suppression of IFN-γ release that reached over 80% with little variability among the patients at 0.1–10 ng/ml TGF-β1 (Fig. 2 B). Overall, these results demonstrate that the addition of TGF-β1 to the initial Ag activation cultures attenuated the ability of activated memory cells to release IFN-γ in response to Ag restimulation.

To understand the molecular mechanism by which TGF-β1 inhibits the acquisition of effector function in memory CD8 T cells, we investigated whether expression of T-bet, a transcription factor critical for Th1 differentiation (13), was altered in memory CD8 T cells derived from TGF-β cultures. The mRNA levels of T-bet were consistently lower in memory CD8 T cells derived from TGF-β activation cultures compared with cultures with no TGF-β (Fig. 2 C). In three of the five patients, there is >2-fold suppression in T-bet expression in TGF-β1 activation cultures. The reduction of T-bet correlates with suppression of IFN-γ secretion, as well as IFN-γ mRNA levels (data not shown). These molecular analyses reveal that TGF-β1 inhibits expression of T-bet in memory CD8 T cells, providing a potential mechanism by which TGF-β1 suppresses gain of effector function by memory CD8 T cells and inhibits generation of effector memory T cells.

Similar to suppression of IFN-γ release, addition of TGF-β1 to the initial activation cultures resulted in reduced secretion of GM-CSF by activated memory CD8 T cells in response to antigenic restimulation (Fig. 3,A). Memory cells preactivated in the absence of exogenous TGF-β1 were able to produce GM-CSF at higher levels (>5000 pg/ml), whereas memory cells preactivated in the presence of TGF-β1 (1 or 10 ng/ml) secreted significantly lower levels of GM-CSF (<2000 pg/ml) (Fig. 3,A, patient 1). The level of suppression in some cases reached the level of background detected for irrelevant peptide (g280). Similarly, addition of TGF-β1 impacted the ability of activated memory cells to produce TNF-α in response to antigenic restimulation (Fig. 3,B). The percent suppression by TGF-β1 (1 ng/ml) was 87% (±6 SE) and 60% (±10 SE) for GM-CSF and TNF-α, respectively (Fig. 3 C). Collectively, these functional analyses reveal that TGF-β1 interferes with the ability of memory CD8 T cells to acquire effector function upon Ag activation, although they exhibit activation and differentiation markers generally associated with the gain of effector function.

FIGURE 3.

Effector cytokine release by tumor Ag-reactive memory CD8 T cells. PBMC were activated and restimulated with T2 cell pulsed with 1 μM g209 or irrelevant peptide, g280, as described in Fig. 2. Twenty-four-hour supernatants were tested for GM-CSF (A) and TNF-α (B). The percent suppressions for each cytokine along with SE were calculated for three patients (C).

FIGURE 3.

Effector cytokine release by tumor Ag-reactive memory CD8 T cells. PBMC were activated and restimulated with T2 cell pulsed with 1 μM g209 or irrelevant peptide, g280, as described in Fig. 2. Twenty-four-hour supernatants were tested for GM-CSF (A) and TNF-α (B). The percent suppressions for each cytokine along with SE were calculated for three patients (C).

Close modal

Given the suppressive impact of TGF-β1 on the acquisition of effector function by activated memory CD8 T cells, we sought to examine the impact of this cytokine on the effector function of already activated effector memory CD8 T cells. To generate fully activated effector memory CD8 T cells, PBMC were activated with the g209-2M peptide and IL-2 for 10–14 days in the absence of exogenous TGF-β1. Activated memory CD8 T cells were cocultured with g209-pulsed T2 cells or melanoma cell lines in the presence of titrated amounts of TGF-β1 for 24 h (Fig. 4). The addition of TGF-β1 reduced the level of IFN-γ secretion by activated memory cells following restimulation with either peptide (Fig. 4, A–C) or HLA-A2+ melanoma cell lines (Fig. 4,D) in a dose-dependent manner. The IFN-γ release was Ag specific because restimulation of T cells with the irrelevant peptide (g280) or HLA-A2 melanoma cell lines was negligible and below the detection limits of ELISA. This suppression was TGF-β1 specific because the anti-TGF-β Ab (1D11) completely abrogated the suppression (Fig. 4,B). We also enriched for CD8 T cells and set up parallel cocultures in the presence of TGF-β1 as described above. TGF-β1 also significantly reduced the level of IFN-γ secretion by purified CD8 T cells (data not shown), suggesting that the suppression was not mediated through non-CD8 T cells. Suppression of IFN-γ was detectable even at a minute concentration (0.001 ng/ml) of TGF-β1 and increased to 30–50% with higher doses (Fig. 4, C and D). However, the suppression levels reached their maximum peak at 1–10 ng/ml TGF-β1. Thus, TGF-β1 can suppress IFN-γ secretion by effector memory CD8 T cells. Melanoma cell lines used in this study did not secrete any detectable levels of active form of TGF-β1.

FIGURE 4.

TGF-β-mediated suppression of IFN-γ release by activated effector memory CD8 T cells. PBMC from immunized patients were activated with g209-2M and IL-2 in the absence of exogenous TGF-β1 for 10–14 days. Activated cells were cocultured with peptide-pulsed T2 cells or melanoma tumor cell lines in the presence of TGF-β1. A, IFN-γ levels secreted by activated memory cells in response to restimulation with peptide-pulsed T2 cells and in the presence of increasing amounts of TGF-β1 were quantified by ELISA. B, Twenty-four-hour cocultures were set up in absence (no TGF-β) or presence of 0.1 ng/ml TGF-β (plus TGF-β) or with 0.1 ng/ml TGF-β neutralized with 1D11 Ab (plus TGF-β and 1D11); IFN-γ release was quantified by ELISA. C, Percent IFN-γ suppression was calculated from six patients in response to peptide-pulsed T2 cells. D, Percent IFN-γ suppression was calculated from three patients in response to HLA-A2+ melanoma cell lines. The cytokine secretions in response to irrelevant peptide or HLA-A2 melanoma tumor cell lines were negligible or undetectable.

FIGURE 4.

TGF-β-mediated suppression of IFN-γ release by activated effector memory CD8 T cells. PBMC from immunized patients were activated with g209-2M and IL-2 in the absence of exogenous TGF-β1 for 10–14 days. Activated cells were cocultured with peptide-pulsed T2 cells or melanoma tumor cell lines in the presence of TGF-β1. A, IFN-γ levels secreted by activated memory cells in response to restimulation with peptide-pulsed T2 cells and in the presence of increasing amounts of TGF-β1 were quantified by ELISA. B, Twenty-four-hour cocultures were set up in absence (no TGF-β) or presence of 0.1 ng/ml TGF-β (plus TGF-β) or with 0.1 ng/ml TGF-β neutralized with 1D11 Ab (plus TGF-β and 1D11); IFN-γ release was quantified by ELISA. C, Percent IFN-γ suppression was calculated from six patients in response to peptide-pulsed T2 cells. D, Percent IFN-γ suppression was calculated from three patients in response to HLA-A2+ melanoma cell lines. The cytokine secretions in response to irrelevant peptide or HLA-A2 melanoma tumor cell lines were negligible or undetectable.

Close modal

Coculture assays from three separate patients demonstrated that secretion of GM-CSF (Fig. 5,A) and TNF-α (Fig. 5,B) was also inhibited by TGF-β1. GM-CSF was released by effector memory CD8 T cells following restimulation by melanoma cell lines (Fig. 5,A, patients 1 and 2) or Ag-pulsed T2 cells (Fig. 5,A, patient 3). The GM-CSF secretion was Ag specific because restimulation with HLA-A2 melanoma cell lines (MEL-888) or T2 cells pulsed with irrelevant peptide (g280) resulted in secretion of negligible amounts of this cytokine (Fig. 5,A). Addition of TGF-β1 during the Ag restimulation significantly reduced (48–82% suppression) the amount of GM-CSF secreted in all three patients regardless of the type of target cells (Fig. 5,A). TGF-β1 also reduced Ag-specific TNF-α secretion in an Ag-specific response for all three patients tested (Fig. 5,B). Although the overall range of percent suppression for TNF-α (15–54%) was lower than GM-CSF (48–82%), addition of TGF-β1 consistently reduced secretion of TNF-α in all three patients and with three different types of stimuli. The suppressive impact of TGF-β1 was less evident in cultures in which activated memory CD8 T cells were not secreting large quantities of TNF-α, even in the absence of exogenous TGF-β1 (Fig. 5,B, patient 2). Although the percent TNF-α suppression (calculated based on the averages of duplicates) for this patient is relatively lower than the other two patients (Fig. 5 B, patient 1, MEL-526, 40% and MEL-624, 51%; and patient 3, 54%), nevertheless, addition of TGF-β1 to cocultures consistently reduced the level of TNF-α secretion. These results collectively demonstrate that TGF-β1 can suppress secretion of effector cytokines by fully activated Ag-specific effector memory CD8 T cells in a short-term (24 h) functional assay; it also further highlights the immunosuppressive impact of TGF-β1 on effector function as reflected by the suppression of effector memory T cells.

FIGURE 5.

TGF-β1 suppresses GM-CSF and TNF-α secretion by activated effector memory CD8 T cells. Effector memory CD8 T cells were activated in the absence of exogenous TGF-β1 for 10–14 days as described in Fig. 5 and restimulated with HLA-A2+ melanoma cell lines (MEL-526 and MEL-624) or T2 cells pulsed with 1 μM g209 peptide in a 24-h coculture. Cell-free supernatants were quantified for GM-CSF (A) and TNF-α (B) levels using proteomic arrays. Cytokine secretions in response to HLA-A2 melanoma cell lines (MEL-888) or T2 pulsed with irrelevant peptide (g280) represent the nonspecific secretion.

FIGURE 5.

TGF-β1 suppresses GM-CSF and TNF-α secretion by activated effector memory CD8 T cells. Effector memory CD8 T cells were activated in the absence of exogenous TGF-β1 for 10–14 days as described in Fig. 5 and restimulated with HLA-A2+ melanoma cell lines (MEL-526 and MEL-624) or T2 cells pulsed with 1 μM g209 peptide in a 24-h coculture. Cell-free supernatants were quantified for GM-CSF (A) and TNF-α (B) levels using proteomic arrays. Cytokine secretions in response to HLA-A2 melanoma cell lines (MEL-888) or T2 pulsed with irrelevant peptide (g280) represent the nonspecific secretion.

Close modal

Because TGF-β1 significantly suppressed cytokine secretion by heterogeneous populations of effector memory CD8 T cells from PBMC, we asked whether CD8 T CTL clones and activated tumor-reactive TIL used in patients’ treatments were also susceptible to TGF-β1-mediated suppression. Treatment TIL generally contain large numbers of CD8 T cells (32); however, some TIL may have a larger ratio of CD4 T cells such as the one presented in Fig. 6. Both CD8 T cell clones (reactive to MART or gp100 melanoma Ags) and treatment TIL produced high titers of IFN-γ following restimulation with A2+ melanoma cell lines or autologous tumor (Fig. 6,A). The background level due to restimulation with HLA-A2 melanoma cell lines was negligible. Addition of TGF-β1 (1 ng/ml) to the cocultures significantly suppressed IFN-γ and GM-CSF secretion by both CD8 T cell clones and TIL (Fig. 6, A and B). TNF-α secretion was also reduced by TGF-β1; however, its impact was more significant on this particular TIL (74% suppression) compared with CD8 clones (27 and 28% suppression) (Fig. 6,C). These experiments were repeated with different CD8 clones, and treatment TIL from additional patients and percent suppression for all three effector cytokines were calculated from a total of four CD8 clones and TIL used for treatment of four melanoma patients (Fig. 7). Although individual responses to tumor Ags are variable among patients, TGF-β1 consistently suppressed cytokine secretion by activated CD8 CTL clones and tumor-reactive TIL examined in this study.

FIGURE 6.

TGF-β1 suppression of effector cytokines secreted by melanoma-reactive CD8 T cell clones and TIL. Human CD8 T cell clones reactive to MART-1 or gp100 melanoma Ags or ex vivo-expanded TIL from a patient before adoptive cell transfer were cocultured with the HLA-A2+ melanoma cell line (MEL-526) or the HLA-A2 melanoma cell line (MEL-888; data not shown) or an autologous tumor for 24 h in the presence or absence of 1 ng/ml TGF-β1. Cell-free supernatants for IFN-γ (A) were analyzed using ELISA, and for GM-CSF (B) and TNF-α (C), cell-free supernatants were analyzed by proteomic arrays. Nonspecific cytokine secretion was undetectable.

FIGURE 6.

TGF-β1 suppression of effector cytokines secreted by melanoma-reactive CD8 T cell clones and TIL. Human CD8 T cell clones reactive to MART-1 or gp100 melanoma Ags or ex vivo-expanded TIL from a patient before adoptive cell transfer were cocultured with the HLA-A2+ melanoma cell line (MEL-526) or the HLA-A2 melanoma cell line (MEL-888; data not shown) or an autologous tumor for 24 h in the presence or absence of 1 ng/ml TGF-β1. Cell-free supernatants for IFN-γ (A) were analyzed using ELISA, and for GM-CSF (B) and TNF-α (C), cell-free supernatants were analyzed by proteomic arrays. Nonspecific cytokine secretion was undetectable.

Close modal
FIGURE 7.

Percent suppression of effector cytokines by TGF-β1. Four CD8 clones reactive to either gp100 or MART-1 Ag and four TIL that were expanded ex vivo before adoptive cell transfer into melanoma patients were cocultured with MEL-526 with or without 1 ng/ml TGF-β1 as described in Fig. 6. The percent suppression for IFN-γ (A), GM-CSF (B), and TNF-α (C) were calculated along with SE.

FIGURE 7.

Percent suppression of effector cytokines by TGF-β1. Four CD8 clones reactive to either gp100 or MART-1 Ag and four TIL that were expanded ex vivo before adoptive cell transfer into melanoma patients were cocultured with MEL-526 with or without 1 ng/ml TGF-β1 as described in Fig. 6. The percent suppression for IFN-γ (A), GM-CSF (B), and TNF-α (C) were calculated along with SE.

Close modal

Although immunization of melanoma patients with tumor Ags can result in the generation of a high frequency of tumor-reactive CD8 T cells, it rarely leads to tumor regression. This lack of clinical effectiveness may in part result from a functional suppression of T cells by the immunosuppressive cytokine milieu present at the tumor site (33) because many tumors, including melanoma, can produce TGF-β1 (34, 35, 36). In the present study, we report that TGF-β1 attenuates the acquisition of effector function by human tumor Ag-specific memory CD8 T cells and suppresses the effector function of fully activated effector memory CD8 T cells, as well as TIL and CTL clones. Our results lend strong support to the premise that antitumor CD8 T cell responses in cancer patients can be hampered by TGF-β1.

We demonstrated that Ag activation of memory CD8 T cells in the presence of TGF-β1 leads to the phenotypic changes attributed to recently activated effector memory cells, yet these cells are functionally impaired. These findings suggest that TGF-β-mediated suppression primarily impacts the mediators of effector function and not the mediators of phenotypic changes, and the mechanism for gain of effector function is uncoupled from the mechanism regulating the induction of phenotypic changes. Thus, these results provide a plausible explanation for the paradigm in which phenotypically activated CD8 T cells are found within a growing tumor.

A mechanism by which TGF-β prevents differentiation of T cells has been reported in murine naive CD4 T cells (11, 12). However, the mechanism of suppression for memory T cells both in murine and human had not been addressed previously. In this study, we showed that TGF-β1 suppressed T-bet expression, which can correlate with the lack of effector function in these memory CD8 T cells. Whether induction of T-bet expression in memory T cells can restore their effector function and overcome the suppressive impact of TGF-β remains unresolved.

In addition to inhibiting the gain of effector function by memory CD8 T cells, TGF-β1 significantly suppressed IFN-γ and GM-CSF secretion by fully activated Ag-specific effector memory CD8 T cells from PBMC, CTL clones, and treatment TIL in 24-h Ag restimulation cultures. Our results demonstrate for the first time that TGF-β1 can suppress the effector function of TIL used for adoptive transfer into melanoma patients by ∼50% during the initial 24 h. Thus, we hypothesize that TGF-β1 can contribute to functional suppression of adoptively transferred TIL in immunotherapy patients. However, it is not surprising that the percent suppression on activated effector memory cells (39%) was significantly (p2 < 0.001) less than memory CD8 T cells (72.8%) because IFN-γ can negate the inhibitory effect of TGF-β (37).

In this study, the TGF-β-mediated suppression of gain of effector function by memory CD8 T cells was mediated in the presence of exogenously added IL-2 (300 IU/ml). Because the increased frequency in Ag-specific CD8 memory T cells was similar in both activation cultures regardless of the addition of TGF-β1 (Table I), we suggest that TGF-β-mediated suppression of acquisition of effector function is uncoupled from the suppressive effect of TGF-β on proliferation as previously reported for murine naive CD4 T cells (8). Interestingly, Wahl and colleagues (38) recently reported that activation of murine naive CD4 T cells in the presence of TGF-β induced Foxp3-expressing regulatory T cells with suppressive function. Whether human memory CD8 T cells activated under similar conditions can likewise adopt a suppressive function remains unresolved now.

The functional relevance associated with the phenotype of TGF-β/Ag-activated CD8 memory T cells is not clear. Our observation on the induction of CCR7 surface expression on memory CD8 T cells by TGF-β1 is consistent with an earlier report by Lanzavecchia and colleagues (39), demonstrating that TGF-β enhanced mRNA levels of CCR7 in human Th1 and Th2 cell lines. It is interesting to speculate that re-expression of CCR7 may allow memory cells to traffic to secondary lymphoid organs because CCR7 expression is correlated with homing and migration to secondary lymphoid tissues (40, 41). A role for TGF-β1 in the generation of the central memory subset has been suggested (42). Given our results, it is possible that TGF-β1 may play a role in the generation and maintenance of central memory T cell pool through its inhibition of effector T cell differentiation and up-regulation of CCR7, which is expressed on central memory CD4 and CD8 T cell subsets (43). In contrast, the inhibition of effector function can contribute to the suppression of an effective antitumor immune response as also suggested by others (42, 44). Therefore, TGF-β1 is a double-edge sword with potent immunoregulatory function.

In summary, we conclude that TGF-β1 impairs the acquisition of effector function by Ag-specific memory CD8 T cells, providing a plausible contributory factor for the inadequacy of current therapeutic cancer vaccines to achieve clinical objectives. Furthermore, TGF-β1 suppresses the antitumor effector function by activated memory CD8 T cells, CD8 clones, and treatment TIL. To the best of our knowledge, this is the first demonstration of the suppressive ability of TGF-β1 on activated effector cells that require minimal activation to exhibit their effector function. Taken together, our results indicate that TGF-β1 leads to functional inactivation of potentially tumor-reactive CD8 T cells as a mechanism of immune evasion in cancer. Therefore, effectiveness of cancer vaccination and immunotherapies may depend on development of strategies to overcome the TGF-β-mediated immunosuppression.

We thank Dr. Nicholas P. Restifo for critical reading of this manuscript, Arnold Mixon and Shawn Farid for flow cytometry analyses, Dr. John Wunderlich for sharing his expertise and reagents, Dr. Udai Kammula for RT-PCR analysis, and Genzyme for providing 1D11 mAb.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

2

Abbreviations used in this paper: TIL, tumor-infiltrating lymphocyte; PI, propidium iodide; FSC, forward scatter; MFI, mean fluorescence intensity.

1
Letterio, J. J., A. B. Roberts.
1998
. Regulation of immune responses by TGF-β.
Annu. Rev. Immunol.
16
:
137
.
2
Shull, M. M., I. Ormsby, A. B. Kier, S. Pawlowski, R. J. Diebold, M. Yin, R. Allen, C. Sidman, G. Proetzel, D. Calvin, et al
1992
. Targeted disruption of the mouse transforming growth factor-β 1 gene results in multifocal inflammatory disease.
Nature
359
:
693
.
3
Kulkarni, A. B., C. G. Huh, D. Becker, A. Geiser, M. Lyght, K. C. Flanders, A. B. Roberts, M. B. Sporn, J. M. Ward, S. Karlsson.
1993
. Transforming growth factor β 1 null mutation in mice causes excessive inflammatory response and early death.
Proc. Natl. Acad. Sci. USA
90
:
770
.
4
Gorelik, L., R. A. Flavell.
2000
. Abrogation of TGF-β signaling in T cells leads to spontaneous T cell differentiation and autoimmune disease.
Immunity
12
:
171
.
5
Lucas, P. J., S. J. Kim, S. J. Melby, R. E. Gress.
2000
. Disruption of T cell homeostasis in mice expressing a T cell-specific dominant negative transforming growth factor β II receptor.
J. Exp. Med.
191
:
1187
.
6
Kehrl, J. H., L. M. Wakefield, A. B. Roberts, S. Jakowlew, M. Alvarez-Mon, R. Derynck, M. B. Sporn, A. S. Fauci.
1986
. Production of transforming growth factor β by human T lymphocytes and its potential role in the regulation of T cell growth.
J. Exp. Med.
163
:
1037
.
7
Brabletz, T., I. Pfeuffer, E. Schorr, F. Siebelt, T. Wirth, E. Serfling.
1993
. Transforming growth factor β and cyclosporin A inhibit the inducible activity of the interleukin-2 gene in T cells through a noncanonical octamer-binding site.
Mol. Cell. Biol.
13
:
1155
.
8
Sad, S., T. R. Mosmann.
1994
. Single IL-2-secreting precursor CD4 T cell can develop into either Th1 or Th2 cytokine secretion phenotype.
J. Immunol.
153
:
3514
.
9
Swain, S. L., G. Huston, S. Tonkonogy, A. Weinberg.
1991
. Transforming growth factor-β and IL-4 cause helper T cell precursors to develop into distinct effector helper cells that differ in lymphokine secretion pattern and cell surface phenotype.
J. Immunol.
147
:
2991
.
10
Ranges, G. E., I. S. Figari, T. Espevik, M. A. Palladino, Jr.
1987
. Inhibition of cytotoxic T cell development by transforming growth factor β and reversal by recombinant tumor necrosis factor α.
J. Exp. Med.
166
:
991
.
11
Gorelik, L., S. Constant, R. A. Flavell.
2002
. Mechanism of transforming growth factor β-induced inhibition of T helper type 1 differentiation.
J. Exp. Med.
195
:
1499
.
12
Gorelik, L., P. E. Fields, R. A. Flavell.
2000
. Cutting edge: TGF-β inhibits Th type 2 development through inhibition of GATA-3 expression.
J. Immunol.
165
:
4773
.
13
Szabo, S. J., S. T. Kim, G. L. Costa, X. Zhang, C. G. Fathman, L. H. Glimcher.
2000
. A novel transcription factor, T-bet, directs Th1 lineage commitment.
Cell
100
:
655
.
14
Zheng, W., R. A. Flavell.
1997
. The transcription factor GATA-3 is necessary and sufficient for Th2 cytokine gene expression in CD4 T cells.
Cell
89
:
587
.
15
Derynck, R., J. A. Jarrett, E. Y. Chen, D. H. Eaton, J. R. Bell, R. K. Assoian, A. B. Roberts, M. B. Sporn, D. V. Goeddel.
1985
. Human transforming growth factor-β complementary DNA sequence and expression in normal and transformed cells.
Nature
316
:
701
.
16
Derynck, R., D. V. Goeddel, A. Ullrich, J. U. Gutterman, R. D. Williams, T. S. Bringman, W. H. Berger.
1987
. Synthesis of messenger RNAs for transforming growth factors α and β and the epidermal growth factor receptor by human tumors.
Cancer Res.
47
:
707
.
17
Bennicelli, J. L., D. T. Guerry.
1993
. Production of multiple cytokines by cultured human melanomas.
Exp. Dermatol.
2
:
186
.
18
Siegel, P. M., J. Massague.
2003
. Cytostatic and apoptotic actions of TGF-β in homeostasis and cancer.
Nat. Rev. Cancer
3
:
807
.
19
Derynck, R., R. J. Akhurst, A. Balmain.
2001
. TGF-β signaling in tumor suppression and cancer progression.
Nat. Genet.
29
:
117
.
20
Torre-Amione, G., R. D. Beauchamp, H. Koeppen, B. H. Park, H. Schreiber, H. L. Moses, D. A. Rowley.
1990
. A highly immunogenic tumor transfected with a murine transforming growth factor type β 1 cDNA escapes immune surveillance.
Proc. Natl. Acad. Sci. USA
87
:
1486
.
21
Kao, J. Y., Y. Gong, C. M. Chen, Q. D. Zheng, J. J. Chen.
2003
. Tumor-derived TGF-β reduces the efficacy of dendritic cell/tumor fusion vaccine.
J. Immunol.
170
:
3806
.
22
Bollard, C. M., C. Rossig, M. J. Calonge, M. H. Huls, H. J. Wagner, J. Massague, M. K. Brenner, H. E. Heslop, C. M. Rooney.
2002
. Adapting a transforming growth factor β-related tumor protection strategy to enhance antitumor immunity.
Blood
99
:
3179
.
23
Gorelik, L., R. A. Flavell.
2001
. Immune-mediated eradication of tumors through the blockade of transforming growth factor-β signaling in T cells.
Nat. Med.
7
:
1118
.
24
Rosenberg, S. A., J. C. Yang, D. J. Schwartzentruber, P. Hwu, F. M. Marincola, S. L. Topalian, N. P. Restifo, M. E. Dudley, S. L. Schwarz, P. J. Spiess, et al
1998
. Immunologic and therapeutic evaluation of a synthetic peptide vaccine for the treatment of patients with metastatic melanoma.
Nat. Med.
4
:
321
.
25
Dudley, M. E., J. Wunderlich, M. I. Nishimura, D. Yu, J. C. Yang, S. L. Topalian, D. J. Schwartzentruber, P. Hwu, F. M. Marincola, R. Sherry, S. F. Leitman, S. A. Rosenberg.
2001
. Adoptive transfer of cloned melanoma-reactive T lymphocytes for the treatment of patients with metastatic melanoma.
J. Immunother.
24
:
363
.
26
Rosenberg, S. A., J. R. Yannelli, J. C. Yang, S. L. Topalian, D. J. Schwartzentruber, J. S. Weber, D. R. Parkinson, C. A. Seipp, J. H. Einhorn, D. E. White.
1994
. Treatment of patients with metastatic melanoma with autologous tumor-infiltrating lymphocytes and interleukin 2.
J. Natl. Cancer Inst.
86
:
1159
.
27
Topalian, S. L., D. Solomon, S. A. Rosenberg.
1989
. Tumor-specific cytolysis by lymphocytes infiltrating human melanomas.
J. Immunol.
142
:
3714
.
28
Powell, D. J., Jr, S. A. Rosenberg.
2004
. Phenotypic and functional maturation of tumor antigen-reactive CD8+ T lymphocytes in patients undergoing multiple course peptide vaccination.
J. Immunother.
27
:
36
.
29
Nielsen, M. B., V. Monsurro, S. A. Migueles, E. Wang, A. Perez-Diez, K. H. Lee, U. Kammula, S. A. Rosenberg, F. M. Marincola.
2000
. Status of activation of circulating vaccine-elicited CD8+ T cells.
J. Immunol.
165
:
2287
.
30
Dutton, R. W., L. M. Bradley, S. L. Swain.
1998
. T cell memory.
Annu. Rev. Immunol.
16
:
201
.
31
Dudley, M. E., J. R. Wunderlich, P. F. Robbins, J. C. Yang, P. Hwu, D. J. Schwartzentruber, S. L. Topalian, R. Sherry, N. P. Restifo, A. M. Hubicki, et al
2002
. Cancer regression and autoimmunity in patients after clonal repopulation with antitumor lymphocytes.
Science
298
:
850
.
32
Dudley, M. E., J. R. Wunderlich, T. E. Shelton, J. Even, S. A. Rosenberg.
2003
. Generation of tumor-infiltrating lymphocyte cultures for use in adoptive transfer therapy for melanoma patients.
J. Immunother.
26
:
332
.
33
Rosenberg, S. A..
2001
. Progress in human tumour immunology and immunotherapy.
Nature
411
:
380
.
34
Rodeck, U., A. Bossler, U. Graeven, F. E. Fox, P. C. Nowell, C. Knabbe, C. Kari.
1994
. Transforming growth factor β production and responsiveness in normal human melanocytes and melanoma cells.
Cancer Res.
54
:
575
.
35
Conrad, C. T., N. R. Ernst, W. Dummer, E. B. Brocker, J. C. Becker.
1999
. Differential expression of transforming growth factor β 1 and interleukin 10 in progressing and regressing areas of primary melanoma.
J. Exp. Clin. Cancer Res.
18
:
225
.
36
Krasagakis, K., D. Tholke, B. Farthmann, J. Eberle, U. Mansmann, C. E. Orfanos.
1998
. Elevated plasma levels of transforming growth factor (TGF)-β1 and TGF-β2 in patients with disseminated malignant melanoma.
Br. J. Cancer
77
:
1492
.
37
Ulloa, L., J. Doody, J. Massague.
1999
. Inhibition of transforming growth factor-β/SMAD signalling by the interferon-γ/STAT pathway.
Nature
397
:
710
.
38
Chen, W., W. Jin, N. Hardegen, K. J. Lei, L. Li, N. Marinos, G. McGrady, S. M. Wahl.
2003
. Conversion of peripheral CD4+CD25 naive T cells to CD4+CD25+ regulatory T cells by TGF-β induction of transcription factor Foxp3.
J. Exp. Med.
198
:
1875
.
39
Sallusto, F., D. Lenig, C. R. Mackay, A. Lanzavecchia.
1998
. Flexible programs of chemokine receptor expression on human polarized T helper 1 and 2 lymphocytes.
J. Exp. Med.
187
:
875
.
40
Campbell, J. J., E. C. Butcher.
2000
. Chemokines in tissue-specific and microenvironment-specific lymphocyte homing.
Curr. Opin. Immunol.
12
:
336
.
41
Forster, R., A. Schubel, D. Breitfeld, E. Kremmer, I. Renner-Muller, E. Wolf, M. Lipp.
1999
. CCR7 coordinates the primary immune response by establishing functional microenvironments in secondary lymphoid organs.
Cell
99
:
23
.
42
Gorelik, L., R. A. Flavell.
2002
. Transforming growth factor-β in T-cell biology.
Nat. Rev. Immunol.
2
:
46
.
43
Sallusto, F., D. Lenig, R. Forster, M. Lipp, A. Lanzavecchia.
1999
. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions.
Nature
401
:
708
.
44
de Visser, K. E., W. M. Kast.
1999
. Effects of TGF-β on the immune system: implications for cancer immunotherapy.
Leukemia
13
:
1188
.