CD4⁺ CD25⁺ regulatory T cells control the induction of antigen-specific CD4⁺ helper T cell responses in cancer patients

NY-ESO-1 is a germ-cell protein that is often expressed by cancer cells, but not by normal somatic cells.¹  In monitoring a large series of cancer patients, frequent humoral responses to NY-ESO-1 were found and they correlated with NY-ESO-1 expression in tumor cells and with the presence of peripheral CD8⁺ T cells against NY-ESO-1,^2,3  suggesting the involvement of CD4⁺ helper T cells in coordinating this response. In fact, we have confirmed that NY-ESO-1–specific CD4⁺ T helper cells are only found in NY-ESO-1 seropositive patients, but not in seronegative patients or healthy individuals.⁴ 

Naturally occurring CD4⁺CD25⁺ regulatory T cells play an important role in maintaining immunologic balance in hosts by suppressing a wide variety of immune responses.^5-8  Although this T-cell population was originally found to suppress the development of autoimmunity,⁵  it has been shown that depletion of this regulatory T-cell population by antibody enhances the antitumor immune responses^9,10  and that stimulation of CD4⁺CD25⁺ regulatory T cells by immunization with self-antigens exacerbates tumor growth in mouse studies.¹¹  In human cancer settings, CD4⁺CD25⁺ regulatory T cells exist at the local tumor site and contribute to growth of tumors in vivo and are associated with an unfavorable prognosis.^12-15 

In contrast to our previous data, it has recently been reported that autoreactive T-cell precursors to NY-ESO-1 are present in most healthy individuals, and that CD4⁺CD25⁺ regulatory T cells play an important role for controlling their activation.¹⁶  These observations prompted us to extend this analysis to cancer patients and assess the impact of CD4⁺CD25⁺ regulatory T cells on the generation of NY-ESO-1–specific CD4⁺ T-cell responses in relation to NY-ESO-1 expression in the tumor and presence of specific serum antibody.

All healthy donors were subjects with no history of autoimmune disease. All cancer patients had NY-ESO-1 expressing melanoma except patients NW2493 with small cell lung cancer and NW1060 with sarcoma. All samples were collected after informed consent as part of a study approved by the Ethics Committee of Landesärztekammer Hessen, Frankfurt.

Tri-Color–conjugated anti-CD4 and anti-CD45RA and fluorescein isothiocyanate–conjugated anti-CD45RO antibodies were purchased from CALTAG (Burlingame, CA). R-phycoerythrin–conjugated anti-CD25 antibody was purchased from Miltenyi Biotec (Auburn, CA). Synthetic peptides of NY-ESO-1_87-98 (LLEFYLAMPFAT), NY-ESO-1_121-132 (VLLKEFTVSGNI), NY-ESO-1_143-154 (RQLQLSISSCLQ), NY-ESO-1_157-170 (SLLMWITQCFLPVF), and HIV P17_37-51 (ASRELERFAVNPGLL)⁴  were obtained from Bio Synthesis (Lewisville, TX). All peptides were shown to be restricted by HLA-DRB1^*04 or HLA-DRB1^*07, except NY-ESO-1_157-170 restricted by HLA-DPB1^*04.^4,17 

NY-ESO-1–specific CD4⁺ T cells were elicited as described previously with some modifications.^4,16  Briefly, CD4⁺ T cells and CD4⁺CD25^- T cells were isolated from peripheral blood mononuclear cells (PBMCs) using CD4⁺CD25⁺ Regulatory T-Cell Isolation Kit (Miltenyi Biotec). In some experiments, CD4⁺CD25^- T cells (> 96% purity) were further separated into CD45RO-depleted T cells (CD4⁺CD25^-CD45RA⁺ T cells) (> 96% purity) or CD45RA-depleted T cells (CD4⁺CD25^-CD45RO⁺ T cells) (> 93% purity) using CD45RO or CD45RA Microbeads (Miltenyi Biotec), respectively. Antigen-presenting cells (APCs) were prepared from non-CD4⁺ cells by allowing them to adhere to tissue culture plates (Corning Inc, Corning, NY) for 2 hours, removing nonadherent cells, and pulsing them with 10 μM of 1 or 2 peptides overnight. After irradiation, 1 × 10⁵ APCs were added to round-bottom 96-well plates (Corning) containing 5 × 10⁴ unfractionated CD4⁺, CD4⁺CD25^-, CD4⁺CD25^-CD45RO⁺, or CD4⁺CD25^-CD45RA⁺ T cells and were fed with 10 U/mL interleukin (IL)–2 (Roche Molecular Biochemicals, Indianapolis, IN) twice per week.

The number of interferon (IFN)–γ–secreting peptide-specific CD4⁺ T cells was assessed by enzyme-linked immunospot (ELISPOT) assays against phytohaemagglutinin (PHA HA15; Murex Diagnostics, Dartford, United Kingdom)–activated CD4⁺ T cells or Epstein Barr virus–transformed B cells pulsed with 10 μM of peptides overnight, as described previously.^4,18 

Presensitized T cells (5 × 10⁴) were cultured with 5 × 10⁴ irradiated CD3-depleted PBMCs pulsed with 10 μM of peptides overnight in wells of round-bottom 96-well plates. Proliferation was evaluated by pulsing with 0.037 MBq/well (1 μCi/well) [³H]-thymidine for the last 18 hours of a 72-hour culture. [³H]-thymidine incorporation was measured by a scintillation counter.

CD4⁺ T cells and CD4⁺CD25^- T cells were isolated from PBMCs and were cultured with APCs pulsed with a series of HLA class II–restricted NY-ESO-1 peptides reported previously.⁴  Fifteen to 20 days later, NY-ESO-1–specific CD4⁺ helper T-cell induction was analyzed by ELISPOT (Figure 1) and proliferation (Figure S1, available on the Blood website; see the Supplemental Figure link at the top of the online article) assays. IFN-γ–secreting NY-ESO-1–specific CD4⁺ T cells with proliferative capacity, CD4⁺ T helper type 1 (Th1) cells, were elicited in 5 of 8 healthy donors, but only in cultures with CD4⁺CD25⁺ T-cell depletion, confirming a previous report.¹⁶ 

Next, we extended this examination to patients with NY-ESO-1–expressing tumors. Th1 cells were induced in 3 of 3 patients with NY-ESO-1 antibody, without need for CD4⁺CD25⁺ T-cell depletion in accordance with our previous results.⁴  In patients without NY-ESO-1 antibody, although NY-ESO-1–specific Th1 cells were not induced from total CD4⁺ T-cell populations, they could be induced in 3 out of 4 NY-ESO-1 seronegative patients after depletion of CD4⁺CD25⁺ T cells (Figures 1 and S1, summarized in Table 1). Taken together, NY-ESO-1–specific CD4⁺ T-cell precursors are present in a wider range of patients than formerly thought.⁴ 

We next asked whether NY-ESO-1–specific CD4⁺ T-cell precursors were derived from similar or from different populations in patients with or without NY-ESO-1 antibody. To address this question, CD4⁺CD25^- T cells were further separated into naive or effector/memory populations according to typical surface-marker molecules CD45RA and CD45RO, respectively (Figure 2A). In patients with NY-ESO-1 antibody, NY-ESO-1–specific Th1 cells were induced from both CD45RA⁺ and CD45RO⁺ populations. In contrast, NY-ESO-1–specific Th1 cells were apparently derived only from CD45RA⁺ population in patients without NY-ESO-1 antibody and healthy donors (Figure 2B). Thus, responses elicited from depletion of CD4⁺CD25⁺ T cells are derived from a naive repertoire. Patients with preexisting immunity to NY-ESO-1 still retain the capacity to elicit additional responses from naive precursors upon CD4⁺CD25⁺ depletion (as in patient NW2457). Expression of NY-ESO-1 in the tumor is not sufficient to naturally prime NY-ESO-1–specific T-cell responses in patients without NY-ESO-1 antibody, since they have no specific precursors with memory phenotype.

To examine the differences between patients with and without NY-ESO-1 antibody, graded amounts of CD4⁺CD25⁺ T cells were added to cultures for in vitro stimulation. NY-ESO-1–specific Th1 cells were induced from CD4⁺CD25^-CD45RO⁺ T-cell population even in the presence of a high percentage of CD4⁺CD25⁺ regulatory T cells. However, induction of NY-ESO-1–specific Th1 cells from CD4⁺CD25^-CD45RA⁺ T-cell population was essentially abrogated in the presence of CD4⁺CD25⁺ regulatory T cells (Figure 2C). It appears that NY-ESO-1–specific CD4⁺ T-cell precursors present in CD4⁺CD25^-CD45RO⁺ T-cell population of patients with NY-ESO-1 antibody are more resistant to CD4⁺CD25⁺ regulatory T cells.

We have previously reported that CD4⁺ T-cell responses are correlated with NY-ESO-1 expression in tumor cells and NY-ESO-1 serum antibody. In this study, we have shown that NY-ESO-1–specific CD4⁺ T-cell precursors are derived from different populations in patients with or without NY-ESO-1 antibody and that differences in naturally occurring antitumor immunity are defined by distinctive CD4⁺CD25⁺ regulatory T-cell sensitivity of each population. It is still unknown if effector/memory T cells become resistant at a single-cell level or whether a relatively high number of precursors within memory population overwhelm suppression. A critical area for future exploration is to explain why some patients can induce memory NY-ESO-1–specific CD4⁺ T-cell responses even in the presence of naturally occurring CD4⁺CD25⁺ regulatory T cells. It is likely that these patients had an effective priming environment (eg, in which toll-like receptor signaling inhibits the generation/activation of CD4⁺CD25⁺ regulatory T cells)^19,20  during the course of tumor development.

Although it has been shown that presence of CD4⁺CD25⁺ regulatory T cells at the local tumor site contributes to tumor growth and is associated with bad prognosis,¹⁴  there is no evidence for a correlation between human antigen-specific antitumor immunity and CD4⁺CD25⁺ regulatory T cells. Here, we have clearly shown that CD4⁺CD25⁺ regulatory T cells influence the induction of NY-ESO-1–specific antitumor immunity. It has been reported that CD4⁺CD25⁺ regulatory T cells require in vitro stimulation to gain their suppressive activity.^6-8,21  It is possible that CD4⁺CD25⁺ regulatory T cells modulating NY-ESO-1 immunity are activated in an antigen-specific manner, comparable to the specific regulatory T cells generated against NY-ESO-1 gene family member LAGE-1 protein.²²  Alternatively, some CD4⁺CD25⁺ regulatory T-cell populations that do not require further in vitro stimulation may exist and suppress induction of immunity.^23,24 

A unique finding in our study is that CD4⁺ T-cell responses to new NY-ESO-1 epitopes are induced from naive populations following CD4⁺CD25⁺ T-cell depletion, even in patients with preexisting immunity to NY-ESO-1. Our data provide an important hint for effective cancer vaccines during antigen priming through down-modulation of CD4⁺CD25⁺ regulatory T-cell function.

We thank Drs E. Sato, T. Kato, and H. Shiku for helpful discussion, and E. Krapivinsky, C. Villalobos, and D. Santiago for technical support.