T_H2 Cytokines from Malignant Cells Suppress T_H1 Responses and Enforce a Global T_H2 Bias in Leukemic Cutaneous T-cell Lymphoma

Cutaneous T-cell lymphomas (CTCL) are a heterogeneous group of non-Hodgkin lymphomas arising from malignant transformation of T cells that home to and populate the skin (1, 2). In leukemic variants of CTCL (L-CTCL), malignant T cells can accumulate in the blood and lymph nodes and also produce widespread inflammatory skin lesions. L-CTCL is often refractory to multiple therapies, and patients often ultimately require hematopoietic stem cell transplantation (3). The median survival for patients with L-CTCL varies with the extent of malignant T-cell burden but is generally 2 to 5 years and patients die most commonly from infection (4–7).

CTCL has been proposed to be a malignancy of 3 separate T-cell populations: FOXP3⁺ regulatory T cell (Treg), T-helper (T_H)-2 T cells, and T_H17 T cells (8–10). Clinically, patients with L-CTCL have abnormalities suggestive of a T_H2-driven immunologic process, including decreased antigen-specific T-cell responses, impaired cell-mediated cytotoxicity, peripheral eosinophilia, and elevated levels of serum immunoglobulin E (IgE) and immunoglobulin A (IgA; refs. 11–13). Prior studies have shown increased levels of T_H2 cytokines and T_H2-associated genes in T cells from patients with L-CTCL (9, 14–17) but a recent report claims that malignant T cells in the disease are actually T_H17 biased (10).

We conducted comprehensive analyses of the cytokine production of benign and malignant T cells from L-CTCL with identifiable malignant T-cell clones. We report here that both malignant and benign T cells in patients with L-CTCL were strongly T_H2 biased, that this bias was intrinsic in malignant T cells but extrinsic in benign cells, and that inhibition of T_H2 cytokines led to recovery of T_H1 responses in benign T cells.

The protocols of this study were conducted in accordance with the Declaration of Helsinki and were approved by the Institutional Review Board of the Partners Human Research Committee (Partners Research Management, Boston, MA). Blood from healthy individuals was obtained as discarded tissue following leukopheresis. Blood and lesional skin from patients with CTCL were obtained from patients seen at the Dana-Farber/Brigham and Women's Cancer Center (Boston, MA) Cutaneous Lymphoma Program. Patients with L-CTCL described in this manuscript met the World Health Organization–European Organisation for Research and Treatment of Cancer criteria for L-CTCL/Sezary Syndrome (18). Patient characteristics are included in Table 1 based on the malignant T-cell burden. In this cohort of patients, expected survival is approximately 5 years (4–7). Peripheral blood mononuclear cell (PBMC) were isolated by ficoll centrifugation, and clonal and nonclonal CD4⁺ T cells from patients with CTCL were isolated using magnetic bead separation (Miltenyi Biotec) after staining with TCR Vβ-specific antibodies (Beckman Coulter Inc.,).

Analysis of T cells was conducted using directly conjugated monoclonal antibodies obtained from BD Biosciences, eBioscience, Biolegend, or R&D Systems. Vβ staining was conducted using the IOtest Beta Mark TCR Vβ Repertoire kit (Beckman Coulter) as per manufacturer's instructions. Isotype-matched negative control antibodies were used to set the gates for positive staining. For analysis of cytokine production, T cells were stimulated with either control medium or 50 ng/mL phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich) and 750 ng/mL ionomycin (Life Technologies) plus 10 μg/mL Brefeldin A (BD) for 4 hours. Cells were surface stained, fixed, permeabilized, stained with anticytokine antibodies, and examined by flow cytometry. Analysis was conducted on Becton Dickinson FACSCanto instruments and data were analyzed using FACSDiva software (V5.1).

The 6-sulfo LacNAc-expressing dendritic cells (slanDC) were isolated as previously described (19). Lipopolysaccharide (LPS; Sigma-Aldrich) was used at 100 ng/mL. For T-cell priming, LPS-stimulated dendritic cells were harvested and cocultured with autologous clonal and nonclonal T cells from patients with CTCL in the presence of 1 ng/mL of staphylococcal enterotoxin B (Sigma-Aldrich). Following 12 days of culture, T cells were restimulated for intracytoplasmatic flow cytometry analysis of cytokine production after treatment with PMA and ionomycin.

A total of 2 × 10⁶/mL PBMC from healthy donors (including lymphocytes, monocytes, basophils, and dendritic cells) were cocultured over 12 days with 1 × 10⁶/mL fluorescently labelled PBMC from patients with L-CTCL in Iscove's modification of Dulbecco's media (Mediatech) supplemented with 20% heat-inactivated fetal calf serum, antibiotics and l-glutamine. 100 IU/mL interleukin (IL)-2 and 15 ng/mL IL-15 (Peprotech) were supplemented every other day. Where indicated, only cell-substrate, but not direct cell–cell interaction was allowed, and Transwell inserts (Corning Inc.) were used to separate L-CTCL from healthy PBMC.

For studies described in Fig. 1, the nonparametric Kruskal–Wallis test and Dunn test were used for comparisons between the 3 groups. For studies described in Figs. 2 to 5, a standard 2-sample Student t test was used. The data were previously tested for normal distribution (Shapiro–Wilk test) and homogeneity of variance. One-way between group ANOVA was used for statistical analysis of differences between 3 or more groups. P values ≤ 0.05 were considered as significant.

CTCL has been proposed to be a malignancy of FOXP3⁺ Tregs, but more recent studies showed that only a subset of patients had malignant T cells with a Treg phenotype (8, 20, 21). We identified 12 patients in whom the malignant T-cell clone could be conclusively identified by staining with commercially available TCR Vβ antibodies and analyzed benign and malignant T cells by flow cytometry (Fig. 1A). This approach allows direct analysis of the phenotype and functional characteristics of the malignant T-cell clone, while at the same time providing similar information about remaining benign T cells.

We studied cytokine production and found that both the benign and malignant T cells in patients with L-CTCL were strongly T_H2 biased. Both benign and malignant T cells from patients with L-CTCL produced markedly higher levels of IL-4 than T cells from healthy individuals (Fig. 1A and B). A mean 42.7% of malignant T cells (SEM 7.8, n = 12) and 28.1% of benign T cells (SEM 7.0, n = 12) produced IL-4 in patients with L-CTCL, compared with 0.65% of T cells from healthy individuals (SEM 0.18, n = 12). IL-13 production was significantly higher in malignant T cells and in the benign T cells of 5 patients, and IL-10 production was also increased in malignant clones. T_H1 responses were globally decreased in both benign and malignant T cells from patients with L-CTCL, with significantly reduced production of IFNγ in both populations (Fig. 1). Production of TNFα was also markedly decreased in both the benign and malignant T cells from patients with L-CTCL compared with normal controls (Fig. 1). In a subset of patients, malignant T-cell clones produced more IL-2 (Fig. 1), consistent with a report that malignant T cells in L-CTCL have a phenotype suggestive of central memory T cells (T_CM), a cell type known to use autocrine production of IL-2 to maintain survival (22, 23). However, given the heterogeneity among patients, this result was not statistically significant.

T_H2 cells can suppress the production of T_H1 cytokines by other T cells in vitro (24, 25). To evaluate whether clonal malignant T cells suppress benign T-cell responses, we separated nonclonal benign from clonal malignant T cells from the blood of patients with L-CTCL using magnetic beads and cultured these T cells separately in the presence of autologous LPS-activated dendritic cell. IFNγ production increased and IL-4 production decreased in nonclonal T cells cultured away from the malignant clone, whereas clonal T cells showed no significant changes in IFNγ or IL-4 production (Fig. 2A and B). Real-time PCR analyses showed that expression of the T_H2-associated transcription factor GATA-3 declined and expression of the T_H1-associated transcription factor t-BET increased in benign T cells after cultured away from the malignant clone (Fig. 2C). In clonal malignant T cells, there was no change in the expression of GATA-3 and t-BET after separate culture. Expression of the T_H17-associated transcription factor RORc was low in both malignant and benign T cells before and after culture. T cells isolated from the skin lesions of patients with L-CTCL had similar responses; IFNγ increased in nonclonal T cells after separate coculture but remained unchanged in malignant T cells (Fig. 2D).

To evaluate the ability of L-CTCL T cells to suppress non-T_H2 cytokine production, we cultured PBMC from healthy individuals with fluorescently labeled PBMC from patients with L-CTCL with a high malignant T-cell burden. Coculture was conducted in direct contact or in Transwell systems, which separated healthy from L-CTCL cells. After coculture with L-CTCL cells, T cells from healthy patients showed reduced production of IFNγ (Fig. 3). Suppression was observed in both direct contact and Transwell cultures, suggesting that a soluble factor was responsible for T_H1 suppression. To determine whether T_H2 cytokines produced by malignant T cells were responsible for the suppression of IFNγ production in healthy T cells, we included neutralizing antibodies to IL-4 and IL-13 in Transwell cocultures of L-CTCL and healthy PBMC. Results showed that neutralization of IL-4 and IL-13 restored the production of IFNγ to baseline levels (Fig. 4).

It has been previously reported that immunomodulatory therapies for CTCL, including IFN-α2b and extracorporeal photopheresis, led to enhanced T_H1 responses in responding patients (26). IFN-α2b and extracorporeal photopheresis, both have immunomodulatory effects beyond their ability to reduce the number of malignant T cells. If T_H2 cytokines produced by malignant T cells are indeed responsible for the decreased T_H1 responses, pruritus, susceptibility to infections, and benign T cell T_H2 skewing observed in these individuals, then a variety of treatment modalities that have in common only the depletion of malignant T cells should all produce enhanced T_H1 responses, regardless of the underlying mechanism of action of the therapy used. To study this issue, we evaluated the cytokine production of benign T cells before and after therapy with a variety of treatment modalities (Fig. 5). We found that successful therapy with variety of modalities, including UVB phototherapy, extracorporeal photopheresis, low-dose alemtuzumab, and systemic chemotherapy with gemcitabine, all had similar effects on the cytokine production of benign T cells. In all therapies and in all patients examined, when the number of circulating malignant T cells declined, benign T cells subsequently produced less IL-4 and more IL-2 and IFNγ (Fig. 5). In addition to the patients shown, similar findings were observed in 6 additional patients treated with a combination of therapies that also included topical corticosteroids, electron beam therapy, and narrow band UVB (data not shown). In all patients included in this analysis, complete remission of at least 4 months was achieved. With respect to the patients shown in Fig. 5, the first patient (Fig. 5A) remains in remission on UVB and topical steroids but photopheresis was discontinued. Following a complete remission with alemtuzumab therapy, the second patient (Fig. 5B) subsequently relapsed with progressive skin disease not controlled by alemtuzumab and ultimately died from progressive disease. The third patient (Fig. 5C) experienced a complete remission following gemcitabine, and underwent stem cell transplantation.

There has been some disagreement about the cytokine production and functional polarization of malignant T cells in L-CTCL. L-CTCL has been proposed to be a uniform malignancy of FOXP3⁺ Tregs, T_H17 cells, and T_H2 cells (8–10). All of the patients in our cohort showed a marked T_H2 bias in malignant T cells, in agreement with earlier studies that the vast majority of patients with L-CTCL have T_H2-biased malignant T cells (9, 14–17, 27). Patients with L-CTCL have characteristics suggestive of a T_H2-driven immunologic process, including elevated levels of IgE and IgA, eosinophilia, and reduced T_H1 responses (11–13). These patients also suffer from marked pruritus of the skin, high rates of colonization with Staphylococcus aureus, and a high susceptibility to cutaneous infections, similar to patients with atopic dermatitis, a prototypic T_H2 disease (28). We did not observe any patients with malignant T cells of a T_H17 phenotype.

Malignant T cells in L-CTCL have a phenotype suggestive of skin tropic T_CM (22). T_CM are highly migratory, recirculate between the blood, skin, and lymph nodes, have a high proliferative potential, are markedly resistant to apoptosis, and have a variety of cytokine production profiles (23). The T_CM phenotype of these cells helps to explain why patients with L-CTCL develop erythroderma, peripheral blood disease, and lymphadenopathy. However, one might expect that malignant T cells would have a heterogeneous pattern of cytokine production that reflects the diverse population of T_CM from which they arise. The fact that the vast majority of patients have strongly T_H2-biased malignant T cells is striking. We found that the T_H2 bias of the malignant T cells was remarkably stable in vitro, suggesting that this bias may result from intrinsic abnormalities of the cell or from autocrine factors, for example, galectin-1 produced by malignant T cells, which has previously been implicated in the T_H2 skewing of the malignant clone (17). It may also be that T_H2 malignant T-cell clones predominate in L-CTCL because they are remarkably effective in suppressing anticancer T_H1 responses and therefore have a survival advantage.

We were surprised to find that the benign T cells remaining in the circulation of patients with L-CTCL were significantly different from those found in normal individuals. There was a striking and consistent T_H2 bias even among the benign T cells in these patients, as well as a reduction in the production of T_H1 cytokines (Fig. 1). These results show that a global T_H2 bias exists in these patients that leads to skewing of the entire T-cell repertoire, both benign and malignant, toward enhanced production of T_H2 cytokines. This overproduction of T_H2 cytokines by both malignant and benign T cells likely contributes to the pruritus, reduced T_H1 responses, and the susceptibility to infections observed in these patients.

The remaining nonclonal benign T cells in patients with L-CTCL would be expected to have a variety of cytokine production capacities. Given the known ability of T_H2 cells to suppress T_H1 responses and to force a T_H2 bias on other T cells (24, 25), we hypothesized that the malignant T cells may suppress the activity of benign T cells, silencing T_H1 responses, as has been suggested previously (26). Indeed, when we separated and cultured nonclonal T cells from the blood and skin lesions of patients away from the malignant clone, we observed improved T_H1 responses (Fig. 2). When we cocultured T cells from healthy individuals with L-CTCL T cells, we saw a marked inhibition of T_H1 responses (Fig. 3, 4). Incubation of cocultures with neutralizing antibodies to IL-4 and IL-13 restored T_H1 responses, suggesting that T_H2 cytokines produced by malignant T cells may be directly responsible for the suppression of T_H1 responses we observed in our patients and in T cells from healthy donors. If so, successful treatment of patients with CTCL should reduce T_H2 and enhance T_H1 responses, as a previous study of 3 patients suggested (26).

To study this question, we evaluated patients with L-CTCL treated with a variety of modalities that had in common only reduction in the number of circulating malignant T cells. Regardless of the underlying mechanism of action of particular therapies, treatments that reduced malignant T-cell burden were invariably associated with enhanced T_H1 responses (Fig. 5). Although circumstantial, our results suggest that T_H2 cytokines produced by malignant T cells play a critical role in suppressing T_H1 responses in our patients.

In summary, we report a marked T_H2 bias in both the benign and malignant T cells in patients with L-CTCL and we find that T_H2 cytokine production by the malignant clone likely suppresses T_H1 responses in these patients. T_H1 responses improved in vitro when the benign T cells were cultured away from the malignant clone, when IL-4 and IL-13 activities were blocked in vitro with neutralizing antibodies, and in patients themselves after treatment with therapies that reduced the number of circulating malignant T cells. The suppression of T_H1 responses in these patients by IL-4 and IL-13 suggests that neutralization of these cytokines may be beneficial in enhancing immune responses. Aeroderm, a recombinant small protein antagonist of the IL-4/IL-13 receptor formulated for skin injection, was previously in clinical trials for the treatment of severe atopic dermatitis (29). By inhibiting T_H2 responses and enhancing T_H1 responses, IL-4/IL-13 antagonists have the potential to enhance immune responses to both skin pathogens and to the malignant T cells themselves.

T.S. Kupper is a consultant/advisory board member of TremRx. No potential conflicts of interest were disclosed by the other authors.

Conception and design: E. Guenova, R. Watanabe, T.S. Kupper, R.A. Clark

Development of methodology: E. Guenova, R. Watanabe, R.A. Clark

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E. Guenova, R. Watanabe, J.A. Desimone, J. Teague, Y. Jiang, M. Dowlatshahi, C. Schlapbach, K. Schaekel, M.C. Tawa, D.E. Fisher, T.S. Kupper, R.A. Clark

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E. Guenova, R. Watanabe, M. Dowlatshahi, A.H. Rook, R.A. Clark

Writing, review, and/or revision of the manuscript: E. Guenova, R. Watanabe, J. Teague, D.E. Fisher, R.A. Clark

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.A. Desimone, Y. Jiang, K. Schaekel, A.H. Rook, D.E. Fisher

Study supervision: R.A. Clark

The authors thank the patients who donated the blood and skin samples that made this work possible.

This work was supported by a generous charitable contribution from Edward P. Lawrence, Esq., a Damon Runyon Clinical Investigator Award (to R.A. Clark) R01 AR056720 (to R.A. Clark) and R03 MH095529 (to R.A. Clark), the SPORE in Skin Cancer P50 CA9368305 NIH/National Cancer Institute (to T.S. Kupper), R01 A1025082 NIH/National Institute of Allergy and Infectious Diseases (to T.S. Kupper), R01 AI097128 (to T.S. Kupper and R.A. Clark), R01 CA122569 (to A.H. Rook), and by a Special Fellow Award from the Leukemia & Lymphoma Society (to R. Watanabe). Grant support GU1271/2-1 (to E. Guenova) was provided by the German Research Foundation (DFG), and salary support for C. Schlapbach was provided by the Swiss National Science Foundation and the Fondation Rene Touraine.

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.