The Role of Tumor Microenvironment in Mycosis Fungoides and Sézary Syndrome

Mycosis fungoides (MF) is the most common subtype of cutaneous T-cell lymphoma (CTCL) that accounts for approximately 60% of CTCL1. The clinical manifestations of MF vary in different stages. Skin lesions may present with scaly patches or plaques in early stage but gradually formed generalized plaques, tumors or erythroderma in the advanced stage. Sézary syndrome (SS) is an aggressive CTCL variant that is characterized by leukemic involvement, which commonly present with widespread erythema2. MF and SS may mimic benign inflammatory diseases, such as eczema, psoriasis, parapsoriasis, vitiligo, folliculitis, and pigmented purpuric dermatoses, making it challenging to make accurate diagnosis3.

There is still some controversy in the relationship between MF and SS. In the past, SS was considered to be an aggressive variant of MF. However, growing evidence from gene expression profiles, microRNA profiling, and cell marker studies suggests that they may be separate entities with different pathogeneses 4, 5, 6, 7, 8, 9, 10. In recent years, genome sequencing technologies have identified some chromosomal abnormalities that have occurred in MF and SS4, 5, 6, 7. There have been losses on chromosomes 1, 5, 9, and 13 and gains on chromosomes 7 and 17 for MF, and losses within chromosome 10 and gains within chromosomes 8 and 17 in SS. One study reported miRNA expression of MF and made a comparison with SS profiles8. The results showed that there is no common upregulated miRNA between the tumor stages MF and SS, and they share only 2 of the same downregulated miRNAs, indicating a difference in miRNA expression between MF and SS.

The pathogenic mechanisms underlying oncogenic transformation include not only intrinsic factors such as tumor gene mutation but also extrinsic events in which environmental factors play critical roles. Within patches, plaques, and tumoral CTCL lesions, nonneoplastic T cells such as macrophages, fibroblasts, dendritic cells (DCs), mast cells (MCs), myeloid-derived suppressor cells (MDSCs) and other cells or cytokines comprise the tumor microenvironment (TME). An increasing amount of evidence suggests that the TME plays important roles in tumor occurrence and development of MF and SS, both of which are regarded as Th2-dominant diseases9. The main role of the TME in the development of MF/SS can be divided into two aspects: one is to support tumor growth and survival, and the other is to suppress antitumoral immune response. In this review, we will focus on chemokines and various cell in TME that affect the characteristic features of MF and SS. Understanding the interactions between malignant tumor cells and TME will help to find new treatments targeted at the microenvironment components for the disease.

Chemokines and their receptors are widely known to play an important role in directing leukocyte movement, facilitating both innate and adaptive immune responses and contributing to the pathogenesis of various diseases, including CTCL10. Malignant T cells in CTCL usually present mature CD4+ T cell phenotypes, accompanied by Th2 or Th17 phenotypes or a T regulatory cells (Tregs) phenotype11, each of which can express different receptors that react differently with the same chemokines.

Chemokine receptor expression on tumor cells differs according to different stages of MF. During early stage, the majority of malignant lymphocytes express CXCR312, 13, an interferon (IFN)-inducible chemokine receptor, which reacts with chemokine CXCL9-11, which are all induced by the Th1 cytokine IFN-γ14. There are numerous cells that express the above chemokines, such as keratinocytes, Langerhans cells and fibroblasts, which are mainly distributed in the epidermis and the superficial dermis. The interaction between the above chemokines and their ligands are potentially important for the epidermotropic phenomenon in early stage of MF15.

A shift from Th1 to Th2 cytokines during the development and progression of MF has been identified. In advanced stage of MF, as malignant T cells tend to express more CCR4 than CXCR3, epidermotropism of tumor cells becomes less evident and tend to infiltrate the dermis15. CCR4 is regarded as a Th2-dominant receptor. Sugaya et al.16 discovered that CCR4 expression is seen in infiltrating malignant T cells in all stages of MF and SS, but the frequency of CCR4 expression is positively related to the tumor stage progression. CCR4 is the receptor for CCL17 and CCL22, where CCL17 is mainly expressed by endothelial cells and CCL22 is primarily secreted by DCs in dermis17, and it has been shown that DCs may attract T cells for cell-cell antigen stimulation of T cells via CCR4 interactions18.

CCR4 represents a promising target for CTCL treatment19. Clinical trials that used mogamulizumab, an anti-CCR4 monoclonal antibody, demonstrated encouraging results in advanced MF patients20, 21, 22. In addition, CCR4 expression was also found in circulating malignant T cells and skin tumor cells in SS patients. Fierro et al.23 found that the leukemic CD4+CD26-subpopulation in SS selectively expressed CCR4 at a high level without an increase in Th1 chemokines such as CXCR3.

CCR10 is overexpressed on infiltrating tumor cells, as well as in the peripheral blood of MF24, 25, while CCL27 is a ligand for CCR10 and is mainly produced by activated keratinocytes. CCL27 is able to selectively attract memory T cells that express CCR10 into inflammatory sites25, 26. Therefore, the CCR10 and CCL27 interaction explains the preferential migration and retention of tumor cells within the skin in MF/SS and in other cancers such as melanoma27.

CCR7 is a lymph node-homing chemokine receptor that has been found to be expressed in the peripheral blood of SS, playing an important role in the pathogenesis of SS28. Picchio et al.29 found that CXCL13 expression is increased in skin lesions and in the lymph nodes of SS patients, while CXCL13 is a selective agonist for CCR7. The addition of CXCL13 to CCL19 or CCL21 strongly enhances the migration of CCR7+SS cells. In addition, the neutralization of the CCR7 receptor causes the impaired chemotaxis of SS cells.

CXCR4 usually combines with CXCL12 specifically, inducing chemotaxis of lymphocytes to inflammatory sites, thus helping to enhance chemotaxis for cancer cells. Daggett et al.30 found that CXCL12 is highly expressed on skin tumor cells MF, demonstrating that the interaction between CXCL12 and CXCR4 is important in the progression of MF. Furthermore, the CXCL12 and CXCR4 interaction also regulates tumor angiogenesis by promoting the production of vascular endothelial growth factor (VEGF)31.

CCL18 was frequently expressed by macrophages and DCs in the dermis in CTCL. Miyagaki et al.32 discovered that circulating CCL18 levels in the peripheral blood of CTCL patients were higher than that of normal controls. They also found that high serum levels of CCL18 often correlated with a poor prognosis. In addition, the interleukin (IL)-4-induced CCL18 transcription was inhibited by the Th1 cytokine IFN-γ, indicating that CCL18 is related to the Th2-dominant TME.

MF and SS are considered to be Th2-type diseases and there is a shift in cytokine expression from Th1 to Th2 profile during the disease progression. Early studies showed that Th1 cytokines, which include IFN-γ, IL-2 and IL-12, comprise a majority of the cytokines expressed in early stage of MF33, while in late stage, tumor cells in MF and SS lesions present with a Th2 phenotype, accompanied with an elevated expression of Th2-prone cytokines such as IL-4, IL-5, IL-10, and IL-1334, 35, 36. By using T-bet as a specific maker for Th1 and GATA3 as a marker for Th2, Hsi et al.37 found that in early stage of MF, T-bet expression was increased and became dominant but as the disease progressed to advanced stage, GATA3 expression gradually became more pronounced and eventually occupied predominance.

Th2 cytokines are able to inhibit Th1 responses and promote a Th2-dominant TME in CTCLs38, making it a possible anticancer therapy by suppressing Th2 cytokines expression and enhancing Th1 response. IFN-γ produced by Th1 cells is capable of improving antitumor responses and has proved to be an effective treatment for CTCL39, 40. Recently, Vieyra-Garcia et al.41 found that nontumor T cells in TME were related to the Th2-recruiting chemokines before PUVA treatment, while Th1-recruiting chemokines were observed after phototherapy, suggesting that PUVA may contribute to a change from the Th2 to the Th1 phenotype in MF.

Tumor-associated macrophages (TAMs) are a critical part of the primary immune response in TME. The macrophages can be divided into two types: one is an antitumorigenic cell (M1) and the other is a protumorigenic cell (M2)42. During the late stage of CTCL, TAMs usually exhibit an M2 phenotype, along with a decreased level of IL-12 but an increased expression of IL-10. TAMs can also produce abundant vascular endothelial growth factor (VEGF), which can drive angiogenesis to support tumor growth and development43. Recent clinical histological studies have demonstrated abundant macrophage infiltration in MF patients. Sugaya et al.44 demonstrated increased CD163+ TAM cells in lesions of CTCL patients compared with healthy individuals. In CTCL, the numbers of CD163+ or CD68+ cells correlated with a poor patient outcome. Moreover, TAMs decreased after effective topical steroid and ultraviolet light treatment, linking TAMs with tumor pathogenesis. Besides, Furudate et al.45 also found that CD206+ macrophages are increased in parallel with tumor progression in MF.

TAMs are capable of regulating tumor progression by producing chemokines. Furudate et al.46 found a decreased expression of CCL17 and CCL18, as well as an increased production of CXCL10 and CXCL11 by M2 macrophages after IFN-α2a and IFN-γ treatment. Furudate et al.45 suggested that increased periostin in the stromal area may promote M2 polarization and TAM recruitment in the early stage of MF. Tanita et al.47 studied the immunomodulatory effects of bexarotene on TAMs in MF patients and results suggested that the clinical benefits are partially attributed to suppressive effects on the production of CCL22 by M2-polarized TAMs. These data suggest that TAM could be a therapeutic target for the treatment of MF.

A phase II clinical trial48 confirmed the clinical effect and safety of a CD30 antibody, brentuximab vedotin, in the treatment of MF and SS. Moreover, the results showed that CD30 was coexpressed in TAMs as well as in tumor cells, indicating that the CD30 antibody may disrupt tumor progression by targeting not only malignant T cells but also TAMs. Wu et al.49 studied the role of macrophages in a CTCL murine model and found that depletion of M2-like TAMs delayed murine T-cell lymphoma tumor development. Recently, by using the same model, Wu et al.50 detected an increased IL-10 level in TME. Compared to wild-type mice, xenografted cutaneous lymphoma in IL-10 knockout mice showed a smaller tumor size and a decreased number of infiltrating microphages, suggesting a critical role of TAMs in pathogenesis and treatment of CTCL. In addition, tumor cells express CD47 to bind signal-regulatory protein on TAMs, and this interaction is able to send a “do-not-eat-me” signal from the tumor cells and thus helps cancer cells to escape from phagocytosis51, 52.

Cancer-associated fibroblasts (CAFs) are a group of fibroblasts that are closely related to tumor progression and migration, mainly by secreting certain specific chemokines and cytokines in TME53. In CTCL, CAFs are related to the change from the Th1- to the Th2-dominant TME along with the tumor progression. Miyagaki et al.54 found that in early stage, dermal fibroblasts expressed high levels of herpesvirus entry mediator (HVEM), CXCL9, CXCL10, and CXCL11, with Th1 cells expressing IFN-γ and LIGHT (a ligand for HVEM), while in late stage, CAFs expressed less HVEM and increased Th2 cells' expression of LIGHT only and not IFN-γ. CCL17 was widely expressed by fibroblasts and DCs, while CXCL9-11 expression was decreased.

Periostin, a glycoprotein that is predominantly secreted by dermal fibroblasts, is able to induce chronic inflammatory responses by stimulating thymic stromal lymphopoietin and then act on DCs to induce a Th2-dominant TME in CTCL55, 56. Takahashi et al.56 identified periostin expression in dermal fibroblasts stimulated by IL-4 or IL-13 in vitro and ex vivo and found that compared to the healthy controls, CAFs expressed significantly increased periostin in CTCL skin lesions. In addition, higher serum periostin levels were also observed in CTCL patients.

CCR3 is the only receptor for eotaxins and is mainly expressed on Th2 cells and eosinophils. Miyagaki et al.57 found a higher expression of eotaxin-3 by fibroblasts from the lesional skin in CTCL than from the normal skin and that an increased expression of eotaxin-3 and CCR3 mRNA was related to a more advanced stage. The study above indicates that eotaxins produced by fibroblasts may interact with CCR3 expressed by Th2 cells and then help to form a Th2-dominant TME in the pathogenesis of CTCL.

DCs enable primary immune responses and play a crucial part in TME. In CTCL, proliferating malignant T cells are in close association with immature DCs. Berger et al.58 reported that CTCL cells appear to retard DC maturation, and when cocultured with immature DCs, CTCL cells can be reproducibly grown in culture.

It is assumed that mature DCs take part in antitumor immune responses while immature DCs induce tumor tolerance59. There are three different DC subsets that are discerned by their receptor expression pattern, which include Langerhans cells, dermal DCs and plasmacytoid DCs. Schwingshackl et al.59 found that all three subsets of DCs were increased and that DC-SIGN+ dermal DCs constituted the majority of the DCs. Among these, most DCs show immature phenotypes in the epidermis and dermis, while the few mature DCs in the dermis are almost all Langerin+ cells.

Recently, Vieyra-Garcia et al.41 found c-Kit+ DCs may activate tumor cells through CD40 and CD40L interactions in CTCL. Their data suggest that c-Kit+OX40L+CD40L+ DCs are able to activate benign T cells to cause inflammation and may transmit tumorigenic signals.

MCs originate from the myeloid stem cells and are part of the innate immune system, which take part in allergies and inflammation, as well as the regulation of adaptive T cell-mediated immune responses60. Recently, MCs were described as regulators in the TME of CTCL with protumorigenic functions. Rabenhorst et al.61 found that MCs were increased in CTCL lesions and had a positive relationship with the tumor stage progression and microvessel density. They also reported delayed development of cutaneous lymphoma in mast cell-deficient mice. Eder et al.62 also identified high MC counts within CTCL, especially in MF, but instead of a higher prominence with disease progression, they found that compared to the IIA and IIB stages, there were higher mast cell counts in the IA and IB stages in MF.

MDSCs are a group of immature myeloid cells that act in cancer tolerance and function as immunosuppressive cells. Geskin et al.63 found that the activity of MDSCs, which is reflected by intracellular reactive oxygen species, was elevated in MF patients of stage IB and above, while the number of MDSCs was not increased. IFN-α2b therapy was related to a decreased function in MDSCs, suggesting that inhibiting MDSCs may be one of the reasons for the effective treatment of IFN. Maliniemi et al.64 found that CD33+ MDSCs express indoleamine 2,3-deoxygenase 1 (IDO1) in all studied CTCL subtypes including MF, while IDO1 induces immune tolerance in the TME and is recognized as a potential therapeutic target.

The role of B cells has been investigated in cancer development, revealing that B cells may mediate both pro- and antitumorigenic effects. On one hand, B cells present antigen and induce tumor-reactive T cell activity, which have antitumorigenic effects. On the other hand, they may promote an immunosuppressive milieu and support tumor growth. In addition, regulatory B cells (Bregs) may upregulate the secretion of IL-10, TGF-β, and granzyme B, leading to the induction of tumor immunosuppressive cells, i.e., Tregs, myeloid-derived suppressor cells, and M2-macrophages, as well as the inhibition of the antitumor activities of cytotoxic T cells. The role of tumor-infiltrating B cells has been reported recently in CTCL. First, an increased B cell density was found in CTCL skin samples. Moreover, B cell infiltration was discovered to be correlated inversely with a progression-free survival. The CD20-targeting antibody rituximab and the type II CD20-directed mAb obinutuzumab are being used in multiple clinical trials for CTCL, and the efficacy of targeting B cells and modulating the TME will be further confirmed by following the patients and analyzing their biological samples.

Table 1 summarized the related cells above and their functions in TME.

Table 1
Different cell types and their characteristics and functions in TME

• Click for larger image
• Click for full table
• Download as Excel file

Exhausted T-cells have the features of the presence of immune checkpoints such as PD-1, LAG-3, CTLA-4, TIM-3 and others65. Kantekure et al.66 found that PD-1 expression was frequently observed in early stage of CTCL and that there was a lower expression of PD-1 at tumor stage, while PD-L1 expression seemed to have a positive coloration with the tumor stage progression. Querfeld et al.65 found that both CD4+ and CD8+ T cells in CTCL expressed more immune checkpoints than those from normal skin. Samimi et al.67 reported that PD-1 was highly expressed on malignant T cells, indicating that the elevated expression of PD-1 plays a role in the immunosuppression of SS. Recent report suggested PD-1-PD-L1 therapies may also function through a direct effect on TAMs, and when PD-L1/PD-1 signaling is blocked by antibodies, CD163+ TAMs activated and polarized into M1-like TAMs68. Since the profiles of chemokines are different between M1 and M2 polarized macrophages, M1-polarized macrophages not only induce cytotoxic reaction against lymphoma cells, but also work together with other immunosuppressive cells69.

CTCL can be challenging to treat in the aggressive late stage, especially with visceral involvement or following relapses. For patients with T4 stage MF, the 10-year overall survival drops to 27% (compared to 86% in the T1 stage), and more than 50% of the observed deaths occur in the first 4 years after diagnosis. Thus, novel therapeutic strategies, especially biological therapy (i.e., immunotherapy), are needed to offer the best promise in managing such cases. Leuchte et al.70 wrote an elegant comprehensive review on innovative treatment concepts in CTCL derived from the growing knowledge of the TME in tumor biology. In general, the successful translation of medications from benchside testing to bedside application has been limited.

In all the listed clinical trials funded by the NIH in the United States, there are thirteen active trials employing biological drugs/medications. These include the immune checkpoint inhibitors and therapeutic monoclonal antibodies targeting CD30, CD52, or ICOS (the T cell inducible costimulator), immune-modulating agents (cytokines (IL-12, IL-15) or drugs (imiquimod/resiquimod, lenalidomide)), and immune cell therapy (allogenic stem cell transplant), most of which modulate the TME.

Therapeutic monoclonal antibodies represent a new advancement in CTCL treatment. While a few of the therapeutic antibodies directly target cancer cells, such as the antibody for KIR3DL2 receptor, which is expressed by tumor cells, most monoclonal antibodies kill tumor cells indirectly though the modulation of the TME. An anti-PD antibody was approved for treat several other malignancies.

CTCL contains an increased proportion of PD-1+ and PD-L1+ cells, which provide the rationale for applying checkpoint inhibitors. In addition, the next-generation sequencing of CTCL cells revealed an outstandingly high mutational burden and a high rate of chromothripsis, both of which predict a superior response to immune checkpoint blockade. However, resistance to anti-PD1 antibody treatment often occurs in cancer patients, and even patients initially responding to treatment can eventually develop acquired resistance. Peranzoni and Donnadieu71 demonstrated that depletion of TAMs restored T cell migration and infiltration into tumor islets and improved the efficacy of anti-PD-1 immunotherapy in a mouse model of squamous cell carcinoma. We have assessed a combined regimen by simultaneously administrating an anti-PD1 antibody and a CCR2 inhibitor that blocks CCR2-mediated chemotaxis of monocytes in a syngeneic mouse implantation model of T-cell lymphoma in skin. CCR2 inhibitor-induced TAM depletion synergized with the anti-PD1 antibody, producing an encouraging antitumor effect, which warrants a future translational study72.

Mogamulizumab a CCR4 antibody, another potent anti-CTCL biological, has completed two clinical trials with results showing the effective targeting of not only malignant T cells but also Tregs and Th2 lymphocytes, which express CCR4. In addition to therapeutic antibodies, the immunomodulatory drug lenalidomide and Toll-like receptor agonists enhance antitumor immune responses by increasing IL-2, IL-12, and IFN-γ and decreasing IL-1, IL-6, and IL-10. The topical formulation in imiquimod and resiquimod would improve compliance and reduce side effects and toxicity to other organs compared to systemic medications. Moreover, small molecule compounds inhibiting signaling pathways such as NF-κB, JAK/STAT, or mTOR also regulate Tregs, DC, NKs and inhibit TGF-β and IL-10 in the TME, which enhance the antitumor effects from their principal mechanisms functioning through tumor intrinsic survival and growth pathways.

The intimate relationship between tumor cells and the TME determines every step of tumorigenesis. Targeting the TME, either through active host immunity or counteracting the activities of cancer-associated partners, provides more options in CTCL clinical management.

The TME has been identified as one of the critical driving factors in tumor occurrence, progression, and invasion in CTCL. In this article, we have reviewed the roles of the TME in MF and SS, including the essential chemokines involved, Th2-dominant microenvironment, T cell exhaustion via immune checkpoints, and the different cell types in the TME. These cell types include TAMs, CAFs, DCs, MCs, and MDSCs.

The interaction between tumor cells and the microenvironment affects the progression of MF and SS. Determining the mechanism and dynamic changes in this process will provide a better understanding of the pathogenesis in CTCL and serve as a basis for potential TME component-targeted treatments or alternative immunotherapies for refractory/relapse MF and SS in the future.