Mac attack: macrophages as key drivers of cutaneous T-cell lymphoma pathogenesis

It has long been recognized that the prognosis of patients with cutaneous T-cell lymphoma (CTCL), particularly mycosis fungoide (MF), is closely linked to tumor staging, namely patch, plaque and tumor stage presentation 1. This key clinical observation then begs the question as to what is the difference in either the malignant T cells or in the microenvironment of the tumor (i.e. the skin) at these various stages that accounts for these strikingly different clinical behaviours? Laying aside the characteristics of the malignant T cells themselves, which have been suggested to be derived from skin-resident memory T cells by Kupper and others 2, a large amount of clinical and experimental work over that last decade has clearly established the influence of the tumor microenvironment on tumor growth and progression 3. Although T cells in tumor microenvironment have drawn most of the attention from researchers, other immune cells play important roles in tumor immune surveillance and tolerance 4. Tumor-associated macrophages (TAMs), the most frequently found leucocyte population within the tumor microenvironment, have received recent attention because they are found in many common cancers, including Hodgkin's lymphoma 5-7. Macrophages exhibit flexibility and plasticity in producing distinct biochemical and cytokine profiles in many types of tumors 8. At later stages of tumor progression, TAMs generally exhibit an M2-like phenotype with low IL-12, but high IL-10, expression. TAMs also produce high levels of VEGF, which would drive angiogenesis with tumors. We and other investigators have observed abundant macrophage infiltration in the skin lesions from patients with MF, the most frequent subtype of CTCL, and have associated this phenomenon with poor clinical prognosis 1, 9, 10 (Fig. 1). Changes in the Th1 vs Th2 environment within different clinical tumor stages of mycosis fungoides have been noted previously by Asadullah and colleagues 11. They showed that an interferon-γ-rich environment was common in patch and plaque stage MF disease, whereas IL-10, a marker of Th2 immunity, was increased in tumor disease. A model suggesting that a switch to a Th2 predominant environment in advanced disease has been espoused in several prominent reviews of this subject 12. What role could macrophages play in that transition? Hoping to answer this question Furudate et al. 13 in this issue examine the presence of macrophages, key cytokines such IL-4, as well as immunoregulatory protein called periostin in the patch, plaque and tumor phases of MF. They use CD206 as a marker for M2 macrophages in immunohistochemical staining of human tissue samples and confirm that macrophages, including CD206 + macrophages, are increased in advanced MF samples. However, it should be noted recent data suggest that M2 macrophages can be divided into at least 4 different subtypes with variable function and cytokine expression patterns 14 and that multiple receptor expression patterns are needed to accurately assign macrophages into certain known subsets. They observe that periostin is primarily expressed near the basement membrane and upper papillary dermis in patch and plaque stage MF but is downregulated in tumor stage MF. IL-4, however, has the opposite expression pattern with higher expression in plaque and tumor stages, an observation that has been made by others. Using in vitro cultured human macrophages that were derived from monocytes via cytokine stimulation, these authors observed that exposure to recombinant periostin increased macrophage transcription of a number of cytokines, including the classic interferon-γ-inducible chemokines, CXCL9, 10 and 11. Because the macrophages used in their experiments were already of an M2 phenotype, albeit incompletely characterized, one could conclude from this data that periostin may reverse the M2 characteristics of those cells, which would be a topic for further study. Of significance, the addition of IL-4 to periostin-containing macrophage cultures downregulated CXCL9-11, suggesting that the IL-4 found in later stages could potentially downregulate chemoattractive signals that would recruit Th1 anti-tumor T cells. Both clinical and experimental data using murine models of CTCL have suggested that macrophages could play a key role in lymphoma progression. Steidl et al. 5 showed that large numbers of CD68 + macrophages in the tumors of Hodgkin's disease portended poor survival and could predict a lack of response to treatment. Sugaya and colleagues 9 performed similar analyses in patients with MF and showed that increasing numbers of CD68 + and CD163 + macrophages in MF lesions were associated with poor survival and that macrophages numbers generally decreased with effective treatment. The spontaneous development of CTCL in murine species has greatly hampered our understanding of the pathomechanisms of CTCL, but the recent development of a novel, inflammation-dependent model of T-cell lymphomagenesis in the skin has resulted in an opportunity to examine a potential role of macrophages in CTCL biology. Wu et al. injected MBL2 T lymphoma cells into the ear skin of mice and induced high grade lymphoma in the skin only after application of di-nitro-fluoro-benzene, a well-known contact allergen sensitizing agent. These tumors were rich in F4/80-positive macrophages. They later showed that removal of macrophages with clodronate-encapsulated liposomes 15, a well-known technique for selectively ‘poisoning’ highly phagocytic cells such as macrophages, markedly reduced the formation of skin T-cell tumors in this model as well as xenograft models using human CTCL tumor cell lines 16. Of note, clodronate treatment reduced both angiogenesis and lymphangiogenesis in this model, reaffirming the potential of macrophages to produce known vascular regulators such as VEGF. Key signalling pathways, including STAT3, were also observed to be down-modulated by clodronate treatment in this model. The clodronate–liposomes had no effect of the viability of known human CTCL cell lines such as Hut78. Thus, their antitumor effect in vivo could mostly be attributed to their impact on macrophages, although effects on other scavenging cells cannot completely be excluded. In addition, supporting evidence for this hypothesis was shown in a recent clinical CTCL trial using the anti-CD30 monoclonal antibody/toxin conjugate brentuximab vedotin. In this study, they showed that M2 CD163+ tumor-associated macrophages (in addition to malignant T cells) abundantly expressed the CD30 molecule. Treatment with brentuximab vedotin resulted in a remarkable clinical response rate of 73%, indicating that brentuximab vedotin may target the tumor-associated macrophages in addition to the malignant T cells and disrupt their tumor promoting function 17. Treatment of T- and B-cell lymphoma in humans has mostly been based on a strategy of altering the growth or survival of the malignant T or B cells through chemotherapy (i.e. gemcitabine, bexarotene, etc.) or immunotherapy (i.e. rituximab or vaccines). The observations described by Furudate confirm recent work suggesting that macrophages could play key pathogenic roles in CTCL. It is intriguing that periostin is expressed mainly in early disease and that it appears to increase expression of typical Th1 chemokines (CXCL9-11), which could attract Th1-dominant antitumor CD8 T cells. Thus, periostin expression could be a positive prognostic biomarker for MF, but this hypothesis was not formally evaluated. Collectively, this work and that of others mentioned above suggest that additional targeting of a supporting cell, namely tumor-associated macrophages, rather than the malignant T cell alone, may be a viable, novel approach to the treatment of cutaneous T- and B-cell lymphomas in the future. Chalid Assaf and Sam T. Hwang wrote the manuscript and Sam T. Hwang created the illustration. The authors of this study do not have any conflict of interests to declare.