Because TAMs are a very heterogeneous population of immune cells, which has been reported to be associated with either anti-tumoral or protumoral functions, assessment of their molecular or functional variety may provide new biomarkers for patient prognosis or to predict treatment responses [6,10]. Accurate prognosis in melanoma is particularly important to identify those patients who might benefit from adjuvant treatment. Researchers have quantitatively evaluated conventional parameters like density, morphology, stroma/tumor nest distribution, as well as more in the deep characterization of TAM polarization and cytokine content. In clinically relevant primary melanoma patient cohorts, we did not find a significant association of TAM density, size, and location, nor CD11c/CD209 subsets with DFS or OS. In prior work, our group identified a particular secretory phenotype of TAMs isolated from melanoma tissues [17]. Melanoma TAMs were also characterized in situ by their high content of CCL20, TNF, and VEGFA by multicolor immunofluorescence in cryopreserved samples; we now show that this particular cytokine profile may be detected and quantified in diagnostic FFPE. Importantly, the high content of CCL20/TNF/VEGFA is strongly associated with a worse prognosis in primary melanoma patients.

Nowadays it is well recognized that the interaction of the cancer cells with other cells or factors in the TME has an impact on tumor biology [16]. A deep analysis of TME may provide clues to understanding these complex relationships and, ultimately, will be of clinical utility for the establishment of novel prognostic tools or therapeutic targets. To this end, we previously used multicolor immunofluorescence microscopy to study in situ diverse immune and stromal components, as well as the expression of chemokines and their receptors in the TME of human melanoma. This non-destructive tissue methodology allowed us to study the spatial distribution of diverse cells, location of specific proteins within cells or tissue areas, and quantification of their relative level of expression. Human tissue studies commonly use a chromogenic IHC approach to evaluate specific proteins or cell markers in situ; however, only single or double-labeling are suitable. Fluorescent IHC and spectral microscopy have the advantage of multiple marker detection; accordingly, it was used to analyze immune infiltrates in primary melanomas, showing that combination assessment of CD8⁺ lymphocytes/CD68⁺ macrophages ratio in the stroma correlated with shortened survival [28]. Researchers used a similar approach, combined with confocal microscopy and image processing, to quantify at the cellular level diverse TAM characteristics in melanoma. Multiple labeling is of particular relevance when evaluating heterogeneous cells like macrophages, which are known to share markers with other myeloid cells or to express molecules in relation to their functional polarization. As an example, CD11c, a marker traditionally associated with dendritic cells (DCs), is also expressed by M1 macrophages [29]. In previous studies, researchers established that TAMs are the most abundant type of myeloid cells in melanoma tissues, as compared with other mononuclear (CCR2⁺ monocytes or CD1c⁺ myeloid DCs) and polymorphonuclear phagocytes [17,19]. In this study, researchers set up conditions for multilabeling and quantitative image analysis of TAM markers in both cryopreserved and paraffin-embedded tissues, which are more convenient for clinical utility. First, as bona fide pan-macrophage markers, researchers showed that CD163, CD68, and CD115 were co-expressed by most cells with macrophage morphology. CD163 did not define any separated subset of TAMs, which is in agreement with its co-expression with both M1 and M2 markers in Th1 and Th2-predominant pathologies [11]. Definition of TAM subpopulations was better achieved by triple staining with a pan-macrophage marker to gate macrophages, and quantitative assessment of the level of expression of CD11c and CD209. Thus, dot-plot analysis of CD11c versus CD209 expression at the single-cell level showed that in control skin most macrophages were CD209⁺CD11c⁻ cells, which may represent tissue-resident macrophages [30]. By contrast, melanoma tissues contained a novel CD11c⁺CD209⁻ subset, maybe representing incoming proinflammatory macrophages, and CD11c⁺CD209⁺ mixed cells. However, no major differences were observed between clinically divergent groups regarding CD11c/CD209 subsets or other TAM parameters as density, morphology, or location.

Cytokine production by macrophages is a hallmark of their polarized function [9]; M1 release large amounts of proinflammatory cytokines (such as IL-12, IL-23, and TNF), whereas M2 are characterized by high expression of IL-10 and proangiogenic growth factors such as IL-8, VEGFA, and VEGFC. Interestingly, CCL20 secreting TAMs were induced via IL-6-regulated macrophage polarization in a mouse model of colon cancer; in which, CCL20 promoted cancer progression by recruiting CCR6⁺ lymphocytes [31]. Macrophage derived-TNF was identified as a crucial melanoma growth factor that contributed to resistance to MAPK pathway inhibitory treatments in a mouse model of Braf^V600E mutated melanoma [32]. Researchers had previously identified the expression of CCL20/TNF/VEGFA cytokines by TAMs in a pilot study with cryopreserved human melanoma tissues [17]. Because FFPE tissues are the current method for patient sample preservation, researchers validated our multiple-labeling methodology to analyze TAM cytokine signature in 83 FFPE samples from a second patient cohort. Importantly, CCL20/TNF/VEGFA cytokine production by TAMs correlated with poor patient survival. Using monocyte-derived macrophages, with either GM-CSF or with M-CSF, which are known to prime towards M1 or M2 phenotypes, we produced in vitro melanoma-conditioned macrophages expressing CCL20 and TNF (M1-primed) or VEGFA (M2-primed). This in vitro culture system of macrophages with melanoma cells, allowed us to explore the pathways underlying the CCL20/TNF/VEGFA cytokine secretory phenotype of TAMs, showing that p53 and NF-κB coregulate tumor-conditioned macrophages. It is known that NF-κB activation drives M2 polarization of TAMs [33] and its inhibition re-educates TAMs towards an anti-tumoral phenotype [34]. In accordance with our results, dual p53 and NF-κB activation in macrophages were described to up-regulate IL6, TNF, and CCL20 [35]. These results identify a unique p53/NF-kB pathway in protumoral TAMs that could be targetable to re-educate TAMs to block their protumoral functions.