Article Text
Abstract
Background Immune suppressive conditions in the melanoma tumor microenvironment (TME) block dendritic cell (DC) development and lead to the accumulation of M2-like macrophages and myeloid-derived suppressor cells (MDSCs). This will effectively hamper T cell priming, recruitment, and effector functions, and so interfere with the efficacy of immunotherapy. Targeting tumor-mediated myeloid suppression represents an interesting therapeutic option to promote the immune attack on tumors. The preclinical human models currently used to study myeloid suppression often fail to reflect the complexity of the TME.
Materials and Methods To study the cross-talk between melanoma and stroma cells and assess its effect on DC differentiation, we therefore established an in vitro three-dimensional (3D) reconstructed organotypic human melanoma-in-skin (Mel-RhS) model, allowing the monitoring of tumor growth and progression for up to six weeks.
Results Significantly higher levels of immune suppressive cytokines (IL-10, M-CSF, VEGF, TGF-beta) were detected in the melanoma model, constructed with the BRAF- and PTEN-mutated SK-MEL-28 cell line, as compared to its control (without melanoma cells). Indeed, Mel-RhS culture supernatants interfered with monocyte-to-DC differentiation, leading to the development of M2-like macrophages with a distinct phenotype (CD14+CD1a-BDCA3+CD163+CD16+PDL1+PDL2+), as established by polychromatic flowcytometry. Correlation matrix heatmap analysis identified IL-10, TGF-beta and M-CSF as the main candidate mediators of this skewing of monocytes to an M2-like state. The use of specific neutralizing antibodies against each of these cytokines prevented the observed DC suppression to varying degrees. t-Distributed Stochastic Neighbor Embedding (t-SNE) identified specific shifts between monocytic subpopulations and modulated expression levels of associated surface markers. Neutralization of M-CSF reduced expression of BDCA3, PD-L2, and PD-L1, while increased CD16; whereas blocking TGF-beta led to a concerted reduction in CD14, CD163, PD-L1, and PD-L2 levels, but, unexpectedly, also of CD80. In contrast, IL-10 neutralization resulted in a decrease of all M2-related markers, while CD80 levels were upregulated. Interestingly, while the SK-MEL-28 cell line did not secrete detectable levels of IL-10 in traditional monolayer cultures, RNA in situ hybridization revealed de novo expression in Mel-RhS in melanoma cells, as well as in keratinocytes and fibroblasts.
Conclusions We conclude that the 3D configuration of the Mel-RhS model results in cross-talk between tumor and stroma, which allows for the delineation of immune suppressive pathways in the melanoma TME. Ultimately, this model could be used as a novel in vitro tool for preclinical testing of immune modulatory therapeutic agents.
Disclosure Information M. Lopez Gonzalez: None. E. Michielon: None. J.L.A. Burm: None. T. Waaijman: None. E.S. Jordanova: None. S. Gibbs: None. T.D. de Gruijl: None.