• Metastatic Melanoma
• PD-L1
• prognostic markers

In the largest study so far reported in the literature, PD-L1 membrane expression was an independent negative prognostic factor in melanoma patients. In the present study, by multivariate analysis, Breslow thickness and PD-L1 membrane positivity were independent risk factors for melanoma-specific death.
This conclusion was reached upon testing PD-L1 expression in a cohort of 81 consecutive MMP treated at a single institution, and comparing two antibodies specific for PD-L1. This is the first study demonstrating the prognostic role of PD-L1 in melanoma patients by using the validated mouse monoclonal 5H1 antibody.
In agreement with a negative prognostic role for PD-L1, our study showed that metastatic melanoma samples express PD-L1 in significantly higher proportions than primary melanoma (40.3% vs 14%). In 17/22 patients the metastatic site resulted positive while primary melanomas were negative, suggesting that PD-L1 expression is acquired during disease progression.
In the second part of the work, we asked whether PD-L1 expression was simply the result of microenvironmental pressures on the tumour cell or whether it was an intrinsic feature, marking a disease subset with specific characteristics. PD-L1+ and PD-L1- subsets of the A375 cell line were stabilized in vitro and compared using gene expression profiling and functional assays. Results were confirmed using xenograft models.
Gene profiling indicated that the PD-L1+ cell subset was characterized by a unique genetic signature, with enhanced expression of genes connected to growth, activation and invasion. A functional comparison confirmed that: i) the PD-L1+ variant showed signs of increased growth and invasion in vitro, ii) these features were enhanced when using three-dimensional cultures and maintained after xenografting in immunocompromised mice.
From the translational standpoint, our data suggest that beside the known effects on immune response modulation, PD-L1 expression marks a subset of melanoma cells characterized by a specific gene expression profile and by increased growth and aggressiveness. Our results suggest that PD-L1 is not mechanistically responsible for the more aggressive phenotype in melanoma cells.
If confirmed, our clinical and experimental data suggest that PD-L1+ melanomas should be considered a disease subset with distinct genetic and morpho-phenotypic features, leading to enhanced aggressiveness and invasiveness.
Future studies will tell whether PD-L1 expression is also a marker of resistance to selected therapies and whether it may be successfully exploited alone or in combination as a target for specific subsets of melanoma patients.