Uveal melanoma (UM) is the most common intraocular malignancy, affecting more than 2,500 adults yearly in the United States. When confined to the eye, UM is often successfully treated with radiation or surgery. However, nearly half of the patients eventually develop metastatic disease, in some cases decades after primary diagnosis. Metastatic UM is highly resistant to current treatments, resulting in a very poor prognosis, with a mean overall survival of less than two years. The liver is nearly always the initial site of metastases. The preferential dissemination to the liver raises the possibility that hepatic microenvironmental factors contribute to metastatic progression and the resistance of UM to available treatments. However, the identification of such factors is limited by the lack of UM preclinical models that exhibit liver metastasis. We aim to elucidate the mechanisms underlying the characteristic hepatotropism in order to develop strategies to intercept metastatic progression and improve therapeutic options for metastatic UM.

Here, we present a metastatic UM model using intraocular, intravenous and subcutaneous injection of patient-derived primary human UM cell lines (MP41 and Mel202) in immunocompromised mice. We performed CycIF (Cyclic Immuno-fluorescence) and scRNAseq on two metastatic sites to evaluate transcriptional differences between the tumor cells deriving from the liver and the lung. Additionally, this model enables in vivo studies of candidate targeted drug treatments such as the PKC inhibitor darovasertib, currently in clinical trial. A parallel approach of in vitro and in vivo CRISPR screens in three cell lines, or mice harboring xenografts respectively, treated with the kinase inhibitor is currently being validated through single gene knockout and mechanistical experiments.

We describe a novel UM model whose metastasis closely resembles metastatic patterns observed in patients. After a latency of two to four months, we consistently see liver macrometastases irrespective of injection site. Surprisingly, we observe single cells in non-hepatic sites such as the lung. Using highly multiplexed spatial profiling (CyCIF tissue imaging), we have further defined the proliferation, dormancy, and oncogenic signaling in the two metastatic sites. Transcriptomic profiling and CRISPR screening of metastatic cells is currently being analyzed.

This model faithfully represents a unique opportunity to dissect the pathways that lead to UM hepatotropism and an extrahepatic dormancy-like phenotype as well as the opportunity of developing more effective therapeutic options in order to improve patient outcomes.