The quantity of programmed cell death-ligand 1 (PD-L1) is regarded as a predicting factor of clinical response to anti-PD-1 axis immunotherapy. However, the expression of PD-L1 and its prognostic value in hepatocellular carcinoma (HCC) patients remain debated. Meanwhile, the molecular features of PD-1's other ligand, namely PD-L2, as well as its correlation with clinicopathological parameters and HCC tumor microenvironment (TME), are still poorly understood. In this study, immunohistochemistry (IHC) data from 304 HCC patients were used to determine the clinicopathological features of PD-L1 and PD-L2 and their correlation with CD8+ T cells in HCC. Moreover, fresh clinical HCC samples were used to identify the immune cell subtypes expressing PD-L1 and PD-L2. By using The Cancer Genome Atlas (TCGA) dataset, we further assessed the correlation between mutation signature, copy number variation (CNV), number of neoepitopes, immune gene expression, immune/stromal cell infiltration to the expression of PD-L1 and PD-L2. While membrane expression of PD-L2 was observed in 19.1% of tumor samples, no obvious expression of PD-L1 was detected on tumor cell membranes. High expression of PD-L2 on tumor membranes and PD-L1 in immune stroma were both significantly associated with poorer overall survival (OS) and disease-free survival (DFS) outcomes. Flow cytometry analysis and immunofluorescence showed that macrophages were the main immune cell subtype expressing both PD-L1 and PD-L2. Moreover, positive expression of PD-Ls was correlated with higher CD8+ T cells infiltration in immune stroma. CNV analysis showed a similarity between PD-L1 and PD-L2 in affecting gene expression. In addition, higher levels of PD-Ls correlated with higher expression of immune related genes, enhanced cytolytic activity, and larger proportions of immune/stromal cell infiltration. Collectively, our study reveals the impact of both PD-L1 and PD-L2 on the HCC tumor microenvironment for the first time, providing insight for new therapeutic options. Copyright © 2019 Liao, Chen, Dai, Richardson, Guo, Yuan, Zeng and Xie.