Emerging advances in cancer therapy have transformed the landscape towards cancer immunotherapy regimens. Recent discoveries have resulted in the development of clinical immune checkpoint inhibitors that are ‘game-changers’ for cancer immunotherapy. Here we show that eEF2K, an atypical protein kinase that negatively modulates the elongation stage of protein synthesis, promotes the synthesis of PD-L1, an immune checkpoint protein which helps cancer cells to escape from immunosurveillance. Ablation of eEF2K in prostate and lung cancer cells markedly reduced the expression levels of the PD-L1 protein. We show that eEF2K promotes the association of PD-L1 mRNAs with translationally active polyribosomes and that translation of the PD-L1 mRNA is regulated by a uORF (upstream open reading-frame) within its 5′-UTR (5′-untranslated region) which starts with a non-canonical CUG as the initiation codon. This inhibitory effect is attenuated by eEF2K thereby allowing higher levels of translation of the PD-L1 coding region and enhanced expression of the PD-L1 protein. Moreover, eEF2K-depleted cancer cells are more vulnerable to immune attack by natural killer cells. Therefore, control of translation elongation can modulate the translation of this specific mRNA, one which contains an uORF that starts with CUG, and perhaps others that contain a similar feature. Taken together, our data reveal that eEF2K regulates PD-L1 expression at the level of the translation of its mRNA by virtue of a uORF in its 5′-region. This, and other roles of eEF2K in cancer cell biology (e.g. in cell survival and migration), may be exploited for the design of future therapeutic strategies.