Aim: Nivolumab, a fully human immunoglobulin G4 programmed death-1 (PD-1) immune checkpoint inhibitor antibody, has activity in melanoma, non-small-cell lung cancer (NSCLC), renal cell carcinoma (RCC), and Hodgkin lymphoma. Nivolumab is approved in the USA and EU for advanced melanoma, NSCLC, and RCC, and relapsed Hodgkin lymphoma in the USA. Programmed death-ligand 1 (PD-L1), a PD-1 ligand, is expressed on mononuclear leukocytes, myeloid cells, and tumor cells. PD-L1 is being investigated as a potential biomarker to predict the association of tumor PD-L1 expression with nivolumab efficacy.
Methods: Bristol-Myers Squibb and Dako previously reported on an automated PD-L1 immunohistochemical (IHC) assay that detects cell surface PD-L1 in formalin-fixed, paraffin-embedded, human tumor tissue specimens using Dako's Autostainer Link 48. The primary antibody for this assay is a rabbit monoclonal antihuman PD-L1 antibody, clone 28-8. Another rabbit monoclonal antihuman PD-L1 antibody, clone E1L3N, was compared with 28-8 for specificity and sensitivity using an identical detection method followed by vendor-recommended detection methods.
Results: Using PD-L1 null clones of L2987 and ES-2 tumor cell lines, both antibodies were specific for detection of PD-L1 on the plasma membrane, although E1L3N also stained cytoplasm in ES-2 knockout cells. Using the identical method, E1L3N was slightly more sensitive than 28-8 based on staining intensities. Using manufacturer-recommended detection methods and predefined scoring criteria for plasma membrane staining of tumor and immune cells, 28-8 demonstrated significantly improved detection compared with E1L3N.
Conclusions: Epitope retrieval and highly sensitive detection reagents are key determinants in IHC detection of PD-L1.