PD-L1 (programmed-death ligand 1; CD274), is a transmembrane protein constitutionally expressed on a variety of cell types, including antigen presenting cells (dendritic cells and histiocytes) and some non-lymphoid tissues (heart and lung). Binding of PD-L1 to PD-1 (programmed-death 1; CD279) expressed by activated T-cells, inhibits their function, causing negative feedback control of immunological reactions, thus impeding inflammation and autoimmunity. Tumour cells may express PD-L1, which binds to PD-1 allowing cancer cells to evade the attack of T-cells. Blockade of the PD-1/PD-L1 pathway has now shown useful in therapy of multiple cancer types, causing durable tumour regressions in a substantial proportion of otherwise treatment refractory cases of melanoma, and carcinomas of e.g., lung, kidney, and urinary tract. Patients without tumour PD-L1 expression can also derive benefit from blocking agents (studies across multiple cancer types demonstrate a pooled response rate of 48% in patients with PD-L1-positive tumours compared to 15% in PD-L1-negative tumours).
PD-L1 expression can currently only be determined by immunohistochemistry, and the PD-L1 IHC 22C3 pharmDx (Dako/agilent) is the only FDA-approved companion diagnostic used to select patients nonsmall lung carcinoma (NSCLC) and gastric/gastroesophageal junction adenocarcinoma for treatment with the anti-PD-1 drug pembrolizumab (Keytruda). For NSCLC with high PD-L1 expression (≥ 50% positive tumour cells) and no EGFR or ALK genomic aberrations, pembrolizumab can be used as first line treatment, while for NSCLC with low expression (≥ 1%-49% positive tumour cells) pembrolizumab can be used as second line treatment. Three other FDA-approved PD-L1 IHC assays, PD-L1 IHC 28-8 pharmDx assay for nivolumab treatment, VENTANA PD-L1 IHC (SP263) assay for durvalumab, and VENTANA PD-L1 IHC (SP142) assay for atezolizumab treatment, are complementary tests for various malignancies providing physicians more information but are not required when deciding treatment. As for the SP142, the scoring is based on PD-L1 immune cells, with a cut-off of 5% for positivity. For SP263, PD-L1 high is defined as tumour cells (TC) and immune cells (IC) ≥25% PD-L1 positive, if IC is >1% of the tumour area, and TC ≥25% and IC=100% positive, if IC was ≤1%.
Conflicting observations regarding PD-L1 as a predictive biomarker of tumour response likely reflects limitations inherent to tumour sampling, with focal expression potentially missed in small biopsies and dynamic PD-L1 expression apparent over time. Different IHC detection methods and antibodies, quality of samples, and pathologist’s scoring are also important factors. Three of the assays (22C3, 28-8 and SP263) have shown a good concordance, while SP142 has shown lower scores when assessing NSCLC. While 22C3 is developed for Dako Autostainer, protocols have been developed by NordiQC allowing for transfer to Omins and BenchMark stainer platforms.
Tonsil and placenta can be used as positive and negative tissue controls. However, tonsil is found to be superior to placenta, as tonsil displayes a range of PD-L1 expression levels. Tonsil displayes the following reaction pattern: No staining reaction in the vast majority of lymphocytes including mantle zone and germinal centre B-cells, no staining reaction in superficial epithelial cells, a weak to moderate, typically punctuated membranous staining reaction of the majority of germinal centre macrophages and finally a moderate to strong staining reaction of the majority of epithelial crypt cells.