(W1230-06-37) PD-L1 Gene Expression Quantification in Single Cell on Tissue for Accurately Predicting Response to ICIs

Purpose: PD-L1 immunohistochemistry (IHC) has been traditionally used for predicting clinical responses to immune checkpoint inhibitors (ICIs). However, there are at least 4 different assays and antibodies used for PD-L1 IHC, each developed with a different ICI. Spatial transcriptomic analysis at the single cell level can be helpful for deeply characterize the complexity and heterogeneity of tumors. We set to test if in-situ hybridization-based RNA scope coupled to single cell quantification is a robust method to determine PD-L1 mRNA expression levels in each cell and furthermore efficacy of predicting response to ICIs as compared to routinely used standardized IHC procedures.

Methods: FFPE sections from retrospectively collected surgically resected NSCLCs (adenocarcinoma and squamous cell carcinoma) tumor with progression disease or complete response from adjuvant pembrolizumab therapy were used. Target genes CD274 (PD-L1) and GZMB (Granzyme B) were assessed. For fluorescent detection, the label probe was conjugated to Opal 570 and 620 respectively. Tissues cells were counterstained with DAPI to detect the nuclei. A Digital pathology software QuPath was used for cell segmentation and dot’s detection for gene expression quantification and further classification.

Results: Difference in expression levels were seen between PD-L1 and GZMB across the tumor with PD-L1 staining mostly located in the tumoral area and Granzyme B in stromal area of the NSCLC tissue. This is consistent with the expected expression of PD-L1 on tumor epithelial cells and Granzyme B on T cells. This heterogeneity in the signal distribution specific to the analyzed tissue correspond to intra-tissue cells metabolism heterogeneity between cell types. This heterogeneity of expression was captured though a digital workflow capable of quantifying the number of cells expressing PD-L1 and GZMB. Among the cells expressing the markers, 96% were positive for PD-L1 for the tumor that did not respond to therapy with the threshold of positivity being set at 1 of dots, that corresponds to 1 transcript, per cell.

Conclusion: Our results demonstrate that PD-L1 was highly expressed in a tumor context that was not responding to PD-1 ICI inhibitor. Measurement of mRNA expression at the single cell level on tissue provides another layer of PD-L1 detection which can be exploited to predict tumor response to ICI suggesting that moving forward with this technology is a viable approach for this dynamic biomarker.

References: Bankhead, P. et al. QuPath: Open-source software for digital pathology image analysis. Scientific Reports (2017). https://doi.org/10.1038/s41598-017-17204-5
ACD: Using QuPath to analyze RNAscope™ , BaseScope™ and miRNAscope™ images (2020). Technical note: https://acdbio.com/qupath-rna-ish-analysis