Background Lung cancer is one of the leading cause of cancer death worldwide. Non-small cell lung cancer (NSCLC) accounts for about 85% of lung cancer. The discovery of immune checkpoints and the development of immune checkpoint inhibitors (ICIs) is revolutionary in cancer immunology. However, only about 20-30% of cancer patients could benefit from ICIs. Recently, secreted and 5' UTR truncated PD-L1 isoforms were shown to mediate resistance to ICI therapies. Immunohistochemistry (IHC) is the current standard PD-L1 detection method to triage patients for ICI therapies but has limitations, including various antibodies with various detection thresholds, challenge to distinguish PD-L1 isoforms, and inability in multiplex analysis of other immune related factors and genetic alterations. Here, we developed a new method to detect the mRNA expression of CD274 (encoding PD-L1), its various splicing variants and multiplex with T-cell activity factors and other genetic drivers by next-generation sequencing (NGS).

Patients and Methods We recruited two patient cohorts. Cohort I included 182 NSCLC (number of stage I, II, III and IV was 73, 48, 59, and 2, respectively) patients with surgery resections. Cohort II included 33 patients who failed first-line therapies and further received anti-PD-1 therapy, prior to which needle biopsies were obtained. Tumor tissues were used for total nucleic acids (DNA and RNA) extraction and subjected to an anchored multiplex PCR NGS panel for one-tube enrichment of 63 genes relevant in NSCLC, including EGFR, KRAS, BRAF, ALK, ROS1, RET, MET, NTRK1/2/3, NRG1 and immune related genes CD274, GZMA (encoding granzyme A) and PRF1 (encoding perforin). We calculated exon-level mRNA expression per copy genomic DNA of the same target, expressed as RNA-to-DNA ratio, without relying on the conventional housekeeping gene normalization. Gene fusions and isoforms were analyzed using our bioinformatics pipeline designed for a clinically established Anchored Multiplex PCR NGS platform (Zheng, et. al, Nat Med, 2014). Kaplan-Meier curve and Cox regression were used for analyses of overall survival (OS) in Cohort I and progression free survival (PSF) in Cohort II.

Results In Cohort I, the mRNA expression of CD274 was correlated with its protein expression assayed by a clinical IHC assay (antibody 22C3; R2 = 0.51, P value < 0.01). ALK mRNA expression were highest among those with ALK fusions. Interestingly, patients with high expression of CD274 exon3 and low expression of exon5 (presumably resembling those secreted and 5-UTR-trancated PD-L1 isoforms) had worst overall survival when compared to patients with other CD274 splicing isoforms and patients with low CD274 expression (P value < 0.01). In Cohort II, the analysis of pre-ICI treatment needle biopsies showed that a high expression of CD274 was associated with response to ICIs and with a better PFS compared to patients with low CD274 expression (P value < 0.01). Further, among patients with high CD274 expression, a high GZMA expression was associated with a trend towards better PFS (P value = 0.06), whereas a high PRF1 expression level was not.

Conclusion The data suggests that PD-L1and GZMA mRNA expression may predict the prognosis and response to immune checkpoint blockade therapies in NSCLC patients, and larger studies are warranted.

Citation Format: Zhengbo Song, Chenyu Lu, Hongxiang Wang, Chunwei Xu, Baifeng Zhang, Wenxian Wang, Juxiang Chen, Zongli Zheng. PD-L1 and granzyme A mRNA expression associated with prognosis and response to immune checkpoint inhibitors in non-small cell lung cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 6490.