e14656 Background: Immuno-oncology revolutionized lung cancer treatment, but predictive biomarkers of response to checkpoint inhibitors (CPI) are still lacking. Only 50% of PD-L1 positive tumors by immunohistochemistry (IHC) respond to anti-PD-L1 treatment. Analytical variability and post-translational modifications (PTM) of the PD-1 signaling associated proteins (e.g.: glycosylation), can explain some of this discrepancy. Plus, the tumor immune microenvironment (TME) is complex and PD-L1 IHC alone is a flawed surrogate its status. Mass spectrometry based technologies have the potential to overcome these challenges by integrating sensitivity, specificity and absolute quantitation of proteins and PTMs in a standardized fashion. Here, we demonstrate the advantages of using anti-peptide antibodies to purify surrogate peptides followed by liquid-chromatography (LC) and multiple reaction monitoring (MRM), herby termed as iMRM, to gain new insight into the TME. Methods: To determine the concentration of PD-L1, PD-1, PD-L2, NT5E, LCK and ZAP70, we used unique and well detectable proteolytic peptides as surrogates. Our previously described protocol (ASCO 2022, #377181) allows robust quantitation of 13 peptides and monitors the glycosylation status of PD-L1, PD-L2, and PD-1. NSCLC samples were either fresh frozen (n = 42) or FFPE (n = 77) and sometimes both (n = 17). PD-L1 quantitation by iMRM was compared to PD-L1 IHC 22C3. Results: This multiplexed iMRM assay successfully quantified the PD-1/PD-L1 axis proteins in 96% of NSCLC patients (61/63). PD-L1 glycosylation levels ranged from 82 to 100% (n = 69, median = 100%, SD = 3.7%). Only PD-1 was significantly different between the fresh frozen (mean = 11±6 amol/µg total protein) and FFPE group (mean = 5±2 amol/µg total protein, ρ = 0.0001), all other peptides showed comparable levels. In our 17 matched FF/FFPE samples, PD-L1 moderately correlated (R = 0.45, ρ = 0.045) and the other peptides did not correlate (R < 0.03). Intra/intertumoral heterogeneity, differences in cellularity and tumor origin likely contributed to this discrepancy. IMRM results correlated moderately (R = 0.56, ρ < 0.01) with PD-L1 IHC. Most of these patients were not CPI-treated, but survival data was available. A trend was noted between the concentration of certain targets and patient survival. Linear regression was used to establish a cut-off for each peptide from which an immunoscore was calculated. The immunoscore could predict long-term survival (accuracy 75.4%) irrespective of tumor staging, grade or subtype and survival was significantly associated with the immunoscore (log-rank ρ = 0.005). Conclusions: Our iMRM workflow provides a new understanding of the TME in NSCLC through the PD-1/PD-L1 axis with an easy-to-read immunoscore. A set of 60 tumors from CPI-treated patients is currently being processed to validate the clinical utility of the assay in relation with CPI-response.
Read full abstract