LAG-3 Protein Expression in Non-Small Cell Lung Cancer and Its Relationship with PD-1/PD-L1 and Tumor-Infiltrating Lymphocytes.

INTRODUCTION: Immunotherapy targeting the programmed death 1 (PD-1)/programmed death ligand 1 (PD-L1) checkpoint has shown promising efficacy in patients with NSCLC. Lymphocyte activating 3 gene (LAG-3) is another important checkpoint, and its role in NSCLC is still not clear. In this study we investigated lymphocyte activing 3 (LAG-3) protein expression; its correlation with PD-1, PD-L1, and tumor-infiltrating lymphocytes (TILs); and its association with survival in NSCLC.

METHODS: The expression of LAG-3 (EPR4392 [Abcam, Cambridge, MA]) protein was assessed in 55 NSCLC cell lines by immunohistochemistry. LAG-3, PD-1 (NAT 105 [Cell Marque, Rocklin, CA]), and PD-L1 (22C3 [Dako, Carpenteria, CA]) protein expression was evaluated by immunohistochemistry, and TIL abundance was scored in 139 surgically resected specimens from patients with NSCLC. We also verified results in samples from 62 patients with untreated NSCLC and detected a correlation between LAG-3 expression and EGFR and KRAS mutation and echinoderm microtubule associated protein like 4 gene (EML4)-anaplastic lymphoma receptor tyrosine kinase gene (ALK) rearrangement.

RESULTS: LAG-3 was not expressed on any of the 55 NSCLC cell lines. However, LAG-3 was expressed on the TILs in 36 patients with NSCLC (25.9%). Sixty patient samples (43.2%) were positive for PD-1 on the TILs, and 25 (18.0%) were positive for PD-L1 on tumor cells. Neither LAG-3 nor PD-1 was expressed on the tumor cells. LAG-3 was overexpressed on the TILs in nonadenocarcinoma compared with in adenocarcinoma (p = 0.031). LAG-3 expression on TILs was significantly correlated with that of PD-1 on TILs (p < 0.001) and PD-L1 on tumor cells (p = 0.041) but not with TIL percentage (p = 0.244). With the logistic regression model, the ORs for LAG-3 were 0.320 (95% confidence interval [CI]: 0.110-0.929) and 4.364 (95% CI: 1.898-10.031) when nonadenocarcinoma was compared with adenocarcinoma and TILs that were negative for PD-1 were compared with those positive for PD-1. Recurrence-free survival was significantly different in patients whose TILs were LAG-3-negative as opposed to LAG-3-positive (1.91 years [95% CI: 0.76-3.06] versus 0.87 years [95% CI: 0.27-1.47] [p = 0.025]). Likewise, LAG-3 status of TILs (negative versus positive) did significantly affect overall survival (OS) (3.04 years [95% CI: 2.76-3.32] versus 1.08 years [95% CI: 0.42-1.74] [p = 0.039]). Using Kaplan-Meier analysis, we found that patients with both PD-L1-negative tumor cells and LAG-3-negative TILs have longer recurrence-free survival than patients who are either PD-L1- or LAG-3-positive or both PD-L1- and LAG-3-positive (2.09 years [95% CI: 0.90-3.28] versus 1.42 years [95% CI: 0.46-2.34] versus 0.67 years [95% CI: 0.00-1.45] [p = 0.007]). In the verification stage, high expression of LAG-3 was also significantly correlated with higher expression of PD-1 on TILs (p = 0.016) and PD-L1 on tumor cells (p = 0.014). There was no correlation between LAG-3 expression and EGFR (p = 0.325) and KRAS mutation (p = 1.000) and ALK fusion (p = 0.562).

CONCLUSIONS: LAG-3 is expressed on TILs in tumor tissues of some patients with NSCLC. Its expression was higher in nonadenocarcinoma and correlated with PD-1/PD-L1 expression. LAG-3 positivity or both LAG-3 and PD-L1 positivity was correlated with early postoperative recurrence. LAG-3 was related to poor prognosis.