Fig 1.
PD-L1 mRNA levels are correlated to surface PD-L1 protein levels in human lung cancer cell lines.
A, Quantitative RT-PCR results of PD-L1 mRNA in human lung cancer cell lines (histological features: 12 adenocarcinomas, 5 squamous cells, 2 small cells, and 1 large cell) is shown as RQ using the ΔΔCt method with BEAS-2B as a reference. Bars and error bars represent the mean and SD of the technical triplicates, respectively. B, The correlation between the mRNA and surface expression levels of PD-L1. A positive correlation was found between the mRNA and protein levels (Pearson’s correlation coefficient, R = 0.389).
Table 1.
Mutational status of human lung cancer cell lines.
Fig 2.
MEK inhibition significantly decreased PD-L1 expression in human lung cancer cell lines.
A, Gating strategy for the determination of PD-L1 levels in lung cancer cell lines stained with FITC anti-PD-L1 mAb. For differentiation between live and dead cells, PI exclusion staining was used. B, (upper) Representative histogram of PD-L1 (shaded histogram) of three KRAS-mutant lung adenocarcinoma cell lines with DMSO or U0126 (20 μM) for 24 h. Empty histogram indicates the isotype control. (upper middle) Relative MFI (PD-L1 MFI / isotype control MFI) of the three cell lines was significantly decreased with U0126. *, p < 0.0001. (lower middle) qRT-PCR also showed a significant decrease of PD-L1 mRNA expression with U0126. *, p < 0.05. Data are the mean of three or four independent experiments. (lower) One representative immunoblot of phospho-ERK.
Fig 3.
Regulation of PD-L1 by MAPK signaling is supported by PMA stimulation, ERK2 RNAi, and KRAS RNAi.
A, Three KRAS-mutant lung adenocarcinoma cell lines were serum-starved for 48 h, pre-conditioned with DMSO or U0126 (20μM) for 1hr, then stimulated with 100 nM PMA for 15 min, and then evaluated for phospho-ERK and PD-L1 mRNA levels. Immunoblot (upper) and qRT-PCR (lower) showed increases in phospho-ERK and PD-L1 mRNA levels 15 min after PMA stimulation, respectively. *, p < 0.05. Data are the mean of three independent experiments. B, Three KRAS-mutant lung adenocarcinoma cell lines were transfected with two siRNAs (#1 and #2) specific for ERK2 or a control siRNA (siCtrl) at 100 nM. Forty-eight hours after transfection, cells were harvested for analysis. (upper) Flow cytometric analysis showed the significant suppression of surface PD-L1 levels, except in NCI-H441. (middle) qRT-PCR showed a significant decrease in PD-L1 mRNA levels in all three cell lines. (lower) Immunoblotting indicated a significant decrease in ERK2 protein levels with the two siRNAs. Data are the mean of three independent experiments. C, PD-L1 expression in NCI-H1373 cells was significantly suppressed with two siRNAs specific for KRAS (#1 and 2) at both surface protein (upper left) and mRNA levels (lower left). Immunoblotting showed that the KRAS knockdown was associated with the inhibition of ERK phosphorylation (right).
Fig 4.
Promoter assay experiments of PD-L1.
A, Diagrammatic representation of the PD-L1 regulatory elements, including PD-L1 promoter fragment and the predicted enhancer containing an AP-1 binding site. Fragments cloned into luciferase constructs are annotated below, with positions relative to the PD-L1 TSS. B, Diagram of luciferase vectors. The empty pGL4.17 vector, the GL4.17 vector with the promoter cloned upstream of the luciferase gene alone (PDL-P), the promoter with the enhancer cloned downstream of the luciferase gene (PDL-P+E). C, PD-L1 promoter- and enhancer-driven luciferase activity in NCI-H1373 with or without MEK inhibition. Firefly luciferase activity normalized by Renilla luciferase activity significantly increased with PDL-P+E compared with that with PDL-P after DMSO treatment (left), while no significant difference was found between PDL-P and PDL-P+E after U0126 treatment (right). Data are representative of three independent experiments with the same results and shown as the mean of quadruplicate experiments. Error bars indicate SD. E, Binding of cJUN to the candidate PD-L1 enhancer region following ChIP-coupled qPCR. Data are representative of three independent experiments with similar results.
Fig 5.
STAT3 signal, but not PI3K signal, partially contributes to ectopic PD-L1 expression in human lung cancer cell lines.
A, STAT3 RNAi reduced PD-L1 expression at both protein (upper) and mRNA (lower) levels in NCI-H1373 and NCI-H441 cells but not in NCI-H358 cells. The experimental conditions were similar to those in Fig 3B. Data are the mean of three independent experiments. B, Immunoblot (upper) and qRT-PCR (lower) of STAT3 showed the decreases in STAT3 protein and mRNA levels, respectively. One representative immunoblot and the mean of three independent qRT-PCR experiments are shown. C, LY294002, a PI3K inhibitor, did not suppress PD-L1 protein expression. Three KRAS-mutant lung adenocarcinoma cell lines were treated with DMSO or LY294002 (40 μM) for 24 h. PD-L1 protein levels did not show any significant changes with LY294002 with regard to both the surface (upper) and total protein (lower) levels. Data are the mean of three independent experiments.
Fig 6.
Supervised cluster analysis and GSEA between PD-L1-high and -low human lung cancer cell lines suggested mRNA upregulation of several molecules related to cell adhesion, motility, cell proliferation, cell signaling, and pathways related to actin cytoskeleton regulation and integrin-mediated cell adhesion.
A, Supervised cluster analysis of 10 PD-L1-high versus 10 PD-L1-low human lung cancer cell lines. Depicted by the CCLE Analysis Tools: Differential Expression (http://www.broadinstitute.org/ccle/data/browseAnalyses). B, qRT-PCR of 11 selected mRNAs in PD-L1-high and -low cell lines. The ΔΔCt method was used for the target mRNA, which was normalized by endogenous β-actin mRNA and reference BEAS-2B cells. RQ data are the mean of 10 cell lines. C, RAC2 and CDA mRNA levels were compared between DMSO and U0126 treatments in the three KRAS-mutant lung adenocarcinoma cell lines. The target mRNAs were normalized with endogenous β-actin mRNAs and DMSO-treated cells as a reference (ΔΔCt method). Data are the mean of three independent experiments. D, GSEA showed significant results for pathways related to actin cytoskeleton regulation by Rho GTPases and to integrin-mediated cell adhesion.