PD-L1 Overexpression With Activating EGFR Mutations in NSCLC
A total of 164 patients with NSCLC (114 with adenocarcinoma and 50 with squamous cell carcinoma) who underwent surgical resection was included in the present study (Table 1). Ninety-one (55%) patients were male and 95 (58%) were never smokers, and the median age of all patients was 66 years (range, 39–82 years). With regard to EGFR mutation status, 30 patients harbored a deletion in exon 19 and 27 patients had an L858R missense mutation in exon 21.
Immunohistochemical staining for PD-L1 was detected at the membrane or in the cytoplasm (or both) of tumor cells and stromal lymphocytes in the surgically resected tumor specimens (Figure 1A). Expression of PD-L1 was significantly higher in tumors from women than in those from men (P < 0.001), in those with an adenocarcinoma histology than in those with a squamous cell carcinoma histology (P < 0.001), in those from never smokers than in those from smokers (P < 0.001), and in those positive for EGFR mutations than in those wild type for EGFR (P < 0.001) (Figure 1B). No significant association was detected between expression of PD-L1 and either patient age (≤66 versus >66 years, P = 0.228) or tumor p stage (I versus II or III, P = 0.207). Given that EGFR mutations are more frequently found in women, never smokers, and individuals with adenocarcinoma, we carried out multivariate analysis to examine which factors were associated with expression of PD-L1 in the present study. Multivariate analysis revealed that EGFR mutation positivity and adenocarcinoma histology were associated with high expression of PD-L1 independently of other patient characteristics (P = 0.027 and P = 0.046, respectively) (Table 2).
(Enlarge Image)
Figure 1.
Relation between immunohistochemical staining for PD-L1 in tumor specimens and other patient characteristics. (A) Representative patterns of PD-L1 immunostaining in two NSCLC tumors each with strong (left panels) or weak (right panels) staining intensity. PD-L1 immunoreactivity was detected at the membrane or in the cytoplasm (or both) of tumor cells and stromal lymphocytes. Original magnification, ×400. (B) Significant association of overall PD-L1 staining (H) score with sex, smoking status, tumor histology, or EGFR mutation status (WT, wild type). Data are presented as box-and-whisker plots, and P values were determined with the Wilcoxon rank-sum test.
Given the positive association between EGFR mutation and high PD-L1 expression in NSCLC tumors, we examined the expression of PD-L1 in NSCLC cell lines whose EGFR mutation status had been previously determined. Three (HCC827, PC9, and H1975) of the 14 cell lines examined harbor EGFR mutations, whereas the other 11 cell lines are wild type for EGFR. Both HCC827 and PC9 harbor an activating in-frame deletion in exon 19 of EGFR, whereas H1975 harbors an activating L858R mutation in exon 21 as well as a secondary mutation (T790M) in exon 20 that contributes to the development of resistance to EGFR-TKIs such as gefitinib and erlotinib. Flow cytometric analysis revealed that the level of PD-L1 expression at the cell surface was significantly higher for cell lines with EGFR mutations than for those with wild-type EGFR (P = 0.023) (Figure 2A). To examine the role of oncogenic EGFR signaling in the regulation of PD-L1 expression, we determined the effect of the EGFR-TKI erlotinib on PD-L1 abundance in NSCLC cell lines. Consistent with our previous observations, immunoblot analysis revealed that EGFR phosphorylation was undetectable in EGFR mutation-negative A549 cells, whereas EGFR mutation-positive PC9, HCC827, and H1975 cells manifested a high basal level of EGFR phosphorylation (Figure 2B). Treatment of PC9 and HCC827 cells with erlotinib resulted in marked downregulation of PD-L1 expression at the cell surface (Figure 2C) as well as inhibition of EGFR phosphorylation (Figure 2B), whereas erlotinib did not affect PD-L1 expression or EGFR phosphorylation in A549 cells. The T790M mutation impedes the binding of gefitinib and erlotinib to the ATP binding pocket of the catalytic domain of EGFR. Importantly, erlotinib had no effect not only on EGFR phosphorylation but also on PD-L1 expression in H1975 cells, which harbor both L858R and T790M mutations (Figure 2B and C). Together, these data thus suggested that the expression of PD-L1 is upregulated by EGFR signaling in EGFR mutation-positive NSCLC cells.
(Enlarge Image)
Figure 2.
Relation between PD-L1 expression and EGFR mutation in NSCLC cell lines. (A) Flow cytometric analysis of PD-L1 expression at the surface of NSCLC cell lines. The mean fluorescence intensity (MFI) for PD-L1 was significantly higher in EGFR mutation-positive cell lines (HCC827, H1975, and PC9) than in cell lines with wild-type EGFR (*P < 0.05, Mann–Whitney U-test). (B and C) Immunoblot analysis of phosphorylated (P) and total (T) forms of EGFR (B) as well as flow cytometric analysis of surface PD-L1 expression (C) in the indicated cell lines after incubation for 48 h in the absence [shaded trace in (C)] or presence [green trace in (C)] of 100 nM erlotinib. H1975 harbors a secondary mutation (T790M) of EGFR that contributes to the development of resistance to EGFR-TKIs such as gefitinib and erlotinib.
The median follow-up time for all 164 patients was 55.6 months (range, 0.8–168.4 months). NSCLC patients with a high expression score for PD-L1 had a significantly shorter OS compared with those with a low expression score (median of 55.9 versus 72.6 months, P = 0.039) (Figure 3). Univariate analysis revealed that, among the characteristics examined, only the expression score for PD-L1 (P = 0.039) and p stage (P < 0.001) were significantly associated with OS (Table 3). Cox regression analysis also revealed that high PD-L1 expression (P = 0.020) and p stage II or III (P < 0.001) were significantly associated with a shorter OS independently of other factors (Table 3).
(Enlarge Image)
Figure 3.
Kaplan–Meier analysis of OS according to PD-L1 expression score in NSCLC patients. The median value (30) of all H scores was a priori chosen as the cutoff point for separating tumors with high or low PD-L1 expression. The P value for the difference between the two curves was determined by the log-rank test.
Results
Patient Characteristics
A total of 164 patients with NSCLC (114 with adenocarcinoma and 50 with squamous cell carcinoma) who underwent surgical resection was included in the present study (Table 1). Ninety-one (55%) patients were male and 95 (58%) were never smokers, and the median age of all patients was 66 years (range, 39–82 years). With regard to EGFR mutation status, 30 patients harbored a deletion in exon 19 and 27 patients had an L858R missense mutation in exon 21.
Immunohistochemical Analysis of PD-L1 Expression in Tumor Specimens
Immunohistochemical staining for PD-L1 was detected at the membrane or in the cytoplasm (or both) of tumor cells and stromal lymphocytes in the surgically resected tumor specimens (Figure 1A). Expression of PD-L1 was significantly higher in tumors from women than in those from men (P < 0.001), in those with an adenocarcinoma histology than in those with a squamous cell carcinoma histology (P < 0.001), in those from never smokers than in those from smokers (P < 0.001), and in those positive for EGFR mutations than in those wild type for EGFR (P < 0.001) (Figure 1B). No significant association was detected between expression of PD-L1 and either patient age (≤66 versus >66 years, P = 0.228) or tumor p stage (I versus II or III, P = 0.207). Given that EGFR mutations are more frequently found in women, never smokers, and individuals with adenocarcinoma, we carried out multivariate analysis to examine which factors were associated with expression of PD-L1 in the present study. Multivariate analysis revealed that EGFR mutation positivity and adenocarcinoma histology were associated with high expression of PD-L1 independently of other patient characteristics (P = 0.027 and P = 0.046, respectively) (Table 2).
(Enlarge Image)
Figure 1.
Relation between immunohistochemical staining for PD-L1 in tumor specimens and other patient characteristics. (A) Representative patterns of PD-L1 immunostaining in two NSCLC tumors each with strong (left panels) or weak (right panels) staining intensity. PD-L1 immunoreactivity was detected at the membrane or in the cytoplasm (or both) of tumor cells and stromal lymphocytes. Original magnification, ×400. (B) Significant association of overall PD-L1 staining (H) score with sex, smoking status, tumor histology, or EGFR mutation status (WT, wild type). Data are presented as box-and-whisker plots, and P values were determined with the Wilcoxon rank-sum test.
Regulation of PD-L1 Expression by EGFR Signaling in EGFR Mutation-positive NSCLC Cells
Given the positive association between EGFR mutation and high PD-L1 expression in NSCLC tumors, we examined the expression of PD-L1 in NSCLC cell lines whose EGFR mutation status had been previously determined. Three (HCC827, PC9, and H1975) of the 14 cell lines examined harbor EGFR mutations, whereas the other 11 cell lines are wild type for EGFR. Both HCC827 and PC9 harbor an activating in-frame deletion in exon 19 of EGFR, whereas H1975 harbors an activating L858R mutation in exon 21 as well as a secondary mutation (T790M) in exon 20 that contributes to the development of resistance to EGFR-TKIs such as gefitinib and erlotinib. Flow cytometric analysis revealed that the level of PD-L1 expression at the cell surface was significantly higher for cell lines with EGFR mutations than for those with wild-type EGFR (P = 0.023) (Figure 2A). To examine the role of oncogenic EGFR signaling in the regulation of PD-L1 expression, we determined the effect of the EGFR-TKI erlotinib on PD-L1 abundance in NSCLC cell lines. Consistent with our previous observations, immunoblot analysis revealed that EGFR phosphorylation was undetectable in EGFR mutation-negative A549 cells, whereas EGFR mutation-positive PC9, HCC827, and H1975 cells manifested a high basal level of EGFR phosphorylation (Figure 2B). Treatment of PC9 and HCC827 cells with erlotinib resulted in marked downregulation of PD-L1 expression at the cell surface (Figure 2C) as well as inhibition of EGFR phosphorylation (Figure 2B), whereas erlotinib did not affect PD-L1 expression or EGFR phosphorylation in A549 cells. The T790M mutation impedes the binding of gefitinib and erlotinib to the ATP binding pocket of the catalytic domain of EGFR. Importantly, erlotinib had no effect not only on EGFR phosphorylation but also on PD-L1 expression in H1975 cells, which harbor both L858R and T790M mutations (Figure 2B and C). Together, these data thus suggested that the expression of PD-L1 is upregulated by EGFR signaling in EGFR mutation-positive NSCLC cells.
(Enlarge Image)
Figure 2.
Relation between PD-L1 expression and EGFR mutation in NSCLC cell lines. (A) Flow cytometric analysis of PD-L1 expression at the surface of NSCLC cell lines. The mean fluorescence intensity (MFI) for PD-L1 was significantly higher in EGFR mutation-positive cell lines (HCC827, H1975, and PC9) than in cell lines with wild-type EGFR (*P < 0.05, Mann–Whitney U-test). (B and C) Immunoblot analysis of phosphorylated (P) and total (T) forms of EGFR (B) as well as flow cytometric analysis of surface PD-L1 expression (C) in the indicated cell lines after incubation for 48 h in the absence [shaded trace in (C)] or presence [green trace in (C)] of 100 nM erlotinib. H1975 harbors a secondary mutation (T790M) of EGFR that contributes to the development of resistance to EGFR-TKIs such as gefitinib and erlotinib.
Survival Analysis in NSCLC Patients
The median follow-up time for all 164 patients was 55.6 months (range, 0.8–168.4 months). NSCLC patients with a high expression score for PD-L1 had a significantly shorter OS compared with those with a low expression score (median of 55.9 versus 72.6 months, P = 0.039) (Figure 3). Univariate analysis revealed that, among the characteristics examined, only the expression score for PD-L1 (P = 0.039) and p stage (P < 0.001) were significantly associated with OS (Table 3). Cox regression analysis also revealed that high PD-L1 expression (P = 0.020) and p stage II or III (P < 0.001) were significantly associated with a shorter OS independently of other factors (Table 3).
(Enlarge Image)
Figure 3.
Kaplan–Meier analysis of OS according to PD-L1 expression score in NSCLC patients. The median value (30) of all H scores was a priori chosen as the cutoff point for separating tumors with high or low PD-L1 expression. The P value for the difference between the two curves was determined by the log-rank test.
SHARE
![Pediatric Supportive Care (PDQ®): Supportive care - Patient Information [NCI]-Overview](https://cdnus.oss-us-west-1.aliyuncs.com/libs/images/NewsJpg/cate_6/072_215x160.jpg)
