Purpose: Although clinical studies have shown promise for targeting programmed cell death protein-1 (PD-1) and ligand (PD-L1) signaling in non–small cell lung cancer (NSCLC), the factors that predict which subtype patients will be responsive to checkpoint blockade are not fully understood.

Experimental Design: We performed an integrated analysis on the multiple-dimensional data types including genomic, transcriptomic, proteomic, and clinical data from cohorts of lung adenocarcinoma public (discovery set) and internal (validation set) database and immunotherapeutic patients. Gene set enrichment analysis (GSEA) was used to determine potentially relevant gene expression signatures between specific subgroups.

Results: We observed that TP53 mutation significantly increased expression of immune checkpoints and activated T-effector and interferon-γ signature. More importantly, the TP53/KRAS comutated subgroup manifested exclusive increased expression of PD-L1 and a highest proportion of PD-L1+/CD8A+. Meanwhile, TP53- or KRAS-mutated tumors showed prominently increased mutation burden and specifically enriched in the transversion-high (TH) cohort. Further analysis focused on the potential molecular mechanism revealed that TP53 or KRAS mutation altered a group of genes involved in cell-cycle regulating, DNA replication and damage repair. Finally, immunotherapeutic analysis from public clinical trial and prospective observation in our center were further confirmed that TP53 or KRAS mutation patients, especially those with co-occurring TP53/KRAS mutations, showed remarkable clinical benefit to PD-1 inhibitors.

Conclusions: This work provides evidence that TP53 and KRAS mutation in lung adenocarcinoma may be served as a pair of potential predictive factors in guiding anti–PD-1/PD-L1 immunotherapy. Clin Cancer Res; 23(12); 3012–24. ©2016 AACR.

Translational Relevance

Programmed cell death ligand 1 (PD-L1) expression, tumor mutational load, and the intensity of CD8+ T-cell infiltrates have recently been proposed as predictive biomarkers for response to PD-1 blockade immunotherapy. However, there are still many treatment responses beyond the explanation of these factors. It is increased need for more effective biomarkers for PD-1 blockade. We demonstrated TP53 and KRAS mutation had remarkable effects on increasing PD-L1 expression, facilitating T-cell infiltration and augmenting tumor immunogenicity. More important, we confirmed that patients with TP53 and/or KRAS mutation showed sensitivity to PD-1 blockade. These findings represent the first demonstration of potential predictive value of TP53 and KRAS mutation for response to PD-1 blockade immunotherapy in lung adenocarcinoma.

Recent clinical trials with anti-programmed cell death 1 (PD-1) and its ligand PD-1 ligand (PD-L1) therapies have shown unprecedented durable responses in patients with non–small cell lung cancer (NSCLC; refs. 1, 2). Unfortunately, only a minority of the total of treated patients respond to the current immunotherapy (3). The factors that determine which patients will be drug sensitive or resistant are not fully understood. Therefore, it has become a primary priority to identify the biomarkers that determine the responsiveness to checkpoint blockade, and to develop strategies that could potentially increase the patient response rates. Encouragingly, recent studies had demonstrated that tumor mutational load (4–6), DNA mismatch repair (MMR) deficiency (7), the intensity of CD8+ T cell infiltrates (8, 9) and intratumoral PD-L1 expression (10, 11) have each been proposed as distinct biomarkers of response to anti–PD-1/PD-L1 therapies. Meanwhile, these factors are functionally interrelated and are often found coordinately in individual tumor specimens (12). This raises the question of whether there exist some other variables simultaneously affect two or more of these above factors so as to provide stronger predictive value for therapeutic outcomes.

The identification of subsets of lung adenocarcinoma with oncogenic drivers has transformed the treatment of NSCLC, particularly for patients whose tumors harbor activating mutations in EGFR. However, the goal of developing specific therapeutic strategies for those bearing activating mutations in KRAS has thus far proven elusive. Meanwhile, mutations in tumor suppressor genes TP53 and STK11 are also common in lung adenocarcinoma and frequently co-occur with KRAS mutations (13–15). Given that activation of specific oncogenic pathways can have broad effects on gene expression, it is reasonable to imagine that the genetic make-up of cancer cells could have major effects on the immune tumor microenvironment (TME), by driving specific immune-related pathways. This could be through induction of immune checkpoints, secretion of specific cytokines or production of chemokines that recruit specific cell types (12). Recent studies had shown TP53 or KRAS mutant NSCLC expressed higher levels of PD-L1 protein compared with corresponding wild-type tumors (16, 17). Meanwhile, it has been demonstrated that loss of TP53 function decreased genomic stability and was associated with defects in DNA damage repair, indicating a higher mutational burden might occur in TP53 mutational tumor (18). Therefore, we speculate that common mutations as TP53 and KRAS in lung adenocarcinoma may be served as effective predictive factors in guiding anti–PD-1/PD-L1 immunotherapy.

Here, in order to systematically address the potential mechanism that TP53/KRAS mutation mediate immune response to lung adenocarcinoma, we describe an integrative analysis that incorporates tumor mutational load, DNA MMR deficiency, intratumoral PD-L1 expression, and content of CD8+ T-cell infiltrates from cohorts of both lung adenocarcinoma repository database analysis and clinical immunotherapeutic patients. Significantly, we uncover TP53/KRAS mutation as a superiority group to anti–PD-1/PD-L1 therapies and highlight a new insight into common mutations in guiding immunotherapy.

Clinical cohorts

The Cancer Genome Atlas (TCGA), GSE72094 and Broad cohorts were retrieved from online data repository. A total of 462 patients were included in the TCGA cohort with mRNA expression profiling and gene mutation data. The GSE72094 cohort recruited 442 patients with detailed mRNA expression data and EGFR/KRAS/TP53/STK11sanger sequencing analysis (19). The Broad cohort contained 183 lung adenocarcinomas and matched normal tissues with detail information about mutation load and mutation spectrum (20). Most of the patients enrolled in the three cohorts were early-stage lung adenocarcinomas. A total of 85 lung adenocarcinomas from the Guangdong Lung Cancer Institute (GLCI), Guangdong General Hospital (GGH) were underwent whole genome sequencing (WGS). Key variables including demographic and clinical information are provided in Supplementary Table S1.

Immunotherapeutic patients

Clinical and mutation data for 34 NSCLC [29 adenocarcinoma (ADC)] patients were retrieved from cbioPortal (http://www.cbioportal.org/study.do?cancer_study_id=luad_mskcc_2015). All patients treated with pembrolizumab (anti–PD-1) from 2012 to 2013 followed the protocol NCT01295827 (KEYNOTE-001). Objective response to pembrolizumab was assessed by investigator-assessed immune-related response criteria (irRC) by a study radiologist (5).

Another group consisted of 20 NSCLC (15 ADC) patients were collected prospectively in the GLCI from August 2015 to August 2016. Eleven of them were treated with pembrolizumab and nine patients were treated with nivolumab. Tumor specimens were obtained for Sanger sequencing and IHC analysis. This study was approved by the Institutional Review Board of GLCI of GGH, and all patients provided specimens with written informed consent. Clinicopathologic and molecular information are provided in Supplementary Table S2.

mRNA expression profiling and reverse phase protein array (RPPA) analysis

For lung adenocarcinomas included in the TCGA cohort, experimental procedures regarding RNA extraction from tumors, mRNA library preparation, sequencing (on the Illumina HiSeq platform), quality control, and subsequent data processing for quantification of gene expression have been previously reported (21). Gene expression data for the GSE72094 lung adenocarcinomas have been deposited in the GEO repository (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE72094). Tumors from the GSE72094 cohort were profiled using a custom Affymetrix GeneChip. The gene expression cutoff value was chosen as median over the entire dataset (TCGA and GSE72094) to ensure all analyses of each gene were based on the same cutoff value (22, 23).

Proteomic analysis was based on RPPA from the TCGA database. The RPPA methodology and data analysis pipeline have been previously described (21). For TCGA, level 3 data were downloaded directly from the TCGA portal and utilized in subsequent analyses.

Mutation data analysis

For the discovery set, somatic mutation data (level 2) of the 462 lung adenocarcinomas were retrieved from the TCGA data portal (https://gdc.cancer.gov/). To assess the mutational load, the number of mutated genes carrying at least one nonsynonymous mutation in the coding region was computed for each tumor. Somatic mutation data of 183 lung adenocarcinomas in Broad cohort was retrieved from cbioPortal (http://www.cbioportal.org/study.do?cancer_study_id=luad_broad). Somatic substitutions and covered bases within their trinucleotide sequence context were analyzed to characterize the mutation spectrum of 183 lung adenocarcinoma. Mutation spectrum for each sample was calculated as the percentage of each of six possible single nucleotide changes (AT>CG, AT>GC, AT>TA, GC>AT, GC>CG, GC>TA) among all single-nucleotide substitutions. The most frequent mutation signatures were C→T transitions and C→A transversions.

For the validation set (GLCI), we conducted whole-exome sequencing of DNA from tumors and matched normal blood from 85 lung adenocarcinoma patients. Enriched exome libraries were sequenced on the HiSeq 2000 platform (Illumina) to >100× coverage. Alignment, base-quality score recalibration and duplicate-read removal were performed, germline variants were excluded, mutations annotated and indels evaluated as previously described (4, 5, 24). Mutations between clinical groups were compared using the Mann–Whitney test.

Gene set enrichment analysis (GSEA)

For GSEA (25), the javaGSEA Desktop Application was downloaded from http://software.broadinstitute.org/gsea/index.jsp. GSEA was used to associate the gene signature with the TP53 or KRAS mutation status (TP53-mut vs. TP53-wt; KRAS-mut vs. KRAS-wt). The genes identified to be on the leading edge of the enrichment profile were subject to pathway analysis. Fold-change values were exported for all genes and analyzed with version 2.2.0 of GSEA, using the GseaPreranked module. The normalized enrichment score (NES) is the primary statistic for examining gene set enrichment results. The nominal P value estimates the statistical significance of the enrichment score. A gene set with nominal P ≤ 0.05 was considered to be significantly enriched in genes.

Immunohistochemistry

Tumor sections were assessed immunohistochemically using PD-L1 (clone: SP142, Spring Bioscience, Inc) and CD8 (clone: C8/144B, Gene Tech (Shanghai) Co. Ltd). The IHC-stained tissue sections were scored separately by two pathologists blinded to the clinical parameters.

PD-L1 expression on tumor cells and immune cell was evaluated using a three-tiered grading system. Strong: ≥50% for tumor cell (TC) or ≥10% for immune cell (IC); weak: 5%–49% for TC or 5%–9% for IC; negative: <5% for TC or IC.

The percentages of CD8+ lymphocytes compared with the total amount of nucleated cells in the stromal compartments were assessed. Scoring cutoff points at 25% or 50% for each core according to the degree of cell densities: low density: <25%; intermediate density: 25% to 49%; high density: ≥50%.

Sanger sequencing

Genomic DNA from each sample was used for sequence analysis of EGFR exons 18–21, KRAS exons 2–3 and TP53 exons 2–11. These exons were amplified by the polymerase chain reaction (PCR) as previously described (19, 26), and the resulting PCR products were purified and labeled for sequencing using the Big Dye 3.1 Kit (Applied Biosystems) according to the manufacturer's protocol.

Statistical analyses

Statistical analyses were conducted using GraphPad Prism (version 7.01) and SPSS version 22.0 (SPSS, Inc.). Scatter dot plot and Box and whisker plots indicate median and 95% confidence interval (CI). Statistical tests were used to analyze the clinical and genomic data, including the Mann–Whitney U, χ2, Fisher exact, and Kruskal–Wallis. Kaplan–Meier curves analysis of progression-free survival (PFS) were compared using the log-rank test. All reported P values are two-tailed, and for all analyses, P ≤ 0.05 is considered statistically significant, unless otherwise specified.

Correlation between TP53 and KRAS mutation and PD-L1 expression in lung adenocarcinoma

To investigate the correlation between common mutations (TP53, KRAS, EGFR, and STK11) and immune checkpoint status in lung adenocarcinoma, we thus initially interrogated RNA sequencing (RNA-Seq) expression data from a repository database including 462 lung adenocarcinomas from The Cancer Genome Atlas (TCGA) and 442 lung adenocarcinomas from GEO repository (GSE72094). Both the TCGA and GEO databases showed significantly increased PD-L1 mRNA expression in the TP53 mutation subgroup than in other gene mutation. Specifically, the TP53 and KRAS comutated group manifested prominent higher PD-L1 expression than other comutation types (Fig. 1A).

We next sought to explore the impact of TP53 and KRAS mutation on PD-L1 expression in both PD-L1 mRNA expression profiling and RPPA analysis based on the TCGA database. The results demonstrated that it was TP53 mutation but not KRAS mutation that boosted PD-L1 expression (Supplementary Fig. S1A and S1B). Significantly, those with co-occurring mutations in TP53 and KRAS revealed the highest PD-L1 expression (both mRNA and protein level) than single gene mutation or wild-type tumors, indicating potential synergistic effect on activating PD-L1 expression (Fig. 1B). To confirm the association between TP53/KRAS mutation and PD-L1 expression as repository data demonstrated, we detected 93 lung adenocarcinoma surgical specimens using an IHC analysis (Fig. 1C; Supplementary Fig. S1C and Table S3) and immunostaining shows TP53/KRAS comutated specimens the strongest staining for the PD-L1 protein (Fig. 1D).

Next, we further analyzed the association between TP53 or KRAS mutation and other non–PD-L1 immune checkpoints. A heatmap depicted the expression level of key immune checkpoints to three groups (TP53, KRAS, and TP53/KRAS; Fig. 1E). The results displayed remarkable increased expression of most checkpoints in the TP53 mutation group while decreased expression in the KRAS mutation group. More interestingly, the TP53/KRAScomutated subgroup manifested exclusive increased expression of PD-L1; however, it showed decreased expression of some other non–PD-L1 immune inhibitory checkpoints, such as Lymphocyte Activating 3 (LAG3) and V-Set Domain Containing T Cell Activation Inhibitor 1 (VTCN1; ref. 27), implying a potential candidate population for anti–PD-1/PD-L1 immunotherapy (Fig. 1F).

TP53 mutation facilitates CD8+ T-cell infiltration and activates T-effector and interferon-γ (IFNγ) associated gene signature

The presence of tumor-infiltrating lymphocytes (TIL) is an important biomarker for predicting responses to PD-L1 blockade therapy. We continue to analyze the correlation between these above common mutations and CD8+ TIL contents in lung adenocarcinoma based on the TCGA database. Our results revealed significantly increased expression of CD8A in TP53 mutation and TP53/KRAS comutated than other groups (Fig. 2A). It has been proposed that four different types of immune tumor microenvironments (TME) exist based on the presence or absence of TIL and PD-L1 expression. To further explore whether TP53 or KRAS mutation would influence the TME, we analyzed the correlation between TP53 or KRAS mutation and TME immune types classified based on PD-L1 and CD8A expression as previously described (28, 29). Positive PD-L1 and CD8A were defined as above-median expression. We identified that the TP53 mutation group displayed a higher proportion of dual positive PD-L1 and CD8A (PD-L1+/CD8A+) than the TP53 wild-type group, while there was no difference between KRAS mutation and wild-type (Fig. 2B and C), suggesting an adaptive immune resistance TME existed in the TP53 mutation population. More importantly, the TP53/KRAS comutated subgroup showed the highest proportion of PD-L1+/CD8A+ than the TP53 or KRAS single mutation and wild-type group (Fig. 2D). These observations were further confirmed by our IHC analysis that TP53/KRAS comutated patients manifested a strong staining of PD-L1 and high intensity of CD8+TILs (Fig. 2E).

Given that TP53 mutation had effects on the TME in lung adenocarcinoma, we subsequently sought to assess the relationship between TP53 mutation and T-effector and IFNγ-associated gene signature, which have previously been associated with activated T cells, immune cytolytic activity, and IFNγ release (30, 31). An integrated heatmap depicting expression levels of T-effector and IFN-γ associated gene signature in tumors with TP53 mutation compared with TP53 wild-type. We identified significant increased expression of both T-effector and IFNγ-associated genes in the TP53 mutation group, while there were no differences between KRAS mutation and wild-type, indicating preexisting immunity within TP53 mutation tumor tissue (Fig. 2F and G).

TP53 and KRAS mutation shows increased mutation burden and distinct mutation spectrum

Recent studies have highlighted the relevance of tumor mutational loads and response to PD-1 blockade (5). We next speculate whether there are some common mutations in lung adenocarcinoma that affect the whole tumor mutational profile and change the tumor antigenicity. We first analyzed the TCGA and Broad databases as discovery set. The TCGA analysis showed significantly increased mutational loads in the TP53 mutation group (median, 325), followed by KRAS (median, 179) and STK11 (median, 132) mutation, EGFR (median, 60) mutation tumor had the lowest mutational loads. Meanwhile, the TP53/KRAS comutated subgroup showed significantly higher mutational loads (median, 358) than other comutated subgroup (Fig. 3A). We then tested these findings using another dataset (Broad), which consisted of 183 lung adenocarcinomas with detailed somatic mutation data, and confirmed that TP53 and KRAS mutation and the TP53/KRAS comutated group had higher mutational loads than other groups (Fig. 3A). To further verify these findings, a total of 85 lung adenocarcinomas from GLCI detected by whole genome sequencing were defined as the validation set. GLCI data manifested the similar results with the discovery set that TP53 mutation and the TP53/KRAS comutated group had higher mutational loads than other groups (Fig. 3B). It is well known that tobacco exposure was responsible for much of the mutagenesis in NSCLC. Multivariate linear regression analysis of mutation count in patients stratified by smoking status manifested that TP53 mutation was an independent factor responsible for increased mutation burden regardless of smoking status, while KRAS mutation showed increased mutation burden only in nonsmokers (Supplementary Table S4).

Previous studies have established the notion that somatic mutations are primarily GC>TA transversions (32). We next investigated whether these above common mutations could affect tumor mutation spectrum by using a TCGA cohort. Transversion-high (TH) and transversion-low (TL) was based on smoking history and GC>AT, GC>TA frequency as previously described (5, 21). We can identify KRAS mutations were significantly enriched in the TH cohort, while EGFR mutations were significantly enriched in the TL group (Supplementary Fig. S2). Consistent with TCGA results, the Broad dataset showed a high rate of transversion/transition (Tv/Ti) in KRAS mutation and the TP53/KRAS comutated group while the lowest rate Tv/Ti in EGFR mutation (Fig. 3C). Notably, TP53 and KRAS mutation was significantly correlated with high somatic mutations, high rate of Tv/Ti and C>A transversion and high smoking index (pack-years; Fig. 3D).

Impact of TP53 and KRAS mutation on the cell cycle, DNA replication, and damage repair–related genes

We sought to determine whether alterations in DNA replication and damage repair–related genes resulted from TP53 or KRAS mutation could account for differential mutation burden and mutation spectrum. GSEA reveals prominent enrichment of signatures relating to cell cycle, DNA replication and DNA repair in both the TP53 and KRAS mutation groups. However, there were distinct differences between these two groups. TP53 mutation predominantly led to acceleration of cell-cycle and DNA replication, which potentially increased mutation probability, for unrepaired DNA damages that do not kill the cell by blocking replication would tend to cause replication errors and thus mutation. KRAS mutation manifested various defects of DNA repair including MMR, nucleotide excision repair (NER), and base excision repair (BER) that greatly enhanced point mutation (Fig. 4A).

Recent studies showed that POLE mutation is associated with disruption of the exonuclease activity required for DNA proofreading and results in a high mutational burden or an “ultramutator” phenotype (33, 34). We identified significantly increased mutation frequencies of POLE in the TP53 mutation group (P = 0.002) while decreased mutation frequencies of POLE in the EGFR and STK11 mutation groups compare with their corresponding wild-type group, indicating TP53 mutation tends to cause DNA replication errors (Fig. 4B).

We next determined the correlation between these common mutations and DNA damage repair–related genes. DNA double-strand breaks (DSB) elicit that DNA damage response largely relies on the activity of ataxia telangiectasia mutated (ATM), which have been found to be mutated in human disorders associated with genome instability (35, 36). The results had revealed that a high frequency of ATM mutation was found predominantly in the KRAS and STK11 mutation groups, and ATM protein analysis further confirmed that waning expression of ATM protein was specifically found in the KRAS and STK11 mutation groups (Fig. 4B and C).

MMR-deficient tumors were recently shown susceptibility to checkpoint blockade immunotherapy (7). Our former GSEA identified that KRAS mutation was negatively correlated with MMR-related gene expression, and we next verified whether KRAS or other genes mutation affected the mutation status and protein expression of MMR-related genes. Four primary MMR-related genes, including MSH2, MSH6, MLH1, and PMS2, were coanalyzed. Consistent with GSEA, high mutation frequency of MMR-related genes was exclusively identified in the KRAS mutation group. Furthermore, the protein of MSH2 and MSH6 was significantly decreased in tumors with KRAS mutation; however, it was increased in tumors with TP53 mutation, suggesting that KRAS mutation might be a potential driver agent to induce MMR deficiency and in consequence produce more neoantigens (Fig. 4B and C).

Patients with TP53 or KRAS mutation, especially co-occurring TP53/KRAS mutations, show favorable clinical benefit to anti–PD-1 treatment

TP53 and KRAS mutation showed remarkable effects on regulating PD-L1 expression, facilitating T-cell infiltration and augmenting tumor immunogenicity. We presumed that patients with these two mutations probably had increased sensitivity to PD-1 blockade immunotherapy. In support of this hypothesis, publicly available trial data (MSKCC, KEYNOTE-001) were reanalyzed. A total of 34 advanced NSCLC (29 ADC) patients were prescribed pembrolizumab from 2012 to 2013 following the NCT01295827 protocol. All tumor tissues underwent whole-exome sequencing. We observed significantly increased nonsynonymous mutation and candidate neoantigen burden in the TP53 or KRAS mutation group compared with the wild-type group (Fig. 5A and B). Consistent with the analysis from mutation burden, there was a strikingly high proportion of TH in the TP53 and KRAS mutation group (Fig. 5C).

Notably, TP53 or KRAS mutation patients obtained a significantly prolonged progression-free survival (PFS) compared with wild-type patients who underwent pembrolizumab treatment (median PFS, TP53-mut vs. KRAS-mut vs. wild-type: 14.5 vs. 14.7 vs. 3.5 months, P = 0.012; Fig. 5D and E). Most of TP53 or KRAS mutation patients enjoyed a durable clinical benefit during treatment, while most of wild-type patients showed no durable benefit (Fig. 5F). More importantly, 4 lung adenocarcinoma patients who concomitantly harbored TP53 and KRAS mutation manifested superior PFS, and all of them had a durable clinical benefit. Meanwhile, these patients displayed a high mutation burden, high rate of transversion, and strong staining of PD-L1 (Fig. 5G).

To further confirm these observations from a public database, we prospectively collected 20 NSCLC (15 ADC) patients who were treated with pembrolizumab (n = 11) or nivolumab (n = 9) from August 2015 to August 2016 in our center (GLCI). All of the patients underwent at least one assessment after baseline. Patients' tissues were used for DNA sequencing for EGFR, KRAS, and TP53, and paraffin-embedded specimens were detected for IHC analysis of PD-L1 and CD8 (Table 1). Eight of the patients showed TP53 mutation and 3 patients showed KRAS mutation. Up to August 25, 2016, 6 patients experienced partial response (PR). Two of them concomitantly harbored TP53 and KRAS mutation, 3 patients with single KRAS or TP53 mutation, and 1 patient without common mutation. Besides, 6 patients were evaluated as progression disease (PD) after 2 or 3 cycles of immunotherapy and 2 of them harbored EGFR mutation (Fig. 6A and Table 1). Patients with TP53 and/or KRAS mutation showed prolonged PFS than both genes negative patients treated with PD-1 inhibitors. Six patients were assessed as PR and five of them had an ongoing response (Fig. 6B).

Next, we focused on 1 patient who had a durable clinical benefit (DCB) with TP53 and KRAS comutation (Fig. 6C). IHC was used for analysis of PD-L1, CD8, and MMR-related genes, including MSH2, MSH6, MLH1, and PMS2. Consistent with our expectation, the patient showed strong staining for both PD-L1 and CD8. Meanwhile, four MMR-related genes displayed different intensity of immunostaining: weak positive for MLH1, moderate positive for MSH6 and PMS2, and strong positive for MSH2, suggesting that a potential possibility of MMR deficiency existed in this tumor (Fig. 6D).

Although the expression of PD-L1 on the surface of tumor cells, as measured by IHC, is recommended as a predictive factor to identify patients who would benefit from PD-1 blockade, not all PD-L1–positive patients respond well (10, 37). The underlying biology of such limitations has not been clearly understood until recent studies, which showed that the presence of TILs and mutational burden correlated with T-effector signature and immunogenic features that supported the response to anti–PD-1/PD-L1 therapy (5, 8, 12, 38, 39). Here, we first identified a group of oncogenic driver (EGFR and KRAS) and tumor suppressor (TP53 and STK11) mutations of lung adenocarcinoma that distinctively affected immune checkpoints expression, T-cell infiltration, and tumor immunogenicity. Specifically, our findings revealed that TP53 mutation remarkably increased PD-L1 expression and facilitated CD8+ T-cell infiltration, and accompanied with a higher proportion of dual positive PD-L1 and CD8A than other mutation groups. Furthermore, TP53-mutated tumors showed prominently increased somatic mutation burden and specifically enriched in the TH subset. Previous studies have classified the TME into four groups on the basis of PD-L1 expression and TIL recruitment. These include type I (PD-L1 positive with TILs driving adaptive immune resistance), type II (PD-L1 negative with no TIL indicating immune ignorance), type III (PD-L1 positive with no TIL indicating intrinsic induction), and type IV (PD-L1 negative with TIL indicating the role of other suppressors in promoting immune tolerance; refs. 28, 29). Significantly, type I is lately thought to be associated with a high mutational burden, PD-L1 amplification, and oncogenic viral infection, which defines a subtype sensitivity to PD-1 blockade (29). These notions, to some extent, support our findings that TP53 mutation represents a state of adaptive immune resistance and a high immunogenicity, which contributes to a probable sensitivity to PD-1 blockade (40). Nevertheless, we could also discover a fact that TP53 mutation equally enhanced some other non–PD-L1 immune–inhibitory checkpoints expression, such as LAG3 and VTCN1, which might serve as potential primary resistance to TP53 mutation patients treated with anti–PD-1/PD-L1 (41).

Recent studies based on a phase III clinical trial have identified that patients who harbored an EGFR mutation displayed unfavorable response to PD-1 blockade than those with a wild-type EGFR (42–44). This may be the first finding in which driver mutation of NSCLC was involved in altering sensitivity to immunotherapy. The most likely explanation is that patients with EGFR mutation were prone to produce a weak immunogenic tumor and an immunosuppressive TME. These perspectives were also confirmed in our study that EGFR mutation showed the lowest mutation burden and lowest rate of Tv/Ti than other mutations. Besides, EGFR mutation did not increase the expression of PD-L1, like others reported, but with a relatively lower expression than TP53 and KRAS mutation (45), which further supported their hyposensitivity to PD-1 blockade. KRAS mutation was the second important oncogenic driver mutation in lung adenocarcinoma. The development of more effective treatment strategies for patients with KRAS mutation is hampered by the biologic and phenotypic heterogeneity of KRAS-mutant tumors. More recently, some studies suggested that patients with activating mutations in KRAS may probably benefit from PD-1 blockade, but the underlying mechanisms remained elusive and most of the researchers attributed this predilection to the association between smoking and the presence of KRAS mutations (5, 42). In this study, we uncovered potential mechanisms that account for this correlation. We discovered a significant increase of mutation load in KRAS-mutant tumors. Particularly, a predominant higher proportion of Tv/Ti was also found in this subgroup. Furthermore, we observed that KRAS mutations defected DNA repair, especially in MMR, which supported the notion that MMR deficiency acted as a favorable agent for PD-1 blockade (7).

It is well known that smoking-related lung cancers were characterized by greater mutation burden, higher rate of transversion, and more frequent KRAS mutation than that occurred in never smokers (21, 32, 46, 47). More recently, studies have demonstrated the association of PD-L1 expression with significant smoking history (48). In our study, we discovered that TP53 mutation, especially TP53/KRAS comutation, showed increased PD-L1 expression and augmented tumor immunogenicity. To confirm whether these correlations are more related to tobacco exposure, a multivariate linear regression analysis of mutation count and PD-L1 expression stratified by smoking status was performed. We demonstrated that TP53 mutation was responsible for increased mutation burden and PD-L1 expression independent of smoking status (Supplementary Tables S4 and S5). Recent studies based on subgroup analysis demonstrated those with a history of current or ever smoking showed much better benefits of PD-1 blockade than non-smokers. So we can imagine current or ever-smoker patients with TP53 and/or KRAS mutation may be the optimal population for PD-1 blockade immunotherapy.

Co-occurring mutations in TP53 and KRAS have recently been defined as a specific cluster associated with activation of antitumor immunity and immune tolerance/escape (49). Interestingly, our study identified TP53 and KRAS comutant tumors manifested exclusive increased expression of PD-L1 and a highest proportion of PD-L1+/CD8A+ than TP53 or KRAS single mutation. Meanwhile, TP53/KRAS dual mutation showed predominant increased mutation burden and enriched in the TH subset. Consistent with these preclinical predictions, the clinical analysis on the base of MSKCC and our center database had further confirmed that those with co-occurring mutations in TP53 and KRAS showed remarkable clinical benefit from pembrolizumab. These results implicated a possibility that TP53 and KRAS mutation played a role with synergistic and complementary in regulating immune biomarkers, which gave rise to a responsive TME with adaptive immune resistance and high immunogenicity. However, these findings were established in a relatively small cohort and even fewer patients with TP53 and KRAS comutation. Based on the preliminary evidence, a prospective study with a larger sample size of TP53/KRAS mutation and PD-L1 expression for response to PD-1 blockade is warranted in the future.

Taken together, the results of this study provided an insight into immune regulation driving by some common mutations of lung adenocarcinoma. We discovered a prominent significance of TP53 and KRAS mutation in boosting PD-L1 expression, facilitating T-cell infiltration, and augmenting tumor immunogenicity. This work provided evidence that TP53 and KRAS mutation in lung adenocarcinoma might be served as a pair of potential predictive factors in guiding PD-1 blockade immunotherapy.

No potential conflicts of interest were disclosed.

Conception and design: Z.-Y. Dong, Y.-L. Wu

Development of methodology: Z.-Y. Dong, J. Su, Z. Xie, H.-H. Yan, S.-P. Wu, Y.-L. Wu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z.-Y. Dong, W.-Z. Zhong, X.-C. Zhang, S.-Y. Liu, H.-Y. Tu, H.-J. Chen, Y.-L. Sun, Q. Zhou, X.-N. Yang, J.-X. Lin, H.-R. Zhai, L.-X. Yan, R.-Q. Liao

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z.-Y. Dong, W.-Z. Zhong, X.-C. Zhang, J. Su, Z. Xie, S.-Y. Liu, H.-Y. Tu, H.-J. Chen, Y.-L. Sun, Q. Zhou, H.-H. Yan, L.-X. Yan

Writing, review, and/or revision of the manuscript: Z.-Y. Dong, W.-Z. Zhong, J. Su, S.-Y. Liu, H.-Y. Tu, H.-J. Chen, Q. Zhou, X.-N. Yang, J.-X. Lin, H.-H. Yan, L.-X. Yan, Y.-L. Wu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z.-Y. Dong, J. Su, Z. Xie, J.-J. Yang, H.-R. Zhai

Study supervision: Z.-Y. Dong, W.-Z. Zhong, X.-C. Zhang, Y.-L. Wu

The authors thank Professor Li Liu and Professor De-Hua Wu's team in Nanfang Hospital for providing lung adenocarcinoma–related public databases (TCGA, Broad and GEO databases) and helping with bioinformatic analysis. The authors also appreciate support from doctors and patients.

This study was supported by the Guangzhou Science and Technology Bureau (grant nos. 2014Y2-00050 and 2014Y2-00545), The National Key Research and Development Program of China (grant no. 2016YFC1303800), the Special Fund for Research in the Public Interest from the National Health and Family Planning Commission of the People's Republic of China (grant 201402031), the Project of National Natural Science Foundation (grant no. 81372285), and the Guangdong Provincial Applied Science and Technology Research and Development Program (grant no. 2016B020237006).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Brahmer
JR
,
Tykodi
SS
,
Chow
LQ
,
Hwu
WJ
,
Topalian
SL
,
Hwu
P
, et al
Safety and activity of anti-PD-L1 antibody in patients with advanced cancer
.
N Engl J Med
2012
;
366
:
2455
65
.
2.
O'Sullivan Coyne
G
,
Madan
RA
,
Gulley
JL
. 
Nivolumab: promising survival signal coupled with limited toxicity raises expectations
.
J Clin Oncol
2014
;
32
:
986
8
.
3.
Carbognin
L
,
Pilotto
S
,
Milella
M
,
Vaccaro
V
,
Brunelli
M
,
Calio
A
, et al
Differential activity of nivolumab, pembrolizumab and MPDL3280A according to the tumor expression of programmed death-ligand-1 (PD-L1): sensitivity analysis of trials in melanoma, lung and genitourinary cancers
.
PLoS One
2015
;
10
:
e0130142
.
4.
Snyder
A
,
Makarov
V
,
Merghoub
T
,
Yuan
J
,
Zaretsky
JM
,
Desrichard
A
, et al
Genetic basis for clinical response to CTLA-4 blockade in melanoma
.
N Engl J Med
2014
;
371
:
2189
99
.
5.
Rizvi
NA
,
Hellmann
MD
,
Snyder
A
,
Kvistborg
P
,
Makarov
V
,
Havel
JJ
, et al
Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer
.
Science
2015
;
348
:
124
8
.
6.
Blank
CU
,
Haanen
JB
,
Ribas
A
,
Schumacher
TN
. 
Cancer Immunology. The "cancer immunogram"
.
Science
2016
;
352
:
658
60
.
7.
Le
DT
,
Uram
JN
,
Wang
H
,
Bartlett
BR
,
Kemberling
H
,
Eyring
AD
, et al
PD-1 blockade in tumors with mismatch-repair deficiency
.
N Engl J Med
2015
;
372
:
2509
20
.
8.
Tang
H
,
Wang
Y
,
Chlewicki
LK
,
Zhang
Y
,
Guo
J
,
Liang
W
, et al
Facilitating T cell infiltration in tumor microenvironment overcomes resistance to PD-L1 blockade
.
Cancer Cell
2016
;
29
:
285
96
.
9.
Dong
ZY
,
Wu
SP
,
Liao
RQ
,
Huang
SM
,
Wu
YL
. 
Potential biomarker for checkpoint blockade immunotherapy and treatment strategy
.
Tumour Biol
2016
;
37
:
4251
61
.
10.
Herbst
RS
,
Soria
JC
,
Kowanetz
M
,
Fine
GD
,
Hamid
O
,
Gordon
MS
, et al
Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients
.
Nature
2014
;
515
:
563
7
.
11.
Shukuya
T
,
Carbone
DP
. 
Predictive markers for the efficacy of anti-PD-1/PD-L1 antibodies in lung cancer
.
J Thorac Oncol
2016
;
11
:
976
88
.
12.
Topalian
SL
,
Taube
JM
,
Anders
RA
,
Pardoll
DM
. 
Mechanism-driven biomarkers to guide immune checkpoint blockade in cancer therapy
.
Nat Rev Cancer
2016
;
16
:
275
87
.
13.
Pao
W
,
Girard
N
. 
New driver mutations in non-small-cell lung cancer
.
Lancet Oncol
2011
;
12
:
175
80
.
14.
Ding
L
,
Getz
G
,
Wheeler
DA
,
Mardis
ER
,
McLellan
MD
,
Cibulskis
K
, et al
Somatic mutations affect key pathways in lung adenocarcinoma
.
Nature
2008
;
455
:
1069
75
.
15.
Gibbons
DL
,
Byers
LA
,
Kurie
JM
. 
Smoking, p53 mutation, and lung cancer
.
Mol Cancer Res
2014
;
12
:
3
13
.
16.
Cortez
MA
,
Ivan
C
,
Valdecanas
D
,
Wang
X
,
Peltier
HJ
,
Ye
Y
, et al
PDL1 regulation by p53 via miR-34
.
J Natl Cancer Inst
2016
;
108
:
djv303
.
17.
Ji
M
,
Liu
Y
,
Li
Q
,
Li
X
,
Ning
Z
,
Zhao
W
, et al
PD-1/PD-L1 expression in non-small-cell lung cancer and its correlation with EGFR/KRAS mutations
.
Cancer Biol Ther
2016
;
17
:
407
13
.
18.
Soussi
T
,
Beroud
C
. 
Assessing TP53 status in human tumours to evaluate clinical outcome
.
Nat Rev Cancer
2001
;
1
:
233
40
.
19.
Schabath
MB
,
Welsh
EA
,
Fulp
WJ
,
Chen
L
,
Teer
JK
,
Thompson
ZJ
, et al
Differential association of STK11 and TP53 with KRAS mutation-associated gene expression, proliferation and immune surveillance in lung adenocarcinoma
.
Oncogene
2016
;
35
:
3209
16
.
20.
Imielinski
M
,
Berger
AH
,
Hammerman
PS
,
Hernandez
B
,
Pugh
TJ
,
Hodis
E
, et al
Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing
.
Cell
2012
;
150
:
1107
20
.
21.
Cancer Genome Atlas Research N
. 
Comprehensive molecular profiling of lung adenocarcinoma
.
Nature
2014
;
511
:
543
50
.
22.
Deng
L
,
Gyorffy
B
,
Na
F
,
Chen
B
,
Lan
J
,
Xue
J
, et al
Association of PDCD1 and CTLA-4 gene expression with clinicopathological factors and survival in non–small-cell lung cancer: results from a large and pooled microarray database
.
J Thorac Oncol
2015
;
10
:
1020
6
.
23.
Choi
YE
,
Meghani
K
,
Brault
ME
,
Leclerc
L
,
He
YJ
,
Day
TA
, et al
Platinum and PARP inhibitor resistance due to overexpression of MicroRNA-622 in BRCA1-mutant ovarian cancer
.
Cell Rep
2016
;
14
:
429
39
.
24.
Ho
AS
,
Kannan
K
,
Roy
DM
,
Morris
LG
,
Ganly
I
,
Katabi
N
, et al
The mutational landscape of adenoid cystic carcinoma
.
Nat Genet
2013
;
45
:
791
8
.
25.
Subramanian
A
,
Tamayo
P
,
Mootha
VK
,
Mukherjee
S
,
Ebert
BL
,
Gillette
MA
, et al
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles
.
Proc Natl Acad Sci U S A
2005
;
102
:
15545
50
.
26.
Silwal-Pandit
L
,
Vollan
HK
,
Chin
SF
,
Rueda
OM
,
McKinney
S
,
Osako
T
, et al
TP53 mutation spectrum in breast cancer is subtype specific and has distinct prognostic relevance
.
Clin Cancer Res
2014
;
20
:
3569
80
.
27.
Anderson
AC
,
Joller
N
,
Kuchroo
VK
. 
Lag-3, Tim-3, and TIGIT: Co-inhibitory receptors with specialized functions in immune regulation
.
Immunity
2016
;
44
:
989
1004
.
28.
Teng
MW
,
Ngiow
SF
,
Ribas
A
,
Smyth
MJ
. 
Classifying cancers based on T-cell infiltration and PD-L1
.
Cancer Res
2015
;
75
:
2139
45
.
29.
Ock
CY
,
Keam
B
,
Kim
S
,
Lee
JS
,
Kim
M
,
Kim
TM
, et al
Pan-cancer immunogenomic perspective on the tumor microenvironment based on PD-L1 and CD8 T-cell infiltration
.
Clin Cancer Res
2016
;
22
:
2261
70
.
30.
Fehrenbacher
L
,
Spira
A
,
Ballinger
M
,
Kowanetz
M
,
Vansteenkiste
J
,
Mazieres
J
, et al
Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial
.
Lancet
2016
;
387
:
1837
46
.
31.
Rooney
MS
,
Shukla
SA
,
Wu
CJ
,
Getz
G
,
Hacohen
N
. 
Molecular and genetic properties of tumors associated with local immune cytolytic activity
.
Cell
2015
;
160
:
48
61
.
32.
Lee
W
,
Jiang
Z
,
Liu
J
,
Haverty
PM
,
Guan
Y
,
Stinson
J
, et al
The mutation spectrum revealed by paired genome sequences from a lung cancer patient
.
Nature
2010
;
465
:
473
7
.
33.
Mehnert
JM
,
Panda
A
,
Zhong
H
,
Hirshfield
K
,
Damare
S
,
Lane
K
, et al
Immune activation and response to pembrolizumab in POLE-mutant endometrial cancer
.
J Clin Invest
2016
;
126
:
2334
40
.
34.
van Gool
IC
,
Eggink
FA
,
Freeman-Mills
L
,
Stelloo
E
,
Marchi
E
,
de Bruyn
M
, et al
POLE proofreading mutations elicit an antitumor immune response in endometrial cancer
.
Clin Cancer Res
2015
;
21
:
3347
55
.
35.
Caron
P
,
Choudjaye
J
,
Clouaire
T
,
Bugler
B
,
Daburon
V
,
Aguirrebengoa
M
, et al
Non-redundant functions of ATM and DNA-PKcs in response to DNA double-strand breaks
.
Cell Rep
2015
;
13
:
1598
609
.
36.
Lima
M
,
Bouzid
H
,
Soares
DG
,
Selle
F
,
Morel
C
,
Galmarini
CM
, et al
Dual inhibition of ATR and ATM potentiates the activity of trabectedin and lurbinectedin by perturbing the DNA damage response and homologous recombination repair
.
Oncotarget
2016
;
7
:
25885
901
.
37.
Garon
EB
,
Rizvi
NA
,
Hui
R
,
Leighl
N
,
Balmanoukian
AS
,
Eder
JP
, et al
Pembrolizumab for the treatment of non-small-cell lung cancer
.
N Engl J Med
2015
;
372
:
2018
28
.
38.
Alexandrov
LB
,
Nik-Zainal
S
,
Wedge
DC
,
Aparicio
SA
,
Behjati
S
,
Biankin
AV
, et al
Signatures of mutational processes in human cancer
.
Nature
2013
;
500
:
415
21
.
39.
Schumacher
TN
,
Schreiber
RD
. 
Neoantigens in cancer immunotherapy
.
Science
2015
;
348
:
69
74
.
40.
Tumeh
PC
,
Harview
CL
,
Yearley
JH
,
Shintaku
IP
,
Taylor
EJ
,
Robert
L
, et al
PD-1 blockade induces responses by inhibiting adaptive immune resistance
.
Nature
2014
;
515
:
568
71
.
41.
Koyama
S
,
Akbay
EA
,
Li
YY
,
Herter-Sprie
GS
,
Buczkowski
KA
,
Richards
WG
, et al
Adaptive resistance to therapeutic PD-1 blockade is associated with upregulation of alternative immune checkpoints
.
Nat Commun
2016
;
7
:
10501
.
42.
Borghaei
H
,
Paz-Ares
L
,
Horn
L
,
Spigel
DR
,
Steins
M
,
Ready
NE
, et al
Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer
.
N Engl J Med
2015
;
373
:
1627
39
.
43.
Herbst
RS
,
Baas
P
,
Kim
DW
,
Felip
E
,
Perez-Gracia
JL
,
Han
JY
, et al
Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial
.
Lancet
2016
;
387
:
1540
50
.
44.
Gainor
JF
,
Shaw
AT
,
Sequist
LV
,
Fu
X
,
Azzoli
CG
,
Piotrowska
Z
, et al
EGFR mutations and ALK rearrangements are associated with low response rates to PD-1 pathway blockade in non-small cell lung cancer: a retrospective analysis
.
Clin Cancer Res
2016
;
22
:
4585
93
.
45.
Chen
N
,
Fang
W
,
Zhan
J
,
Hong
S
,
Tang
Y
,
Kang
S
, et al
Upregulation of PD-L1 by EGFR activation mediates the immune escape in EGFR-driven NSCLC: implication for optional immune targeted therapy for NSCLC patients with EGFR mutation
.
J Thorac Oncol
2015
;
10
:
910
23
.
46.
Govindan
R
,
Ding
L
,
Griffith
M
,
Subramanian
J
,
Dees
ND
,
Kanchi
KL
, et al
Genomic landscape of non-small cell lung cancer in smokers and never-smokers
.
Cell
2012
;
150
:
1121
34
.
47.
Pfeifer
GP
,
Denissenko
MF
,
Olivier
M
,
Tretyakova
N
,
Hecht
SS
,
Hainaut
P
. 
Tobacco smoke carcinogens, DNA damage and p53 mutations in smoking-associated cancers
.
Oncogene
2002
;
21
:
7435
51
.
48.
Huynh
TG
,
Morales-Oyarvide
V
,
Campo
MJ
,
Gainor
JF
,
Bozkurtlar
E
,
Uruga
H
, et al
Programmed cell death ligand 1 expression in resected lung adenocarcinomas: association with immune microenvironment
.
J Thorac Oncol
2016
;
11
:
1869
78
.
49.
Skoulidis
F
,
Byers
LA
,
Diao
L
,
Papadimitrakopoulou
VA
,
Tong
P
,
Izzo
J
, et al
Co-occurring genomic alterations define major subsets of KRAS-mutant lung adenocarcinoma with distinct biology, immune profiles, and therapeutic vulnerabilities
.
Cancer Discov
2015
;
5
:
860
77
.

Supplementary data