Neutrophil in the suppressed immune microenvironment: Critical prognostic factor for lung adenocarcinoma patients with KEAP1 mutation

Abstract

Purpose: It is still unclear whether KEAP1 mutation is detrimental to immunotherapy of lung adenocarcinoma (LUAD) patients, we try to analyse the exact changes in the TME in LUAD patients with KEAP1 mutations and to identify key factors influencing prognosis.

Experimental design: A total of 1,029 patients with lung squamous carcinoma (LUSC) or LUAD with data obtained from The Cancer Genome Atlas were included in this study. The TME and OS of patients with LUAD stratified by mutant versus wild-type KEAP1 status were comprehensively measured. Moreover, we classified LUAD patients with KEAP1 mutations into three subtypes, by unsupervised consensus clustering. We further analysed the TME, OS, commutated genes and metabolic pathways of different subgroups. A total of 40 LUAD patients underwent immunotherapy were collected and classified into mutant KEAP1 group and wild-type KEAP1 group. We also conducted immunohistochemical staining in KEAP1-MT groups.

Result: Suppressed TME was observed not only in LUAD patients but also in LUSC patients. LUAD patients with mutant KEAP1 underwent immunotherapy had worse PFS than wild-type KEAP1. Unsupervised consensus clustering analysis suggested that the three subtypes of patients exhibited different densities of neutrophil infiltration and had different OS results: cluster 2 patients had significantly higher levels of neutrophils had significantly worse prognoses than those of patients in clusters 1 and 3 and patients with wild-type KEAP1. Univariate and multivariate Cox analyses proved that a high density of neutrophils was significantly associated with worse OS and immunohistochemical staining proved that shorter PFS showed high density of neutrophils.

Conclusion: KEAP1 mutation significantly suppresses the tumour immune microenvironment in LUAD patients. LUAD patients with mutant KEAP1 underwent immunotherapy had worse PFS than with wild-type KEAP1. Neutrophils may play an important role in the prognosis of LUAD patients with KEAP1 mutations and may provide a promising therapeutic target.

Keywords: KEAP1; NSCLC; lung adenocarcinoma; neutrophil; tumour microenvironment.

FIGURE 1

The level of ssGSEA of immune cells in lung cancer samples in TCGA is tested by Wilcox between Keap1 mutation group (KEAP1-MT) and wild group (KEAP1-WT). (A). Comparison of immune infiltration levels between Keap1 mutant group and wild group in combined LUAD and LUSC samples. (B). Comparison of immune infiltration levels between KEAP1-MT and KEAP1-WT in LUAD samples. (C). Comparison of immune infiltration levels between KEAP1-MT and KEAP1-WT in LUSC samples. (“ns” means p ≥ 0.05; “*“means p < 0.05; “**“means p < 0.01; “***” means p < 0.001).

FIGURE 2

Kaplan Meier curve shows the difference of overall survival rate among different groups. (A–C). survival curves between KEAP1 mutant samples and KEAP1 wild-type samples in lung cancer samples, LUAD samples and LUSC samples. (D). Survival curves of three types of clusters and KEAP1 wild-type samples in LUADs. (E). Survival curves of two types of clusters in LUSCs.

FIGURE 3

Unsupervised consistency clustering is carried out through concensusclusterplus. (A). Consistency clustering matrix of LUAD mutation samples. (B). Consistency clustering CDF diagram of LUAD mutation samples. (C). Lollipop chart showed different KEAP1 mutation sites in LUADs. (D). The subtype expression results obtained by LUAD consistent clustering and the gene mutation of the top 10 mutation variables in cluster2.

FIGURE 4

Cox regression analysis of immune infiltration level and consistency clustering on OS and the mechanism of poor prognosis of cluster2. (A). Univariate Cox regression of the immune infiltration level and subgroups of consistency clustering. (B). Multivariate Cox regression of the immune infiltration level and subgroups of consistency clustering. (C). The enrichment and analysis of related differential genes between cluster2 and non-cluster2 group. (D). Chi-square test results of TOP10 mutant genes among different subtypes. (E). Comparison of immune cell infiltration levels between KEAP1 mutant samples and wild samples. (“ns” means p ≥ 0.05; “*“means p < 0.05; “**“means p < 0.01; “***” means p < 0.001).

FIGURE 5

Kaplan Meier curve shows the difference of overall survival rate between high and low neutrophil in tumor immune microenvironment in different groups. (A). The difference of OS between high and low neutrophil in tumor immune microenvironment in LUADs. (B). The difference of OS between high and low neutrophil in tumor immune microenvironment in KEAP1 wt. (C). The difference of OS between high and low neutrophil in tumor immune microenvironment in KEAP1 mt. (D). Correlation heat map of KEAP1 mutation differential genes and cluster2 differential genes intersection genes, type 2 T helper cell, neutrophil marker genes and genes related to nitrogen metabolism pathway.

FIGURE 6

(A). Sector diagram reveals difference of Best of response (BOR) among KEAP1-MT group and KEAP1-WT group of CZHR cohorts. (B). Kaplan Meier curve shows the difference of progression-free survival (PFS) among KEAP1-MT group and KEAP1-WT group of CZHR cohorts. (C). Immunohistochemical staining of tissue from lung cancer reveals different density of neutrophile granulocyte in KEAP1-MT group with different PFS.

FIGURE 7

Two KEAP1 mutated patients with different density of neutrophile granulocyte received standard therapy of albumin paclitaxel + carboplatin + Coreda for 3 months and showed totally different responds. (A1,A2). Photograph of PD-L1 and tumour proportion score (TPS) (B1,B2). Immunohistochemical staining of tissue from lung cancer reveals density of neutrophile granulocyte. (C1, C2). Chest CT of pulmonary tumor at initial diagnosis. (D1,D2). Chest CT of pulmonary tumor after receiving standard therapy of albumin paclitaxel + carboplatin + Coreda for 3 months.