Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Activated Mast Cells Combined with NRF2 Predict Prognosis for Esophageal Cancer

Activated Mast Cells Combined with NRF2 Predict Prognosis for Esophageal Cancer Hindawi Journal of Oncology Volume 2023, Article ID 4211885, 13 pages https://doi.org/10.1155/2023/4211885 Research Article Activated Mast Cells Combined with NRF2 Predict Prognosis for Esophageal Cancer 1 1 1 2 1 Xinxin Guo, Weitao Shen, Mingjun Sun, Junjie Lv , and Ran Liu Key Laboratory of Environmental Medicine Engineering, Ministry of Education, School of Public Health, Southeast University, Nanjing 210009, China Cancer Institute of Fudan University, Fudan University, Shanghai 200032, China Correspondence should be addressed to Junjie Lv; lylvjunjie@163.com and Ran Liu; ranliu@seu.edu.cn Received 10 September 2022; Revised 12 December 2022; Accepted 20 December 2022; Published 4 January 2023 Academic Editor: Shuanglin Qin Copyright © 2023 Xinxin Guo et al. Tis is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background. Esophageal cancer (EC) had the sixth-highest mortality rate of all cancers due to its poor prognosis. Immune cells and mutation genes infuenced the prognosis of EC, but their combined efect on predicting EC prognosis was unknown. In this study, we comprehensively analyzed the immune cell infltration (ICI) and mutation genes and their combined efects for predicting prognosis in EC. Methods. Te CIBERSORT and ESTIMATE algorithms were used to analyse the ICI scape based on the TCGA and GEO databases. EC tissues and pathologic sections from Huai’an, China, were used to verify the key immune cells and mutation genes and their interactions. Results. Stromal/immune score patterns and ICI/gene had no statistical signifcance in overall survival (OS) (p> 0.05). Te combination of ICI and tumor mutation burden (TMB) showed that the high TMB and high ICI score group had the shortest OS (p � 0.004). We recognized that the key mutation gene NRF2 was signifcantly diferent in the high/low ICI score subgroups (p � 0.002) and positivity with mast cells (MCs) (p< 0.05). Trough experimental validation, we found that the MCs and activated mast cells (AC-MCs) were more infltration in stage II/III (p � 0.032; p � 0.013) of EC patients and that NRF2 expression was upregulated in EC (p � 0.045). AC-MCs combined with NRF2 had a poor prognosis, according to survival analysis (p � 0.056) and interactive analysis (p � 0.032). Conclusions. We presume that NRF2 combined with AC-MCs could be a marker to predict prognosis and could infuence immunotherapy through regulating PD-L1 in the EC. Recently, immunotherapy had been proven to have 1. Introduction prospective results for EC therapy; however, the immuno- With the rapid growth and aging of the world’s population, therapy’s efectiveness was afected by the complex tumor cancer will be the main reason for the rising burden of microenvironment (TME), so not all patients are benefted st disease in the 21 century. Esophageal cancer (EC) is the from these therapeutic interventions [4]. Te majority of sixth leading cause and has the eighth highest incidence rate research studies indicated that tumor-associated immune in the world. In China, 90% of EC is esophageal squamous cells, especially innate immune cells such as macrophages, dendritic cells, and mast cells (MCs), were related to im- cell carcinoma (ESCC), and esophageal adenocarcinoma (EAC) is more common in western countries [1]. Traditional munotherapy and tumoral responses [5–7]. MCs were bone technologies such as radiotherapy, chemotherapy, surgery, marrow-derived cells which could be recruited into tumor and trimodality are the main therapy methods for EC [2], tissue by SCF, chemokine factors, and so on. Hypoxia, the but the fve-year survival rate is still less than 15% [3]. Hence, accumulation of (lactic acid, adenosine, PGE , IFN-c, etc.) many researches were aimed to fnding meaningful thera- and low pH in TME could activate MCs discharge particles peutic and prognostic biomarkers for EC in order to im- to pro- and antitumoral by IgE/FcεRI pathway [8–10]. prove the prognosis and prolong the lives of patients. Activated mast cells (AC-MCs) have been recognized as an 2 Journal of Oncology important prognostic indicator and immune therapy target between diferent datasets. Because the clinical information for cancers [11]. in GEO is limited, we only use the clinical features from TCGA when analyzing the results, which refer to the clinical Te prognosis was afected by the complex immune cell infltration (ICI) in TME. Recently, some researchers created information. In addition, we collected 33 ESCC patients’ models according to immune cells and diferential expres- tissues and 30 ESCC pathological sections who had not sion genes (DEGs) to predict prognosis. Apart from this, received therapy from Huai an First Hospital, in 2021. Te somatic mutation genes also infuenced a patient’s prognosis detailed information about the patient is listed in Table 1. and immunotherapy response [12]. TP53 mutations afected Tis study was performed in accordance with the principles the immunophenotype in gastric cancer and infuenced the of the Helsinki Declaration and was performed, reviewed, patient’s prognosis [13]. In addition, some clinical trials also and approved by the Ethics Committee of Zhongda Hospital indicated that KEAP1/NRF2 mutations can be regarded as of Southeast University; the grant number is predictive markers for immunotherapy and prognosis 2021ZDKYSB004. makers for cancer [14]. Tumor mutation burden (TMB) is defned as the total number of somatic gene coding errors, 2.2. Estimation of Stromal and Immune Scores. Te base substitutions, gene insertions, or gene deletions de- “CIBERSORT” algorithm is a deconvolution algorithm and tected per million bases. Some research studies suggest that was used to quantify the infltration level of the distinct im- TMB is associated with the emergence of neoantigens which mune cells based on the input reference gene sets and repeated trigger antitumor immunity [15, 16]. Tumor patients with 1000 times to ensure stability. Te “ESTIMATE” algorithm was higher TMB had higher survival rates [17, 18]. A few somatic used to calculate the immune scores, stromal scores, and es- mutations in tumor DNA can be translated into neo- timate scores by the “estimate” R package. At the same time, we antigens, which could be present on the surface of cells in the analyzed the prognostic value of immune stores and stromal form of the major histocompatibility complex and recog- scores and their relationship with clinical features. nized by the immune system [19]. However, the combined efects of ICI and TMB in predicting prognosis remained unknown. 2.3.ICIClusters. We used the R packages “biomanager” and In this study, we established multiple immune score “consensus” to divide the samples into diferent clusters models and TMB to predict prognosis and immunotherapy. according to the immune cells’ relative fraction levels in EC. Our results indicated that the combined immune score with And the prognostic values in diferent ICI groups were TMB was related to prognosis and PD-L1, and we recog- indicated by the “survival” and “survminer” R packages. Te nized the key mutation gene NRF2. We also found that immune cells in the difernt clusters were reshaped by NRF2 was related to AC-MCs. Based on these results, we “ggpubr” package. Results were visualized through heat analyzed the combined efects of NRF2 and AC-MCs for maps by the “pheatmap” R package. prognosis by TCGA database and experiment verifcation. Our results showed that there is an interaction between 2.4. DEGs Associated with the ICI Phenotype and Gene NRF2 and MCs, especially the higher NRF2, which had Clusters. DEGs in diferent ICI clusters were determined by a worse survival rate. In total, we thought NRF2 combined setting the signifcance cutof to p< 0.05 (adjust) and with AC-MCs could be used to predict the prognosis for EC logFC>1, which was performed by the “limma” R package. and provide a new direction for the prognostic study of According to DEGs, the samples were divided into diferent esophageal cancer. types using the “biomanager” and “consensus cluster plus” R packages. Immune cells in diferent gene clusters were an- 2. Materials and Methods alyzed by “ggpubr.” We also analyzed the prognostic value of diferent gene clusters as indicated by the “survival” and 2.1.ECDatasetsandSamples. A total of 524 EC samples were “survminer” R packages. downloaded from the TCGA-GDC database (https://portal. gdc.cancer.gov/) and the GEO database (https://www.ncbi. nlm.nih.gov/geo/). Te RNA sequencing (RNA-seq; frag- 2.5. ICI Scores. First, unsupervised clustering was used to ments per kilobase million value) data and the clinical in- deal with the samples in TCGA and GEO according to DEG formation (BCR-XML) including futime, survival state, age, values, which were positively or negatively correlated with the cluster signature and described as ICI gene signatures A gender, grade, stage, and the TNM stage system were downloaded from TCGA-EC. Te microarray data and B, respectively. Second, the “Boruta” R package was used for dimension reduction of the ICI gene signatures A and B (GSE68698, GSE69925, and GSE161533) were downloaded from the GEO. To increase the readability of the data, the and to extract feature genes. Tird, principal component 1 FPKM values were transformed into TPMs (transcripts per was extracted as the signature score by using the principal kilobase million), which were identical to the results of component analysis (PCA). Finally, the formula that defned microarrays, and clinical information (BCR-XML) was the ICI score of each patient was transformed into a matrix. Te “limma” R package and the ICIscore � 􏽐 PC1A − 􏽐 PC1B, and we divided the ICI score “sva” R package were used to merge the RNA array. Te into a high ICI score group and a low ICI score group. “ComBat” algorithm was used to decrease the likelihood of According to the ICI score, the functional enrichment an- batch efects from diferent biological and technical biases alyses of GO and KEGG pathways were analyzed using the Journal of Oncology 3 Table 1: Te relationship between MCs with clinicopathological features of ESCC. Clinicopathological N MC (x ± s) P N FcεR1G (x ± s) P features Gender Male 8 25.87± 15.22 0.299 21 2.09± 6.46 0.69 Female 13 33.07± 33.4681 4 1.04± 1.99 Age ≤65 16 24.37± 17.52 0.033 9 0.41± 0.61 0.15 >65 5 49.40± 46.39 16 2.77± 7.37 Diferentiation High diferentiation 4 43.00± 11.91 5 0.24± 0.22 Middle diferentiation 10 32.10 ± 37.69 0.446 14 3.07± 7.87 0.568 Low diferentiation 7 20.57± 13.22 6 0.64± 0.46 T1-T2 7 42.28± 42.13 0.160 7 1.24± 1.31 0.345 T3-T4 14 24.35± 16.29 8 2.19± 7.01 N0 13 20.23± 14.60 0.119 10 0.99± 1.15 0.179 N1–N3 8 46.75± 36.95 15 2.54± 7.67 M0 — — — 22 0.79± 0.97 <0.001 M1 — — 3 10.24± 17.24 Stage I 4 26.25± 19.25 0.032 — — II 8 15.00± 10.46 12 1.05± 1.06 0.095 III 8 50.12± 34.36 13 2.72± 8.26 “clusterProfler” R package for the feature genes in the high extract RNA from tumor tissue and para-tumor tissue. Te ICI score group and the low ICI score group. In order to RAN was cDNA obtained by reverse transcription according know the prognostic signifcance of the ICI score, we also to the protocol (Vazyme, China). SYBR green was used to analyze the connection between clinical features and ICI complete the related expression. Te Q-PCR procedure fol- score based on the TCGA database. lowed the protocol (Vazyme, China). Te primer sequences ° ° ° ° were as follows: 95 C 3 min, 95 C 30 s, 60 C 15 s, 72 C, 30 s for 40 cycles, and solubility curve. Te primer sequence is listed in 2.6. Somatic Alteration Data Analysis. Te related somatic Table S1. In addition, tissues were addedto RIPA and lysed in mutation datasets for EC were downloaded from the an ultrasound machine. After being divided by SDS-PAGE, TCGA-GDC database. Tumor mutation burden (TMB) is the proteins were transferred onto PVDF membranes and defned as the total number of somatic coding errors in then blocked with 5% skim milk for 2 h, subsequently in- genes, base substitutions, and gene insertion and deletion cubated with primary antibodies of NRF2 (1 :1000), TPSB2 errors in EC. Te “ggpubr” R package was used to analyze (1 :1000), and GAPDH (1 : 5000) overnight at 4 C and next the TMB for high ICI scores and low ICI scores. Te mu- incubated with secondary antibodies for 1 h at room tem- tation genes with high ICI scores and low ICI scores were perature. Te target protein was visualized by the ECL Gel identifed through the “maftool” R package, and the top 30 Image System and analyzed by the software Image J. genes with the highest mutation frequency were listed. 2.9. Statistical Analysis. All statistical analyses were ac- 2.7. Toluidine Blue Staining. Toluidine blue staining was complished with R version 4.0.3, GraphPad Prism 8, and used to detect the number and distribution of MCs in ESCC. SPSS version 25.0. Te comparison between the two groups Parafn-embedded tissues were dewaxed in diferent con- was tested by the Wilcoxon test and T test; otherwise, it was centrations of alcohol, subsequently stained with toluidine tested by Kruskal–Wallis H and ANOVA. Te survival blue (Solarbio, China) for 15 min, and washed with PBS 3 curves for the subtypes were accomplished with the times. Photomicrographs of ten felds were taken at diferent Kaplan–Meier plotter. Te chi-square test was used to an- magnifcations using the camera (ZEISS, Germany), and the alyze the correction between the ICI score subtypes and mean value was used to describe the number and distri- somatic mutation frequency. Te chi-square test was used to bution of MCs in EC. Te AC-MCs rate was calculated by analyze the classifed variable. And the correlation analysis the ratio of the AC-MCs number to the total MCs number. was completed by Pearson’s analysis. Univariate and mul- tivariate Cox regression models were used to analyze the 2.8. Q-PCR and Western Blot Analyzed the Expression of MCs prognosis. Te interaction of NRF2 and AC-MC was ana- Related Genes and NRF2. We analyzed the relative genes in lyzed by interactive analysis. All analyses were two-tailed, ESCC tumor tissue and para-tumor tissue. Trizol was used to and p< 0.05 was regarded as the statistically signifcant level. 4 Journal of Oncology patients were divided into two groups (high ICI score and 3. Results low ICI score). We analyze the prognostic value of the ICI 3.1. Te Characteristic of ICI in the TME of EC. 22 human score. Te survival rate in the two ICI score groups has no immune cells were calculated through the CIBERSORT al- statistical diference (Figure 2(c)), but statistical analysis gorithm according to the TCGA and GEO databases and showed that survival status and the TN stage system were found to have diferential expression in tumor tissues and related to ICI score (Figures S2E–S2G). Meanwhile, we para-carcinoma tissues. Tese results suggested that the rel- analyzed the main pathways in high ICI scores, such as ative fractions of Tregs and resting MCs in the tumor tissue adherens junction, cell cycle, Hedgehog signaling pathway, were lower than those in para-carcinoma tissue, but the TGF-β signaling pathway, and Wnt signaling pathway, while naive CD4 T cells, activated CD4 memory T cells, M0 the main pathways in the low ICI score were the B cell macrophages, activated DCs, activated MCs, and neutrophils receptor signaling pathway, drug metabolism cytochrome in the tumor tissues were higher compared with para- P450, intestinal immune network for IgA production, pri- carcinoma tissues (Figure 1(a)). Te “corrplot” R package mary immunodefciency, and T cell receptor according to was used to generate a correlation coefcient heatmap to KEGG (Figure 2(i)). Functional enrichment analysis sug- visualize the landscape of 22 immune cells’ interactions in gested that the main functions of the high ICI score group in TME (Figure 1(b)). Additionally, the ESTIMATE algorithm the biological process were response to virus, type I in- was used to calculate the immune scores and stromal scores terferon signal pathway, and response to tumor necrosis according to the levels of immune cells in EC. According to factor, but in the low ICI score group were extracellular the clinical information from the TCGA database, we ex- matrix organization and endodermal cell diferentiation. Te plored the relationship between clinical features and estimate main functions enriched in the cellular component of the scores. Tese results suggested that the immune scores and high ICI score group were the extracellular matrix immu- stromal scores were not associated with survival time, but nological synapse, membrane raft, anchored component of clinical stage and Tstage were related to stromal scores, and T membrane, and apical plasma membrane, while in the low stage was related to immune scores (Figures 1(c)–1(f)). ICI score group, they were the endoplasmic reticulum lu- men, extracellular matrix, and fbrillar collagen trimer (Figures S2A–S2D). Tese results suggested that the ICI 3.2. Diferent Patterns Were Used to Predict the Prognosis. score may be related to the prognosis of EC. We analyzed the prognosis value of stromal scores and immune scores, but the results suggested that the score patterns were unrelated to prognosis (Figures S1(A) and 3.3. Combine ICI Score with TMB Predict Prognosis. Most S1(B)). So, we try to create new patterns according to im- evidence indicated that TMB could be used to evaluate the predictive prognosis [20, 21]. In the study, we analyzed the mune cells and DEGs to predict the prognosis. First, the ICI types were divided into three clusters (Figure S1(C)). relationship of TMB with the somatic mutation landscape in the However, the three ICI clusters have no signifcant survival EC and ICI scores, but the result showed that the TMB showed diference in EC (Figure 2(a)). Ten, we constructed another no signifcant diferences between the two groups (Figure 3(a)). subtype according to DEG (Figure S1(D)). Similarly, dif- Ten, we divided the samples into high/low TMB, and the result ferent gene clusters were unrelated to prognosis suggested that the survival rate of low TMB was higher than that (Figure 2(b)). However, the ICI clusters and gene clusters of high TMB (Figure 3(b)). At the same time, when we com- were all related to PD-L1 (Figures 2(d) and 2(e)). So, we bined the ICI score with the TMB, we found that the survival analyzed the immune cells and DEGs in the clusters, and we rate in the group with a low TMB and a low ICI score was the found that PD-L1 was more highly expressed in cluster longest, whereas the group with a high TMB and a high ICI score C. Te main immune cells in cluster C were CD8 T cells, was the shortest (Figure 3(c)). Furthermore, we assessed the distribution of somatic variants in EC driver genes between the CD4 T follicular helper cells, T cells gamma delta, NK cells, M1 macrophages, DCs, Tregs, and MCs (Figures 2(g) and high/low ICI score subgroups. Te top 30 genes with the highest alteration frequency were further analyzed (Figures S3A and 2(h)). At the same time, we also analyzed the relationship between the immune cells in cluster C and PD-L1, and the S3B). We also analyzed that the relative expression of mutation genes in high/low ICI score subgroups; DYNC2H1, OBSCN, results suggested that PD-L1 is positively related to CD8 T cells and DCs but negatively related to Tregs (Figure 2(f)). DNAH11, PIK3CA, MUCSB, NRF2, ARID1A, SACS, LRRK2, Te heatmap delineated the transcriptomic profle of all NOTCH1, and SMAD4 were all signifcantly diferent between DEGs in three gene clusters and gene types (Figures S1E and the high/low ICI score subgroups (Figure 3(d)). After analyzing S1F). To achieve quantitative indicators of the ICI landscape the role of these genes in TMB, we found that NRF2 was related in EC patients, PCA was used to calculate two aggregate to TMB (Figure 3(e)). We further analyzed the prognosis value scores according to the ICI score A from ICI signature gene of NRF2 and indicated that the expression of NRF2 was related to the survival status, TNM system, and grade Figures S(3C-3F). A and the ICI score B from ICI signature gene B (Table S2). In this research, the individual score of patients was com- However, univariate variables and multivariate Cox regression models were used to investigate the relationship between the puted through the ISA and ISB of each patient. All the Journal of Oncology 5 ** P<0.05 0.3 ** ** ** 0.2 2000 B cells naive B cells memory Plasma cells T cells CD8 ** T cells CD4 naive 1000 * T cells CD4 memory resting 0.1 T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes 0.0 Macrophages M0 -1000 Macrophages M1 Macrophages M2 Dendritic cells resting Dentritic cells activated Mast cells resting -2000 Mast cells activated Eosinophils T1 T2 T3 T4 Neutrophils StromalScore ImmuneScore -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Normal EC (a) (b) (c) P<0.001 p=0.289 p=0.032 1000 2000 0 1000 -1000 -1000 -2000 -1000 -2000 -3000 -2000 -3000 T1 T2 T3 T4 stageI stage II stageIII stageIV stageI stage II stageIII stageIV (d) (e) (f) Figure 1: Te landscape of ICI in the TME of EC. (a) Te immune cells in EC tissue and para-cancer tissue. (b) Te landscape of 22 immune cells’ interactions in TME. (c and d) Association of immune scores with T stage (c) and clinical stage (d). (e and f) Association of stromal ∗ ∗∗ ∗∗∗ scores with T stage (e) and clinical stage (f). p< 0.05; p< 0.01; p< 0.001. 1.00 1.00 1.00 0.75 0.75 0.75 0.50 0.50 0.50 0.25 p=0.801 0.25 p=0.936 0.25 p=0.180 0.00 0.00 Time (years) 0123456 0.00 Number at risk A 20 15 1 0 0 0 0 Time (years) Number at risk 012345 6 B 25 18 8 2 1 0 0 C 23 15 4 2 2 2 0 A 50 35 6 1 0 0 0 Time (years) 67 41 19 11 5 1 0 0123456 Number at risk 42 29 8 3 2 2 0 Time (years) High 57 43 5 0123456 11 1 1 0 22 10 2 Time (years) Low 102 62 6 0 012 3 4 56 ICI cluster Time (years) A Gene cluster ICI score High Low (a) (b) (c) *** *** *** *** 12.5 ns 12.5 10.0 10.0 PD-L1 * p < 0.05 T cells CD8 7.5 T cells follicular helper 7.5 ** p < 0.01 T cells regulatory (Tregs) Correlation T cells gamma delta 1.0 NK cells resting 5.0 0.5 NK cells activated 5.0 Macrophages M1 0.0 Dendritic cells resting -0.5 Dendritic cells activated Mast cells resting -1.0 2.5 AB C Mast cells activated AB C ICI cluster Gene cluster (d) (e) (f) Figure 2: Continued. expression fraction level PD-L1 expression Survival probability stromalscore B cells naive ICI cluster B cells memory Plasma cells T cells CD8 T cells CD4 naive T cells CD4 memory resting T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes Macrophages M0 Macrophages M1 Macrophages M2 Dendritic cells resting Dentritic cells activated Mast cells resting Mast cells activated Eosinophils Neutrophils PD-L1 expression Survival probability Gene cluster ImmuneScore StromalScore immunescore Neutrophils Eosinophils Mast cells activated Mast cells resting Dentritic cells activated Dendritic cells resting Macrophages M2 Macrophages M1 Macrophages M0 Monocytes NK cells activated NK cells resting T cells gamma delta T cells regulatory (Tregs) T cells follicular helper T cells CD4 memory activated T cells CD4 memory resting T cells CD4 naive T cells CD8 Plasma cells B cells memory B cells naive Survival probability ICI score stromalscore immunescore PD-L1 T cells CD8 T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Macrophages M1 Dendritic cells resting Dendritic cells activated Mast cells resting Mast cells activated 6 Journal of Oncology 12 12 *** ns ** ns ns ns ns ** * ns *** ** ** *** *** ** *** ns*n ns *** ** * *** *** *** ***sns *** *** *** *** *** *** *** ** *** *** ** *** *** *** *** *** ** *** ICI cluster Gene cluster (g) (h) 0.5 0.0 -0.5 High ICI score Low ICI score KEGG_ADHERENS_JUNCTION KEGG_B_CELL_RECEPTOR_SIGNALING_PATHWAY KEGG_CELL_CYCYLE KEGG_DRUG_METABOLISM_CYTOCHROME_P450 KEGG_HEDGEHOG_SIGNALING_PATHWAY KEGG_INTESTINAL_IMMUNE_NETWORK_FOR_IGA_PRODUCTION KEGG_PRIMARY_IMMUNODEFICIENCY KEGG_T_CELL_RECEPTOR_SIGNALING_PATHWAY KEGG_TGF_BETA_SIGNALING_PATHWAY KEGG_WNT_SIGNALING_PATHWAY (i) Figure 2: Diferent prognostic models constructed based on immune cells and DEGs. (a) Kaplan–Meier curves for overall survival of EC with diferent ICI clusters. (b) Kaplan–Meier curves for overall survival of EC with diferent gene clusters. (c) Kaplan–Meier curves for overall survival of EC with diferent ICI scores. (d) Te expression of PD-LA in ICI clusters. (e) Te expression of PD-L1 in ICI clusters. (f) Te relationship between PD-L1 and immune cells. (g) Te fraction of tumor immune cells in three ICI clusters. (h) Te fraction of tumor ∗ ∗∗ immune cells in three gene clusters. (i) Enrichment plots showing signaling pathways in high/low ICI scores. p< 0.05; p< 0.01; ∗∗∗ p< 0.001. NRF2 mutation and the overall survival of EC patients, and the intracytoplasmic, and AC-MCswerecharacterized by many result revealed that the NRF2 mutation was not an independent blue dye particles surrounding the cells. Our results showed prognostic factor for OS in EC (Table S3). that MCs were mainly in the muscular layer (p< 0.05) squamous epithelium (0.67± 3.46, Figure 4(a)-A/B/C), tumor nest (4.48± 9.63, Figure 4(a)-D/E/F), and muscu- laris propria (36.33± 37.84, Figure 4(a)-G/H/I). We also 3.4. Combining AC-MCs with NRF2 Could Predict Prognosis. calculated the rate of AC-MCs in EC tissue. Te MCs in Considering the relationship between the TMB and ICI patients in stage III were higher than those in patients in I scores, what follows is the relationship between NRF2 and and II (Table 1). We also analyzed the related gene, which immune cells. We found that NRF2 was only related to could activate MCs. Tese results showed that FcεR1A MCs (Table S4). Next, we attempt to assess the combined (Figure 4(b), p< 0.005), NRF2 (Figure 4(b), p< 0.05), efect of NRF2 and MCs for predicting prognosis in EC. We FcεR1G (Figure S4D, p< 0.000), and PD-L1(Figure S4E, collected 30 EC patients’ tissue slices and their clinical p< 0.05) were all upregulated in tumor tissue (Figures S4D information to analyze the number and distribution of and S4E), and the protein level of NRF2 and TPSB2 was MCs/AC-MCs and their prognosis value. MCs were also higher expressed in tumor tissue (Figure 4(d)–E). Te characterized by blue densely basophilic granules in the Scale of fraction B cells naive B cells memory Plasma cells T cells CD8 T cells CD4 naive T cells CD4 memory resting T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes Macrophages M0 Enrichment Score Macrophages M1 Macrophages M2 Dendritic cells resting Dendritic cells activated Mast cells resting Mast cells activited Eosinophils Neutrophils StromalScore ImmuneScore Scale of fraction B cells naive B cells memory Plasma cells T cells CD8 T cells CD4 naive T cells CD4 memory resting T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes Macrophages M0 Macrophages M1 Macrophages M2 Dendritic cells resting Dendritic cells activated Mast cells resting Mast cells activited Eosinophils Neutrophils StromalScore ImmuneScore Journal of Oncology 7 0.065 1.00 0.75 0.50 p=0.004 0.25 0.00 01 23456 Low High Time (years) Number at risk ICI score H–TMB 31 15 4 1 1 0 0 L–TMB 127 89 29 14 6 3 0 Low 0 123 4 5 6 Time (years) High H–TMB L–TMB (a) (b) 12.5 *** *** ns ** *** *** *** ns ns ** * ns *** ns *** ** 1.00 0.75 10.0 0.50 7.5 0.25 p=0.015 5.0 0.00 0 12345 6 2.5 Time (years) Number at risk H–TMB+H–ICI score 5 4 1 0 0 0 0 H–TMB+L–ICI score 26 11 3 1 1 0 0 L–TMB+H–ICI score 40 32 6 2 1 1 0 L–TMB+L–ICI score 87 57 23 12 5 2 0 0 123456 ICI score Time (years) Low H–TMB+H–ICI score L–TMB+H–ICI score High H–TMB+L–ICI score L–TMB+L–ICI score (c) (d) p=0.035 (e) Figure 3: Interaction between the ICI score and the TMB. (a) TMB diference in the high ICI score and low ICI score. (b) Kaplan–Meier curves for high and low TMB groups of the TCGA-EC cohort. (c) Kaplan–Meier curves for patients in the TCGA-EC cohort stratifed by both TMB and ICI scores. (d) Te relative expression level in the high and low ICI score groups. (e) Te value of TMB for NRF2 mutations ∗ ∗∗ ∗∗∗ and non-NRF2 mutations. p< 0.05; p< 0.01; p< 0.001. expression of NRF2 was related to FcεR1A (Figure S4E, unrelated to OS (Figures S4A and S4B), but the group with r = 0.515), and PD-L1 was related to FcεR1G (Figure S4G, low NRF2 and high FcεR1G was the lowest malignant r = 0.468). We divided the expression of NRF2 and FcεR1G (Figure S4C). Most importantly, there is an interaction into two groups by median, and interaction analysis was between NRF2 and MC (Figure 4(f)). Hence, we thought used to explore the interaction of NRF2 and AC-MCs with that a combination of NRF2 and AC-MCs could be TNM. Cox results suggested that NRF2 and MCs were all a prognosis maker for EC. Survival probability Tumor Burden Mutation TMB (mutation) NRF2 mutation Gene expression DYNC2H1 OBSCN Survival probability APOB NRF2 non-mutation DNAH11 MUC4 PIK3CA MUC5B MUC16 IVL NFE2L2 ARID1A WDFY4 SACS LRRK2 NOTCH1 SMAD4 8 Journal of Oncology (a) p=0.046 p=0.005 -5 -5 -10 -15 -10 ESCC Normal ESCC Normal (b) (c) Figure 4: Continued. relative expression level of FCER1G relative expression level of NRF2 Journal of Oncology 9 1.5 ** NRF2 ** TPSB2 1.0 GAPDH NT N T N T 0.5 0.0 NT N T NRF2 TPSB2 (d) (e) Estimated Marginal Means of TNM 2.0 1.8 1.6 1.4 1.2 1.0 NRF2 FccR1G (f) Figure 4: Te infltration of MCs in EC and the related gene expression as well as its relationships. (a) Te MCs’ infltration of EC tissues. (A/B/C) Squamous epithelium. (D/E/F) Tumor nest. (G/H/I) Muscularis propria. Te black arrow represents undegranulated MCs, and the red arrow represents granulated MCs. (b) Expression of FCER1G. (c) Expression of NRF2. (d) Statistical analysis of TPSB2 and NRF2 ∗ ∗∗ protein expression. (e) Protein expression levels of NRF2 and TPSB2. (f) Interaction of NRF2 and AC-MCs. p< 0.05; p< 0.01. immunosuppression and promote tumor survival and 4. Discussion progression [22–24]. In this study, we analyzed the ICI Te majority of studies have demonstrated that the het- landscape of EC according to the TCGA and GEO databases. erogeneous TME and TMB participated in tumor pro- Our results indicated that CD4 T cells, M0 macrophages, gression, prognosis, and therapeutic for EC. However, AC-MCs, and activated DCs were increased, but the Tregs clarifying the modulation of TME and TMB as well as their and resting MCs were decreased in tumor tissue. Tregs combination efects during EC remains a challenge. Our suppress the activation and proliferation of multiple types of + + study comprehensively described the ICI landscape and immunocompetent cells such as CD4 T cells, CD8 T cells, somatic mutation gene landscape and constructed diferent B cells, NK, and T cells, as well as suppressive immuno- patterns to quantify the ICI and TMB by the “CIBERSORT” reaction [25]. CD4 T cells could increase the secretion of and “ESTIMATE” algorithms to predict prognosis and the IL4, IL2 promoting breast cancer progression, and the relationship with PD-L1. We found that the combined mature dendritic cells induced the proliferation of CD4 T immune fltration cells and tumor mutation burden could cells [26, 27]. AC-MCs could produce VEGF, PDGF, MMP9, predict the prognosis for EC. At the same time, we recog- and PGE2 to promote angiogenesis and tumor migration nized the key mutation genes NRF2 and immune cells (mast [28]. Moreover, AC-MCs’ secreted cytokines could also cells), which played an important role in predicting prog- infuence the development and function of T cells and B nosis. We verifed the combined role of NRF2 and mast cells cells [29]. Apart from evaluating the infltration of single in EC patient and found that combined NRF2 and MCs immune cells, we also attempt to quantify the ICI landscape would be a prognostic target and provide new insight into to evaluate the prognosis through built-score patterns. In the prognosis of EC. previous studies, the ESTIMATE algorithm has been used to Multiple pieces of evidence have demonstrated that analyze the immune scores and stromal scores, and it has dysfunctional immune cells in the TME lead to been suggested that the risk model is benefcial for the early relative expression of related genes Estimated Marginal Means 10 Journal of Oncology showed that the high TMB and low ICI group had the worst identifcation of high-risk patients to formulate an in- dividualized treatment project and improve the possibility of OS. Meanwhile, these results indicated that TP53, TTN, MUC16, LRP1B, and SYNE1 were high-frequency muta- an immunotherapy response [30, 31]. In our study, based on the stromal scores and immune scores, we divided the tions in EC. Especially, NRF2 was not only a high-frequency patients into high-score and low-score groups. We found mutation gene in EC but also signifcantly diferent in ICI that the survival probability in the two groups did not score groups. Tere was a study that reported that NRF2/ signifcantly change, but the stromal scores were higher in KEAP1 mutations correlate with higher TMB value/PD-L1 stage III, and the higher the immune scores and stromal expression and potentiate improved clinical outcomes with scores, the higher the T stage. At the same time, we divided immunotherapy [45]. An NRF2 mutation could disrupt the the samples into three parts based on the infltrated immune weak binding of Keap1 with the NRF2-DLG motif and cells. Our results demonstrated that the immune cells which activate NRF2 to promote tumor progression [46]. Con- sidering the complexity of mutations, we only detected the have immunosuppressive function were focused on ICI cluster C. PD-L1, a key immune checkpoint, was higher in expression of NRF2 and the prognosis value in EC. In our result, the NRF2 was upregulated in EC, but not an in- ICI cluster C. Previous evidence had shown that immune cell-related genes could predict disease progression and dependent prognostic biomarker, which was diferent from immunotherapeutic responses [32, 33]. Based on the previous research studies [47]; the reason probably was that immune-related gene in EC, we divided the patient into the number of patients was not enough. Meanwhile, we three ICI gene clusters. Te results suggested that ICI gene analyzed the relationship between NRF2 and immune cells cluster C had a more favorable immune-activated type with and found that NRF2 was related to MCs. Other studies the highest density of CD8 T cells, M1 macrophages, acti- indicated that NRF2 could activate MCs, IgG/FcεRI pro- vated DCs, and CD4 T follicular helper cells [34–36]. moted the phosphorylation of Lyn and activated Syk/PI3K, Additionally, the expression of PD-L1 was highest in ICI LAT/p38, and LAT/Raf-1/ERK1/2 pathways, and the AKT- Nrf2 and p38MAPK-Nrf2 signal pathways play an important gene cluster C. Hence, the patients in ICI gene cluster C might have a better immune response. Te outcome of our role in hypersensitivity induced by MCs [48, 49]. Hence, we analyzed the MCs in EC tissue and found that MCs were analysis was in accordance with the previous study, which indicated that ICI clusters and ICI gene clusters in EC might irrelevant to OS. Surprisingly, the combination of NRF2 with MCs could afect prognosis. In addition, previous infuence the expression of PD-L1 [37]. In recent years, gene clusters related to immune response studies indicated that MCs could express PD-L1 and play and proliferation were used to predict the outcome of a crucial role in immunosuppression [50]. Our results also cancers and identify high-risk patients; the distant indicated that FcεR1G could activate MCs, and the AC-MCs metastasis-free survival in high-score immune gene was were positively related to PD-L1, but the mechanism by higher than low-score in breast cancer [38]. Te prognosis which activated MCs regulated PD-L1-induced immuno- value of the ICI score was calculated by the “Boruta algo- suppression deserves deep research. Terefore, we thought combining NRF2 with MCs would be used to predict rithm” based on the immune cell-related gene, which has been proven in head and neck squamous cell carcinoma [39]. prognosis. However, whether the NRF2-activated MCs are involved in immune suppression in EC needs further study. In the current study, we assessed the prognosis value of the ICI score in EC and found that there was no signifcant In summary, we comprehensively analyzed the ICI diference in OS in high/low ICI scores , but the ICI score landscape and TMB of EC and found that high ICI and high was higher in alive, no lymph node metastasis samples. TMB had worse prognoses. We also recognized key mu- Trough KEGG, our results indicated that the high ICI score tation genes and immune cells and analyzed the common mainly regulates the hedgehog signaling pathway, TGF-β prognostic value of NRF2 with MCs by experiment verif- signaling pathway, Wnt signaling pathway, and so on. cation and database analysis. Nevertheless, several limita- Hedgehog signaling could be induced by activated T cells tions in our study should be considered. First, due to the and NK cells and participate in immunotherapy [40]. Re- limited patient information from TCGA, a larger sample size and sufcient information were required for further proof of pression of the Wnt signaling pathway would decrease the expression of PD-L1 and increase the immune-killing efect our results. Second, the role of NRF2 and MCs participated in immune regulation and tumor progression in EC needs of NK cells [41]. Te TGF-β/EMT signaling pathway infuenced the expression of PD-L1 and promoted immune further experimental study. Tird, it is not enough to clarify escape. All these results demonstrated that the ICI score was diferent patterns and MCs that could infuence immuno- related to the PD-L1 but was not an independent prognosis therapeutic efectiveness only by analyzing the relationship marker for EC [37]. with PD-L1. In all, we found that high ICI and high TMB Te majority of studies demonstrated that TMB was could afect the prognosis, and the combination of NRF2 related to prognosis and could be a marker for predicting the with AC-MCs had a worse prognosis and could be an ef- fective prognostic factor for EC. efectiveness of immune checkpoint inhibitors in cancer [42, 43]. Mutation genes related to TMB were crucial prognostic biomarkers for cancers [32, 44]. In our study, we Abbreviations analyzed the somatic mutation landscape according to TCGA. Our results indicated that the high TMB level had EC: Esophageal cancer a poor OS, and the combination of the TMB with ICI scores EAC: Esophageal adenocarcinoma Journal of Oncology 11 Table S4: correlation of NRF2 with immune cells in patients ESCC: Esophageal squamous cell carcinoma from the TCGA database and GEO database using the TME: Tumor microenvironment Pearson correlation coefcient. Figure S1: diferent patterns NK: Natural killer were used to predict the prognosis. Figure S2: the ICI score DCs: Dendritic cells predicts the prognosis and the GO analysis. Figure S3: the Tregs: Regulatory T cells mutation genes in diferent ICI score groups and the MCs: Mast cells prognosis of NRF2. Figure S4: the prognostic value of NRF2 AC-MCs: Activated mast cells and activated mast cells. (Supplementary Materials) TCGA: Te Cancer Genome Atlas database GEO: Gene Expression Omnibus ICI: Immune cell infltration References TMB: Tumor mutation burden [1] F. Z. Wang, L. Zhang, Y. Xu, Y. Xie, S Li, and S. Li, NFE2L2/ Nuclear factor erythroid 2-related factor 2 “Comprehensive analysis and identifcation of key driver NRF2: genes for distinguishing between esophageal adenocarcinoma RYR2: Ryanodine receptor 2 and squamous cell carcinoma,” Frontiers in Cell and De- KEAP1: Kelch-like ECH-associated protein 1 velopmental Biology, vol. 9, Article ID 676156, 2021. TPMs: Transcripts per kilobase million [2] A. A. Watkins, J. A. Zerillo, and M. S. Kent, “Trimodality DEGs: Diferentially expressed genes approach for esophageal malignancies,” Surgical Clinics of PCA: Principal component analysis North America, vol. 101, no. 3, pp. 453–465, 2021. KEGG: Kyoto Encyclopedia of Genes and Genomes [3] J. C. Lu, H. Tao, Y. Q. Zhang et al., “Extent of prophylactic GO: Gene Ontology postoperative radiotherapy after radical surgery of thoracic esophageal squamous cell carcinoma,” Diseases of the ISA: ICI signature gene A Esophagus, vol. 21, no. 6, pp. 502–507, 2008. ISB: ICI signature gene B. [4] H. D. Teixeira Farinha, A. Digklia, D. Schizas, N. Demartines, M. Schafer, ¨ and S. Mantziari, “Immunotherapy for esophageal Data Availability cancer: state-of-the art in 2021,” Cancers, vol. 14, no. 3, p. 554, Te datasets supporting our results are available in the [5] W. Z. Zhang, Z. Zhao, and F. Li, “Natural killer cell dys- TCGA and GEO database as well as data sources in the function in cancer and new strategies to utilize NK cell po- method. Te data of our cohort are provided in tables, and tential for cancer immunotherapy,” Molecular Immunology, further inquiries can be directed to the corresponding vol. 144, pp. 58–70, 2022. authors. [6] I. Plesca, L. Muller, J. P. Bottcher, H. Medyouf, R. Wehner, and M. Schmitz, “Tumor-associated human dendritic cell subsets: phenotype, functional orientation, and clinical rele- Conflicts of Interest vance,” European Journal of Immunology, vol. 52, no. 11, pp. 1750–1758, 2022. Te authors declare that they have no conficts of interest. [7] D. S. Villalobos, I. G. Martinez-Aguilar, M. Ibarra-Sanchez et al., “Mast cell-tumor interactions: molecular mechanisms of Authors’ Contributions recruitment, intratumoral communication and potential therapeutic targets for tumor growth,” Cells, vol. 11, no. 3, Ran Liu, Junjie Lv, and Xinxin Guo designed the study. Weitao Shen and Mingjun Sun analyzed the data. Xinxin [8] G. Varricchi, M. R. Galdiero, S. Lofredo et al., “Are mast cells Guo wrote the manuscript. and Ran Liu provided funding MASTers in cancer?” Frontiers in Immunology, vol. 8, p. 424, support. All the authors have read and reviewed the manuscript. [9] K. Mukai, M. Tsai, H. Saito, and S. J. Galli, “Mast cells as sources of cytokines, chemokines, and growth factors,” Im- Acknowledgments munological Reviews, vol. 282, no. 1, pp. 121–150, 2018. [10] F. A. Redegeld, Y. Yu, S. Kumari, N. Charles, and U. Blank, Te authors thank Dr. Jiajia Xu from Department of Pa- “Non-IgE mediated mast cell activation,” Immunological Reviews, vol. 282, no. 1, pp. 87–113, 2018. thology, Zhongda Hospital Southeast University, for helping [11] H. W. N. Choi, M. Naskar, H. K. Seo, and H. W. Lee, “Tumor- us analyze the pathological section. Tis work was supported associated mast cells in urothelial bladder cancer: optimizing by the National Natural Science Foundation of China (grant immuno-oncology,” Biomedicines, vol. 9, no. 11, p. 1500, 2021. nos. 82173479 and 81872579) and the Postgraduate Research [12] A. Desrichard, A. Snyder, and T. A. Chan, “Cancer neo- and Practice Innovation Program of Jiangsu Province (grant antigens and applications for immunotherapy,” Clinical no. KYCX22-0303). Cancer Research, vol. 22, no. 4, pp. 807–812, 2016. [13] K. Z. Nie, Z. Zheng, Y. Wen et al., “Construction and vali- Supplementary Materials dation of a TP53-associated immune prognostic model for gastric cancer,” Genomics, vol. 112, no. 6, pp. 4788–4795, Table S1: the primer sequences of related genes. Table S2: the DEGs in ICI gene signatures A and B. Table S3: association [14] W. C. M. Dempke and M. Reck, “KEAP1/NRF2 (NFE2L2) with overall survival and clinicopathological characteristics mutations in NSCLC - fuel for a superresistant phenotype?” in patients from the TCGA database using Cox regression. Lung Cancer, vol. 159, pp. 10–17, 2021. 12 Journal of Oncology [15] M. M. Gubin, M. N. Artyomov, E. R. Mardis, and [31] J. Gao, T. Tang, B. Zhang, and G. Li, “A prognostic signature R. D. Schreiber, “Tumor neoantigens: building a framework based on immunogenomic profling ofers guidance for for personalized cancer immunotherapy,” Journal of Clinical esophageal squamous cell cancer treatment,” Frontiers in Investigation, vol. 125, no. 9, pp. 3413–3421, 2015. Oncology, vol. 11, Article ID 603634, 2021. [16] M. Yarchoan, A. Hopkins, and E. M. Jafee, “Tumor muta- [32] F. L. Zhang, Y. Liu, Y. Yang, and K. Yang, “Development and tional burden and response rate to PD-1 inhibition,” New validation of a fourteen- innate immunity-related gene pairs England Journal of Medicine, vol. 377, no. 25, pp. 2500-2501, signature for predicting prognosis head and neck squamous cell carcinoma,” BMC Cancer, vol. 20, no. 1, p. 1015, 2020. [17] R. M. Samstein, C. H. Lee, A. N. Shoushtari et al., “Tumor [33] Y. Fu, S. Sun, J. Bi, C. Kong, and L. Yin, “A novel immune- mutational load predicts survival after immunotherapy across related gene pair prognostic signature for predicting overall multiple cancer types,” Nature Genetics, vol. 51, no. 2, survival in bladder cancer,” BMC Cancer, vol. 21, no. 1, p. 810, pp. 202–206, 2019. 2021. [18] L. Ke, S. Li, and H. Cui, “Te prognostic role of tumor [34] J. Hamanishi, M. Mandai, M. Iwasaki et al., “Programmed cell mutation burden on survival of breast cancer: a systematic death 1 ligand 1 and tumor-infltrating CD8+ T lymphocytes review and meta-analysis,” BMC Cancer, vol. 22, no. 1, p. 1185, are prognostic factors of human ovarian cancer,” Proceedings 2022. of the National Academy of Sciences, vol. 104, no. 9, [19] T. A. Chan, M. Yarchoan, E. Jafee et al., “Development of pp. 3360–3365, 2007. [35] H. Law, V. Venturi, A. Kelleher, and C. M. L. Munier, “Tfh tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic,” Annals of Oncology, vol. 30, cells in health and immunity: potential targets for systems no. 1, pp. 44–56, 2019. biology approaches to vaccination,” International Journal of [20] J. S. Wang, J. Song, Z. Liu, T. Zhang, and Y. Liu, “High tumor Molecular Sciences, vol. 21, no. 22, p. 8524, 2020. mutation burden indicates better prognosis in colorectal [36] M. Yang, Y. Cao, Z. Wang, T. Zhang, Y. Hua, and Z. Cai, cancer patients with KRAS mutations,” Frontiers Oncology, “Identifcation of two immune subtypes in osteosarcoma vol. 12, Article ID 1015308, 2022. based on immune gene sets,” International Immuno- [21] Y. Zheng, M. Yao, and Y. Yang, “Higher tumor mutation pharmacology, vol. 96, Article ID 107799, 2021. burden was a predictor for better outcome for nsclc patients [37] Y. H. Chen, X. Huang, L. Chen et al., “Characterization of the treated with PD-1 antibodies: a systematic review and meta- immune infltration landscape and identifcation of prog- analysis,” SLAS Technology, vol. 26, no. 6, pp. 605–614, 2021. nostic biomarkers for esophageal cancer,” Molecular Bio- [22] H. Sun, T. G. Martin, J. Marra et al., “Individualized genetic technology, pp. 1–23, 2022. makeup that controls natural killer cell function infuences the [38] M. Callari, V. Cappelletti, F. D’Aiuto et al., “Subtype-specifc efcacy of isatuximab immunotherapy in patients with metagene-based prediction of outcome after neoadjuvant and multiple myeloma,” Journal for ImmunoTerapy of Cancer, adjuvant treatment in breast cancer,” Clinical Cancer Re- vol. 9, no. 7, Article ID e002958, 2021. search, vol. 22, no. 2, pp. 337–345, 2016. [23] L. Shao, Q. Yu, R. Xia et al., “B7-H3 on breast cancer cell [39] X. Zhang, M. Shi, T. Chen, and B. Zhang, “Characterization of MCF7 inhibits IFN-gamma release from tumour-infltrating the immune cell infltration landscape in head and neck T cells,” Pathology, Research & Practice, vol. 224, Article ID squamous cell carcinoma to aid immunotherapy,” Molecular Terapy - Nucleic Acids, vol. 22, pp. 298–309, 2020. 153461, 2021. [24] M. Li, D. Xie, X. Tang et al., “Phototherapy facilitates tumor [40] H. M. Onishi, T. Morisaki, A. Kiyota et al., “Te Hedgehog recruitment and activation of natural killer T cells for potent inhibitor cyclopamine impairs the benefts of immunotherapy with activated T and NK lymphocytes derived from patients cancer immunotherapy,” Nano Letters, vol. 21, no. 14, pp. 6304–6313, 2021. with advanced cancer,” Cancer Immunology Immunotherapy, [25] M. Miyara and S. Sakaguchi, “Natural regulatory T cells: vol. 62, no. 6, pp. 1029–1039, 2013. mechanisms of suppression,” Trends in Molecular Medicine, [41] R. Deng, C. Zuo, Y. Li et al., “Te innate immune efector ISG12a promotes cancer immunity by suppressing the ca- vol. 13, no. 3, pp. 108–116, 2007. [26] C. Aspord, A. Pedroza-Gonzalez, M. Gallegos et al., “Breast nonical Wnt/β-catenin signaling pathway,” Cellular and cancer instructs dendritic cells to prime interleukin 13- Molecular Immunology, vol. 17, no. 11, pp. 1163–1179, 2020. [42] M. C. Guo, Z. Chen, Y. Li et al., “Tumor mutation burden secreting CD4+ T cells that facilitate tumor development,” Journal of Experimental Medicine, vol. 204, no. 5, pp. 1037– predicts relapse in papillary thyroid carcinoma with changes 1047, 2007. in genes and immune microenvironment,” Frontiers in En- [27] H. Qi, J. G. Egen, A. Y. C. Huang, and R. N. Germain, docrinology, vol. 12, Article ID 674616, 2021. [43] X. Liu, R. Qiu, M. Xu et al., “KMT2C is a potential biomarker “Extrafollicular activation of lymph node B cells by antigen- bearing dendritic cells,” Science, vol. 312, no. 5780, of prognosis and chemotherapy sensitivity in breast cancer,” pp. 1672–1676, 2006. Breast Cancer Research and Treatment, vol. 189, no. 2, [28] D. A. Ribatti, T. Annese, and R. Tamma, “Adipocytes, mast pp. 347–361, 2021. cells and angiogenesis,” Romanian Journal of Morphology and [44] Z. Guo, X. Yan, C. Song et al., “FAT3 mutation is associated Embryology, vol. 61, no. 4, pp. 1051–1056, 2021. with tumor mutation burden and poor prognosis in esoph- [29] Y. A. Mekori and D. D. Metcalfe, “Mast cell–T cell in- ageal cancer,” Frontiers in Oncology, vol. 11, Article ID teractions,” Te Journal of Allergy and Clinical Immunology, 603660, 2021. vol. 104, no. 3, pp. 517–523, 1999. [45] X. Y. Xu, Y. Yang, X. Liu et al., “NFE2L2/KEAP1 mutations [30] M. Liu, F. Li, B. Liu et al., “Profles of immune cell infltration correlate with higher tumor mutational burden value/PD-L1 and immune-related genes in the tumor microenvironment of expression and potentiate improved clinical outcome with esophageal squamous cell carcinoma,” BMC Medical Geno- immunotherapy,” Te Oncologist, vol. 25, no. 6, pp. 955–963, mics, vol. 14, no. 1, p. 75, 2021. 2020. Journal of Oncology 13 [46] L. W. Cheng, H. Wang, S. Li, Z. Liu, and C. Wang, “New insights into the mechanism of Keap1-Nrf2 interaction based on cancer-associated mutations,” Life Sciences, vol. 282, Article ID 119791, 2021. [47] Q. Ju, X. Li, H. Zhang, S. Yan, Y. Li, and Y. Zhao, “NFE2L2 is a potential prognostic biomarker and is correlated with im- mune infltration in brain lower grade glioma: a pan-cancer analysis,” Oxidative Medicine and Cellular Longevity, vol. 2020, Article ID 3580719, 26 pages, 2020. [48] X. Wang, Z. Lv, B. Han et al., “Te aggravation of allergic airway infammation with dibutyl phthalate involved in Nrf2- mediated activation of the mast cells,” Science of the Total Environment, vol. 789, Article ID 148029, 2021. [49] C. L. Xu, L. Li, C. Wang et al., “Efects of G-Rh2 on mast cell- mediated anaphylaxis via AKT-Nrf2/NF-κB and MAPK- Nrf2/NF-κB pathways,” Journal of Ginseng Research, vol. 46, no. 4, pp. 550–560, 2022. [50] Y. Lv, Y. Zhao, X. Wang et al., “Increased intratumoral mast cells foster immune suppression and gastric cancer pro- gression through TNF-alpha-PD-L1 pathway,” Journal for ImmunoTerapy of Cancer, vol. 7, no. 1, p. 54, 2019. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Oncology Hindawi Publishing Corporation

Activated Mast Cells Combined with NRF2 Predict Prognosis for Esophageal Cancer

Loading next page...
 
/lp/hindawi-publishing-corporation/activated-mast-cells-combined-with-nrf2-predict-prognosis-for-fsFLz8KrUB

References (50)

Publisher
Hindawi Publishing Corporation
ISSN
1687-8450
eISSN
1687-8469
DOI
10.1155/2023/4211885
Publisher site
See Article on Publisher Site

Abstract

Hindawi Journal of Oncology Volume 2023, Article ID 4211885, 13 pages https://doi.org/10.1155/2023/4211885 Research Article Activated Mast Cells Combined with NRF2 Predict Prognosis for Esophageal Cancer 1 1 1 2 1 Xinxin Guo, Weitao Shen, Mingjun Sun, Junjie Lv , and Ran Liu Key Laboratory of Environmental Medicine Engineering, Ministry of Education, School of Public Health, Southeast University, Nanjing 210009, China Cancer Institute of Fudan University, Fudan University, Shanghai 200032, China Correspondence should be addressed to Junjie Lv; lylvjunjie@163.com and Ran Liu; ranliu@seu.edu.cn Received 10 September 2022; Revised 12 December 2022; Accepted 20 December 2022; Published 4 January 2023 Academic Editor: Shuanglin Qin Copyright © 2023 Xinxin Guo et al. Tis is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background. Esophageal cancer (EC) had the sixth-highest mortality rate of all cancers due to its poor prognosis. Immune cells and mutation genes infuenced the prognosis of EC, but their combined efect on predicting EC prognosis was unknown. In this study, we comprehensively analyzed the immune cell infltration (ICI) and mutation genes and their combined efects for predicting prognosis in EC. Methods. Te CIBERSORT and ESTIMATE algorithms were used to analyse the ICI scape based on the TCGA and GEO databases. EC tissues and pathologic sections from Huai’an, China, were used to verify the key immune cells and mutation genes and their interactions. Results. Stromal/immune score patterns and ICI/gene had no statistical signifcance in overall survival (OS) (p> 0.05). Te combination of ICI and tumor mutation burden (TMB) showed that the high TMB and high ICI score group had the shortest OS (p � 0.004). We recognized that the key mutation gene NRF2 was signifcantly diferent in the high/low ICI score subgroups (p � 0.002) and positivity with mast cells (MCs) (p< 0.05). Trough experimental validation, we found that the MCs and activated mast cells (AC-MCs) were more infltration in stage II/III (p � 0.032; p � 0.013) of EC patients and that NRF2 expression was upregulated in EC (p � 0.045). AC-MCs combined with NRF2 had a poor prognosis, according to survival analysis (p � 0.056) and interactive analysis (p � 0.032). Conclusions. We presume that NRF2 combined with AC-MCs could be a marker to predict prognosis and could infuence immunotherapy through regulating PD-L1 in the EC. Recently, immunotherapy had been proven to have 1. Introduction prospective results for EC therapy; however, the immuno- With the rapid growth and aging of the world’s population, therapy’s efectiveness was afected by the complex tumor cancer will be the main reason for the rising burden of microenvironment (TME), so not all patients are benefted st disease in the 21 century. Esophageal cancer (EC) is the from these therapeutic interventions [4]. Te majority of sixth leading cause and has the eighth highest incidence rate research studies indicated that tumor-associated immune in the world. In China, 90% of EC is esophageal squamous cells, especially innate immune cells such as macrophages, dendritic cells, and mast cells (MCs), were related to im- cell carcinoma (ESCC), and esophageal adenocarcinoma (EAC) is more common in western countries [1]. Traditional munotherapy and tumoral responses [5–7]. MCs were bone technologies such as radiotherapy, chemotherapy, surgery, marrow-derived cells which could be recruited into tumor and trimodality are the main therapy methods for EC [2], tissue by SCF, chemokine factors, and so on. Hypoxia, the but the fve-year survival rate is still less than 15% [3]. Hence, accumulation of (lactic acid, adenosine, PGE , IFN-c, etc.) many researches were aimed to fnding meaningful thera- and low pH in TME could activate MCs discharge particles peutic and prognostic biomarkers for EC in order to im- to pro- and antitumoral by IgE/FcεRI pathway [8–10]. prove the prognosis and prolong the lives of patients. Activated mast cells (AC-MCs) have been recognized as an 2 Journal of Oncology important prognostic indicator and immune therapy target between diferent datasets. Because the clinical information for cancers [11]. in GEO is limited, we only use the clinical features from TCGA when analyzing the results, which refer to the clinical Te prognosis was afected by the complex immune cell infltration (ICI) in TME. Recently, some researchers created information. In addition, we collected 33 ESCC patients’ models according to immune cells and diferential expres- tissues and 30 ESCC pathological sections who had not sion genes (DEGs) to predict prognosis. Apart from this, received therapy from Huai an First Hospital, in 2021. Te somatic mutation genes also infuenced a patient’s prognosis detailed information about the patient is listed in Table 1. and immunotherapy response [12]. TP53 mutations afected Tis study was performed in accordance with the principles the immunophenotype in gastric cancer and infuenced the of the Helsinki Declaration and was performed, reviewed, patient’s prognosis [13]. In addition, some clinical trials also and approved by the Ethics Committee of Zhongda Hospital indicated that KEAP1/NRF2 mutations can be regarded as of Southeast University; the grant number is predictive markers for immunotherapy and prognosis 2021ZDKYSB004. makers for cancer [14]. Tumor mutation burden (TMB) is defned as the total number of somatic gene coding errors, 2.2. Estimation of Stromal and Immune Scores. Te base substitutions, gene insertions, or gene deletions de- “CIBERSORT” algorithm is a deconvolution algorithm and tected per million bases. Some research studies suggest that was used to quantify the infltration level of the distinct im- TMB is associated with the emergence of neoantigens which mune cells based on the input reference gene sets and repeated trigger antitumor immunity [15, 16]. Tumor patients with 1000 times to ensure stability. Te “ESTIMATE” algorithm was higher TMB had higher survival rates [17, 18]. A few somatic used to calculate the immune scores, stromal scores, and es- mutations in tumor DNA can be translated into neo- timate scores by the “estimate” R package. At the same time, we antigens, which could be present on the surface of cells in the analyzed the prognostic value of immune stores and stromal form of the major histocompatibility complex and recog- scores and their relationship with clinical features. nized by the immune system [19]. However, the combined efects of ICI and TMB in predicting prognosis remained unknown. 2.3.ICIClusters. We used the R packages “biomanager” and In this study, we established multiple immune score “consensus” to divide the samples into diferent clusters models and TMB to predict prognosis and immunotherapy. according to the immune cells’ relative fraction levels in EC. Our results indicated that the combined immune score with And the prognostic values in diferent ICI groups were TMB was related to prognosis and PD-L1, and we recog- indicated by the “survival” and “survminer” R packages. Te nized the key mutation gene NRF2. We also found that immune cells in the difernt clusters were reshaped by NRF2 was related to AC-MCs. Based on these results, we “ggpubr” package. Results were visualized through heat analyzed the combined efects of NRF2 and AC-MCs for maps by the “pheatmap” R package. prognosis by TCGA database and experiment verifcation. Our results showed that there is an interaction between 2.4. DEGs Associated with the ICI Phenotype and Gene NRF2 and MCs, especially the higher NRF2, which had Clusters. DEGs in diferent ICI clusters were determined by a worse survival rate. In total, we thought NRF2 combined setting the signifcance cutof to p< 0.05 (adjust) and with AC-MCs could be used to predict the prognosis for EC logFC>1, which was performed by the “limma” R package. and provide a new direction for the prognostic study of According to DEGs, the samples were divided into diferent esophageal cancer. types using the “biomanager” and “consensus cluster plus” R packages. Immune cells in diferent gene clusters were an- 2. Materials and Methods alyzed by “ggpubr.” We also analyzed the prognostic value of diferent gene clusters as indicated by the “survival” and 2.1.ECDatasetsandSamples. A total of 524 EC samples were “survminer” R packages. downloaded from the TCGA-GDC database (https://portal. gdc.cancer.gov/) and the GEO database (https://www.ncbi. nlm.nih.gov/geo/). Te RNA sequencing (RNA-seq; frag- 2.5. ICI Scores. First, unsupervised clustering was used to ments per kilobase million value) data and the clinical in- deal with the samples in TCGA and GEO according to DEG formation (BCR-XML) including futime, survival state, age, values, which were positively or negatively correlated with the cluster signature and described as ICI gene signatures A gender, grade, stage, and the TNM stage system were downloaded from TCGA-EC. Te microarray data and B, respectively. Second, the “Boruta” R package was used for dimension reduction of the ICI gene signatures A and B (GSE68698, GSE69925, and GSE161533) were downloaded from the GEO. To increase the readability of the data, the and to extract feature genes. Tird, principal component 1 FPKM values were transformed into TPMs (transcripts per was extracted as the signature score by using the principal kilobase million), which were identical to the results of component analysis (PCA). Finally, the formula that defned microarrays, and clinical information (BCR-XML) was the ICI score of each patient was transformed into a matrix. Te “limma” R package and the ICIscore � 􏽐 PC1A − 􏽐 PC1B, and we divided the ICI score “sva” R package were used to merge the RNA array. Te into a high ICI score group and a low ICI score group. “ComBat” algorithm was used to decrease the likelihood of According to the ICI score, the functional enrichment an- batch efects from diferent biological and technical biases alyses of GO and KEGG pathways were analyzed using the Journal of Oncology 3 Table 1: Te relationship between MCs with clinicopathological features of ESCC. Clinicopathological N MC (x ± s) P N FcεR1G (x ± s) P features Gender Male 8 25.87± 15.22 0.299 21 2.09± 6.46 0.69 Female 13 33.07± 33.4681 4 1.04± 1.99 Age ≤65 16 24.37± 17.52 0.033 9 0.41± 0.61 0.15 >65 5 49.40± 46.39 16 2.77± 7.37 Diferentiation High diferentiation 4 43.00± 11.91 5 0.24± 0.22 Middle diferentiation 10 32.10 ± 37.69 0.446 14 3.07± 7.87 0.568 Low diferentiation 7 20.57± 13.22 6 0.64± 0.46 T1-T2 7 42.28± 42.13 0.160 7 1.24± 1.31 0.345 T3-T4 14 24.35± 16.29 8 2.19± 7.01 N0 13 20.23± 14.60 0.119 10 0.99± 1.15 0.179 N1–N3 8 46.75± 36.95 15 2.54± 7.67 M0 — — — 22 0.79± 0.97 <0.001 M1 — — 3 10.24± 17.24 Stage I 4 26.25± 19.25 0.032 — — II 8 15.00± 10.46 12 1.05± 1.06 0.095 III 8 50.12± 34.36 13 2.72± 8.26 “clusterProfler” R package for the feature genes in the high extract RNA from tumor tissue and para-tumor tissue. Te ICI score group and the low ICI score group. In order to RAN was cDNA obtained by reverse transcription according know the prognostic signifcance of the ICI score, we also to the protocol (Vazyme, China). SYBR green was used to analyze the connection between clinical features and ICI complete the related expression. Te Q-PCR procedure fol- score based on the TCGA database. lowed the protocol (Vazyme, China). Te primer sequences ° ° ° ° were as follows: 95 C 3 min, 95 C 30 s, 60 C 15 s, 72 C, 30 s for 40 cycles, and solubility curve. Te primer sequence is listed in 2.6. Somatic Alteration Data Analysis. Te related somatic Table S1. In addition, tissues were addedto RIPA and lysed in mutation datasets for EC were downloaded from the an ultrasound machine. After being divided by SDS-PAGE, TCGA-GDC database. Tumor mutation burden (TMB) is the proteins were transferred onto PVDF membranes and defned as the total number of somatic coding errors in then blocked with 5% skim milk for 2 h, subsequently in- genes, base substitutions, and gene insertion and deletion cubated with primary antibodies of NRF2 (1 :1000), TPSB2 errors in EC. Te “ggpubr” R package was used to analyze (1 :1000), and GAPDH (1 : 5000) overnight at 4 C and next the TMB for high ICI scores and low ICI scores. Te mu- incubated with secondary antibodies for 1 h at room tem- tation genes with high ICI scores and low ICI scores were perature. Te target protein was visualized by the ECL Gel identifed through the “maftool” R package, and the top 30 Image System and analyzed by the software Image J. genes with the highest mutation frequency were listed. 2.9. Statistical Analysis. All statistical analyses were ac- 2.7. Toluidine Blue Staining. Toluidine blue staining was complished with R version 4.0.3, GraphPad Prism 8, and used to detect the number and distribution of MCs in ESCC. SPSS version 25.0. Te comparison between the two groups Parafn-embedded tissues were dewaxed in diferent con- was tested by the Wilcoxon test and T test; otherwise, it was centrations of alcohol, subsequently stained with toluidine tested by Kruskal–Wallis H and ANOVA. Te survival blue (Solarbio, China) for 15 min, and washed with PBS 3 curves for the subtypes were accomplished with the times. Photomicrographs of ten felds were taken at diferent Kaplan–Meier plotter. Te chi-square test was used to an- magnifcations using the camera (ZEISS, Germany), and the alyze the correction between the ICI score subtypes and mean value was used to describe the number and distri- somatic mutation frequency. Te chi-square test was used to bution of MCs in EC. Te AC-MCs rate was calculated by analyze the classifed variable. And the correlation analysis the ratio of the AC-MCs number to the total MCs number. was completed by Pearson’s analysis. Univariate and mul- tivariate Cox regression models were used to analyze the 2.8. Q-PCR and Western Blot Analyzed the Expression of MCs prognosis. Te interaction of NRF2 and AC-MC was ana- Related Genes and NRF2. We analyzed the relative genes in lyzed by interactive analysis. All analyses were two-tailed, ESCC tumor tissue and para-tumor tissue. Trizol was used to and p< 0.05 was regarded as the statistically signifcant level. 4 Journal of Oncology patients were divided into two groups (high ICI score and 3. Results low ICI score). We analyze the prognostic value of the ICI 3.1. Te Characteristic of ICI in the TME of EC. 22 human score. Te survival rate in the two ICI score groups has no immune cells were calculated through the CIBERSORT al- statistical diference (Figure 2(c)), but statistical analysis gorithm according to the TCGA and GEO databases and showed that survival status and the TN stage system were found to have diferential expression in tumor tissues and related to ICI score (Figures S2E–S2G). Meanwhile, we para-carcinoma tissues. Tese results suggested that the rel- analyzed the main pathways in high ICI scores, such as ative fractions of Tregs and resting MCs in the tumor tissue adherens junction, cell cycle, Hedgehog signaling pathway, were lower than those in para-carcinoma tissue, but the TGF-β signaling pathway, and Wnt signaling pathway, while naive CD4 T cells, activated CD4 memory T cells, M0 the main pathways in the low ICI score were the B cell macrophages, activated DCs, activated MCs, and neutrophils receptor signaling pathway, drug metabolism cytochrome in the tumor tissues were higher compared with para- P450, intestinal immune network for IgA production, pri- carcinoma tissues (Figure 1(a)). Te “corrplot” R package mary immunodefciency, and T cell receptor according to was used to generate a correlation coefcient heatmap to KEGG (Figure 2(i)). Functional enrichment analysis sug- visualize the landscape of 22 immune cells’ interactions in gested that the main functions of the high ICI score group in TME (Figure 1(b)). Additionally, the ESTIMATE algorithm the biological process were response to virus, type I in- was used to calculate the immune scores and stromal scores terferon signal pathway, and response to tumor necrosis according to the levels of immune cells in EC. According to factor, but in the low ICI score group were extracellular the clinical information from the TCGA database, we ex- matrix organization and endodermal cell diferentiation. Te plored the relationship between clinical features and estimate main functions enriched in the cellular component of the scores. Tese results suggested that the immune scores and high ICI score group were the extracellular matrix immu- stromal scores were not associated with survival time, but nological synapse, membrane raft, anchored component of clinical stage and Tstage were related to stromal scores, and T membrane, and apical plasma membrane, while in the low stage was related to immune scores (Figures 1(c)–1(f)). ICI score group, they were the endoplasmic reticulum lu- men, extracellular matrix, and fbrillar collagen trimer (Figures S2A–S2D). Tese results suggested that the ICI 3.2. Diferent Patterns Were Used to Predict the Prognosis. score may be related to the prognosis of EC. We analyzed the prognosis value of stromal scores and immune scores, but the results suggested that the score patterns were unrelated to prognosis (Figures S1(A) and 3.3. Combine ICI Score with TMB Predict Prognosis. Most S1(B)). So, we try to create new patterns according to im- evidence indicated that TMB could be used to evaluate the predictive prognosis [20, 21]. In the study, we analyzed the mune cells and DEGs to predict the prognosis. First, the ICI types were divided into three clusters (Figure S1(C)). relationship of TMB with the somatic mutation landscape in the However, the three ICI clusters have no signifcant survival EC and ICI scores, but the result showed that the TMB showed diference in EC (Figure 2(a)). Ten, we constructed another no signifcant diferences between the two groups (Figure 3(a)). subtype according to DEG (Figure S1(D)). Similarly, dif- Ten, we divided the samples into high/low TMB, and the result ferent gene clusters were unrelated to prognosis suggested that the survival rate of low TMB was higher than that (Figure 2(b)). However, the ICI clusters and gene clusters of high TMB (Figure 3(b)). At the same time, when we com- were all related to PD-L1 (Figures 2(d) and 2(e)). So, we bined the ICI score with the TMB, we found that the survival analyzed the immune cells and DEGs in the clusters, and we rate in the group with a low TMB and a low ICI score was the found that PD-L1 was more highly expressed in cluster longest, whereas the group with a high TMB and a high ICI score C. Te main immune cells in cluster C were CD8 T cells, was the shortest (Figure 3(c)). Furthermore, we assessed the distribution of somatic variants in EC driver genes between the CD4 T follicular helper cells, T cells gamma delta, NK cells, M1 macrophages, DCs, Tregs, and MCs (Figures 2(g) and high/low ICI score subgroups. Te top 30 genes with the highest alteration frequency were further analyzed (Figures S3A and 2(h)). At the same time, we also analyzed the relationship between the immune cells in cluster C and PD-L1, and the S3B). We also analyzed that the relative expression of mutation genes in high/low ICI score subgroups; DYNC2H1, OBSCN, results suggested that PD-L1 is positively related to CD8 T cells and DCs but negatively related to Tregs (Figure 2(f)). DNAH11, PIK3CA, MUCSB, NRF2, ARID1A, SACS, LRRK2, Te heatmap delineated the transcriptomic profle of all NOTCH1, and SMAD4 were all signifcantly diferent between DEGs in three gene clusters and gene types (Figures S1E and the high/low ICI score subgroups (Figure 3(d)). After analyzing S1F). To achieve quantitative indicators of the ICI landscape the role of these genes in TMB, we found that NRF2 was related in EC patients, PCA was used to calculate two aggregate to TMB (Figure 3(e)). We further analyzed the prognosis value scores according to the ICI score A from ICI signature gene of NRF2 and indicated that the expression of NRF2 was related to the survival status, TNM system, and grade Figures S(3C-3F). A and the ICI score B from ICI signature gene B (Table S2). In this research, the individual score of patients was com- However, univariate variables and multivariate Cox regression models were used to investigate the relationship between the puted through the ISA and ISB of each patient. All the Journal of Oncology 5 ** P<0.05 0.3 ** ** ** 0.2 2000 B cells naive B cells memory Plasma cells T cells CD8 ** T cells CD4 naive 1000 * T cells CD4 memory resting 0.1 T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes 0.0 Macrophages M0 -1000 Macrophages M1 Macrophages M2 Dendritic cells resting Dentritic cells activated Mast cells resting -2000 Mast cells activated Eosinophils T1 T2 T3 T4 Neutrophils StromalScore ImmuneScore -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Normal EC (a) (b) (c) P<0.001 p=0.289 p=0.032 1000 2000 0 1000 -1000 -1000 -2000 -1000 -2000 -3000 -2000 -3000 T1 T2 T3 T4 stageI stage II stageIII stageIV stageI stage II stageIII stageIV (d) (e) (f) Figure 1: Te landscape of ICI in the TME of EC. (a) Te immune cells in EC tissue and para-cancer tissue. (b) Te landscape of 22 immune cells’ interactions in TME. (c and d) Association of immune scores with T stage (c) and clinical stage (d). (e and f) Association of stromal ∗ ∗∗ ∗∗∗ scores with T stage (e) and clinical stage (f). p< 0.05; p< 0.01; p< 0.001. 1.00 1.00 1.00 0.75 0.75 0.75 0.50 0.50 0.50 0.25 p=0.801 0.25 p=0.936 0.25 p=0.180 0.00 0.00 Time (years) 0123456 0.00 Number at risk A 20 15 1 0 0 0 0 Time (years) Number at risk 012345 6 B 25 18 8 2 1 0 0 C 23 15 4 2 2 2 0 A 50 35 6 1 0 0 0 Time (years) 67 41 19 11 5 1 0 0123456 Number at risk 42 29 8 3 2 2 0 Time (years) High 57 43 5 0123456 11 1 1 0 22 10 2 Time (years) Low 102 62 6 0 012 3 4 56 ICI cluster Time (years) A Gene cluster ICI score High Low (a) (b) (c) *** *** *** *** 12.5 ns 12.5 10.0 10.0 PD-L1 * p < 0.05 T cells CD8 7.5 T cells follicular helper 7.5 ** p < 0.01 T cells regulatory (Tregs) Correlation T cells gamma delta 1.0 NK cells resting 5.0 0.5 NK cells activated 5.0 Macrophages M1 0.0 Dendritic cells resting -0.5 Dendritic cells activated Mast cells resting -1.0 2.5 AB C Mast cells activated AB C ICI cluster Gene cluster (d) (e) (f) Figure 2: Continued. expression fraction level PD-L1 expression Survival probability stromalscore B cells naive ICI cluster B cells memory Plasma cells T cells CD8 T cells CD4 naive T cells CD4 memory resting T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes Macrophages M0 Macrophages M1 Macrophages M2 Dendritic cells resting Dentritic cells activated Mast cells resting Mast cells activated Eosinophils Neutrophils PD-L1 expression Survival probability Gene cluster ImmuneScore StromalScore immunescore Neutrophils Eosinophils Mast cells activated Mast cells resting Dentritic cells activated Dendritic cells resting Macrophages M2 Macrophages M1 Macrophages M0 Monocytes NK cells activated NK cells resting T cells gamma delta T cells regulatory (Tregs) T cells follicular helper T cells CD4 memory activated T cells CD4 memory resting T cells CD4 naive T cells CD8 Plasma cells B cells memory B cells naive Survival probability ICI score stromalscore immunescore PD-L1 T cells CD8 T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Macrophages M1 Dendritic cells resting Dendritic cells activated Mast cells resting Mast cells activated 6 Journal of Oncology 12 12 *** ns ** ns ns ns ns ** * ns *** ** ** *** *** ** *** ns*n ns *** ** * *** *** *** ***sns *** *** *** *** *** *** *** ** *** *** ** *** *** *** *** *** ** *** ICI cluster Gene cluster (g) (h) 0.5 0.0 -0.5 High ICI score Low ICI score KEGG_ADHERENS_JUNCTION KEGG_B_CELL_RECEPTOR_SIGNALING_PATHWAY KEGG_CELL_CYCYLE KEGG_DRUG_METABOLISM_CYTOCHROME_P450 KEGG_HEDGEHOG_SIGNALING_PATHWAY KEGG_INTESTINAL_IMMUNE_NETWORK_FOR_IGA_PRODUCTION KEGG_PRIMARY_IMMUNODEFICIENCY KEGG_T_CELL_RECEPTOR_SIGNALING_PATHWAY KEGG_TGF_BETA_SIGNALING_PATHWAY KEGG_WNT_SIGNALING_PATHWAY (i) Figure 2: Diferent prognostic models constructed based on immune cells and DEGs. (a) Kaplan–Meier curves for overall survival of EC with diferent ICI clusters. (b) Kaplan–Meier curves for overall survival of EC with diferent gene clusters. (c) Kaplan–Meier curves for overall survival of EC with diferent ICI scores. (d) Te expression of PD-LA in ICI clusters. (e) Te expression of PD-L1 in ICI clusters. (f) Te relationship between PD-L1 and immune cells. (g) Te fraction of tumor immune cells in three ICI clusters. (h) Te fraction of tumor ∗ ∗∗ immune cells in three gene clusters. (i) Enrichment plots showing signaling pathways in high/low ICI scores. p< 0.05; p< 0.01; ∗∗∗ p< 0.001. NRF2 mutation and the overall survival of EC patients, and the intracytoplasmic, and AC-MCswerecharacterized by many result revealed that the NRF2 mutation was not an independent blue dye particles surrounding the cells. Our results showed prognostic factor for OS in EC (Table S3). that MCs were mainly in the muscular layer (p< 0.05) squamous epithelium (0.67± 3.46, Figure 4(a)-A/B/C), tumor nest (4.48± 9.63, Figure 4(a)-D/E/F), and muscu- laris propria (36.33± 37.84, Figure 4(a)-G/H/I). We also 3.4. Combining AC-MCs with NRF2 Could Predict Prognosis. calculated the rate of AC-MCs in EC tissue. Te MCs in Considering the relationship between the TMB and ICI patients in stage III were higher than those in patients in I scores, what follows is the relationship between NRF2 and and II (Table 1). We also analyzed the related gene, which immune cells. We found that NRF2 was only related to could activate MCs. Tese results showed that FcεR1A MCs (Table S4). Next, we attempt to assess the combined (Figure 4(b), p< 0.005), NRF2 (Figure 4(b), p< 0.05), efect of NRF2 and MCs for predicting prognosis in EC. We FcεR1G (Figure S4D, p< 0.000), and PD-L1(Figure S4E, collected 30 EC patients’ tissue slices and their clinical p< 0.05) were all upregulated in tumor tissue (Figures S4D information to analyze the number and distribution of and S4E), and the protein level of NRF2 and TPSB2 was MCs/AC-MCs and their prognosis value. MCs were also higher expressed in tumor tissue (Figure 4(d)–E). Te characterized by blue densely basophilic granules in the Scale of fraction B cells naive B cells memory Plasma cells T cells CD8 T cells CD4 naive T cells CD4 memory resting T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes Macrophages M0 Enrichment Score Macrophages M1 Macrophages M2 Dendritic cells resting Dendritic cells activated Mast cells resting Mast cells activited Eosinophils Neutrophils StromalScore ImmuneScore Scale of fraction B cells naive B cells memory Plasma cells T cells CD8 T cells CD4 naive T cells CD4 memory resting T cells CD4 memory activated T cells follicular helper T cells regulatory (Tregs) T cells gamma delta NK cells resting NK cells activated Monocytes Macrophages M0 Macrophages M1 Macrophages M2 Dendritic cells resting Dendritic cells activated Mast cells resting Mast cells activited Eosinophils Neutrophils StromalScore ImmuneScore Journal of Oncology 7 0.065 1.00 0.75 0.50 p=0.004 0.25 0.00 01 23456 Low High Time (years) Number at risk ICI score H–TMB 31 15 4 1 1 0 0 L–TMB 127 89 29 14 6 3 0 Low 0 123 4 5 6 Time (years) High H–TMB L–TMB (a) (b) 12.5 *** *** ns ** *** *** *** ns ns ** * ns *** ns *** ** 1.00 0.75 10.0 0.50 7.5 0.25 p=0.015 5.0 0.00 0 12345 6 2.5 Time (years) Number at risk H–TMB+H–ICI score 5 4 1 0 0 0 0 H–TMB+L–ICI score 26 11 3 1 1 0 0 L–TMB+H–ICI score 40 32 6 2 1 1 0 L–TMB+L–ICI score 87 57 23 12 5 2 0 0 123456 ICI score Time (years) Low H–TMB+H–ICI score L–TMB+H–ICI score High H–TMB+L–ICI score L–TMB+L–ICI score (c) (d) p=0.035 (e) Figure 3: Interaction between the ICI score and the TMB. (a) TMB diference in the high ICI score and low ICI score. (b) Kaplan–Meier curves for high and low TMB groups of the TCGA-EC cohort. (c) Kaplan–Meier curves for patients in the TCGA-EC cohort stratifed by both TMB and ICI scores. (d) Te relative expression level in the high and low ICI score groups. (e) Te value of TMB for NRF2 mutations ∗ ∗∗ ∗∗∗ and non-NRF2 mutations. p< 0.05; p< 0.01; p< 0.001. expression of NRF2 was related to FcεR1A (Figure S4E, unrelated to OS (Figures S4A and S4B), but the group with r = 0.515), and PD-L1 was related to FcεR1G (Figure S4G, low NRF2 and high FcεR1G was the lowest malignant r = 0.468). We divided the expression of NRF2 and FcεR1G (Figure S4C). Most importantly, there is an interaction into two groups by median, and interaction analysis was between NRF2 and MC (Figure 4(f)). Hence, we thought used to explore the interaction of NRF2 and AC-MCs with that a combination of NRF2 and AC-MCs could be TNM. Cox results suggested that NRF2 and MCs were all a prognosis maker for EC. Survival probability Tumor Burden Mutation TMB (mutation) NRF2 mutation Gene expression DYNC2H1 OBSCN Survival probability APOB NRF2 non-mutation DNAH11 MUC4 PIK3CA MUC5B MUC16 IVL NFE2L2 ARID1A WDFY4 SACS LRRK2 NOTCH1 SMAD4 8 Journal of Oncology (a) p=0.046 p=0.005 -5 -5 -10 -15 -10 ESCC Normal ESCC Normal (b) (c) Figure 4: Continued. relative expression level of FCER1G relative expression level of NRF2 Journal of Oncology 9 1.5 ** NRF2 ** TPSB2 1.0 GAPDH NT N T N T 0.5 0.0 NT N T NRF2 TPSB2 (d) (e) Estimated Marginal Means of TNM 2.0 1.8 1.6 1.4 1.2 1.0 NRF2 FccR1G (f) Figure 4: Te infltration of MCs in EC and the related gene expression as well as its relationships. (a) Te MCs’ infltration of EC tissues. (A/B/C) Squamous epithelium. (D/E/F) Tumor nest. (G/H/I) Muscularis propria. Te black arrow represents undegranulated MCs, and the red arrow represents granulated MCs. (b) Expression of FCER1G. (c) Expression of NRF2. (d) Statistical analysis of TPSB2 and NRF2 ∗ ∗∗ protein expression. (e) Protein expression levels of NRF2 and TPSB2. (f) Interaction of NRF2 and AC-MCs. p< 0.05; p< 0.01. immunosuppression and promote tumor survival and 4. Discussion progression [22–24]. In this study, we analyzed the ICI Te majority of studies have demonstrated that the het- landscape of EC according to the TCGA and GEO databases. erogeneous TME and TMB participated in tumor pro- Our results indicated that CD4 T cells, M0 macrophages, gression, prognosis, and therapeutic for EC. However, AC-MCs, and activated DCs were increased, but the Tregs clarifying the modulation of TME and TMB as well as their and resting MCs were decreased in tumor tissue. Tregs combination efects during EC remains a challenge. Our suppress the activation and proliferation of multiple types of + + study comprehensively described the ICI landscape and immunocompetent cells such as CD4 T cells, CD8 T cells, somatic mutation gene landscape and constructed diferent B cells, NK, and T cells, as well as suppressive immuno- patterns to quantify the ICI and TMB by the “CIBERSORT” reaction [25]. CD4 T cells could increase the secretion of and “ESTIMATE” algorithms to predict prognosis and the IL4, IL2 promoting breast cancer progression, and the relationship with PD-L1. We found that the combined mature dendritic cells induced the proliferation of CD4 T immune fltration cells and tumor mutation burden could cells [26, 27]. AC-MCs could produce VEGF, PDGF, MMP9, predict the prognosis for EC. At the same time, we recog- and PGE2 to promote angiogenesis and tumor migration nized the key mutation genes NRF2 and immune cells (mast [28]. Moreover, AC-MCs’ secreted cytokines could also cells), which played an important role in predicting prog- infuence the development and function of T cells and B nosis. We verifed the combined role of NRF2 and mast cells cells [29]. Apart from evaluating the infltration of single in EC patient and found that combined NRF2 and MCs immune cells, we also attempt to quantify the ICI landscape would be a prognostic target and provide new insight into to evaluate the prognosis through built-score patterns. In the prognosis of EC. previous studies, the ESTIMATE algorithm has been used to Multiple pieces of evidence have demonstrated that analyze the immune scores and stromal scores, and it has dysfunctional immune cells in the TME lead to been suggested that the risk model is benefcial for the early relative expression of related genes Estimated Marginal Means 10 Journal of Oncology showed that the high TMB and low ICI group had the worst identifcation of high-risk patients to formulate an in- dividualized treatment project and improve the possibility of OS. Meanwhile, these results indicated that TP53, TTN, MUC16, LRP1B, and SYNE1 were high-frequency muta- an immunotherapy response [30, 31]. In our study, based on the stromal scores and immune scores, we divided the tions in EC. Especially, NRF2 was not only a high-frequency patients into high-score and low-score groups. We found mutation gene in EC but also signifcantly diferent in ICI that the survival probability in the two groups did not score groups. Tere was a study that reported that NRF2/ signifcantly change, but the stromal scores were higher in KEAP1 mutations correlate with higher TMB value/PD-L1 stage III, and the higher the immune scores and stromal expression and potentiate improved clinical outcomes with scores, the higher the T stage. At the same time, we divided immunotherapy [45]. An NRF2 mutation could disrupt the the samples into three parts based on the infltrated immune weak binding of Keap1 with the NRF2-DLG motif and cells. Our results demonstrated that the immune cells which activate NRF2 to promote tumor progression [46]. Con- sidering the complexity of mutations, we only detected the have immunosuppressive function were focused on ICI cluster C. PD-L1, a key immune checkpoint, was higher in expression of NRF2 and the prognosis value in EC. In our result, the NRF2 was upregulated in EC, but not an in- ICI cluster C. Previous evidence had shown that immune cell-related genes could predict disease progression and dependent prognostic biomarker, which was diferent from immunotherapeutic responses [32, 33]. Based on the previous research studies [47]; the reason probably was that immune-related gene in EC, we divided the patient into the number of patients was not enough. Meanwhile, we three ICI gene clusters. Te results suggested that ICI gene analyzed the relationship between NRF2 and immune cells cluster C had a more favorable immune-activated type with and found that NRF2 was related to MCs. Other studies the highest density of CD8 T cells, M1 macrophages, acti- indicated that NRF2 could activate MCs, IgG/FcεRI pro- vated DCs, and CD4 T follicular helper cells [34–36]. moted the phosphorylation of Lyn and activated Syk/PI3K, Additionally, the expression of PD-L1 was highest in ICI LAT/p38, and LAT/Raf-1/ERK1/2 pathways, and the AKT- Nrf2 and p38MAPK-Nrf2 signal pathways play an important gene cluster C. Hence, the patients in ICI gene cluster C might have a better immune response. Te outcome of our role in hypersensitivity induced by MCs [48, 49]. Hence, we analyzed the MCs in EC tissue and found that MCs were analysis was in accordance with the previous study, which indicated that ICI clusters and ICI gene clusters in EC might irrelevant to OS. Surprisingly, the combination of NRF2 with MCs could afect prognosis. In addition, previous infuence the expression of PD-L1 [37]. In recent years, gene clusters related to immune response studies indicated that MCs could express PD-L1 and play and proliferation were used to predict the outcome of a crucial role in immunosuppression [50]. Our results also cancers and identify high-risk patients; the distant indicated that FcεR1G could activate MCs, and the AC-MCs metastasis-free survival in high-score immune gene was were positively related to PD-L1, but the mechanism by higher than low-score in breast cancer [38]. Te prognosis which activated MCs regulated PD-L1-induced immuno- value of the ICI score was calculated by the “Boruta algo- suppression deserves deep research. Terefore, we thought combining NRF2 with MCs would be used to predict rithm” based on the immune cell-related gene, which has been proven in head and neck squamous cell carcinoma [39]. prognosis. However, whether the NRF2-activated MCs are involved in immune suppression in EC needs further study. In the current study, we assessed the prognosis value of the ICI score in EC and found that there was no signifcant In summary, we comprehensively analyzed the ICI diference in OS in high/low ICI scores , but the ICI score landscape and TMB of EC and found that high ICI and high was higher in alive, no lymph node metastasis samples. TMB had worse prognoses. We also recognized key mu- Trough KEGG, our results indicated that the high ICI score tation genes and immune cells and analyzed the common mainly regulates the hedgehog signaling pathway, TGF-β prognostic value of NRF2 with MCs by experiment verif- signaling pathway, Wnt signaling pathway, and so on. cation and database analysis. Nevertheless, several limita- Hedgehog signaling could be induced by activated T cells tions in our study should be considered. First, due to the and NK cells and participate in immunotherapy [40]. Re- limited patient information from TCGA, a larger sample size and sufcient information were required for further proof of pression of the Wnt signaling pathway would decrease the expression of PD-L1 and increase the immune-killing efect our results. Second, the role of NRF2 and MCs participated in immune regulation and tumor progression in EC needs of NK cells [41]. Te TGF-β/EMT signaling pathway infuenced the expression of PD-L1 and promoted immune further experimental study. Tird, it is not enough to clarify escape. All these results demonstrated that the ICI score was diferent patterns and MCs that could infuence immuno- related to the PD-L1 but was not an independent prognosis therapeutic efectiveness only by analyzing the relationship marker for EC [37]. with PD-L1. In all, we found that high ICI and high TMB Te majority of studies demonstrated that TMB was could afect the prognosis, and the combination of NRF2 related to prognosis and could be a marker for predicting the with AC-MCs had a worse prognosis and could be an ef- fective prognostic factor for EC. efectiveness of immune checkpoint inhibitors in cancer [42, 43]. Mutation genes related to TMB were crucial prognostic biomarkers for cancers [32, 44]. In our study, we Abbreviations analyzed the somatic mutation landscape according to TCGA. Our results indicated that the high TMB level had EC: Esophageal cancer a poor OS, and the combination of the TMB with ICI scores EAC: Esophageal adenocarcinoma Journal of Oncology 11 Table S4: correlation of NRF2 with immune cells in patients ESCC: Esophageal squamous cell carcinoma from the TCGA database and GEO database using the TME: Tumor microenvironment Pearson correlation coefcient. Figure S1: diferent patterns NK: Natural killer were used to predict the prognosis. Figure S2: the ICI score DCs: Dendritic cells predicts the prognosis and the GO analysis. Figure S3: the Tregs: Regulatory T cells mutation genes in diferent ICI score groups and the MCs: Mast cells prognosis of NRF2. Figure S4: the prognostic value of NRF2 AC-MCs: Activated mast cells and activated mast cells. (Supplementary Materials) TCGA: Te Cancer Genome Atlas database GEO: Gene Expression Omnibus ICI: Immune cell infltration References TMB: Tumor mutation burden [1] F. Z. Wang, L. Zhang, Y. Xu, Y. Xie, S Li, and S. Li, NFE2L2/ Nuclear factor erythroid 2-related factor 2 “Comprehensive analysis and identifcation of key driver NRF2: genes for distinguishing between esophageal adenocarcinoma RYR2: Ryanodine receptor 2 and squamous cell carcinoma,” Frontiers in Cell and De- KEAP1: Kelch-like ECH-associated protein 1 velopmental Biology, vol. 9, Article ID 676156, 2021. TPMs: Transcripts per kilobase million [2] A. A. Watkins, J. A. Zerillo, and M. S. Kent, “Trimodality DEGs: Diferentially expressed genes approach for esophageal malignancies,” Surgical Clinics of PCA: Principal component analysis North America, vol. 101, no. 3, pp. 453–465, 2021. KEGG: Kyoto Encyclopedia of Genes and Genomes [3] J. C. Lu, H. Tao, Y. Q. Zhang et al., “Extent of prophylactic GO: Gene Ontology postoperative radiotherapy after radical surgery of thoracic esophageal squamous cell carcinoma,” Diseases of the ISA: ICI signature gene A Esophagus, vol. 21, no. 6, pp. 502–507, 2008. ISB: ICI signature gene B. [4] H. D. Teixeira Farinha, A. Digklia, D. Schizas, N. Demartines, M. Schafer, ¨ and S. Mantziari, “Immunotherapy for esophageal Data Availability cancer: state-of-the art in 2021,” Cancers, vol. 14, no. 3, p. 554, Te datasets supporting our results are available in the [5] W. Z. Zhang, Z. Zhao, and F. Li, “Natural killer cell dys- TCGA and GEO database as well as data sources in the function in cancer and new strategies to utilize NK cell po- method. Te data of our cohort are provided in tables, and tential for cancer immunotherapy,” Molecular Immunology, further inquiries can be directed to the corresponding vol. 144, pp. 58–70, 2022. authors. [6] I. Plesca, L. Muller, J. P. Bottcher, H. Medyouf, R. Wehner, and M. Schmitz, “Tumor-associated human dendritic cell subsets: phenotype, functional orientation, and clinical rele- Conflicts of Interest vance,” European Journal of Immunology, vol. 52, no. 11, pp. 1750–1758, 2022. Te authors declare that they have no conficts of interest. [7] D. S. Villalobos, I. G. Martinez-Aguilar, M. Ibarra-Sanchez et al., “Mast cell-tumor interactions: molecular mechanisms of Authors’ Contributions recruitment, intratumoral communication and potential therapeutic targets for tumor growth,” Cells, vol. 11, no. 3, Ran Liu, Junjie Lv, and Xinxin Guo designed the study. Weitao Shen and Mingjun Sun analyzed the data. Xinxin [8] G. Varricchi, M. R. Galdiero, S. Lofredo et al., “Are mast cells Guo wrote the manuscript. and Ran Liu provided funding MASTers in cancer?” Frontiers in Immunology, vol. 8, p. 424, support. All the authors have read and reviewed the manuscript. [9] K. Mukai, M. Tsai, H. Saito, and S. J. Galli, “Mast cells as sources of cytokines, chemokines, and growth factors,” Im- Acknowledgments munological Reviews, vol. 282, no. 1, pp. 121–150, 2018. [10] F. A. Redegeld, Y. Yu, S. Kumari, N. Charles, and U. Blank, Te authors thank Dr. Jiajia Xu from Department of Pa- “Non-IgE mediated mast cell activation,” Immunological Reviews, vol. 282, no. 1, pp. 87–113, 2018. thology, Zhongda Hospital Southeast University, for helping [11] H. W. N. Choi, M. Naskar, H. K. Seo, and H. W. Lee, “Tumor- us analyze the pathological section. Tis work was supported associated mast cells in urothelial bladder cancer: optimizing by the National Natural Science Foundation of China (grant immuno-oncology,” Biomedicines, vol. 9, no. 11, p. 1500, 2021. nos. 82173479 and 81872579) and the Postgraduate Research [12] A. Desrichard, A. Snyder, and T. A. Chan, “Cancer neo- and Practice Innovation Program of Jiangsu Province (grant antigens and applications for immunotherapy,” Clinical no. KYCX22-0303). Cancer Research, vol. 22, no. 4, pp. 807–812, 2016. [13] K. Z. Nie, Z. Zheng, Y. Wen et al., “Construction and vali- Supplementary Materials dation of a TP53-associated immune prognostic model for gastric cancer,” Genomics, vol. 112, no. 6, pp. 4788–4795, Table S1: the primer sequences of related genes. Table S2: the DEGs in ICI gene signatures A and B. Table S3: association [14] W. C. M. Dempke and M. Reck, “KEAP1/NRF2 (NFE2L2) with overall survival and clinicopathological characteristics mutations in NSCLC - fuel for a superresistant phenotype?” in patients from the TCGA database using Cox regression. Lung Cancer, vol. 159, pp. 10–17, 2021. 12 Journal of Oncology [15] M. M. Gubin, M. N. Artyomov, E. R. Mardis, and [31] J. Gao, T. Tang, B. Zhang, and G. Li, “A prognostic signature R. D. Schreiber, “Tumor neoantigens: building a framework based on immunogenomic profling ofers guidance for for personalized cancer immunotherapy,” Journal of Clinical esophageal squamous cell cancer treatment,” Frontiers in Investigation, vol. 125, no. 9, pp. 3413–3421, 2015. Oncology, vol. 11, Article ID 603634, 2021. [16] M. Yarchoan, A. Hopkins, and E. M. Jafee, “Tumor muta- [32] F. L. Zhang, Y. Liu, Y. Yang, and K. Yang, “Development and tional burden and response rate to PD-1 inhibition,” New validation of a fourteen- innate immunity-related gene pairs England Journal of Medicine, vol. 377, no. 25, pp. 2500-2501, signature for predicting prognosis head and neck squamous cell carcinoma,” BMC Cancer, vol. 20, no. 1, p. 1015, 2020. [17] R. M. Samstein, C. H. Lee, A. N. Shoushtari et al., “Tumor [33] Y. Fu, S. Sun, J. Bi, C. Kong, and L. Yin, “A novel immune- mutational load predicts survival after immunotherapy across related gene pair prognostic signature for predicting overall multiple cancer types,” Nature Genetics, vol. 51, no. 2, survival in bladder cancer,” BMC Cancer, vol. 21, no. 1, p. 810, pp. 202–206, 2019. 2021. [18] L. Ke, S. Li, and H. Cui, “Te prognostic role of tumor [34] J. Hamanishi, M. Mandai, M. Iwasaki et al., “Programmed cell mutation burden on survival of breast cancer: a systematic death 1 ligand 1 and tumor-infltrating CD8+ T lymphocytes review and meta-analysis,” BMC Cancer, vol. 22, no. 1, p. 1185, are prognostic factors of human ovarian cancer,” Proceedings 2022. of the National Academy of Sciences, vol. 104, no. 9, [19] T. A. Chan, M. Yarchoan, E. Jafee et al., “Development of pp. 3360–3365, 2007. [35] H. Law, V. Venturi, A. Kelleher, and C. M. L. Munier, “Tfh tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic,” Annals of Oncology, vol. 30, cells in health and immunity: potential targets for systems no. 1, pp. 44–56, 2019. biology approaches to vaccination,” International Journal of [20] J. S. Wang, J. Song, Z. Liu, T. Zhang, and Y. Liu, “High tumor Molecular Sciences, vol. 21, no. 22, p. 8524, 2020. mutation burden indicates better prognosis in colorectal [36] M. Yang, Y. Cao, Z. Wang, T. Zhang, Y. Hua, and Z. Cai, cancer patients with KRAS mutations,” Frontiers Oncology, “Identifcation of two immune subtypes in osteosarcoma vol. 12, Article ID 1015308, 2022. based on immune gene sets,” International Immuno- [21] Y. Zheng, M. Yao, and Y. Yang, “Higher tumor mutation pharmacology, vol. 96, Article ID 107799, 2021. burden was a predictor for better outcome for nsclc patients [37] Y. H. Chen, X. Huang, L. Chen et al., “Characterization of the treated with PD-1 antibodies: a systematic review and meta- immune infltration landscape and identifcation of prog- analysis,” SLAS Technology, vol. 26, no. 6, pp. 605–614, 2021. nostic biomarkers for esophageal cancer,” Molecular Bio- [22] H. Sun, T. G. Martin, J. Marra et al., “Individualized genetic technology, pp. 1–23, 2022. makeup that controls natural killer cell function infuences the [38] M. Callari, V. Cappelletti, F. D’Aiuto et al., “Subtype-specifc efcacy of isatuximab immunotherapy in patients with metagene-based prediction of outcome after neoadjuvant and multiple myeloma,” Journal for ImmunoTerapy of Cancer, adjuvant treatment in breast cancer,” Clinical Cancer Re- vol. 9, no. 7, Article ID e002958, 2021. search, vol. 22, no. 2, pp. 337–345, 2016. [23] L. Shao, Q. Yu, R. Xia et al., “B7-H3 on breast cancer cell [39] X. Zhang, M. Shi, T. Chen, and B. Zhang, “Characterization of MCF7 inhibits IFN-gamma release from tumour-infltrating the immune cell infltration landscape in head and neck T cells,” Pathology, Research & Practice, vol. 224, Article ID squamous cell carcinoma to aid immunotherapy,” Molecular Terapy - Nucleic Acids, vol. 22, pp. 298–309, 2020. 153461, 2021. [24] M. Li, D. Xie, X. Tang et al., “Phototherapy facilitates tumor [40] H. M. Onishi, T. Morisaki, A. Kiyota et al., “Te Hedgehog recruitment and activation of natural killer T cells for potent inhibitor cyclopamine impairs the benefts of immunotherapy with activated T and NK lymphocytes derived from patients cancer immunotherapy,” Nano Letters, vol. 21, no. 14, pp. 6304–6313, 2021. with advanced cancer,” Cancer Immunology Immunotherapy, [25] M. Miyara and S. Sakaguchi, “Natural regulatory T cells: vol. 62, no. 6, pp. 1029–1039, 2013. mechanisms of suppression,” Trends in Molecular Medicine, [41] R. Deng, C. Zuo, Y. Li et al., “Te innate immune efector ISG12a promotes cancer immunity by suppressing the ca- vol. 13, no. 3, pp. 108–116, 2007. [26] C. Aspord, A. Pedroza-Gonzalez, M. Gallegos et al., “Breast nonical Wnt/β-catenin signaling pathway,” Cellular and cancer instructs dendritic cells to prime interleukin 13- Molecular Immunology, vol. 17, no. 11, pp. 1163–1179, 2020. [42] M. C. Guo, Z. Chen, Y. Li et al., “Tumor mutation burden secreting CD4+ T cells that facilitate tumor development,” Journal of Experimental Medicine, vol. 204, no. 5, pp. 1037– predicts relapse in papillary thyroid carcinoma with changes 1047, 2007. in genes and immune microenvironment,” Frontiers in En- [27] H. Qi, J. G. Egen, A. Y. C. Huang, and R. N. Germain, docrinology, vol. 12, Article ID 674616, 2021. [43] X. Liu, R. Qiu, M. Xu et al., “KMT2C is a potential biomarker “Extrafollicular activation of lymph node B cells by antigen- bearing dendritic cells,” Science, vol. 312, no. 5780, of prognosis and chemotherapy sensitivity in breast cancer,” pp. 1672–1676, 2006. Breast Cancer Research and Treatment, vol. 189, no. 2, [28] D. A. Ribatti, T. Annese, and R. Tamma, “Adipocytes, mast pp. 347–361, 2021. cells and angiogenesis,” Romanian Journal of Morphology and [44] Z. Guo, X. Yan, C. Song et al., “FAT3 mutation is associated Embryology, vol. 61, no. 4, pp. 1051–1056, 2021. with tumor mutation burden and poor prognosis in esoph- [29] Y. A. Mekori and D. D. Metcalfe, “Mast cell–T cell in- ageal cancer,” Frontiers in Oncology, vol. 11, Article ID teractions,” Te Journal of Allergy and Clinical Immunology, 603660, 2021. vol. 104, no. 3, pp. 517–523, 1999. [45] X. Y. Xu, Y. Yang, X. Liu et al., “NFE2L2/KEAP1 mutations [30] M. Liu, F. Li, B. Liu et al., “Profles of immune cell infltration correlate with higher tumor mutational burden value/PD-L1 and immune-related genes in the tumor microenvironment of expression and potentiate improved clinical outcome with esophageal squamous cell carcinoma,” BMC Medical Geno- immunotherapy,” Te Oncologist, vol. 25, no. 6, pp. 955–963, mics, vol. 14, no. 1, p. 75, 2021. 2020. Journal of Oncology 13 [46] L. W. Cheng, H. Wang, S. Li, Z. Liu, and C. Wang, “New insights into the mechanism of Keap1-Nrf2 interaction based on cancer-associated mutations,” Life Sciences, vol. 282, Article ID 119791, 2021. [47] Q. Ju, X. Li, H. Zhang, S. Yan, Y. Li, and Y. Zhao, “NFE2L2 is a potential prognostic biomarker and is correlated with im- mune infltration in brain lower grade glioma: a pan-cancer analysis,” Oxidative Medicine and Cellular Longevity, vol. 2020, Article ID 3580719, 26 pages, 2020. [48] X. Wang, Z. Lv, B. Han et al., “Te aggravation of allergic airway infammation with dibutyl phthalate involved in Nrf2- mediated activation of the mast cells,” Science of the Total Environment, vol. 789, Article ID 148029, 2021. [49] C. L. Xu, L. Li, C. Wang et al., “Efects of G-Rh2 on mast cell- mediated anaphylaxis via AKT-Nrf2/NF-κB and MAPK- Nrf2/NF-κB pathways,” Journal of Ginseng Research, vol. 46, no. 4, pp. 550–560, 2022. [50] Y. Lv, Y. Zhao, X. Wang et al., “Increased intratumoral mast cells foster immune suppression and gastric cancer pro- gression through TNF-alpha-PD-L1 pathway,” Journal for ImmunoTerapy of Cancer, vol. 7, no. 1, p. 54, 2019.

Journal

Journal of OncologyHindawi Publishing Corporation

Published: Jan 4, 2023

There are no references for this article.