Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Numerous studies reveal the clinical significance of tumor microenvironment (TME) in multiple cancers. The association between TME in oral squamous cell carcinoma (OSCC) and clinical outcomes remains unsolved.

OBJECTIVE:

This study aims to exhibit the TME of OSCC and identified the prognostic marker.

METHODS:

Gene expression profile and clinical data OSCC patients were from the TCGA database. The validated stage data was from the Gene Expression Omnibus (GSE65858). Immune/stromal scores of each patient were calculated by ESTIMATE algorithm. Biological functional prediction was conducted. Prognostic genes identified by survival analysis. Nomogram and Receiver operating characteristic curves were employed to test the predicting power. TIMER database was applied to evaluated the immune infiltrates.

RESULTS:

Lower immune scores were observed in male patients ( $P=$ 0.0107) and different primary tumor sites of oral cavity with different stromal scores ( $P=$ 0.0328). The Differentially expressed genes (DEGs) were involved in immune related pathways. HGF gene (hepatocyte growth factor) was prognostic related and with a better prognostic performance when combined with clinical features (AUC ${}_{\text{TCGA}}=$ 0.638, AUC ${}_{\text{GEO}}=$ 0.714). HGF was significantly related with B cell, CD4 $ï ¼ ‹$ T cell, CD8 $+$ T cell, macrophage, neutrophils, and dendritic cell infiltration.

CONCLUSION

: The current study analyzed the TME and presented immune related prognostic biomarkers for OSCC.

Keywords

Prognostic genes cancer oral tumor microenvironment

1. Introduction

Oral squamous cell carcinoma (OSCC) is one of the most common types of cancers in the world. The risk factors including tobacco and alcohol consumption or betel nut chewing contribute to the high incidence of OSCC [1]. Despite the advances in OSCC patients’ treatment, such as surgery, chemotherapy, radiotherapy or biologic therapy, the overall survival rate for OSCC remains poor in the recent decades [2]. In addition, OSCC patients have a poor quality of life due to the dysfunction in chewing and swallowing [3].

Tumor microenvironment (TME) refers to the cellular environment in which tumors or cancer stem cells exist. Apart from the tumor cells, the TME includes surrounding blood vessels, other non-malignant cells, the extracellular matrix (ECM) and signaling molecules. TME is a complex integrated system, which can be divided into immune microenvironment dominated by immune cells and non-immune remission dominated by fibroblasts [4]. It has been reported that TME is associated with tumor survival and recurrence [5], such as glioblastoma [6], colon cancer [7] and prostate cancer [8] by estimation of stromal and immune cells in malignant tumor tissues using expression data (ESTIMATE) algorithm [9], which was applied for calculating the immune and stromal cells in tissues. However, up to now, no studies have been conducted on the association between tumor microenvironment and prognosis in oral squamous cell carcinoma.

With the development of data sharing, transcriptome profiles and clinical data of cancer patients can be easily obtained from some public database such as The Cancer Genome Atlas (TCGA) database. TCGA were launched in 2005 using genome analysis technologies to understand the genetics of cancers, which helping to improve the diagnosis, treatment and prevention of cancers [10]. Here, we conducted the current study to identify prognostic genes which were associated with OSCC. A total of 331 OSCC patients with gene expression profiles and clinical data were collected to evaluate the TME by ESTIMATE, then patients were divided into two groups based on the median of immune scores as well as stromal scores to access the association of survival and differentially expressed genes (DEGs). Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) pathway enrichment analysis were applied to investigate the biological functions of identified genes. Survival analysis was used to explore the prognostic genes and prognostic genes were validated in Gene Expression Omnibus (GEO) database. In addition, receiver operating characteristic curve (ROC) and nomogram were applied to access the prognostic value of the identified key gene and clinical characteristics. Furthermore, the TIMER database was applied to evaluate the association between the key gene HGF (hepatocyte growth factor) and immune infiltration, the whole work flow was exhibited in Supplementary Fig. 1.

We present the following article in accordance with the MDAR checklist.

2. Methods and materials

2.1 Data collection

Level 3 gene transcriptome profiles for 331 OSCC patients were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/), the RNA expressions for OSCC polymorphism were collected by using the Illumina HiSeq 2000 RNA Sequencing platform. The following corresponding clinical characteristics were extracted: sex, age, cancer type, topography, lymph node, metastasis (TNM) based on American Joint Committee on Cancer (AJCC), pathological stage and primary site. Then, we download the survival results of OSCC patients through the Genomic Data Commons (GDC) tool from TCGA portal(Supplementary Table 1 and Supplementary Table 2). Every patients from TCGA database were complete the form of enrollment and we collected “days_to_death” as the time of follow-up, when the patient was deceased. Otherwise, the patients was alive, we collected “days_to_last_followup” as the time of follow-up. Immune scores and stromal scores were calculated by applying ESTIMATE algorithm of which R script was downloaded from the website (https://bioinformatics. mdanderson.org/estimate/rpackage.html). The inclusion criteria of the validation dataset (1) the sample size of OSCC patients must over 30; (2) the expression profiling was from human OSCC tissues; (3) the patients were primary OSCC; (4) patients with follow up information. 83 OSCC patients (GSE65858) employed as validation were obtained from Gene Expression Omnibus (GEO) database. The RNA expressions of OSCC were collected by using Illumina HumanHT-12 V4.0 expression bead chip. Clinical data of survival outcomes were also downloaded from GEO database in the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/geo/).

2.2 Screening for differentially expressed genes (DEGs)

In groups divided into by immune scores and stromal scores, the transcriptome data were transformed by log2 (x $+$ 1) for normalization and analyzed by R package (limma), the threshold value was according with both $|$ log ${}_{2}$ change (FC) $|$ $>$ 1 and false discovery rate (FDR) $<$ 0.05, screening for the differentially expressed genes. Volcano plot executed by Graph prism 8.0 software acted as the exhibition of the analysis results. The venn graphs were generated by online tool (https://bioinfogp.cnb.csic.es/tools/venny/).

Figure 1.

(A) Immune and stromal scores were associated with the clinical characteristics of OSCC patients. Scatter plots revealed the distribution of immune scores in sex ( $P=$ 0.0107), stromal scores in sex ( $P=$ 0.3850), immune scores in age ( $P=$ 0.1829), stromal scores in age ( $P=$ 0.9785) (B) immune scores in stage ( $P=$ 0.0436), stromal scores in stage ( $P=$ 0.6441), immune scores in grade ( $P=$ 0.0521), stromal scores in grade ( $P=$ 0.2363) (C) immune scores in primary site ( $P=$ 0.1497) Lower stromal scores were significantly associated with oropharynx cancer than other cases. (D) Results of K-M analysis in immune scores ( $P=$ 0.887) and stromal scores ( $P=$ 0.995) shown no statistical significance (K-M, Kaplan-Meier; tongue: Other and unspecified parts of tongue; oral cavity: Other and ill-defined sites in lip, oral cavity and pharynx; mouth: Other and unspecified parts of mouth).

Figure 2.

(A) Differentially expressed genes between the high immune scores-group and the low immune scores-group and (B) differentially expressed genes between the high stromal scores- group and the low stromal scores-group were exhibited in volcano plot. (C) Venn diagram showed the intersection of differentially expressed genes which were upregulated, (D) and which were downregulated.

Figure 3.

KEGG and GO analysis. (A) Pathway enrichment in KEGG analysis, (B) Results of CC, (C) MF and (D) BP in GO analysis revealed the relationship between differentially expressed genes and functional pathways.

2.3 GO analysis and KEGG

Online tool DAVID (https://david.ncifcrf. gov/) was exploited to construct the gene ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis. Bar charts were used to reveal the top 10 pathway annotation. The results of Cellular Component (CC), Molecular Function (MF), and Biological Process (BP) were also shown in bar charts performed by Graph prism 8.0 software.

2.4 Nomogram and ROC curves

Cox proportional hazards regression models were used for univariate analyses and to build a nomogram. The predictors analyzed included age, gender, TMN stage, tumor grade and the expression value of HGF. The predicted 1- and 5-year survival probabilities were calculated for patients with the nomogram. Receiver operating characteristic (ROC) curves were drawn to compare single gene (HGF) model with gene (HGF) combined with clinical characteristics model, and the areas under the curve (AUCs) at the 1- and 5-year survival probabilities were calculated to assess the discriminative power of the model. The risk scores (RS) were calculated for ROC curves by considering the expression value of HGF, age, gender, TMN stage and tumor grade with their corresponding coefficient in the formular of risk score $=\sum$ ( $\beta i\times$ Expi or Valuei) ( $i=$ the number of the impact factor).

2.5 Timer database analysis

The deconvolution algorithm provided by the TIMER (tumor immune estimation resource) database (https://ci strome.shinyapps.io/timer/) was used to explore the specific association between HGF and immune cells along with the association between tumor purity and HGF. The immune cells involved in the analysis include: B cell, CD4 $+$ T cell, CD8 $+$ T cell, macrophage, neutrophil, dendritic cell.

2.6 Statistical analysis

The R software (version 4.0.5) was used to execute all the statistical analyses and $P<$ 0.05 was statistically significant for survival analysis and correlation analyses. Unpaired $t$ test was used to explore the distribution of AJCC-TMN, pathological grade, sex in immune scores and stromal scores. Ordinary one-way Analysis of Variance (ANOVA) was used to explore the distribution of age and primary site in immune scores, and stromal scores. Kaplan-Meier (K-M) analysis based on the immune scores and stromal scores of OSCC patients was performed through survival package. Log-rank test and univariate Cox proportional hazards regression analyses were used to select prognostic genes. The forest plot was used to display the results of the analysis. R packages (survival and survivalROC) were applied to evaluate the prognostic value of screened gene.

3. Results

3.1 Immune scores and stromal scores are associated with OSCC subtypes

We divided stage into early stage (stage I and II) and advanced stage (stage III and IV),the immune scores represented the distribution of immune cells in the immune microenvironment, we found that the lower immune scores were observed in male patients ( $P=$ 0.0107, Fig. 1A) and advanced stages (Stage III and Stage IV, $P=$ 0.0436, Fig. 1B). Moreover, the stromal scores represented the distribution of stromal cells in the immune microenvironment and the stromal scores were different distributed among different primary tumor sites of oral ( $P=$ 0.0328, Fig. 1C). However, there was no significant difference in overall survival between the immune scores and the stromal scores (Fig. 1D).

Figure 4.

(A) 72 genes predicting the significant poor overall survival selected by univariate cox proportional hazards regression analyses (TCGA cohort). (B) Survival analysis of hub genes. The K-M curves revealed that the top 9 hub genes and HGF were significantly associated with survival outcomes (TCGA cohort). (C) HGF was proved to be associated with poor prognosis in GEO cohort.

Figure 5.

(A) prognostic nomogram for OS prediction. (B) ROC curve (single gene-based model and gene combined with clinical characteristics model) indicating 1-year and 5-year overall survival in the TCGA cohort. (C) ROC curve (single gene-based model and gene combined with clinical characteristics model) indicating 1-year and 5-year overall survival in the GEO cohort. ROC, receiver operator characteristic; AUC, area under the curve; OS, overall survival.

Figure 6.

(A–C) HGF was associated with immune cells infiltration. Immune cells included B cell, CD4 $+$ T cell, CD8 $+$ T cell, macrophage, neutrophils, and dendritic cell.

3.2 DEGs screening based on immune scores and stromal scores

OSCC patients were divided into low-level group ( $n=$ 165) and high-level group ( $n=$ 166) according to median immune scores and stromal scores respectively. A total of 10397 genes were collected in to explore the DEGs, the volcano maps showed the DEGs. Based on immune scores, 988 DEGs were screened out, including 890 up-regulated and 98 down-regulated (Fig. 2A, Supplementary Table 3). Similarly, for stromal scores, 1676 DEGs which contained 1657 upregulated genes and 19 downregulated genes (Fig. 2B, Supplementary Table 4). A total of 533 DEGs were identified in both immune scores group and stromal scores group, including 524 upregulated genes and 9 downregulated genes (Fig. 2C and D).

3.3 GO and KEGG analysis of DEGs

The KEGG pathway analysis of the 533 DEGs showed on Fig. 3A, cytokine-cytokine receptor interaction in the top 10 pathways. Gene ontology (GO) terms including cellular component, molecular function and biological process and pathway analyses were listed in Fig. 3B to D. The GO analysis revealed that DEGs might be linked to positive regulation of T cell proliferation, T-cell co-stimulation, MHC class II protein complex and MHC class II receptor activity.

3.4 Identification of prognostic gene based on TCGA database and validated on GEO database

A total of 72 DEGs were associated with overall survival which were identified by univariate cox regression analyses and the results were listed in Fig. 4A and Supplementary Table 5. We exhibited top 9 significant DEGs in log-rank test (AFF3, CCL22, CCR7, CD79A, COL29A1, CTSG, FAIM2, MS4A2, RSPO1) and HGF (Fig. 4B). HGF which was associated with overall survival in GEO database as well ( $P=$ 0.0114, Fig. 4C).

3.5 Predicting prognosis of OSCC by nomogram and ROC curve

We constructed a nomogram including age, gender, TMN stage, tumor grade and the expression value of HGF to predict OS at 1 and 5 years after diagnosis in the TCGA cohort (Fig. 5A). The area under the curve (AUC) of clinical characteristics combined HGF was 0.638 and 0.674 corresponding to 1-year and 5-year OS (Fig. 5B) in the TCGA cohort, basing on the risk score (RS) calculated as: RS $=$ 0.077*gender $+$ 0.894*stage $+$ 0.539*grade $+$ 0.306*HGF $+-$ 0.482*age. It was improved when only taken HGF into consideration (AUC of 1-year OS $=$ 0.568, AUC of 5-year OS $=$ 0.524, RS $=$ 0.259*HGF). In the GEO cohort, the AUC of 1-year OS (HGF and clinical characteristics model) were 0.714 (RS $=-$ 0.334*gender $+$ 0.913*stage $+$ 0.933*HGF $+$ 0.076*age), it was improved as well when only taken HGF into consideration (AUC of 1-year OS $=$ 0.601, RS $=$ 0.953*HGF Fig. 5C).

3.6 Key gene were associated with immune cell infiltration

We explored the relationship between HGF and six kinds of immune cells in OSCC patients by TIMER database. The results illuminated that HGF was significantly related with B cell, CD4 $+$ T cell, CD8 $+$ T cell, macrophage, neutrophils, and dendritic cell infiltration (Fig. 6A–C).

4. Discussion

The current study by integrating the public databases TCGA and GEO to evaluated the TME of OSCC, a total of 533 DEGs were identified in immune cells and stromal cells. Further functional enrichment analysis showed that those DEGs were involved in immune related pathways including the positive regulation of T cell proliferation pathway, the T-cell co-stimulation pathway, the MHC class II protein complex pathway and the MHC class II receptor activity related pathway. HGF was identified as a candidate gene closely related to the overall survival of OSCC by survival related analysis and the prognostic value of HGF was accessed by nomogram and ROC analysis. The TIMER database analyses also provide evidence for the strong correlation between HGF and immune cells infiltration, indicating the significant prognostic value of HGF in OSCC.

The TME is the cellular environment around tumor cells including endothelial cells, lymphocytes, macrophages, NK cells, other cells of the immune system, fibroblasts, mesenchymal stem cells (MSCs), and the extracellular matrix (ECM), contributing to the development, growth, and metastatic spread of malignancies [11]. The immune cell infiltrates in the environment are associated with better outcomes in early-stage cancer patients [12, 13]. The stromal cells build the growth and survival capacity of tumor cells during tumor initiation and progression by producing growth factors, chemokines and cytokines [14].

Hepatocyte growth factor (HGF), also called scatter factor, is a heterodimer protein that is normally synthesized by mesenchymal cells [15]. Combined with its sole physiological ligand: MET(c-MET, mesenchymal-epithelial transition factor) [16], HGF contributed to both the early and later stages of the cancer development by directly or indirectly stimulating proliferation, invasion, metastasis processes and angiogenesis [17, 18]. Studies have demonstrated that overexpression of HGF and/or MET is characteristic of several epithelial cell derived cancers and for some is an independent prognostic factor associated with aggressive nature of the tumor type and poor clinical outcome [19]. Wislez et al. reported that neutrophils may be attracted to the microenvironment generated by tumors such as bronchioloalveolar carcinoma, releasing bioactive HGF and promoting the invasion of cancer cells [20].

The four functional pathways including: positive regulation of T cell proliferation, T-cell co-stimulation, MHC class II protein complex and MHC class II receptor activity, which are important components of immune response to cancer [21]. A number of studies have shown that CD27/CD70, the members of the tumor necrosis factor receptor (TNFR) family, can enhance tumor or graft rejection, with co-stimulatory effects on T cells enhancing both NK-dependent and T cell-dependent mechanisms of tumor elimination [22]. The MHC class II protein complex transfer antigens derived by cancer cells to tumor-infiltrating CD4 T cells, stimulating direct cytotoxicity toward tumor cells [23]. The positive regulation of T cell proliferation in tumors are related to improved prognosis [24], for the cytotoxic activity of CD8 $+$ T cells alone [25], the CD8 $+$ T cells with the helper function of tumor-reactive CD4 $+$ T cells [26], even the anti-tumor properties of CD4 $+$ T cells in the absence of CD8 $+$ T cells [27].

Finally, there are still some limitations existing in our study. Firstly, the lack of the clinical samples may affect the validation of the results. Secondly, less than 1/3 patients provided information on treatment, the effect of treatment on prognosis could not be evaluated. Moreover, further functional experiments of HGF should be conducted to understand the potential mechanism in the process of the carcinogenesis of OSCC.

Author contribution

Interpretation or analysis of data: Longbiao Zhu and Xinyu Zhang.

Preparation of the manuscript: Donglin Yan and Yan Chen.

Revison for important intellectual content: Donglin Yan and Xinyu Zhang.

Supervision: Longbiao Zhu and Jing Han.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-210432.

sj-docx-1-cbm-10.3233_CBM-210432.docx - Supplemental material

Supplemental material, sj-docx-1-cbm-10.3233_CBM-210432.docx

Footnotes

Acknowledgments

This work was supported by National Natural Science Foundation of China (81602920, 81702686), The young talents program of Jiangsu Cancer Hospital.

References

Hayes

R.B.

Ahn

Fan

Peters

B.A.

Yang

Agalliu

Burk

R.D.

Ganly

Purdue

M.P.

Freedman

N.D.

Gapstur

S.M.

and Pei

, Association of oral microbiome with risk for incident head and neck squamous cell cancer, JAMA Oncol 4 (2018), 358–365.

Pai

Bamodu

O.A.

Lin

Y.K.

Lin

C.S.

Chu

P.Y.

Chien

M.H.

Wang

L.S.

Hsiao

Yeh

C.T.

and Tsai

J.T.

, CD47-SIRPalpha Signaling Induces Epithelial-Mesenchymal Transition and Cancer Stemness and Links to a Poor Prognosis in Patients with Oral Squamous Cell Carcinoma, Cells 8 (2019).

Harrington

K.J.

Ferris

R.L.

Blumenschein

, Jr. Colevas

A.D.

Fayette

Licitra

Kasper

Even

Vokes

E.E.

Worden

Saba

N.F.

Kiyota

Haddad

Tahara

Grunwald

Shaw

J.W.

Monga

Lynch

Taylor

DeRosa

Morrissey

Cocks

Gillison

M.L.

and Guigay

, Nivolumab versus standard, single-agent therapy of investigator’s choice in recurrent or metastatic squamous cell carcinoma of the head and neck (CheckMate 141): Health-related quality-of-life results from a randomised, phase 3 trial, Lancet Oncol 18 (2017), 1104–1115.

DeBerardinis

R.J.

, Tumor microenvironment, metabolism, and immunotherapy, N Engl J Med 382 (2020), 869–871.

Fane

and Weeraratna

A.T.

, How the ageing microenvironment influences tumour progression, Nat Rev Cancer 20 (2020), 89–106.

Jia

Xue

Yang

and Liu

, Mining TCGA database for genes of prognostic value in glioblastoma microenvironment, Aging (Albany NY) 10 (2018), 592–605.

Unterleuthner

Neuhold

Schwarz

Janker

Neuditschko

Nivarthi

Crncec

Kramer

Unger

Hengstschlager

Eferl

Moriggl

Sommergruber

Gerner

and Dolznig

, Cancer-associated fibroblast-derived WNT2 increases tumor angiogenesis in colon cancer, Angiogenesis 23 (2020), 159–177.

Sun

Shi

Yuan

Xia

Xiang

Quan

Shi

and Jiang

, Prognostic value of genes and immune infiltration in prostate tumor microenvironment, Front Oncol 10 (2020), 584055.

Yoshihara

Shahmoradgoli

Martinez

Vegesna

Kim

Torres-Garcia

Trevino

Shen

Laird

P.W.

Levine

D.A.

Carter

S.L.

Getz

Stemke-Hale

Mills

G.B.

and Verhaak

R.G.

, Inferring tumour purity and stromal and immune cell admixture from expression data, Nat Commun 4 (2013), 2612.

10.

Jager

M.J.

Brouwer

N.J.

and Esmaeli

, The cancer genome atlas project: An integrated molecular view of uveal melanoma, Ophthalmology 125 (2018), 1139–1142.

11.

Frediani

J.N.

and Fabbri

, Essential role of miRNAs in orchestrating the biology of the tumor microenvironment, Mol Cancer 15 (2016), 42.

12.

Nyamundanda

Fontana

and Sadanandam

, Is the tumour microenvironment a critical prognostic factor in early-stage colorectal cancer? Ann Oncol 30 (2019), 1538–1540.

13.

Fridman

W.H.

Pagès

Sautès-Fridman

and Galon

, The immune contexture in human tumours: impact on clinical outcome, Nat Rev Cancer 12 (2012), 298–306.

14.

Shi

Lin

and Wang

, Tumour-associated mesenchymal stem/stromal cells: Emerging therapeutic targets, Nat Rev Drug Discov 16 (2017), 35–52.

15.

Price

E.W.

Carnazza

K.E.

Carlin

S.D.

Cho

Edwards

K.J.

Sevak

K.K.

Glaser

J.M.

de Stanchina

Janjigian

Y.Y.

and Lewis

J.S.

, (89)Zr-DFO-AMG102 Immuno-PET to Determine Local Hepatocyte Growth Factor Protein Levels in Tumors for Enhanced Patient Selection, J Nucl Med 58 (2017), 1386–1394.

16.

Chen

and Zhu

, The HGF inhibitory peptide HGP-1 displays promising in vitro and in vivo efficacy for targeted cancer therapy, Oncotarget 6 (2015), 30088–30101.

17.

Comoglio

P.M.

and Boccaccio

, Scatter factors and invasive growth, Semin Cancer Biol 11 (2001), 153–165.

18.

Jiang

W.G.

Martin

T.A.

Parr

Davies

Matsumoto

and Nakamura

, Hepatocyte growth factor, its receptor, and their potential value in cancer therapies, Crit Rev Oncol Hematol 53 (2005), 35–69.

19.

Cecchi

Rabe

D.C.

and Bottaro

D.P.

, Targeting the HGF/Met signaling pathway in cancer therapy, Expert Opin Ther Targets 16 (2012), 553–572.

20.

Wislez

Rabbe

Marchal

Milleron

Crestani

Mayaud

Antoine

Soler

and Cadranel

, Hepatocyte growth factor production by neutrophils infiltrating bronchioloalveolar subtype pulmonary adenocarcinoma: Role in tumor progression and death, Cancer Res 63 (2003), 1405–1412.

21.

Haabeth

O.A.

Tveita

A.A.

Fauskanger

Schjesvold

Lorvik

K.B.

Hofgaard

P.O.

Omholt

Munthe

L.A.

Dembic

Corthay

and Bogen

, How Do CD4(+) T Cells Detect and Eliminate Tumor Cells That Either Lack or Express MHC Class II Molecules? Front Immunol 5 (2014), 174.

22.

Watts

T.H.

, TNF/TNFR family members in costimulation of T cell responses, Annu Rev Immunol 23 (2005), 23–68.

23.

Pisapia

Barba

Cortese

Cicatiello

Morelli

and Del Pozzo

, EBP1 protein modulates the expression of human MHC class II molecules in non-hematopoietic cancer cells, Int J Oncol 47 (2015), 481–489.

24.

Galon

Angell

H.K.

Bedognetti

and Marincola

F.M.

, The continuum of cancer immunosurveillance: Prognostic, predictive, and mechanistic signatures, Immunity 39 (2013), 11–26.

25.

Dudley

M.E.

Wunderlich

J.R.

Robbins

P.F.

Yang

J.C.

Hwu

Schwartzentruber

D.J.

Topalian

S.L.

Sherry

Restifo

N.P.

Hubicki

A.M.

Robinson

M.R.

Raffeld

Duray

Seipp

C.A.

Rogers-Freezer

Morton

K.E.

Mavroukakis

S.A.

White

D.E.

and Rosenberg

S.A.

, Cancer regression and autoimmunity in patients after clonal repopulation with antitumor lymphocytes, Science 298 (2002), 850–854.

26.

Boon

Coulie

P.G.

Van den Eynde

B.J.

and van der Bruggen

, Human T cell responses against melanoma, Annu Rev Immunol 24 (2006), 175–208.

27.

Greenberg

P.D.

Kern

D.E.

and Cheever

M.A.

, Therapy of disseminated murine leukemia with cyclophosphamide and immune Lyt-1+, 2- Tcells, Tumor eradication does not require participation of cytotoxic T cells, J Exp Med 161 (1985), 1122–1134.

Identification of prognostic genes in oral squamous cell carcinoma microenvironment

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION

Keywords

1. Introduction

2. Methods and materials

2.1 Data collection

2.2 Screening for differentially expressed genes (DEGs)

2.4 Nomogram and ROC curves

2.5 Timer database analysis

2.6 Statistical analysis

3. Results

3.1 Immune scores and stromal scores are associated with OSCC subtypes

3.3 GO and KEGG analysis of DEGs

3.4 Identification of prognostic gene based on TCGA database and validated on GEO database

3.5 Predicting prognosis of OSCC by nomogram and ROC curve

3.6 Key gene were associated with immune cell infiltration

4. Discussion

Author contribution

Supplementary data

sj-docx-1-cbm-10.3233_CBM-210432.docx - Supplemental material

Footnotes

Acknowledgments

References