| Objective: In the past 10 years,immunotherapy seems to be a new hope to conquer head and neck tumors.At present,the formulation of immunotherapy plan and the prediction of curative effect for head and neck squamous cell carcinoma mainly rely on the biomarkers of immunotherapeutic effect recommended by FDA,such as PD-L1 and mismatch repair dysfunction(DMMR)or microsatellite instability(MSI-H).However,the accuracy of these single immunotherapeutic effect or prediction biomarkers in predicting the curative effect of patients still needs to be further improved.Compared with single immune efficacy index,researchers prefer to describe the characteristics of immune microenvironment of tumor patients through more dimensional information,and then predict the immune efficacy and prognosis of patients.The purpose of this study is to screen important biomarkers that affect the prognosis of patients with head and neck squamous cell carcinoma by bioinformatics analysis technology,and build a new algorithm to establish a prediction model that can identify immune subtypes of patients with head and neck squamous cell carcinoma,to screen and describe the biological characteristics of immunotherapy beneficiaries more accurately.Methods: Part I Screening of prognostic signaling pathways and construction of immune subtype prediction model for head and neck squamous cell carcinoma based on multi omics 1.In this study,the important signal pathway and biological markers that affect the prognosis of head and neck squamous cell carcinoma were screened by multi-layer histochemical prognosis analysis.The research methods are as follows: 1)The data of HNSCC multi-layer histochemistry(including transcriptional data,gene mutation data,gene copy number change data and gene methylation data)and clinical prognosis data were downloaded.The transcriptional data were TCGA level3 counts,The formula is standardized to TPM data,and the proportion of expression 0 is more than 20%,or the gene with the median value of 0 is deleted,and the gene type is annotated according to Vega standard;the methylation data is annotated with VEP software;the copy number change data is evaluated by gistic2 software according to CNV.All data are normalized and standardized.2)The signal pathway and biomarkers that affect the prognosis of head and neck squamous cell carcinoma were screened by MOSClip multi-omics histography.3)The biological function of cells was annotated by the Reactome database.4)The clinicopathological parameters,survival status and total survival period of the patients were obtained by retrospective collection of head and neck squamous cell carcinoma patients.5)The expression of IL2 RG in the tumor tissues of head and neck squamous cell carcinoma was verified by immunohistochemistry.6)The level of LCK methylation in tumor tissue of head and neck squamous cell carcinoma was verified by methylation specific PCR.2.The prediction model of identifying immune subtypes was constructed by using interleukin-2(IL-2)related genes 1)The immune subtype related genes were screened by 1000 lasso logistic analysis.2)Univariate and multivariate Cox regression models were used to screen prognostic genes of head and neck squamous cell carcinoma.3)The unsupervised cluster analysis adopts the consistent cluster analysis,which is implemented by R package Consensus Cluster Plus.4)Single sample gene enrichment analysis(ssgsea)was used to evaluate tumor related signaling pathways and tumor immune cell infiltration abundance.5)In order to verify the relationship between gene expression in tumor and microenvironment,the purity of tumor was evaluated by estimate algorithm.6)Non negative matrix factorization and bootstrapping algorithm were used to construct IRIM immune subtype prediction model.Part II Biological significance of IRIM model in tumor immunotyping of head and neck squamous cell carcinoma Objective to evaluate the immune microenvironment of head and neck squamous cell carcinoma by IRIM algorithm,and explore the relationship between IRIM evaluation of inflammatory characteristics and different immune cell function status.1)In this study,single cell sequencing data gse103322 of head and neck squamous cell carcinoma was used.The "double cell" and low-quality cells were deleted by cell quality evaluation.2)The normalization data function in Seurat package was used to standardize the gene expression,and the inter batch difference of data was corrected by harmony package.3)The first 25 principal components are selected for t SNE and Umap dimension reduction analysis.By DBClustdimension function,the subpopulations with less than 30 cluster cells were deleted.4)There are two ways to annotate cell types:(1)annotate by Single R package;(2)annotate by biomarkers of various types of genes.5)The "Findvariablegenes" function in Seurat package was used to calculate the differential genes among different cell subpopulations.6)The biological functions of cell subsets were analyzed by single cell function enrichment analysis(ssgsea).7)R software was used for statistical analysis.The quantitative data of normal distribution was expressed as mean ± standard deviation.T test was used for comparison between groups.The quantitative data of non-normal distribution was expressed as median and quarterback distance.Wilcoxon test was used for comparison between groups.The mean value of quantitative variables in accordance with normal distribution is selected as cut-off value,and the median value of quantitative variables in accordance with normal distribution is selected as cut-off value.Kaplan Meier method was used to generate survival curves of each subgroup in each data set.Log rank test was used to determine the statistical significance of the differences.Landmark analysis was used for the data with intersecting survival curves.P < 0.05,the difference was statistically significant.Results: 1.The prognostic factors of HNSCC were screened by multi-layer omics data,including TCGA transcriptome data,gene mutation data,methylation data and copy number change data.The results showed that there were 45 pathways associated with the prognosis of head and neck squamous cell carcinoma.These 45 signaling pathways play an important role in human immune function in head and neck squamous cell carcinoma.2.Prognosis related pathways and gene modules in prognosis related pathways were screened according to p value.The results showed that: the signal pathway of interlinukin-2 family and its gene coexpression module 1 were most closely related to the prognosis of head and neck squamous cell carcinoma.3.The Kaplan Meier curve was used to evaluate the signal pathway and its molecular module 1.The results showed that low gene expression and high methylation level in this signaling pathway were protective factors for the prognosis of head and neck squamous cell carcinoma.4.By cluster analysis,we found that the genes related to the interlinukin-2 family signaling pathway could divide the head and neck squamous cell carcinoma patients into two subgroups.5.Wilcoxon test showed that there were significant differences in T cell content,B cell content,WNT pathway score,IFNG pathway score and PD-L1 expression between the two subgroups.Based on previous studies,the two subgroups showed the characteristics of immune inflammatory type and non-immune inflammatory type respectively.6.In order to verify the role of Interlukin-2 family signaling related genes in immunophenotyping,we tested them in imvigor210 data set.The results showed that the gene expression levels of IL21 R,IL2RG,LCK,INPP5 D,JAK3 and PIK3R3 could indicate that the patients were of immune inflammatory type and were related to the efficacy of immunotherapy.7.In order to better describe the relationship between the above six key genes and immune subtypes,we designed IRIM algorithm and constructed the immune subtype prediction model of head and neck squamous cell carcinoma(IRIM model).The results showed that: in the head and neck squamous cell carcinoma,if three or more of the above six key genes play a role,the head and neck squamous cell carcinoma patients can be identified as "immune inflammatory type",AUC was 0.8146,the accuracy was 74.3%.8.Other immune related data sets(GSE35640,GSE78220,GSE63557 and GSE91061)were used to verify the results.The results show that IRIM model has the function of indicating immune efficacy in other data sets.9.The six key genes were found to be expressed in tumor microenvironment by estimate and correlation analysis,so we verified them in different cancer species.The results show that irim model can identify "immune inflammatory" patients in lung squamous cell carcinoma,breast cancer,cervical squamous cell carcinoma,esophageal squamous cell carcinoma and other solid tumors.10.In order to verify the relationship between these six key genes and the prognosis of head and neck squamous cell carcinoma patients,we screened the representative genes IL2 RG and LCK by principal component analysis.11.GEPIA database analysis showed that IL2 RG was a protective factor for the prognosis of head and neck squamous cell carcinoma(P = 0.015).12.Methods: 79 cases of head and neck squamous cell carcinoma were collected retrospectively.The overall survival status and overall survival time of the patients were followed up.The results of immunohistochemical staining of IL2 RG showed that IL2 RG was a protective factor for the prognosis of head and neck squamous cell carcinoma.The patients with positive results of IL2 RG immunohistochemistry had a good prognosis(P = 0.0017).MSP test was used to evaluate the methylation status of LCK in patients with head and neck squamous cell carcinoma in our center,and the relationship between the methylation status of LCK and the overall survival time of patients was analyzed by combining with the results of IL2 RG immunohistochemistry.The results showed that the long-term survival time of patients with LCK unmethylation status combined with IL2 RG positive was significantly better than that of patients with LCK methylation status combined with IL2 RG negative,and the difference was statistically significant(P = 0.0075)13.The IRIM model score was closely related to T cells in head and neck squamous cell carcinoma.14.Through T cell subgroup analysis,IRIM model score can mainly indicate the infiltrative T cell aggregation in tumor microenvironment of patients with head and neck squamous cell carcinoma.15.The poor prognosis of interleukin-2 family signaling pathway is mainly related to CD8 + T cell depletion.Conclusion: 1.The signal pathway of Interlukin-2 family is a risk factor of head and neck squamous cell carcinoma.The level of IL2 RG transcriptome and the level of LCK methylation are involved in the influence of Interlukin-2 family signaling on the prognosis of head and neck squamous cell carcinoma.High expression of IL2 RG is a protective factor in patients with head and neck squamous cell carcinoma,and hypermethylation of LCK is a risk factor in patients with head and neck squamous cell carcinoma.2.IRIM model is an important prediction model for the immunophenotyping of head and neck squamous cell carcinoma.It can screen the patients with inflammatory subtype of head and neck squamous cell carcinoma and effectively predict the therapeutic effect of immune checkpoint inhibitors. |
PDF Full Text Request