Construction Of Spatial Transcriptomics Atlas And Study On Tumor Microenvironment Heterogeneity In Cervical Cancer

Research background and purpose:Cervical cancer is one of the most common malignant tumors in women.Cervical adenocarcinoma and squamous cell carcinoma are the two most common histological types.They differ in clinical manifestations,treatment response,transcriptome,and tumor microenvironment.The differences in tumor microenvironment will affect tumor treatment response and prognosis,and can be used to predict the patients’ response to immunotherapy.It is necessary to comprehend the complex nature of the microenvironment of cervical cancer in view of the heterogeneity of its response to immunotherapy.Spatial transcriptomics is currently a new technology in the field of single cell sequencing for the study of tumor microenvironment.But there are few research papers focusing on the differences of the microenvironment of various histological types of cervical cancer.Therefore,the purpose of this study is to apply the spatial transcriptomic techniques to create the geographic map of cervical cancer,to investigate the spatial characteristics of microenvironment,and to explore the intra-and intertumoral heterogeneity of tumor microenvironment in cervical squamous cell carcinoma and adenocarcinoma.Methods:The tumor samples from one case of cervical adenocarcinoma and one case of squamous cell carcinoma were obtained.The fresh frozen OCT embedded tissue samples were sliced as frozen sections at 10μm thickness.The sections were then tested for RNA quality and went through the process of methanol fixation,H&E staining and brightfield imaging.Two pathologists reviewed the images and confirmed that the sections included areas of invasive tumor and stromal areas.Permeabilization,reverse transcription,library construction and sequencing were then carried out in order.The bioinformatics analysis.methods include unsupervised dimensionality reduction,clustering or classification,manual annotations based on literature and single-cell database,gene function enrichment analysis,pseudotime analysis and cell-cell communication analysis(CellPhoneDB for ligand-receptor analysis).we also validated the results with external data from TCGA-CESC.Results:The study included one case of cervical adenocarcinoma,Silva pattern C(named as sample A)and one case of poor-differentiated squamous cell carcinoma(named as sample S).The tumor invasion areas are present,and the clear demarcations between the tumoral and stromal areas can be observed in both samples.Based on the results of unsupervised clustering and manual annotation,sample A was divided into tumoral area,immune infiltration area and stromal area;sample S was divided into tumoral area,immune infiltration area,stromal area and normal epithelial area.The number of genes detected and UMI in the tumoral area were significantly higher than those in the stromal area,indicating the presence of active transcription in tumor cells.We identified several cell types of cervical cancer,including adenocarcinoma cell type Ⅰ,adenocarcinoma cell type Ⅱ,malignant squamous cell,normal epithelial cell,smooth muscle cell,cancer-associated fibroblast,fibroblast,endothelial cell,plasma cell,dendritic cell,monocyte macrophage cell,T cell,NK cell and mast cell.There are few tumor infiltrating lymphocytes including T cells in the tumoral areas of both samples,suggesting the presence of immunosuppression.Besides,in sample S,the normal epithelial cells differentiate into tumor cells,and the expression of key node genes MUC5B,PTGDS,S100A8,MT2A and LGALS1 shows a downward trend during the differentiation.In sample A,the two subtypes of adenocarcinoma cells may have a sequential relationship in cell differentiation,that is,the clump-distributed cell type Ⅰ would evolve into the dispersed cell type II during the occurrence and development of the disease.The key node genes were ACTG1,APOE;FTH1,GNAS,IF127,NTS,etc.There are cell-cell interactions between tumor cells and other cell types in the microenvironment,especially stromal cells.FAM3C/HLA-C,APP/CD74,COPA/CD74,MIF/CD74 are the most significant ligand-receptor pairs in both histological types.The external data validation shows that the cancer-associated fibroblasts(CAF)with ACTA2,COL5A1,MMP2,SERPINE1 as markers have a prognostic value in overall survival.Patients with high expression levels of CAF markers have a significantly worse prognosis(HR=1.90,P=0.0074).Conclusions:In this study,we constructed the spatial transcriptome maps of cervical squamous cell carcinoma and adenocarcinoma.We found that the two histological types were significantly different in components of tumor microenvironment and the proportions of various cell types.The intratumoral heterogeneity of cervical adenocarcinoma is greater compared with squamous lesion,which may contribute to its unique clinical characteristics and worse treatment response.The infiltration of CAF with ACTA2,COL5A1,MMP2 and SERPINE1 as cell markers is associated with poor prognosis.CAF has significant interactions with tumor cells through FAM3C/HLA-C,APP/CD74,COPA/CD74 and MIF/CD74,which may lead to immunosuppression in cervical cancer.This study will help us better understand the pathological mechanisms and developmental process of cervical cancer,and provide possible novel targets for immunotherapy.