Single-cell Transcriptomic Analysis Of Tumor Microenvironment In Hepatocellular Carcinoma With Portal Vein Tumor Thrombus

BackgroundHepatocellular carcinoma(HCC)is the most common type of primary liver cancer.Due to the biological characteristics of HCC and the anatomical traits of the liver,it is easy to develop portal vein tumor thrombus(PVTT),resulting in poor prognosis.Although a treatment system has been established in recent years with surgery as the main pillar,the overall treatment effect still needs to be improved.Therefore,in-depth study of the mechanisms related to the progression of HCC,the search for more effective therapeutic targets and efficacy predictive markers,and the development of individualized therapeutic measures remain the focuses of current clinical research.HCC heterogeneity can be manifested between patients,between different lesions in the same patient and within the same lesion,and tumor heterogeneity is one of the main reasons that affects treatment response.Previous bulk sequencing methods from whole samples yielded results that averaged the gene expression of all cells within the sample,masking heterogeneity between different cell subpopulations.The newly emerged singlecell sequencing technology enables us to distinguish cell types in samples,discover new cell subpopulations,reveal different biological functions of subsets,explore changes in cell states and intercellular interactions at the single cell level,further reveal intra-tumor heterogeneity and tumor microenvironment trait from a higher resolution,and dissect the molecular characteristics inside tumors to guide precise tumor treatment.ObjectiveThe aim of this study was to evaluate the heterogeneity in HCC patients with PVTT,the similarities and differences of the tumor microenvironment between the tumor and PVTT,and explore common and unique potential therapeutic targets.Methods and participantsSingle-cell transcriptome sequencing was performed on tumors and paired PVTT samples from five HCC patients with PVTT.Tumor cells,stromal cells and immune cells were sorted into different populations to analyze the number changes,gene expression differences and cellular communication network of each subpopulation in different sample sources.Quantitative real-time polymerase chain reaction,Western blotting,construction of cell lines stably expressing target gene,cell co-culture,subcutaneous tumor-bearing experiments in nude mice,immunohistochemical staining,multiplex immunofluorescence staining were used to study target genes and related cell subsets.ResultsWe obtained a total of 35,177 single cell transcriptome datasets to construct a panoramic cellular atlas of this patient population.Tumor cells were divided into four subgroups with different drug target expression,distinctive metabolic activation patterns and special transcription factor regulation patterns.Likewise,different stromal cell subsets had different metabolic activation patterns.Through clonal evolution analysis of cell subsets,we found that tumor-associated fibroblast,vein endothelial cell and neutrophil have different differentiation directions.By performing a multigene prognostic analysis using cell subpopulation marker genes,we found that the complement-related subpopulation of neutrophils may be associated with excellent prognosis.There was a significant increase of tumor-associated fibroblasts in PVTT,and the inflammatory-associated subpopulation shifted to the dominant subpopulation.After comparing the tumor microenvironments of tumors and PVTT,we found that the numbers of some key cell subsets,gene expression patterns,and metabolic activation traits changed in different tumor microenvironments.The tumor cells were dominated by highly aggressive and indolent subsets,but the number of highly aggressive subsets increased in PVTT.Gene expression associated with lenvatinib targets was elevated in some tumor subpopulations in PVTT.The expression of proangiogenic genes was elevated in most subpopulations of stromal cells in PVTT.Among endothelial cells,we found a class of complement-associated venous endothelial cells that occurs almost exclusively within the tumor.This subpopulation was the highest BMP4-expressing subpopulation of all endothelial cells,and recombinant BMP4 could enhance leukocyte transendothelial migration.The reduction in the number of this subpopulation in PVTT may be related to the decrease of neutrophils in this environment.In terms of immune cells,the neutrophil subpopulations had lower infiltration in PVTT,but some of the subpopulations upregulated proinflammatory factors,and the inflammatory state of the microenvironment was not correspondingly reduced.The overall number of T cells was increased in PVTT.Similar to NK cell,most cytotoxicity-related subpopulations were decreased in number or upregulated immunosuppression-related genes,reducing their antitumor effects.In conclusion,compared with the tumor,the immunosuppressive state in PVTT was aggravated.By analyzing the activation of metabolism-related gene sets,we observed that the metabolic status of the most immune cells was unchanged in the tumor and in PVTT.In the most tumor cells and stromal cell subpopulations,gene sets related to glycolysis,lipid uptake or lipid synthesis were significantly activated in PVTT,and the alterations in the metabolic processes and metabolites of the above cells may be involved in remodeling the tumor microenvironment and exacerbating immunosuppression.By using Cell Phone DB database for ligand-receptor analysis between different cell populations,we identified several potential positive feedback loops involving tumor cell,stromal cell,and immune cell in the tumor and PVTT environments,driving the progression of different tumor subpopulations in different tumor microenvironments.Finally,we selected CFHR3,one of the inert tumor subgroup marker genes,for functional and mechanistic studies.High expression of this gene was found to be associated with good prognosis in patients with HCC using public database.By analyzing clinical samples,it was found that this gene was expressed higher in adjacent peritumoral tissues than in tumor.In vivo and in vitro experiments we confirmed that overexpression of this gene inhibited tumor cell growth,migration,invasion,epithelial-mesenchymal transition and tumor angiogenesis.By investigating the mechanism,we found that overexpression of this gene suppressed epithelial-mesenchymal transition of tumor cell by inhibiting p38 phosphorylation.ConclusionsUsing single-cell transcriptome sequencing technology,we depicted a panoramic cellular atlas of patients with HCC and PVTT,revealing tumor heterogeneity in tumor tissue and PVTT,as well as changes in the tumor microenvironment.Increased immunosuppressive status in PVTT compared to tumors.By studying the Marker gene of specific tumor subgroups,we found that CFHR3 plays a tumor suppressor role in HCC.Overexpression of this gene can inhibit HCC cells growth,migration,invasion,epithelialmesenchymal transition and tumor angiogenesis.This study provided some potential targets for the treatment of HCC.