| Background:Oral Squamous Cell Carcinoma(OSCC)is the most common oral and maxillofacial tumor and is highly aggressive and heterogeneous.Although the exact origin of OSCC remains unknown,there are many risk factors,including UV radiation exposure,alcohol and tobacco consumption,and HPV infection.In recent years,significant advances have been made in the development of novel therapeutic strategies for OSCC,including the use of immunotherapy,targeted therapies,and gene therapy.The prognosis of OSCC is correlated with clinicopathologic grading,and patients may have significantly different prognoses despite having similar clinicopathologic characteristics.Therefore,predicting the prognosis of OSCC patients must be"tailored"and individualized.Recently,a number of efforts have been focused on the development of prognostic models for oral squamous carcinoma,but no consensus has been reached.With the continuous development of bioinformatics,a variety of biological markers have been identified for the detection and study of biological processes,which have an important role in disease diagnosis,prediction and treatment in particular.Prognostic modeling is a method of analysis based on biological markers,clinicopathological features and other important information aimed at predicting patient prognosis,i.e.patient survival status,metastasis rate,etc.The construction of prognostic models can help doctors develop more accurate treatment plans and improve patient outcomes and survival rates.With the continuous development of biology and medical technology,prognostic modeling has become one of the important tools for personalized medicine and has a wide application prospect.Currently in oral squamous carcinoma,researchers have constructed various prognostic models by various biologic lineage markers,but still no consensus has been reached.Tumor-associated fibroblasts(cancer-associated fibroblasts,CAFs)are essential for tumor development.CAFs can remodel extracellular matrix formation,promote angiogenesis,facilitate tumor metastasis and participate in tumor immune escape,all of which accelerate tumor progression.Therefore,the study of TAFs is important for the creation of novel cancer therapies.CAFs have long been considered as an attractive therapeutic target.However,clinical trials using CAFs as therapeutic targets have not yet achieved satisfactory clinical results.Notably,the phenotype and function of CAFs are widely heterogeneous,and the pro-tumorigenicity of CAFs and their interactions with other subpopulations may change dynamically with tumor progression.The advent of single-cell RNA sequencing(sc-RNA seq)has greatly improved our understanding of the heterogeneity of CAFs.Single-cell transcriptome sequencing is a method of sequencing transcripts from individual cells,which provides a high-resolution view of gene expression in complex cell populations.Its ability to identify and characterize different cell subpopulations in tumor tissues is important for studying cell differentiation and development and identifying rare cell populations,but it does not provide information on the spatial location of individual cells while providing sequencing data.Space transcriptome sequencing,however,is a state-of-the-art sequencing technology that can analyze gene expression at independent loci in cancer pathology sections while preserving the spatial localization of the loci.The combination of single-cell transcriptome sequencing and spatial transcriptome sequencing allows for the resolution of high-resolution views of cell subpopulations while preserving spatial location information in the sections.The combination of the two sequencing approaches is an important tool for resolving heterogeneity in tumors and their tumor microenvironment,identifying rare cell subtypes,and discovering novel tumor markers.Spatiotemporal heterogeneity of cancer-associated fibroblasts(CAFs)is a novel feature of oral squamous cell carcinoma(OSCC).However,the mechanisms regulating the phenotypic transition in different subpopulations of CAFs remain elusive.In the present study,analysis of single-cell RNA-seq data using null-open publication showed that expression of platelet-derived growth factor receptorsα(PDGFRα)could functionally specify inflammatory CAFs(inflammatory CAFs,i CAFs)subpopulation in OSCC patients.In tumor specimens and patient-derived organoid(PDOs)models,PDGFRα~-staining-positive fibroblasts(i.e.,PDGFRα~+CAFs)and corresponding fibroblasts(i.e.,PDGFRα~-CAFs)exhibit different spatial locations and biological characteristics.Spontaneously,PDGFRα~+CAFs are endowed with higher proliferative potential and can secrete more oncogenic cytokines such as IL6 and IL8.however,PDGFRα~-CAFs show enhanced collagen matrix contraction but are more prone to senescence.after PDGFAA treatment,PDGFRα~+CAFs are phenotypically and functionally transformed into myofibroblasts(myofibroblasts,m CAFs),accumulateα-SMA expression,and enhance oncogenic properties in a PDOs model.Mechanistically,PDGFAA cooperates with TWIST1to transcriptionally regulate the expression of acta2,a gene encodingα-SMA,in a subpopulation of PDGFRα~+CAFs.In tumor regeneration experiments with co-implanted CAFs and OSCC cells,tumor growth dynamics as well as expression of a marker for myofibroblasts(i.e.,α-SMA)in transplanted tumors was reduced when TWIST1 was silenced in CAFs.In conclusion,we confirmed that the PDGF-TWIST1 signaling axis can regulate myofibroblast activation in i CAFs in OSCC and targeting this phenotypic shift may be beneficial for OSCC patients.Objective:1.The subgroups and specific surface markers of CAFs in oral squamous cell carcinoma were identified.The spatial distribution of different CAFs in tissues was defined based on pathological sections and spatial transcriptomic data.At the same time,PDGFRα~+CAFs and PDGFR~-CAFS were separated by immunomagnetic cell sorting technique according to the specific surface markers in i CAFs subpopulation,and the sorting effect was evaluated.Meanwhile,the functional heterogeneity of PDGFRα+CAFs and PDGFR~-CAFs was characterized by integrated fibroblast organoid model(FAO).2.To study the differences in biological behaviors of PDGFRα~+CAFs and PDGFR~-CAFs,including proliferative activity,collagen shrinkage ability,invasion characteristics,etc.3.The differentiation trajectory and pseudo timing of CAFs subpopulation were reconstructed to predict the timing of differentiation of CAFs cells.Specific binding receptors for i CAFs and other cell subsets in the tumor microenvironment were identified based on single-cell sequencing data.Final verificationRole of PDGFAA/PDGFRαin phenotypic transformation of PDGFRα~+CAFs into myofibroblasts.4.Find and identify the most potential transcription factors in i CAFs.To clarify its role in the phenotypic transformation of PDGFRα~+CAFs into myofibroblasts.5.To construct a novel prognostic model for predicting the prognosis of oral squamous carcinoma patients and guiding the clinical application of common chemotherapeutic agents.Methods:1.Identify subgroups of CAFs and their specific surface markers in oral squamous cancer tissues.Combined with case sections as well as spatial transcriptomic data,the spatial distribution of different CAFs in the tissues was defined.Flow cytometry,immunoblotting and immunofluorescence were also applied to identify the sorting efficiency and the expression of subpopulation markers of CAFs in the sorted cells.Finally,a patient-derived organoid(PDO)model was applied to characterize the functional heterogeneity of PDGFRα~+and PDGFRα~-CAFs.2.To further characterize the functional heterogeneity of PDGFRα~+and PDGFRα~-CAFs,Edu staining was applied to identify the proliferative potential of PDGFRα~+and PDGFRα~-CAFs;3D invasion assay was performed to identify the invasive ability of PDGFRα~+and PDGFRα~-CAFs under 3D culture conditions;to assess force-mediated collagen contraction,collagen contraction assay was performed with the same assess the contractility of PDGFRα~+and PDGFRα~-CAFs in type I collagen.Simultaneousβ-galactosidase staining was applied to label senescent cells in PDGFRα~+and PDGFRα~-CAFs;conditioned media isolated from PDGFRα~+and PDGFRα~-CAFs,as well as the corresponding PFs,were used for incubation assays in patient-derived organoids.Finally,human cytokine microarray C5 was used to detect differences in cytokines in the conditioned media of PDGFRα~+and PDGFRα~-CAFs.3.The Monocle 3 package was used to reconstruct the differentiation trajectory and proposed chronology of each subgroup of CAFs,and the results were visualized as UMAP plots for inferring the chronology of cellular differentiation of CAFs.Combined with cellchat and cellphone DB algorithms,cell-cell communication links in the tumor microenvironment were inferred for identifying specific binding ligand receptors for i CAFs and other cell subpopulations in the TME.Exogenous PDGFAA was also added to PDGFRα~+and PDGFRα~-CAFs to verify the phenotypic transformation by immunoblotting and collagen contraction assays.Finally,changes in integration morphology and markers were observed by addition of exogenous PDGFAA in an organoid model of fibroblast attachment.4.To investigate the mechanisms regulating the phenotypic transformation of PDGFRα~+CAFs,SCENIC analysis was performed to identify the most potential transcription factors in i CAFs.the GEPIA online dataset(GEPIA,http://gepia.cancer-pku.cn/)was used to validate the expression of transcription factors and markers of CAFs in head and neck squamous carcinoma samples expression correlations,where p<0.05 was defined as having a significant difference.The difference in TWIST1 expression in PDGFRα~+and PDGFRα~-CAFs was first assessed.The expression of TWIST1 and its upstream kinases(including AKT and ERK)in PDGFRα~+post-treated CAFs by PDGFAA was again assessed.Changes in the expression of each marker in CAFs were also clarified by down-regulating TWIST1 expression.The changes in collagen contractility in PDGFRα~+and PDGFRα~-CAFs after the knockdown of TWIST1 expression were assessed by collagen contraction assay.To identify the effect of TWIST1 on cell proliferation,Edu staining was applied to identify the proliferative potential of cells;a 3D invasion assay was performed to identify the changes in the invasion ability of PDGFRα~+CAFs under 3D culture conditions after knockdown of TWIST1 expression.Changes in integration morphology and markers were also observed by integrating CAFs from control and TWIST1knockdown groups in an organoid model of fibroblast attachment.Finally,to obtain more direct evidence supporting the role of TWIST1 in regulatingα-SMA expression,a luciferase reporter assay was performed using p GL3.CAFs transfected with TWIST1 mixed with oral squamous carcinoma cells were implanted into nude mice for assessing their effect on tumor growth.5.Pearson’s correlation coefficient was used to calculate the correlation between Lnc RNA and EMT-related genes."LASSO regression was performed to construct prognostic models.Based on the median risk score,the TCGA samples with oral squamous carcinoma were divided into high risk and low risk groups.Results:1.Single-cell sequencing results identified three subgroups of CAFs,of which PDGFRA was a specific surface marker for i CAFs.Immunomagnetic bead sorting was applied to sort the CAFs into PDGFRα~-CAFs and PDGFRα~+CAFs.immunoblotting,immunofluorescence and flow cytometry identified stable sorting results.In oral squamous carcinoma tissue samples,m CAFs were confirmed to be close to the tumor and i CAFs were far from the tumor,while spatial transcriptome sequencing results verified the spatial location relationship of CAFs subpopulations.More importantly,morphological analysis showed that the admixture of PDGFRα~-CAFs triggered the formation of invasive frontal structures in the organoid compared to precursor cells such as paraneoplastic fibroblasts(PFs).Otherwise,the admixture of PDGFRα~+CAFs could lead to an alternative growth pattern of tumor-like organs with colonies dominated by outgrowth structures compared to PFs.2.PDGFRα~-CAFs and PDGFRα~+CAFs have different biological behaviors.edu staining results confirmed an increased frequency of proliferating cells in the PDGFRα~+CAFs subpopulation compared to the PDGFRα~-CAFs population.In combination with cell cycle analysis in single-cell sequencing,the G1 phase was significantly higher in i CAFs than in other CAFs subpopulations,suggesting an active material metabolism.3D invasion assays confirmed that PDGFRα~+CAFs were more efficient than PDGFRα-CAFs in invasion.Meanwhile,PDGFRα-CAFs showed stronger collagen contraction than PDGFRα~+CAFs,andβ-galactosidase staining showed that PDGFRα~-CAFs were a relatively senescent subtype of CAFs.Conditioned media isolated from PDGFRα~+and PDGFRα~-CAFs,as well as the corresponding PFs,were used in organoid incubation assays.The results clearly showed that PDGFRα~+CAFs-derived conditioned media exhibited the most efficient efficiency in promoting tumor-like organ colony size compared to PFs or PDGFRα~-CAFs.To further validate the inflammatory properties of PDGFRα~+and PDGFRα~-CAFs,we collected conditioned media of PDGFRα~+and PDGFRα~-CAFs for cytokine microarray assays.The results confirmed that several oncogenic cytokines,including IL6,CXCL1,HGF,VEGFA,and IL8,were significantly enriched in PDGFRα~+CAFs.Overall,these results suggest that PDGFRα~+CAFs have an enhanced inflammatory cell phenotype.3.Cell Chat package to display the cell-cell interaction strength and the number of interactions and to identify the most important signaling pathways.Among them,i CAFs are the major efferent cell subpopulation of the collagen signaling pathway,cxcl signaling pathway,TGF signaling pathway,and wnt signaling pathway;meanwhile,i CAFs are the most important afferent cell subpopulation of PDGF signaling pathway.cellphone DB 2 results again suggest that PDGFAA-PDGFA interactions in the tumor microenvironment are highly confined to a subset of i CAFs.Also the results of the proposed time series analysis confirmed that PDGFRαappeared only at the early stage of differentiation of CAFs.PDGFRα~+and PDGFRα~-CAFs were further investigated after exogenous PDGFAA treatment.The results of immunoblotting assays showed that PDGFRα~+CAFs showed increasedα-SMA expression after PDGFAA treatment.In addition,PDGFRα~-CAFs did not show any changes under similar conditions.Consistent with this,the results of the collagen contraction assay showed that PDGFAA treatment promoted the collagen contractility of PDGFRα~+CAFs;whereas PDGFRα~-CAFs showed no significant change in collagen contractility after PDGFAA treatment.In the organoid model of fibroblast attachment,PDGFAA treatment promoted the formation of invasive frontal structures in the organoid and increased the expression ofα-SMA in the organoid.Taken together,it is shown that PDGFAA treatment promotes myofibroblast differentiation of PDGFRα~+CAFs.Application of PDGF inhibitor-Avapritinib treatment resulted in a dramatic degradation of infiltrative frontal structures in the organoid,suggesting that maintenance of PDGF signaling contributes to myofibroblast activation in OSCC.Also in the organoid model integrating PDGFRα~+CAFs,Avapritinib treatment inhibited the formation of outgrowth colony structures and suppressed the expression of CD44 in the organoid.Finally,the application of Avapritinib reduced the weight and volume of xenografts.At the same time,CD44~+and ki-67~+cells were significantly downregulated.4.SCENIC transcription factor analysis identified TWIST1 as a transcription factor specifically activated in i CAFs,and NR2F2 and MEF2D motifs were highly activated in m CAFs.Correlation analysis confirmed that the expression of TWIST1 positively correlated with the expression of several CAFs markers,including PDGFRA,PDGFA,ACTA2,FAP,and VIM.immunoblotting assays confirmed that the expression of TWIST1 and the phosphorylation levels of Akt and ERK were elevated in PDGFRα+CAFs compared to PDGFRα-CAFs(Figure4C).Also,we found that PDGFAA treatment promoted the expression ofα-SMA and the phosphorylation levels of Akt and ERK in PDGFRα~+CAFs,especially in a time-dependent and dose-dependent manner.In addition,PDGFAA treatment did not affect TWIST1 expression in PDGFRα~-CAFs.Furthermore,knockdown of TWIST1 reducedα-SMA expression in CAFs after PDGFAA treatment.Consistent with this,the results of collagen contraction assays showed that knockdown of Twist1 reduced collagen contractility in PDGFRα~+CAFs without affecting the collagen contractility of PDGFRα~+CAFs.EDU staining results confirmed that the EDU~+rate was lower in twist1 knockdown cells in PDGFRα~+CAFs compared with controls.In addition,we analyzed the 3D invasion properties of PDGFRα~+CAFs,and the control group had higher invasion efficiency than the sh-TWIST1 group.In the organoid model of fibroblast attachment,downregulation of Twist1 reduced the expression ofα-SMA and CD44 in organoids,suggesting that Twist1 plays an important role in the co-culture system of CAFs and organoids.The luciferase reporter assay confirmed that the transcriptional activity ofαSMA promoter was significantly reduced after sh-Twist1 vector introduction into CAFs.In conclusion,their results suggest that TWIST is essential for activation of PDGFRα+CAFs by myofibroblasts.The xenograft model showed that downregulation of TWIST in CAFs significantly retarded the growth dynamics of co-implanted xenograft xenografts.More interestingly,the frequency of the typical m CAFs markerα-SMA was dramatically reduced in xenograft cells when silencing TWIST in CAFs.At the same time CD44~+and ki-67~+cells were significantly downregulated in the sh-Twist1 group.5.The prognostic model we constructed is an independent predictor of prognosis for oral squamous carcinoma patients and can be used to predict the prognostic status of patients.Also patients with high risk score were more sensitive to paclitaxel and docetaxel;patients with low risk score were more sensitive to methotrexate and cisplatin.Finally,the low-risk group was more likely to benefit from immunotherapy.Conclusion:We provide evidence that PDGFRαexpression can functionally specify a subpopulation of CAFs with inflammatory cell properties,enhanced proliferative capacity,and preserved myofibroblast differentiation potential,and that targeting the PDGFAA-TWIST1 axis in CAFs regulates the transformation of i CAFs to myofibroblasts in OSCC,and that targeting this phenotypic shift may benefit oral squamous cancer patients.Also,the novel prognostic model constructed can be used to predict patient prognosis and the use of common chemotherapeutic agents. |
PDF Full Text Request