Single-Cell Transcriptomics of Glioblastoma Reveals Pericytes Contributing to Blood–Brain Tumor Barrier and Tumor Progression

doi:10.21203/rs.3.rs-3561475/v1

Download PDF

Research Article

Single-Cell Transcriptomics of Glioblastoma Reveals Pericytes Contributing to Blood–Brain Tumor Barrier and Tumor Progression

https://doi.org/10.21203/rs.3.rs-3561475/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background:Pericytes compose blood–brain barrier (BBB) and may affects the blood-brain tumor barrier function (BBTB) in glioblastoma (GBM), which eventually affect chemotherapy efficiency and tumor progression of the disease. However, the expression signatures and detailed effect of these pericyte are still unclear. We aimed to identify these pericytes affecting BBTB in GBM, and to explore their clinical effect and underlying functions.

Methods: Pericytes in GBM were identified from single-cell RNA sequencing (scRNA-seq) da ta from nine GBM samples by hallmarks and expression signatures of pericytes from previous researches, and cell cluster functional enrichments were also used in identification. CellPhoneDB were used to explore interactions between target pericyte and other cells. One in-house glioma clinical cohort was collected and included for survival analysis. CGGA, REMBRANDT, and GSE16011 GBM cohorts were used to develop and validate a tumor pericyte risk score (TPRS) for prognosis prediction.

Results: This study demonstrated that a group of pericytes concerning BBTB function exist in GBM, and PTH1R may be a specific biomarker for them. The pericytes of BBTB function interact with other cell in GBM mainly through extracellular matrix (ECM)-integrin signaling pathways. Also, when comparing with normal pericytes, pericytes in GBM has up-regulation of several ECM gene networks (THY1, COL3A1, COL4A1, TIMP1, FN1 etc.) related to poor prognosis and basement membrane formation. TPRS based on those differentially expressed genes (DEGs) has significant predictive value in GBM patients.

Conclusions:Pericytes of BBTB function in GBM exist and PTH1R potentially serving as a hallmark for them. These cells have close functional relationships with BBTB and ECM-integrin signing pathways, and may have predictive value for GBM patients.

Glioblastoma

Pericyte

Blood–brain tumor barrier

Prognosis

Glioblastoma (GBM) is the most common primary malignant brain tumor [1]. Typical treatment of GBM consists of extensive surgical resection, chemotherapy with temozolomide and adjuvant radiotherapy. Yet, despite this, clinical effectiveness is restricted and the relative survival rate at one year is 42.5%[2]. The pronounced drug resistance of the tumor during postoperative therapy is reflected in the poor prognosis and high recurrence rate of GBM [3]. The poor prognosis of GBM is attributed to the lack of effective chemotherapeutic agents and the unsuccessful drug delivery across the blood-brain barrier (BBB)[4].

The BBB is a cellular barrier comprising specialized brain endothelial cells (ECs), pericytes, astrocytes, and perivascular macrophages [5]. The BBB primarily regulates the passage of substances from the blood to the brain to maintain optimal neuronal activity. However, it restricts most drug compounds, except for highly lipid-soluble small molecules with less than 400 daltons. As a result, many anti-cancer drugs are ineffective for treating GBM[6]. Overwhelming clinical evidence from studies using MR and positron emission tomography (PET) imaging show significant tumor burden with functional BBB in GBM tissue: a considerable proportion of the PET-defined tumor volume and MRS-defined metabolically detectable tumors extend beyond the contrast-enhancing lesion in most GBM cases [7, 8]. Studies on brain metastases have shown that highly effective targeted therapies that are BBB impenetrable have "brain-only failures". Failures of targeted therapies of brain metastases suggests that poor drug delivery into GBM tissues with an intact BBB may limit efficacy of chemotherapy in these regions[9]. Abnormal vessel formation and the blood-brain tumor barrier (BBTB) occur as a result of VEGF expression and angiogenesis, stimulated by excessive metabolic demands of GBM cells [10, 11]. The BBTB hinders the delivery of sufficient therapeutic drugs into GBM tissue due to aberrant expression of transcytotic proteins in tumor vessels [12]. These findings indicate that overcoming the BBTB while preserving the normal BBB may be feasible in treating GBM.

Pericytes are important component of the BBB, defined by their anatomical location surrounding the walls of blood vessels[13–15]. They communicate with endothelial cells (ECs) through physical contact via their long processes surrounding the blood vessel walls and paracrine signaling[16, 17]. The coverage ratio of pericytes to ECs is 1/3 in the brain [18–20], underscoring the crucial role of pericytes in the BBB function. High pericyte coverage in GBM is associated with poor chemotherapy outcomes in GBM patients [21]. But pericytes is never a precise definition, they has different functions from supporting, participating development of blood vessel[22, 23], regulating blood flow and vessel permeability, immune functions[24], affecting coagulation[25], pinocytosis/phagocytosis[26], and communicating with endothelial cells and astrocyte. Pericytes carry out various functions in different locations [27], with diverse expression of hallmarks[28] and originating from different sources[29]. Recent studies have revealed pericytes to exhibit significant flexibility in both their hallmarks and functions [27, 30]. However, even recognized pericyte markers such as αSMA, NG2, PDGFR-β, CD13, and desmin are generally nonspecific; NG2, for example, may also be expressed by oligodendrocyte precursors and microglia[31, 32]. PDGFR-β is expressed in smooth muscle cells[33, 34], fibroblasts[35], and mesenchymal stem cells[36]. These markers are not necessary for pericytes identification either: capillary pericytes are negative for αSMA and positive for desmin, whereas venular pericytes express both [37]. Identifying more specific hallmarks for pericytes with distinct functions is necessary to demonstrate the complex perivascular microenvironment.

It has been observed that tumors with lower pericyte coverage rates are more susceptible to radiation and chemotherapy[38]. Research by Lin Cheng et al shows that most pericytes in GBM are derived from glioblastoma stem cells (GSCs)[39], indicating that pericytes in GBM may have distinct hallmarks and functions. Identifying these hallmarks and functions could reveal novel therapeutic options for tackling BBTB. The objective of this study is to identify pericytes that contribute to the BBTB integrity in GBM and investigate their impact on prognosis.

Patient Specimens and Public Patient Cohorts

Nine fresh GBM tissues for single-cell RNA sequencing (scRNA-seq) and pathological wax blocks of 50 GBM/LGG (low-grade gliomas) samples were collected from patients admitted to Xiangya Hospital (Fig. 1a). Informed written consent was obtained from all participants, and the research was approved by the Ethics Committee on Scientific Research of Central South University Xiangya Hospital (approval number: 202103150). GBM bulk RNA sequencing datasets (CGGA, REMBRANDT, and GSE16011[40]) with full clinical information and normalized RNA-Seq data were downloaded from GlioVis (http://gliovis.bioinfo.cnio.es/) and Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). Differentially expressed genes (DEGs) of various clusters of normal brain pericyte were obtained from the supplementary data of a previous study[41], and the GBM para-carcinoma/tumor pairing scRNA-seq data were obtained from GSE162631[42].

Acquisition of scRNA-Seq datasets

ScRNA-seq was performed using the droplet-based 10x Genomics platform. Rigorous quality control and filtering according to the standard procedure of "Seurat" R package in R studio version 3.6.3 was carried out as described previously[43]. A Seurat file containing transcriptomic data for 9,3343 cells derived from nine GBM samples was produced. Please refer to Data S1 for detailed methods and supplementary information.

Identification of tumor-related pericytes

Two approaches were employed to identify potential tumor pericytes. Firstly, various pericyte markers were collected through literature review [20, 30, 39, 44–47], which are commonly shared with various cells, including fibroblasts, smooth muscle cells (SMCs), endothelial cells, mesenchymal stem cells and oligodendrocyte precursors. Brain pericyte expression signatures were gathered from previous studies which described the human brain vasculature using snRNA-seq data. Gene set variation analysis (GSVA) and gene set enrichment analysis (GSEA) were subsequently conducted on each stromal cell sub cluster. After assessment using two methods, potential tumor pericytes were selected from the stromal cell sub clusters and new hallmarks were identified. Specifically, the top 50 up-regulated differentially expressed genes (DEGs) were extracted from potential tumor pericytes and normal brain pericytes, whilst genes from GO terms related to capillary permeability were included. Overlapping genes were obtained across these lists to identify those expressed consistently in pericytes for BTBB functions. Morphological characteristics of pericytes are crucial in the identification, which is cells with anatomical localization surrounding blood vessel walls. So, nominated hallmark of tumor pericytes were co-stained with CD31(A recognized hallmark of endothelial cells) in GBM samples, to conform the tumor pericytes identity. Multiple immunofluorescence assays were carried out on GBM samples using CD31, ANPEP, and Desmin as pericyte hallmarks to confirm their specificity. To determine the characteristics of pericyte functions in GBM, we analysed scRNA-seq data from 4 GBM samples, including the tumor core and paired surrounding peripheral tissues [42] (GSE162631) from Gene Expression Omnibus (GEO) to determine the DEGs between normal pericytes and GBM pericytes, further analysis and research were based on these GBM pericytes expression signatures.

Cell-cell interaction analysis

Our analysis of cell-cell interactions involved the use of CellPhoneDB v2.1.0[48] and focused on known receptor-ligand interactions between two cell types or subtypes. To determine the null distribution of the average ligand-receptor expression levels of the interacting clusters, the cluster labels of all cells were randomly permuted 1000 times. Individual ligand or receptor expression was evaluated using a cutoff based on the average log gene expression distribution for all genes across all cell types. Significant cell-cell interactions were identified by a p-value < 0.05 and an average log expression > 0.1, and we visualized these interactions using the R package circlize v0.4.10.

Immunohistochemistry (IHC)

IHC was carried out as described previously[49]. The tissue sections were incubated with 0.3% hydrogen peroxide for 10 min at room temperature and washed twice with PBS. Then blocked with 5% goat serum for 10 min, the sections were washed and incubated overnight with a primary antibody at 4°C. Secondary antibody (1:400, Abcam, Cambridge, MA, USA) was added and incubation was carried out at 37°C for 30 min. Then the sections were developed by a DAB substrate kit (Sangon Biotech, Shanghai, China) and counterstained with hematoxylin (Sangon Biotech). The immunohistochemical score was calculated by IOD/Area. The following primary antibody was used: Rabbit polyclonal to Collagen IV (Abcam, ab6586, 1:100 dilution); Rabbit polyclonal to Fibronectin (Abcam, ab2413, 1:100 dilution).

Immunofluorescence (IF) Staining

Free-floating sections were first blocked with a buffer containing 10% goat serum, 1% bovine serum albumin, and 0.3% Triton X-100 in PBS for 1 hour, followed by overnight incubation with selected antibodies in blocking buffer at 4°C. The next day, sections were washed three times in PBS and incubated with appropriate fluorescent secondary antibodies (Jackson ImmunoResearch) diluted 1:250 in blocking buffer for 1 hour at 25°C. Slides were visualized using a fluorescence microscope. The following primary antibody was used: Rabbit monoclonal [EPR25414-249] to ATP1A3 + alpha 1 Sodium Potassium ATPase + ATP1A2 + ATP1A4 (Abcam, ab300507, 1:100 dilution); GATA1(Ab-142) Antibody (SAB, 21041-2, 1:100 dilution); PTHR1 Antibody (Affinity, DF6589, 1:100 dilution); Mouse monoclonal [JC/70A] to CD31 (Abcam, ab9498, 1:200 dilution). Results were documented using an FV500 upright confocal microscope (ZESIS, AXlO, Oberkochen, Germany).

Development of a tumor pericyte risk score (TPRS)

First, we analyzed the prognostic value of pericyte-associated DEGs in three independent cohorts using univariate Cox regression. Prognostically relevant DEGs with prognostic value in all three cohorts were screened using Venn diagrams. Subsequently, risk models were constructed in the CGGA cohort using a similar approach as before[50]. Briefly, the prognosis relevant DEGs were downscaled for further screening using the least absolute shrinkage and selector operation (LASSO) algorithm. Then the final model was constructed, and coefficients determined using stepwise Cox regression. The modeling equations are as follows:

TPRS = Σ 𝛽𝑖 ∗ 𝑅𝑁𝐴𝑖

where βi is the coefficient of the gene in the risk model, and 𝑅𝑁𝐴𝑖 is the gene expression value. Specifically, patients were classified into high and low TPRS groups based on the median TPRS. The Kaplan–Meier method was applied, and the log-rank test was used to statistically compare the groups to estimate the prognostic significance of the TPRS. The statistical performance of the TPRS was assessed using the tROC R package. Then two independent cohorts (REMBRANDT, and GSE16011) were used as external validation sets to validate the performance of the TPRS algorithm.

Immunological characterization

The ESTIMATE algorithm[51] was used as described in previous research[52] to calculate overall immune and stromal cell abundance and quantified tumor purity for each sample. A set of pan-cancer T-cell inflammation-associated signature was collected from the Ayers et al, which could define pre-existing cancer immunity and predict clinical response to immune checkpoint inhibitor (ICI)[53]. In addition, a gene set for the cancer immune cycle including 7 steps[54] was collected. A gene set of 28 immune cells was collected from the research by Charoentong et al[55]. A set of gene profiles positively associated with clinical response to ICI was collected from the research by Mariathasan[56]. The enrichment scores of the above gene sets were quantified by Single Sample Gene Set Enrichment Analysis (ssGSEA) using the R package "GSVA" as described in the previous study[57].

Statistical analysis

All statistical data analyses were implemented using R software, version 3.6.3. Correlations between variables were measured by Pearson or Spearman coefficients. T-test was used rendering continuous variables of normal distribution, and the Mann-Whitney U test was applied for other situations. Categorical variables were compared by the chi-squared test/Fisher’s exact test. Survival curves for prognostic analyses were generated by the Kaplan-Meier method, and the log-rank test was applied to estimate statistical significance. Significance level was set at P < 0.05, and all statistical tests were two-sided.

Identify the pericytes concerning BTBB functions in GBM.

Nine GBM fresh tumor samples were selected for scRNA-seq (Table 1), and 9,3343 cells passed quality control and were included in this study. Graph-based clustering analysis, using known lineage markers, identified seven cell populations, with their characteristics shown on the heatmap (Fig. 1b-c). Highly heterogeneous population of cancer cells were observed, due to the GBM heterogeneity among samples. Pericytes are part of the stromal cells by definition. Stromal cells make comprise 2.38% of the total cell count in the 9 specimens, including 2217 cells, and further clustering divided the stromal cells into 12 sub clusters with their hallmarks shown in the bubble diagram (Fig. 1d-e).

Table 1

Clinical information of specimens selected to perform scRNA-seq.
Sample	Age	Gender	GFAP	Ki67	P53	IDH1	Olig2	EGFR	MGMT
G3	66	M	+	15%+	wt	-	+	+
G4	64	M	+	30%+	wt	-	+	-	+
G12	58	M	+	10%+	+	-	+	+	-
G14	65	F	+	30%+	wt	-	+	-
G15	69	F	+	25%+	wt	-	+	2+
G17	73	M	+	10%+	+	-	+	1+
G18	72	M	+	70%+	+	-	+	1+
G19	56	F	+	40%+	wt	-	+	-	+
G21	46	M	+	30%+	wt	-	-	+	+
G22	67	M	+	10%+	+	-	+	+	-

The results indicate that pericyte markers, from previous studies, display discrete distributions within sub clusters (Fig. 2a), suggesting significant heterogeneity among various pericyte research. BMX, which has been identified as a therapeutic target in GBM pericytes[44], exhibits low expression in all clusters. Due to the crossed and discrete marker expression, pericyte characteristics are not distinct. Garcia FJ et al[41] investigated human brain vasculature using single-nucleus RNA-sequencing (snRNA-seq) data, and we collected gene expression signatures from five clusters annotated as pericytes (Pericyte 1, Pericyte 2, Post Mortem Pericyte 1, Post Mortem Pericyte 2, Post Mortem Pericyte 3). Pericyte 1 and Pericyte 2 were annotated from ex vivo freshly resected surgical human brain tissue, while Post Mortem Pericyte 1, Post Mortem Pericyte 2, and Post Mortem Pericyte 3 were annotated from post mortem frozen human brain tissue. We used the top 50 DEGs of five pericytes clusters for the GSVA in our study. Cluster 1, 6, and 11 exhibit distinct pericyte characteristics, with cluster 6 demonstrating significant enrichment in all five pericyte expression signatures (Fig. 2b). Thus, cluster 1, 6, and 11 are potential pericyte clusters. To further confirm our findings, we conducted GSEA in cluster 6, resulting in significant enrichment outcomes in five different pericyte expression signatures (Fig. 2c).

To further investigate the function of different pericytes in GBM, we conducted Gene Ontology (GO) functional enrichment analysis on the top genes of each stromal cell cluster. In GO term analysis, we identified different biological functions of pericytes such as contraction, cell adhesion, immune functions and vascular processes that were enriched in distinct clusters (Fig. s1). Cluster 6 is characterised by BP terms that relate to vascular function and permeability, such as "vascular process in circulatory system", "transport across the blood-brain barrier", and "vascular transport". The CC terms in this cluster pertain to basal plasma membrane and cell adhesion, for example, "basal plasma membrane", "basolateral plasma membrane", "collagen-containing extracellular matrix", "cell-substrate junction", and "focal adhesion". The MF terms in cluster 6 are associated with transmembrane transporter activities (Fig. 3a). The GO functional enrichment analysis demonstrated significantly enriched functions relating to the structure and permeability of the BBB in cluster 6. This indicates a close correlation between cluster 6 and BTBB functions in GBM.

Additionally, cluster 1 is enriched in pathways relating to cellular adhesion, biosynthesis, and regulation of cellular morphology. It is worth noting that cluster 8 exhibits pathway enrichment with strong characteristics of smooth muscle cells (SMCs) (Fig s1), and cluster 8 highly expresses most pericyte markers (DES, αSMA, CALD1, Ang1), and GSEA findings for pericyte expression signatures are insignificant (Fig s2). The conflicting findings from GSVA, GO functional enrichment analysis, and biomarker expression emphasize the lack of specificity in pericyte markers within literary reviews.

Pertinent hallmark for pericyte of BTBB functions.

The results above demonstrate stromal cell sub cluster 6 to be pericyte of BTBB functions, to determine potential hallmarks for them, the top 50 up-regulated DEGs of cluster 6 and normal brain pericytes and GO terms genes related to capillary permeability were extracted. After intersection, only three genes remained: ATP1A2, SLC6A1, and PTH1R (Fig. 3b). These three genes were found to be expressed specifically in cluster 6 of stromal cells (Fig. 3c).

To verify those potential hallmarks, IF co-staining was conducted using histological sections obtained from nine scRNA-seq GBM samples. CD31 was selected as a hallmark for identifying endothelial cells in co-staining with potential hallmarks. The results indicate that cells marked by PTH1R, ATP1A2, and SLC6A1 exhibit morphological cohesion with the surrounding blood vessels of GBM tissue (Fig. 3d). Specifically, cells marked by PTH1R displayed significant specificity in the IF co-staining among the three markers and conform to the classic morphological definition of pericytes, which are cells located anatomically surrounding the walls of blood vessels. Therefore, the findings indicate that PTH1R can be regarded as a novel and highly specific hallmark of pericytes with respect to BTBB functions in GBM.

Multiple IF of PTH1R, CD31, and recognized hallmark of pericytes (ANPEP and Desmin) were conducted in GBM tissue to verify the specificity of PTH1R as a hallmark of pericytes related to BTBB functions in the brain. The results demonstrate that PTH1R has analogous specificity to ANPEP when labeling pericytes. PTH1R and ANPEP both have significantly higher specificity than Desmin (Fig. 3e). There is a significant negative correlation between higher expression of PTH1R and prognosis in the CGGA GBM cohort (Fig. 3f), and the expression of PTH1R is significantly lower in high grade glioma than in low grade glioma, as depicted in Fig. 3g.

PTH1R ⁺ Pericytes communicate with other cells by ECM-integrin signaling pathways and secreting heparin-binding growth factor.

To investigate the cell–cell interaction network among the cell types identified in our present work, we used the CellPhoneDB analysis and pericytes showed active interactions with other cell types, and they showed especially strong interactions with other stromal cells (Fig. 4a). The results identified crucial pathways within ECM-integrin signaling, including FN1, COL1A2, COL4A1 and other collagen-related genes enriched in PTH1R⁺ pericytes interacting with multiple integrin complexes in other cell populations, such as aVb1 complex, a4b1 complex, aVb5 complex, a5b1 complex, a1b1 complex, a10b1 complex, and others. Of significant interest to us are the pathways enriched between PTH1R⁺ pericytes and tumor cells/ECs. The FN1-aVb1 and FN1-a5b1 pair exhibit a high level of enrichment in the interaction between PTH1R⁺ pericytes and ECs, with FN1-aVb1 also showing significant enrichment between PTH1R⁺ pericytes and tumor cells. Collagen-integrin signalling pathways are primarily activated between PTH1R⁺ pericytes and other stromal cells. In addition, CD74 enriched in ECs interacts with APP in PTH1R⁺ pericyte as indicated (Fig. 4b).

Additionally, PTN was enriched in PTH1R + Pericytes via paracrine interactions with PTPRZ1 in various cell types, including tumor cells, oligodendrocytes, other stromal cells and multiple immune cells in GBM (Fig. s3).

To confirm the findings of the CellPhoneDB analysis, we examined the relationships of the interaction pairs above in the CGGA GBM cohort. We observed that all pairs exhibited significant positive correlations, particularly the ECM-integrins interaction pairs (Fig. 4c).

Difference between PTH1R ⁺ Pericytes in GBM and normal brain tissue.

Clustering was conducted and visualised on scRNA-seq data from 4 GBM samples containing the tumor core and paired surrounding peripheral tissues. Identification of cell types in different populations was conducted using marker genes provided by the original research (Fig. 5a). Stromal cells were subdivided into sub clusters (Fig. 5b). According to GSVA and GSEA, cluster 8 of the stromal cells exhibited expression signatures similar to pericytes in both previous research and our study (Fig. 5c-d). The pericyte hallmarks identified above - ATP1A2, SLC6A1, and PTH1R - exhibit a distinct expression in this cluster (Fig. 5e). Conversely, previously acknowledged pericyte hallmarks - RGS5, PDGFRβ, and NOTCH3 - demonstrate low specificity in the findings (Fig. 5f). Differential gene expression analyses were performed between pericytes in GBM and non-malignant tissue, followed by functional enrichment analysis. The findings indicate significant enrichment in pathways such as the complex of collagen trimers, granulocyte chemotaxis, basement membrane, endodermal cell differentiation, and collagen-containing extracellular matrix (Fig. 5g), with the proportion of pathways exhibited in a pie chart (Fig. 5h). The top 10 DEGs, including THY1, COL3A1, COL4A1, TIMP1, COL4A2, COL1A1, MGP, MIR4435-2HG, and FN1, are mostly ECM-related genes. These results demonstrate substantial up-regulation and interactions among ECM components in pericytes of GBM. COL4A1 and COL4A2 encode alpha proteins for collagen IV and are hub genes that combine pathways for collagen trimers and the basement membrane pathway. This indicates that additional collagen IV generated by pericytes is a structural characteristic of BTBB in GBM. The CellPhoneDB analysis has revealed interactions between ECM-related genes and the integrin family as a major pattern of PTH1R⁺ pericytes interacting with other cells in GBM tissue. The analysis of DEGs highlights active ECM-integrin family interactions led by PTH1R⁺ pericytes in GBM.

We then correlated the high and low expression groups of DEGs with the prognosis in the CGGA GBM cohort. All Top 10 DEGs, except THY1 (Fig. 6a), exhibited significant negative correlations. Collagen IV is an essential component of the blood-brain barrier. There is a substantial up-regulation of genes encoding collagen IV in GBM, which corresponds with the prognosis (HR = 0.67 Wilcoxon p value = 0.0151 in the COL4A2 group; HR = 0.61 Wilcoxon p value = 0.0021 in the COL4A1 group). These findings suggest that Collagen IV has potential value for prognosis prediction. FN1 is recognized as the hub gene in the ECM-integrin interaction network. The significant up-regulation of FN1 in GBM correlates with the prognosis (HR = 0.68 Wilcoxon p value = 0.0447). As such, Collagen IV and FN1 are selected as genes of interest for further validation.

30 samples of GBM and 20 LGG samples, all with complete follow-up data, underwent IHC staining to demonstrate collagen IV and FN1 expression levels. The expression levels of collagen IV and FN1 in GBM and LGG samples were presented using a violin plot, calculated with the median value of "IOD/Area" (collagen IV GBM: 0.0274 ± 0.01; LGG: 0.0202 ± 0.012 P = 0.026), (Fibronectin GBM: 0.018 ± 0.013; LGG: 0.0153 ± 0.008 P = 0.411) (Fig. 6b). The IHC results are indicated in Fig. 6c. Subsequent Kaplan-Meier survival analysis revealed a negative correlation between up-regulation of collagen IV and fibronectin and both progression-free survival (PFS) and overall survival (OS) in relation to GBM diagnosis. This correlation is illustrated in Fig. 6d, highlighting the potential of these markers as indicators of poor prognosis.

The Predictive Value of TPRS in GBM Patients

The significant prognostic value of tumor pericyte DEGs in GBM demonstrated above, we expect to build a robust prognostic model based on these DEGs to facilitate individualized therapy for GBM. Firstly, we identified tumor pericyte DEGs with significant prognostic value in all three independent cohorts (CGGA, REMBRANDT and GSE16011) (Fig. 7a). Subsequently, we identified eight key prognostic-related DEGs in the CGGA cohort using LASSO regression (Fig. 7b-c). Finally, we created a prognostic model using the stepwise Cox regression, which included three genes (Fig. 7d).

By using the median value of TPRS as the risk cut-off, patients were separated into high and low-TPRS groups. The low TPRS group demonstrated significantly longer OS in three dependent cohorts (Fig. 7e), and the predictive efficacy of the TPRS was confirmed using ROC curves (Fig. 7f). These results indicate that TPRS can be a useful prognostic model for GBM.

Correlation of TPRS and Immune characteristics in CGGA cohort

Pericytes and other stromal cells, in addition to diverse immune cells, have a crucial function in tumor immunity and potentially impact the choice of varied immunotherapies, such as mRNA vaccines[58]. To investigate the link between TPRS and immune characteristics in the cohorts, we analysed the correlation of TPRS with immune score, stromal score, and tumor purity in the CGGA cohort. As illustrated in Fig. 8a, we identified significant positive correlations between immune score and stromal score, whereas tumor purity was negatively correlated. Furthermore, Fig. 8b demonstrated a positive correlation with T cell infiltrated score. Enrichment of 28 immune cells was also assessed in the CGGA cohort using the ssGSEA method (Fig. 8c)[55]. Analysis revealed significant differences in immune cell components between TPRS high/low groups, with the TPRS high group scoring higher for activated CD4 T cells, central memory CD4 T cells, gamma delta T cells, memory B cells, and type 2 T helper cells, while the TPRS low group scored higher for activated B cells, eosinophils, immature B cells, monocytes, neutrophils, and type 17 T helper cells (Fig. 8d).

The cancer immunity cycle demonstrates various functions of immunomodulators objectively. In the TPRS high group, the release of cancer cell antigens (Step 1) and the infiltration of immune cells into the tumor (Step 5) were more activated. In contrast, dendritic cell recruiting (Step 4), NK cell recruiting (Step 4), and Th2 cell recruiting (Step 4) were more activated in the TPRS low group (Fig. 8e). Additional findings indicate that TPRS had negative correlations with most immune checkpoints, excluding CD276, HAVCR2, LAIR1, and LGALS3. However, TPRS was positively correlated with most pathways predicted for immunotherapy (Fig. 8f-g).

GBM is highly malignant brain tumor with a poor prognosis. The exclusive nature of the BBB in brain tissue determines the distinct characteristics of this cancer. Drug resistance remains a crucial obstacle in chemotherapy, leading to poor prognosis and high recurrence rates. Since the majority of drug compounds cannot penetrate the BBB and reach the tumor site, drug resistance in GBM is exacerbated. Several approaches have been explored over the last few decades to improve the delivery of chemotherapeutic agents to brain tumors, including osmotic BBB disruption, bradykinin receptor-mediated BBB opening, and inhibition of drug efflux transporters, as well as exploiting receptor-mediated transport systems[4].

Pericytes are an important component of the BBB, and emerging studies reveal that pericytes in tumors regulate vascular stability and permeability[59, 60]. Pericyte coverage is linked to both tumor invasiveness and immune infiltration, sequentially affecting immune surveillance and promoting tumor metastasis[61–63]. The BBTB, formed by tumor mesenchymal cells, comprises a barrier that is variably distinct from the BBB, suggesting the feasibility of novel strategies to overcome the BBTB. In GBM tissue, pericytes are composed of BBTB and tumor microenvironment (TME), contributing to both tumor metastasis and therapeutic resistance[64–67]. However, there is no precise functional definition or specific hallmarks for pericytes in the brain, and there is a lack of research on pericytes in GBM tissue. It is hence crucial to clarify the characteristics of the pericytes in GBM to obtain a comprehensive understanding of BBTB function and to theoretically target BBTB to alleviate drug resistance in GBM.

Utilizing scRNA-seq data from nine GBM patients and examining brain pericyte expression signatures from previous studies, we localized and focused on a group of pericytes involved in BTBB function. This group of pericytes demonstrates a significant enrichment in functions related to the BTBB structure and permeability. The hallmarks of pericytes from the previous study exhibited low specificity in our scRNA-seq data. We observed that many other stromal cells showed expression of pericyte markers, particularly in smooth muscle cells (SMCs). Identifying the cells based on hallmarks is challenging in this scenario. The current study's analysis of pericyte markers discovered that they belong to different cell clusters from previous research. Although ACTA1, KCNJ8, NG2, PDGFRβ, ABCC1, DLK1, and PROM1 are indicative of pericytes, they are expressed in different pericyte clusters (DLK1 and PROM1 in cluster 1; ACTA1 and KCNJ8 in cluster 6; NG2, PDGFRβ, and ABCC1 in cluster 11). We concentrated on cluster 6, which is a group of pericytes strongly associated with BTBB functions. We demonstrated and confirmed ATP1A2, SLC6A1, and particularly PTH1R as reliable and specific markers for this type of pericytes when it comes to BTBB functions (PTH1R + Pericytes). Subsequently, the co-staining of these genes with CD31 via immunofluorescence proved that their anatomical distribution is adjacent to EC. The results proved these genes mark brain pericytes, and PTH1R has the most specificity.

PTH1R encodes the receptor for parathyroid hormone (PTH) and parathyroid hormone-like hormone (PTHLH). This gene is associated with GO terms related to the structure of the BBB, including the basolateral plasma membrane, basal plasma membrane and basal part of the cell. While extensively studied in areas such as skeletal development, osteoporosis[68, 69] and Ca2 + homeostasis, its role in BTBB function or GBM diagnosis/treatment has not been investigated. Down-regulation of the PKA/pAKT/HIF1α/VEGF pathway is linked to reduced angiogenesis during fracture healing[70], with lack of PTH being a contributing factor. Additionally, PTH1R signaling, working in conjunction with various downstream pathways, regulates MSCs to drive diverse biological activities in bone marrow[71]. This study presents evidence of a potential positive correlation between PTH1R expression and BTBB integrity. Nonetheless, further research is necessary to verify this hypothesis and to determine the specific mechanisms by which PTH1R or PTH1R⁺ pericytes influence BTBB integrity.

Recent findings suggest that pericytes are an important target for overcoming the BTBB. TGF-β inhibition reduces the presence of αSMA-positive pericytes, leading to a relaxation in pericyte alignment that allows for enhanced nanoparticle penetration into tumor tissue[72]. Specific elimination of pericytes in vivo in GBM leads to the collapse of endothelial walls and vessel lumens, disruption of neovasculature and impediment of cancer progression[39]. The studies above suggest potential for improving prognosis through the targeting of pericytes in GBM. The optimal goal is to surmount the BBTB while retaining the regular functions of BBB, as BBB function is crucial for cerebral metabolism. To accomplish accurate and effective treatment for GBM, comprehensive features of GBM pericytes and BBTB are essential.

Upon comparing normal brain tissue from tumor-adjacent normal tissue in paired samples, we discovered that PTH1R⁺ Pericytes in GBM tissue demonstrated significant up-regulation of ECM-related genes. These genes include THY1, COL3A1, COL4A1, TIMP1, COL4A2, COL1A1, MGP, and FN1. We observed that the up-regulation of these genes has a significant correlation with the grade of glioma and poor prognosis in GBM. Collagen IV was identified as the most abundant collagen in the basement membrane. Collagen IV forms a network in the basement membrane, interacting with the integrin family, G protein-coupled receptors, and discoidin domain receptors (DDR)[73, 74]. This network-forming collagen is one of the four main component proteins of the basement membrane (collagen IV, laminin, nidogen and perlecan), which contributes to the integrity of the BBB[75]. Mutations in COL4A1/2 have been associated with brain malformation and intracerebral hemorrhage [76–78]. T This study has presented evidence of collagen IV having a significant impact on GBM BBTB function. Further research should be undertaken to confirm or refute the observed effect.

FN1 is responsible for the production of fibronectin, which is a soluble dimeric glycoprotein involved in numerous cell adhesion and migration processes such as embryogenesis, wound healing, blood coagulation, host defense, and metastasis. CellPhoneDB analysis revealed interactions between FN1 and integrin family as a main pattern of PTH1R + Pericytes interacting with other cells in GBM tissue, particularly FN1-aVb1 and FN1-a5b1 in PTH1R + Pericyte-EC interactions, and FN1-aVb1 in BBTB-Pericyte-tumor cell interactions. The study found that FN1 is highly expressed in advanced gliomas and is significantly linked to GBM prognosis. The increase in FN1 indicates active FN1-integrin family interactions governed by pericytes in GBM.

Integrins are present on every cell type throughout the body. They consist of α and β subunits, forming heterodimeric transmembrane proteins, with 24 complex formations discovered [79]. Integrins bind to their corresponding ECM proteins and function in cell proliferation, differentiation, migration and maintenance [80, 81]. In PTH1R + pericytes, several integrins are enriched in the interactions with ECM proteins such as collagen IV, collagen I, and fibronectin. These integrins are involved in the function of the BBB. β1-integrins interact with the ECM proteins laminin, collagen IV, and perlecan and are present in the basal membrane of all brain microvessels[82]. Reducing β1-integrin expression markedly reduced interendothelial claudin-5 expression by confluent or subconfluent endothelial cells in culture. This also led to increased cerebral microvessel permeability in vitro and in vivo, demonstrating the essential role of β1-integrin function in preserving a fully functioning BBB[83]. α5β1 integrin is found on a variety of different cell types, particularly on ECs and astrocytes. It is chiefly recognized as a fibronectin receptor. Numerous studies have investigated the role of α5β1 in cancer, as it is significantly upregulated in multiple tumors, with functions related to angiogenesis, tumor development and progression[84]. In solid-phase binding assays, α5β1 shows a strong interaction with VEGF receptor 1. After treatment with integrin α5 antibodies or knockout, processes such as cell crosstalk, migration, proliferation, and adhesion are blocked[85, 86]. The α5β1 integrin has been a target for therapeutic intervention in inhibiting angiogenesis for decades due to its potent angiogenic functions. The α5β1 integrin is also linked to resistance to temozolomide in GBM by inhibiting p53 signaling[87]. The α1β1 integrin is present on endothelial cells/astrocytes and binds to the ECM components of collagen IV and perlecan. Moreover, the αv integrins play a crucial role in tumor growth and angiogenesis. The lack of Integrin αv in the brain leads to the detachment of astrocytes from the ECM, which in turn increases the permeability of the BBB. The targeting of αv/α5 integrins to suppress tumor angiogenesis has advanced. Clionitids has recently undergone clinical efficacy testing for the treatment of GBM in combination with temozolomide radiochemotherapy in a phase III clinical study. Although the trials did not reach OS endpoints[88], they nonetheless indicate significant progress in understanding and utilizing ECM-integrin signaling in GBM. The active interaction of the FN1-integrin family, which is led by PTH1R⁺ pericytes in GBM, not only contributes to the strong barrier function of the BBTB, but is also associated with tumor invasiveness, angiogenesis, progression and drug resistance.

In summary, PTH1R⁺ pericytes exert a significant impact on BBTB functions. Moreover, we detected an increase in the expression of ECM related genes in these cells in GBM, which display negative correlations with prognosis. Lastly, we designed a TPRS to forecast GBM prognosis using the RNA-expressionprofile.

Nonetheless, this study has several limitations. Due to technical limitations, we were unable to separate the pericytes we identified in GBM using a flow cytometer for further experimental validation. Additionally, while the efficiency of TPRS has been validated in other cohorts, it has not yet been validated in prospective cohorts.

Through scRNA-seq data, we identified a type of pericyte in GBM, highly related to BBTB function, with PTH1R potentially serving as a hallmark for them. These pericytes produce an excess of ECM compounds, such as collagen IV and fibronectin, enhance ECM-integrin interactions in GBM. These compounds and interactions may promote basement membrane formation, and then contribute to BBTB function. This study offers new perspectives and directions for exploring the BBTB function and the GBM tumor microenvironment.

BBB: blood–brain barrier; BBTB: blood-brain tumor barrier function; GBM: glioblastoma; scRNA-seq: single-cell RNA sequencing; ECs: endothelial cells;

PET: positron emission tomography; GSCs: glioblastoma stem cells; LGG: low-grade gliomas; CGGA: Chinese Glioma Genome Atlas; GEO: Gene Expression Omnibus;

DEGs: Differentially expressed genes; GSVA: Gene set variation analysis; GSEA: gene set enrichment analysis; SMCs: smooth muscle cells; IHC: Immunohistochemistry;

IF: Immunofluorescence; TPRS: tumor pericyte risk score; ICI: immune checkpoint inhibitor; ssGSEA: Sample Gene Set Enrichment Analysis; TME: tumor microenvironment; ECM: extracellular matrix; PTH: parathyroid hormone; PTHLH : parathyroid hormone-like hormone; DDR: discoidin domain receptor; OS: Overall survival;

Ethics approval and consent to participate

This research was approved by the Ethics Committee on Scientific Research of Central South University Xiangya Hospital (approval number: 202103150).

Consent for publication

Written informed consent was obtained from all participants.

Availability of data and materials

Normalized RNA-Seq dataof CGGA, REMBRANDT, GSE162631, and GSE16011, were downloaded from GlioVis (http://gliovis.bioinfo.cnio.es/) and Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). Our RNA-Seq data was analyzed by Singleron Biotechnologies CO., Ltd. (Nanjing, China).

Competing interests

The authors declare that they have no conflicts of interest.

Funding

This work was supported by the National Natural Science Foundation of China (Grant Number 81802974) and grants from the National Key Technology Research and Development Program of the Ministry of Science and Technology of China (Grant Number 2014BAI04B01). Sponsoring foundations exert no effect on the writing of this manuscript.

Authors' contributions

Conception and design: Yuzhe Li, Changwu Wu, and Qing Liu.

Acquisition and analysis of data (provided tissue microarray, statistical

analysis, biostatistics, etc.): Yuzhe Li, Changwu Wu, and Minghua Wu.

Writing, review, and/or revision of the manuscript: Yuzhe Li, Xinmiao Long, Xiangyu Wang, Wei Gao, Kun Deng, Bo Xie, and Sen Zhang.

Study supervision: Qing Liu and Minghua Wu.

Theauthors read and approved the final manuscript.

Acknowledgements

Not applicable.

Authors' information

¹Department of Neurosurgery, Xiangya Hospital, Central South University, Changsha, Hunan 410008, China.

²Cancer Research Institute of Central South University, Changsha, Hunan 410008, China.

³Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100037, China.

Alexander BM, Cloughesy TF. Adult Glioblastoma. J Clin Oncol. 2017;35(21):2402–9.
Ostrom QT, et al. CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2014–2018. Neuro Oncol. 2021;23(12 Suppl 2):iii1–iii105.
Wu W, et al. Glioblastoma multiforme (GBM): An overview of current therapies and mechanisms of resistance. Pharmacol Res. 2021;171:105780.
van Tellingen O, et al. Overcoming the blood-brain tumor barrier for effective glioblastoma treatment. Drug Resist Updat. 2015;19:1–12.
Langen UH, Ayloo S, Gu C. Development and Cell Biology of the Blood-Brain Barrier. Annu Rev Cell Dev Biol. 2019;35:591–613.
Pardridge WM. Drug and gene targeting to the brain with molecular Trojan horses. Nat Rev Drug Discov. 2002;1(2):131–9.
Matsuo M, et al. Impact of [11C]methionine positron emission tomography for target definition of glioblastoma multiforme in radiation therapy planning. Int J Radiat Oncol Biol Phys. 2012;82(1):83–9.
Pafundi DH, et al. Biopsy validation of 18F-DOPA PET and biodistribution in gliomas for neurosurgical planning and radiotherapy target delineation: results of a prospective pilot study. Neuro Oncol. 2013;15(8):1058–67.
Toyokawa G, et al. Insights into brain metastasis in patients with ALK + lung cancer: is the brain truly a sanctuary? Cancer Metastasis Rev. 2015;34(4):797–805.
Plate KH, Scholz A, Dumont DJ. Tumor angiogenesis and anti-angiogenic therapy in malignant gliomas revisited. Acta Neuropathol. 2012;124(6):763–75.
Hardee ME, Zagzag D. Mechanisms of glioma-associated neovascularization. Am J Pathol. 2012;181(4):1126–41.
Strickland LA, et al. Plasmalemmal vesicle-associated protein (PLVAP) is expressed by tumor endothelium and is upregulated by vascular endothelial growth factor-A (VEGF). J Pathol. 2005;206(4):466–75.
Paiva AE, et al. Endothelial Cells as Precursors for Osteoblasts in the Metastatic Prostate Cancer Bone. Neoplasia. 2017;19(11):928–31.
Prazeres P, et al. Perivascular cell αv integrins as a target to treat skeletal muscle fibrosis. Int J Biochem Cell Biol. 2018;99:109–13.
Birbrair A, et al. Skeletal muscle pericyte subtypes differ in their differentiation potential. Stem Cell Res. 2013;10(1):67–84.
Caporarello N et al. Pericytes in Microvessels: From Mural Function to Brain and Retina Regeneration. Int J Mol Sci, 2019. 20(24).
Díaz-Flores L, et al. Microvascular pericytes: a review of their morphological and functional characteristics. Histol Histopathol. 1991;6(2):269–86.
Giannoni P, et al. The pericyte-glia interface at the blood-brain barrier. Clin Sci (Lond). 2018;132(3):361–74.
Attwell D, et al. What is a pericyte? J Cereb Blood Flow Metab. 2016;36(2):451–5.
Armulik A, Genové G, Betsholtz C. Pericytes: developmental, physiological, and pathological perspectives, problems, and promises. Dev Cell. 2011;21(2):193–215.
Guerra DAP, et al. Targeting glioblastoma-derived pericytes improves chemotherapeutic outcome. Angiogenesis. 2018;21(4):667–75.
Enge M, et al. Endothelium-specific platelet-derived growth factor-B ablation mimics diabetic retinopathy. Embo j. 2002;21(16):4307–16.
Hellström M, et al. Lack of pericytes leads to endothelial hyperplasia and abnormal vascular morphogenesis. J Cell Biol. 2001;153(3):543–53.
Andreotti JP, et al. The role of natural killer cells in the uterine microenvironment during pregnancy. Cell Mol Immunol. 2018;15(11):941–3.
Kim JA, et al. Brain endothelial hemostasis regulation by pericytes. J Cereb Blood Flow Metab. 2006;26(2):209–17.
Rustenhoven J, et al. TGF-beta1 regulates human brain pericyte inflammatory processes involved in neurovasculature function. J Neuroinflammation. 2016;13:37.
Asada N, et al. Differential cytokine contributions of perivascular haematopoietic stem cell niches. Nat Cell Biol. 2017;19(3):214–23.
Birbrair A, et al. Role of pericytes in skeletal muscle regeneration and fat accumulation. Stem Cells Dev. 2013;22(16):2298–314.
Dias Moura Prazeres PH, et al. Pericytes are heterogeneous in their origin within the same tissue. Dev Biol. 2017;427(1):6–11.
Santos GSP, et al. Pericyte Plasticity in the Brain. Neurosci Bull. 2019;35(3):551–60.
Nishiyama A et al. Lineage, fate, and fate potential of NG2-glia. Brain Res, 2016. 1638(Pt B): p. 116–128.
Wohl SG, et al. In situ dividing and phagocytosing retinal microglia express nestin, vimentin, and NG2 in vivo. PLoS ONE. 2011;6(8):e22408.
Lindahl P, et al. Pericyte loss and microaneurysm formation in PDGF-B-deficient mice. Science. 1997;277(5323):242–5.
Winkler EA, Bell RD, Zlokovic BV. Pericyte-specific expression of PDGF beta receptor in mouse models with normal and deficient PDGF beta receptor signaling. Mol Neurodegener. 2010;5:32.
Öhlund D, et al. Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer. J Exp Med. 2017;214(3):579–96.
Crisan M, et al. Perivascular cells for regenerative medicine. J Cell Mol Med. 2012;16(12):2851–60.
Morikawa S, et al. Abnormalities in pericytes on blood vessels and endothelial sprouts in tumors. Am J Pathol. 2002;160(3):985–1000.
Bergers G, et al. Benefits of targeting both pericytes and endothelial cells in the tumor vasculature with kinase inhibitors. J Clin Invest. 2003;111(9):1287–95.
Cheng L, et al. Glioblastoma stem cells generate vascular pericytes to support vessel function and tumor growth. Cell. 2013;153(1):139–52.
Gravendeel LA, et al. Intrinsic gene expression profiles of gliomas are a better predictor of survival than histology. Cancer Res. 2009;69(23):9065–72.
Garcia FJ, et al. Single-cell dissection of the human brain vasculature. Nature. 2022;603(7903):893–9.
Xie Y et al. Key molecular alterations in endothelial cells in human glioblastoma uncovered through single-cell RNA sequencing. JCI Insight, 2021. 6(15).
Chen Z, et al. Single-cell RNA sequencing highlights the role of inflammatory cancer-associated fibroblasts in bladder urothelial carcinoma. Nat Commun. 2020;11(1):5077.
Zhou W, et al. Targeting Glioma Stem Cell-Derived Pericytes Disrupts the Blood-Tumor Barrier and Improves Chemotherapeutic Efficacy. Cell Stem Cell. 2017;21(5):591–603e4.
Stallcup WB. The NG2 Proteoglycan in Pericyte Biology. Adv Exp Med Biol. 2018;1109:5–19.
Cheng Q et al. CALD1 Modulates Gliomas Progression via Facilitating Tumor Angiogenesis. Cancers (Basel), 2021. 13(11).
Oishi K, Kamiyashiki T, Ito Y. Isometric contraction of microvascular pericytes from mouse brain parenchyma. Microvasc Res. 2007;73(1):20–8.
Efremova M, et al. CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes. Nat Protoc. 2020;15(4):1484–506.
Wu C, et al. Overexpression of the phospholipase A2 group V gene in glioma tumors is associated with poor patient prognosis. Cancer Manag Res. 2019;11:3139–52.
Wu C, et al. Liquid biopsy-based identification of prognostic and immunotherapeutically relevant gene signatures in lower grade glioma. J Big Data. 2023;10(1):19.
Yoshihara K, et al. Inferring tumor purity and stromal and immune cell admixture from expression data. Nat Commun. 2013;4:2612.
Wu C, et al. Pan-cancer analyses reveal molecular and clinical characteristics of cuproptosis regulators. iMeta. 2023;2(1):e68.
Ayers M, et al. IFN-γ-related mRNA profile predicts clinical response to PD-1 blockade. J Clin Invest. 2017;127(8):2930–40.
Chen DS, Mellman I. Oncology meets immunology: the cancer-immunity cycle. Immunity. 2013;39(1):1–10.
Charoentong P, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell Rep. 2017;18(1):248–62.
Mariathasan S, et al. TGFβ attenuates tumor response to PD-L1 blockade by contributing to exclusion of T cells. Nature. 2018;554(7693):544–8.
Wu C, et al. Tumor antigens and immune subtypes of glioblastoma: the fundamentals of mRNA vaccine and individualized immunotherapy development. J Big Data. 2022;9(1):92.
Xie L, et al. Phenolic immunogenic cell death nanoinducer for sensitizing tumor to PD-1 checkpoint blockade immunotherapy. Biomaterials. 2021;269:120638.
Östman A, Corvigno S. Microvascular Mural Cells in Cancer. Trends Cancer. 2018;4(12):838–48.
Song S, et al. PDGFRbeta + perivascular progenitor cells in tumors regulate pericyte differentiation and vascular survival. Nat Cell Biol. 2005;7(9):870–9.
Hong J et al. Role of Tumor Pericytes in the Recruitment of Myeloid-Derived Suppressor Cells. J Natl Cancer Inst, 2015. 107(10).
Tian L, et al. Mutual regulation of tumor vessel normalization and immunostimulatory reprogramming. Nature. 2017;544(7649):250–4.
Cooke VG, et al. Pericyte depletion results in hypoxia-associated epithelial-to-mesenchymal transition and metastasis mediated by met signaling pathway. Cancer Cell. 2012;21(1):66–81.
Gao D, et al. Myeloid progenitor cells in the premetastatic lung promote metastases by inducing mesenchymal to epithelial transition. Cancer Res. 2012;72(6):1384–94.
Bonde AK, et al. Intratumoral macrophages contribute to epithelial-mesenchymal transition in solid tumors. BMC Cancer. 2012;12:35.
Meads MB, Gatenby RA, Dalton WS. Environment-mediated drug resistance: a major contributor to minimal residual disease. Nat Rev Cancer. 2009;9(9):665–74.
Barcellos-Hoff MH, Lyden D, Wang TC. The evolution of the cancer niche during multistage carcinogenesis. Nat Rev Cancer. 2013;13(7):511–8.
Santa Maria C, et al. Interplay between CaSR and PTH1R signaling in skeletal development and osteoanabolism. Semin Cell Dev Biol. 2016;49:11–23.
Zhai X, et al. Molecular insights into the distinct signaling duration for the peptide-induced PTH1R activation. Nat Commun. 2022;13(1):6276.
Ding Q, et al. Lack of endogenous parathyroid hormone delays fracture healing by inhibiting vascular endothelial growth factor–mediated angiogenesis. Int J Mol Med. 2018;42(1):171–81.
Kuznetsov SA, et al. The interplay of osteogenesis and hematopoiesis: expression of a constitutively active PTH/PTHrP receptor in osteogenic cells perturbs the establishment of hematopoiesis in bone and of skeletal stem cells in the bone marrow. J Cell Biol. 2004;167(6):1113–22.
Kano MR, et al. Improvement of cancer-targeting therapy, using nanocarriers for intractable solid tumors by inhibition of TGF-beta signaling. Proc Natl Acad Sci U S A. 2007;104(9):3460–5.
Chew C, Lennon R. Basement Membrane Defects in Genetic Kidney Diseases. Front Pediatr. 2018;6:11.
Paavola KJ, et al. Type IV collagen is an activating ligand for the adhesion G protein-coupled receptor GPR126. Sci Signal. 2014;7(338):ra76.
Patino MG, et al. Collagen: an overview. Implant Dent. 2002;11(3):280–5.
Favor J et al. Type IV procollagen missense mutations associated with defects of the eye, vascular stability, the brain, kidney function and embryonic or postnatal viability in the mouse, Mus musculus: an extension of the Col4a1 allelic series and the identification of the first two Col4a2 mutant alleles. Genetics, 2007. 175(2): p. 725 – 36.
Kuo DS, et al. Allelic heterogeneity contributes to variability in ocular dysgenesis, myopathy and brain malformations caused by Col4a1 and Col4a2 mutations. Hum Mol Genet. 2014;23(7):1709–22.
Jeanne M, et al. COL4A2 mutations impair COL4A1 and COL4A2 secretion and cause hemorrhagic stroke. Am J Hum Genet. 2012;90(1):91–101.
Takada Y, Ye X, Simon S. The integrins Genome Biol. 2007;8(5):215.
Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell. 1992;69(1):11–25.
Iwamoto DV, Calderwood DA. Regulation of integrin-mediated adhesions. Curr Opin Cell Biol. 2015;36:41–7.
Yamada KM, Miyamoto S. Integrin transmembrane signaling and cytoskeletal control. Curr Opin Cell Biol. 1995;7(5):681–9.
Osada T, et al. Interendothelial claudin-5 expression depends on cerebral endothelial cell-matrix adhesion by β(1)-integrins. J Cereb Blood Flow Metab. 2011;31(10):1972–85.
Schaffner F, Ray AM, Dontenwill M. Integrin α5β1, the Fibronectin Receptor, as a Pertinent Therapeutic Target in Solid Tumors. Cancers (Basel). 2013;5(1):27–47.
Li L, et al. An angiogenic role for the α5β1 integrin in promoting endothelial cell proliferation during cerebral hypoxia. Exp Neurol. 2012;237(1):46–54.
Wang H, Keiser JA. Vascular endothelial growth factor upregulates the expression of matrix metalloproteinases in vascular smooth muscle cells: role of flt-1. Circ Res. 1998;83(8):832–40.
Janouskova H, et al. Integrin α5β1 plays a critical role in resistance to temozolomide by interfering with the p53 pathway in high-grade glioma. Cancer Res. 2012;72(14):3463–70.
Stupp R, et al. Cilengitide combined with standard treatment for patients with newly diagnosed glioblastoma with methylated MGMT promoter (CENTRIC EORTC 26071 – 22072 study): a multicentre, randomised, open-label, phase 3 trial. Lancet Oncol. 2014;15(10):1100–8.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Single-Cell Transcriptomics of Glioblastoma Reveals Pericytes Contributing to Blood–Brain Tumor Barrier and Tumor Progression

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

Patient Specimens and Public Patient Cohorts

Acquisition of scRNA-Seq datasets

Identification of tumor-related pericytes

Cell-cell interaction analysis

Immunohistochemistry (IHC)

Immunofluorescence (IF) Staining

Development of a tumor pericyte risk score (TPRS)

TPRS = Σ 𝛽𝑖 ∗ 𝑅𝑁𝐴𝑖

Immunological characterization

Statistical analysis

Results

The Predictive Value of TPRS in GBM Patients

Correlation of TPRS and Immune characteristics in CGGA cohort

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1