Sanger Box was an online database used to analyze FPR3 expression across cancers. Expression data for the FPR3 gene were obtained from individual cancer samples. The analysis of the associations between FPR3 levels and overall survival (OS), progression-free interval (PFI), and disease-specific survival (DSS) in cancer patients was conducted using the Sanger Box tool.

The examination of the relationships between FPR3 and immune checkpoint elements, encompassing inhibitory and stimulatory factors, along with major histocompatibility complexes (MHC), was carried out using the TISIDB database. Recent research has focused on predictive biomarkers such as high TMB and high MSI to evaluate immunotherapy efficacy (Chan et al., 2019). Spearman’s rank assessed the association of FPR3 expression with TMB or MSI, followed by the utilization of the R “fmsb” package to produce a radar chart illustrating these results. Furthermore, UMAP plots depicting the expression profiles of FPR3 across various cell types were acquired from the TISCH database (Sun et al., 2021).

This research also investigated the genomic alterations in FPR3 using the cBioPortal database. Additionally, we employed GSCA (Liu et al., 2018) to examine the copy number variation (CNV) of FPR3. The analysis of the associations between FPR3 levels and overall survival (OS), progression-free interval (PFI), and disease-specific survival (DSS) in cancer patients was conducted using the Sanger Box tool.

The present study focused on four distinct cohorts comprising individuals diagnosed with gliomas. Data on gene expression and pertinent clinical details were sourced from reliable databases, including the TCGA and CGGA databases (Zhao et al., 2021), as well as the Rembrandt (Gusev et al., 2018) and Gravendeel databases (Gravendeel et al., 2009), accessed through their respective online portals. The Rembrandt and Gravendeel databases were accessed via the Gliovis website (Bowman et al., 2017). The TCGA dataset served as our primary resource for investigating the prognostic implications of FPR3 in glioma. To ensure the validity and robustness of our findings, we further utilized three additional cohorts to validate and reinforce the results.

The Wilcoxon test was used to examine the levels of FPR3 and its corresponding clinical features. Variables such as age, sex, histologic grade, isocitrate dehydrogenase (IDH) mutation status, 1p19q codeletion status, and MGMT promoter (MGMTp) methylation status were evaluated. The R package “ComplexHeatmap” was employed to create a heatmap that compared the distribution of clinical features across various FPR3 expression groups. Box plots illustrated the variations in FPR3 expression among different clinical categories. The “Survminer” package in R was utilized to build the Cox model and perform sequential variable selection. These findings were validated using the CGGA cohort.

The nomogram is a widely employed tool in oncology, providing personalized prognostic information for cancer patients. It offers easily understandable risk assessments (Iasonos et al., 2008). A representative nomogram model was developed using the “RMS” package in R. This model integrated the FPR3 transcription level with traditional clinical factors such as grade, age, and IDH status. Each patient’s composite score was transformed by the nomogram into the predicted probability of outcome duration. The effectiveness and predictive capability of the nomogram were evaluated using statistical methods, including the concordance index (C-index), time-dependent receiver operating characteristic (tROC) curve, calibration curve, and decision curve analysis (DCA).

The median value was used to dichotomize FPR3 expression, creating a categorical dependent variable. Differentially expressed genes (DEGs) in the TCGA dataset were pinpointed through the Wilcoxon rank test, applying screening criteria of a false discovery rate (FDR) < 0.05 and |log fold change (logFC)| > 1. To ensure accuracy, a minimum average expression level threshold of 1 for the raw DEGs was established. DEGs that met the specified criteria were chosen for additional analysis. Visualizations were generated using R, with a volcano plot illustrating all DEGs and a heatmap depicting the expression profiles of the top 50 DEGs. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) tools were utilized, and the results were displayed with bubble charts. Discrepancies between two biological states were assessed for statistical significance using Gene Set Enrichment Analysis (GSEA). with each analysis involving 1,000 gene set permutations. FPR3 levels were used as a distinguishing phenotype indicator. Enriched networks for each phenotype were evaluated based on their minimal p-value.

We measured FPR3 expression in 28 TIL categories using the Tumor Immunology and Signaling Database (TISIDB) (Ru et al., 2019) to investigate FPR3-associated immune infiltration. The TISIDB database integrates diverse data sources pertinent to tumor immunology. FPR3 and TILs were examined using Spearman correlation analysis. TISIDB was implemented to analyze FPR3 levels in LGG immune subgroups. FPR3 expression was predominantly detected in three distinct subtypes, namely C3 (characterized by inflammation), C4 (characterized by lymphocyte depletion), and C5 (characterized by immunological quiescence). We computed stromal, immune, and ESTIMATE scores for each sample utilizing the ESTIMATE (Estimation of Stromal and Immune Cells in Malignant Tumor Tissues Using Expression) methodology. Furthermore, the proportions of 22 tumor immune cell types were assessed using CIBERSORTx. Subsequently, the TCGA dataset was analyzed with the R package “ggplot2” to investigate the relationship between FPR3 and immune checkpoints. GEPIA2 also identified statistically significant relationships between immune cell gene markers and FPR3 expression.

Glioma-related RNA sequences were obtained from the TCGA dataset. Subsequently, the IC50 values of widely used chemotherapeutic drugs were computed using the R package “oncoPredict,” which predicts chemotherapy response by leveraging tumor gene expression profiles. These IC50 values were then compared across subgroups with high and low FPR3 expression to determine therapy efficacy.

The Department of Neurosurgery supplied glioma and non-tumor specimens spanning the period from 2020 to 2023. Following the surgical removal of the tissue, the samples were promptly frozen at a temperature of −80°C and then transferred to a container containing liquid nitrogen for extended preservation. Surgically resected tumor tissue specimens were sectioned and processed in the pathology department. Histological features were assessed using light microscopy. Immunohistochemical staining was performed to detect specific markers, such as IDH mutations and ATRX deletions. These analyses were essential for confirming the diagnosis and ensuring that the tumor grading adhered to the WHO CNS5 diagnostic criteria. Out of the 63 samples initially obtained, 41 samples with inadequate information were eliminated, leaving 22 samples with a clear diagnosis for further study.

Initially, the tissue slices underwent dewaxing and subsequent hydration through a series of ethanol solutions of increasing concentrations. This process was followed by antigen retrieval using a sodium citrate solution. In order to suppress endogenous peroxidase activity, a 30% hydrogen peroxide and methanol solution was employed. The transparencies were subsequently incubated with a primary antibody solution overnight (anti-FPR3, orb39488, Biorbyt, Cambridge, United Kingdom). Staining was performed using the DAB kit from Beyotime, China, and hematoxylin. Visual examination of the slides was conducted using a Leica laboratory microscope manufactured in Germany.

U251 human glioblastoma cell lines were obtained from the cell bank of the Chinese Academy of Sciences (Shanghai, China). In DMEM medium with 10% foetal bovine serum (Gibco, China) at 37°C and 5% CO₂, U251 cells were cultivated daily. All experiments used logarithmic-growing cells. Sangon Bioengineering (Shanghai, China) chemically synthesised FPR3 siRNA. Lipofectamine 3000 from Invitrogen was used to transfect FPR3 siRNA and negative control (NC) cells as specified. The following siRNA sequence were used in this study:

FPR3-siRNA: 5′-AAC AAC AUC UCU UUG AGC C-3′

U251 cells were divided into two groups: NC group and Si-FPR3 group. After 24 h of transfection, trypsin digestion was resuspended and counted by cell counter, and 96-well plates were inoculated with 1,000 cells per well, and 5 replicate wells were repeated. After incubation for 1.5 h in the absence of light, the plates were set at 37°C and OD450 quantification was performed.

U251 cells were inoculated into six-well plates 200,000 cells per well and incubated overnight in cell culture incubator, waiting for them to grow to 80–90 confluence, lipo3000 was transfected and after 22 h of transfection, they were changed to serum-free DMEM medium and after 2 hours of starvation, the wells were scribed using a 200 µL pipette. After washing the cells with PBS, DMEM medium containing 1% serum was added to the cells, and cell migration photographs were taken under the microscope at 0, 12, and 24 h, respectively.

The verification of FPR3 protein expression levels was conducted using Western blotting. Initially, the samples underwent treatment with RIPA lysis buffer, and the measurement of protein content was conducted with a BCA test kit (Beyotime, China). Subsequently, an equivalent quantity of protein (20 mg) was loaded into lanes, followed by electrophoresis separation and transfer of the proteins to a PVDF membrane (Merck Millipore, Germany). Primary antibodies were incubated on the membrane after blocking it with a 5% milk solution. The main antibodies used were anti-FPR3 (orb39488, Biorbyt, Cambridge, United Kingdom) and anti-α-tubulin (66,031-1-lg, proteintech, China) diluted at 1:1,000. The incubation was carried out overnight at a temperature of 4°C. After incubation, the secondary antibody was applied at room temperature for an hour. This stage detected the targeted proteins using the enhanced chemiluminescence kit (Servicebio, China).

Bioinformatics statistical analyses were conducted using R version 4.1.0, with significance levels set at P < 0.05 to determine statistical significance.

FPR3 expression was elevated in the majority of tumor tissues, with a marked decrease in LUSC, ALL, and ACC compared to normal tissues. Noteworthy, no notable variances were observed in WT, BLCA, PCPG, KICH, and CHOL tissues (Figure 1A). We then evaluated the predictive value of FPR3 for OS, PFI, and DSS (Figure 1B). FPR3 expression levels predicted numerous cancer outcomes. Specifically, FPR3 was identified as a protective factor for SKCM but a risk index for GBMLGG, UVM, LGG, LAML, KIPAN, TGCT, and GBM for OS. The upregulation of FPR3 was associated with shorter DSS in GBMLGG, LGG, UVM, and GBM, while SKCM exhibited longer DSS. For PFI, FPR3 expression was a risk factor in GBMLGG, PRAD, GBM, and LGG.

Single-cell sequencing serves as a potent technique for scrutinizing oncogene expression at the individual cell level. Through scRNA sequencing, we were able to identify FPR3 expression in malignant cells as well as immune cells like CD8Tex cells, dendritic cells (DC), and macrophages within BRCA, glioma, GBM, LIHC, NSCLC, and SKCM samples (Figure 1C). Figure 1D displays the expression of FPR3 in the single-cell datasets GSE131928, GSE111360, and GSE140228.

In various malignancies such as BLCA, BRCA, COAD, GBM, and LUAD, FPR3 expression correlated positively with immune checkpoint factor expression. Among these immunoinhibitors (Supplementary Figure S1A), FPR3 exhibited the strongest correlations with CSF1R, HAVCR2, PDCD1LG2, and TIGIT. Regarding immunostimulators, FPR3 demonstrated the highest correlations with CD80, CD86, and TNFSF13B (Supplementary Figure S1B). Furthermore, FPR3 was strongly correlated with most MHC molecules (Supplementary Figure S1C). Supplementary Figure S1D shows that FPR3 and TMB are positively linked in UCEC, THYM, SARC, PRAD, OV, LGG, COAD, and BRCA, but negatively correlated in THCA, TGCT, PAAD, LUSC, LIHC, and GBM. UCEC and COAD exhibit a positive correlation between FPR3 expression and MSI, whereas TGCT, STAD, SKCM, OV, LUSC, LUAD, LIHC, LGG, and KIRP exhibit a negative correlation (Supplementary Figure S1E).

Overall, our comprehensive analysis of FPR3 alterations in various cancers utilizing the TCGA dataset and COSMIC database revealed intriguing insights. The FPR3 gene exhibited alterations in 2% of patients across 32 studies and 10,967 samples (Supplementary Figure S2A), with 98 missense, 10 truncating, and 1 fusion mutations identified in amino acids 0–353. Notably, the R137 H/C mutation site was prevalent in UCEC instances (Supplementary Figure S2B). FPR3 genetic mutation in TCGA tumour samples revealed unique patterns. The greatest number of FPR3 alterations (>5%) was observed in patients with UCS, with “Multiple Alterations” being the most prevalent category of alteration (Supplementary Figure S2C). Among the five cancer types, SKCM, UCEC, COADREAD, LIHC, and LUAD, “Mutation” occupied the most common type of FPR3 alterations. In LGG, the FPR3 alteration was primarily characterised by “Deep Deletion”. Supplementary Figure S2D shows Pan-Cancer FPR3 CNV states. Heterozygous Amplification and Heterozygous Deletion are the two main types of CNV in all cancers, The former as the main type of cancer include ACC, UCS, BLCA, KICH, CESC, DLBC, GBM, BRCA and LUSC, while the former as the main type of cancer include SARC, LUAD,TGCT,OV and LGG. The COSMIC database results demonstrated that the major form of FPR3 mutations was missense mutation (35.56%), and the primary SNV was C>T (38.91%) (Supplementary Figure S2E).

By integrating GTEx data with TCGA, we performed a differential expression analysis. The investigation demonstrated that glioma tissues exhibited significantly elevated levels of FPR3 expression in comparison to normal tissues (p < 0.001). This finding was further corroborated by data from the Rembrandt and Gravendeel datasets (Figure 2A). The cohort with high FPR3 expression displayed a higher incidence of grade III tumors, IDH wild-type status, unmethylated MGMTp, and 1p19q non-codeletion (p < 0.01) (Figure 2B). Analysis of the TCGA dataset showed that individuals over the age of 40, with grade III tumors, wild-type IDH status, and 1p19q non-codeletion, had elevated levels of FPR3 expression (p < 0.001) (Figure 2C). This discovery was further confirmed by utilizing the CGGA_693 validation dataset (Figure 2D). These outcomes highlight a robust association between elevated FPR3 expression and aggressive features in gliomas.

In order to assess the prognostic significance of FPR3 in glioma, samples from TCGA and CGGA were categorized into groups based on their FPR3 expression levels relative to the median. The group with elevated FPR3 levels comprised a higher proportion of patients. In the TCGA dataset, elevated FPR3 expression significantly correlated with reduced OS. (p < 0.001). Comparable findings were noted in the validation datasets from CGGA (Figure 3A). Additionally, time-ROC curve analysis of FPR3 expression in the TCGA database yielded area under the curve (AUC) values exceeding 0.65, consistent with the findings from the CGGA data (Figure 3B). In glioma patients, increased FPR3 expression may predict poorer outcomes. To identify independent prognostic determinants, Cox regression analysis with one or more variables was conducted. Within the TCGA cohort, FPR3 showed an independent correlation with OS (p < 0.05). Additionally, tumor grade, patient age, and IDH status also independently influenced prognosis (Figure 3C). The CGGA_325 validation dataset was used to confirm these findings (Figure 3D). In conclusion, these findings indicate that FPR3 holds promise as a prognostic biomarker for assessing glioma patients.

A nomogram was developed to precisely forecast survival outcomes (Figure 4A). This nomogram incorporated age, grade, IDH status, and FPR3 expression as parameters, which were determined through stepwise Cox multivariate regression analysis. Following this, risk scores were computed utilizing the model. Subsequently, individuals in the training dataset were categorized into two separate groups according to their risk scores. The K-M survival curve demonstrated a notable discriminative capacity of the nomogram. (p < 0.001) (Figure 4B). A calibration curve evaluated the agreement between predicted and observed risks (Figure 4C). The calibration curve showed satisfactory alignment between the nomogram-derived and measured OS rates. Furthermore, the model demonstrated exceptional predictive capabilities for OS, as evidenced by the higher C-index and AUC values obtained through time-ROC analysis (Figures 4D–F). These results surpassed those achieved by other clinical and molecular characteristics. Significantly, the DCA curve offered proof endorsing the favorable predictive precision of the nomogram over 1-, 3-, and 5-year periods (Figure 4G).

The Wilcoxon test identified 1,892 differentially expressed genes. A heatmap was created to visually represent the top 50 genes that exhibited significant correlation with FPR3 expression in gliomas (Figure 5A). This analysis revealed 1,112 genes positively associated with FPR3 and 780 genes negatively associated with FPR3. Functional enrichment analysis was performed on genes that positively linked with FPR3, and the results were successfully shown using a bubble chart (Figure 5B). The GO term annotations revealed that the genes are implicated in various biological processes, including leukocyte-mediated immunity and immune response-regulating signaling pathways. KEGG study linked these genes to neuroactive ligand-receptor interaction, phagosome formation, and cell adhesion molecules (Figure 5C). GSEA techniques were employed to identify signaling pathways associated with FPR3 in gliomas. Increase in FPR3 expression was identified to be pivotal in various processes, including the signaling pathways of B-cell receptors, T-cell receptors, and Toll-like receptors (Figure 5D).

Correlations exist between the FPR3 gene expression and 28 different classifications of TILs. Supplementary Figure S3A illustrates a significant association between FPR3 expression levels and TILs across various human cancers, with a particular emphasis on gliomas. Additionally, a TISIDB analysis of FPR3 levels in LGG immune subgroups revealed that the C3 subtype exhibited the highest expression levels, while the C5 subtype showed reduced expression (Supplementary Figure S3B). The ESTIMATE algorithm also compared glioma immune infiltration patterns to FPR3 expression. The findings suggested that groups with high expression levels exhibited elevated immunological, stromal, and ESTIMATE scores in contrast to those with low expression levels (Supplementary Figure S3C). FPR3 was examined in connection to immune checkpoint target genes. The FPR3 protein positively correlated with CD48, CD28, CD80, CD86, and PDCD1 (Figure 5E). Moreover, FPR3 expression exhibited connections with macrophage abundance, Tem_CD8 abundance, Myeloid-derived suppressor cell abundance, mast cell abundance and regulatory T cell abundance in GBM (r > 0.65, p < 0.05) (Supplementary Figure S3D). Similarly, this trend was observed in LGG. Figures 5F, G shows intergroup differences in 22 immune cell subpopulations. In the high FPR3 expression group, there were elevated levels of resting memory CD4 T cells, regulatory T cells, M0 and M1 macrophages, and neutrophils, whereas the low FPR3 expression group displayed increased proportions of naive B cells, plasma cells, and follicular helper T cells. We also examined FPR3 expression and gene markers associated with distinct TILs (Table 1). In general, the expression of FPR3 was significantly associated with gene markers for B cells, T cells, and M1 and M2 macrophages.

We investigated the link between FPR3 expression and drug sensitivities using common medicines or chemicals. Through a comparative assessment of IC50 values in two different groups, we found that the cohort with elevated FPR3 expression had significantly lower IC50 values for Temozolomide, AMG-319, Bortezomib, Cediranib, Dasatinib, Entospletinib, Savolitinib, and 5-Fluorouracil compared to the group with lower expression levels. In contrast, the high-expression group demonstrated higher IC50 values for SB505124 and Vorinostat than the low-expression group (Supplementary Figure S3E).

The levels of FPR3 protein in glioma were verified using immunohistochemistry and Western blot. The results consistently indicated an upregulation of FPR3 protein expression in glioma specimens compared to the paired adjacent specimens (Figures 6A, B).

To investigate the biological role of FPR3 in GBM, its expression was significantly reduced in U251 cells transfected with siRNA-FPR3 (Figure 6C). The CCK-8 assay revealed that FPR3 suppression hindered GBM cell proliferation compared to the NC cells (Figure 6D). Additionally, the wound healing assay further demonstrated that the migratory capacity of GBM cells was significantly reduced in the FPR3 knockdown cells (Figures 6E, F). These findings suggest that dysregulated FPR3 expression may affect both the proliferation and migration of GBM cells.