Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo) is a public functional genomics data repository of high-throughput gene expression, gene chip and microarray data. Two gene expression datasets (GSE118186 and GSE66743) were downloaded from GEO (18). GSE118186 contains SUZ12-wild-type (SUZ12-WT, 2 cases) and SUZ12-mutant (SUZ12-Mut, 2 cases) RNA-Seq data and H3K27me3-related chromatin immunoprecipitation sequencing (ChIP-Seq) data for ipNF05.5 cells. GSE66743 is a gene chip dataset containing data for 30 MPNST patients, including the survival times of the patients.

To identify differentially expressed genes (DEGs) between SUZ12-WT and SUZ12-Mut cells, we used Trim Galore and FastQC to process and optimize the raw data. The hg38.fa file was downloaded from the UCSC Genome Browser as the reference genome sequence. After the index was constructed with Salmon, quantitative gene expression analysis was carried out on 4 samples. With p adj ≤ 0.05 and Log|FC|≥2 as the screening criteria, DESeq2 was used to perform differential gene expression analysis on 2 samples of SUZ12-WT and 2 samples of SUZ12-Mut cells. The hg38.fa file was used as the reference genome sequence to construct an index with Bowtie to complete the ChIP-Seq gene alignment. SAMtools was used to convert the compared SAM files into BAM files and sort them. DeepTools was used to convert the BAM files to BW files to complete the visualization of transcription start sites (TSSs).

To understand the functions of the DEGs, we used R software to perform Gene Ontology (GO) enrichment analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis and gene set enrichment analysis (GSEA). Set H (hallmark gene sets; 50 gene sets), a predefined GSEA dataset, was used in this study. The GO enrichment analysis results show the top 10 pathways in descending order, the KEGG enrichment analysis results show the top 20 pathways in descending order, and the GSEA results show activated and inhibited pathways.

The DEGs were uploaded to STRING (https://string-db.org/) to construct the PPI network. According to the network’s maximal clique centrality (MCC, associated median value), CytoHubba in Cytoscape was used to identify the top 10 core genes in the network and annotate the functions of the core genes.

A total of 28 MPNST patients in GSE66743 were included in the survival analysis, and 2 MPNST patients with unknown survival times were removed. The patients were divided into high-risk groups and low-risk groups based on the average core gene expression levels. To analyze the relationships between the hub genes and survival, Cox proportional hazards regression was used for survival analysis with R software. Genes with P<0.05 were defined as key genes.

The gene bodies and promoter regions of key genes were searched in the NCBI database, and the BW file obtained by ChIP-Seq analysis was imported into Integrative Genomics Viewer (IGV). The H3K27me3 distribution in key gene promoter regions and across the genome were compared between SUZ12-WT and SUZ12-Mut cells.

Eight MPNST specimens were obtained during surgical resection from patients at Beijing Tiantan Hospital between 2019 and 2021. Two of these cases were sporadic MPNSTs, and six were NF1-associated MPNSTs. The histopathological diagnosis was assessed according to the 2016 WHO Classification of Tumors of the Central Nervous System by two independent neuropathologists. Part of each specimen was snap frozen and stored in liquid nitrogen, and the remainder was fixed with 4% formalin for histopathological and immunohistochemical examination. All procedures were approved by the Institutional Review Board of the Beijing Tiantan Hospital Ethics Committee.

The human MPNST cell line sNF96.2 and pNF cell line ipNF05.5 were purchased from the American Type Culture Collection (ATCC, USA). All cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA) and maintained at 37°C in 5% CO₂. Forskolin, 3-isobutyl-1-methylxanthine (IBMX) and H89 were purchased from MedChem Express (MCE, USA). Cells were treated with 25 µM forskolin and 100 µM IBMX for 20 min; this treatment is designated F/I in the text and figures. The protein kinase A (PKA) inhibitor H89 was applied at 10 µM 20 min prior to treatment with F/I.

For immunohistochemistry (IHC), sections were sequentially incubated with appropriate primary antibodies (specific for SUZ12, H3K27me3, ADCY1 and p-ERK) according to the corresponding instructions, incubated with the secondary antibody and diaminobenzidine (Zhongshan Bio Corp, China), counterstained with hematoxylin, and visualized under a light microscope.

SUZ12 and NC siRNA were chemically synthesized with the following sequences: NC antisense: 5′-ACGUGA CACGUUCGGAGAATT-3′, SUZ12 siRNA antisense 1: 5′-UAUUGGUGCUAUGAGAUUCCGTT-3′, antisense 2: 5′-UUCUAGUGGCAAGAGGUUUGGTT-3′, and antisense 3: 5′-UAGAUGAAGCAUGAAGUUUCGTT-3′ (Sangon Biotech, China). The recombinant plasmid SUZ12-pENTER was constructed by Vigene Bio, China. siRNAs and plasmids were transfected into cells at 70-80% confluence with Lipofectamine RNAiMAX or Lipofectamine 3000, respectively, according to the manufacturer’s instructions (Invitrogen, USA).

Total RNA was extracted with TRIzol (Life Technology, USA), and reverse transcription was performed using a high capacity reverse transcription kit (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. Real-time PCR (RT-PCR) was performed using Power SYBR Green Master Mix (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. RT-PCR data were analyzed by the 2-^ΔΔCt method. The following primer sequences were used: SUZ12 (F: 5′-CAA ACT GAA GCA AGA GAT GAC C-3′; R: 5′-GCT ATG GCA GAG TTT AAG ATG C-3′).

Total cell lysates were prepared using RIPA lysis buffer supplemented with PMSF (Solarbio, China). The protein concentration was determined using a BCA protein assay kit (Biosharp, China). Proteins were resolved by SDS-PAGE and electrotransferred to PVDF membranes (Millipore, USA). Anti-ERK2 (sc-1647), anti-adenylate cyclase 1 (sc-365350), and anti-p-ERK (sc-7383) antibodies were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, USA). Anti-SUZ12 (D39F6) antibodies were purchased from Cell Signaling Technology (CST, USA). The Rap1 activation assay was performed with an Active Rap1 Detection Kit (Cell Signaling Technology, USA) according to the instructions.

Statistical methods used for bioinformatic analysis are detailed in each section of the experimental methods. Statistical analysis was performed using GraphPad Prism version 8. The results were presented as the mean ± standard deviation (SD) of independent triplicate experiments. Statistical differences between groups were assessed using a Student’s t-test analysis. *P<0.05 was considered statistically significant, and ***P<0.001 and ****P<0.0001 were considered highly significant.

After normalization of the microarray results, DEGs were identified. A total of 704 genes were upregulated and 125 genes were downregulated in ipNF05.5 SUZ12-Mut cells compared with SUZ12-WT cells ( Figure 1A ). A heat map was constructed to show the names of top 50 genes ( Figure 1B ), and all DEGs were displayed in Supplementary Table 1 .

To analyze the biological classification of the DEGs, functional and pathway enrichment analyses were performed using GO, KEGG and GSEA methods. GO analysis showed that in the DEGs were significantly enriched in the biological process (BP) terms axonogenesis, regulation of cellular response to growth factor stimulus and regionalization. In addition, the DEGs were enriched mainly in the cell component (CC) terms synaptic membrane, collagen-containing extracellular matrix, and glutamatergic synapse and in the molecular function (MF) terms DNA-binding transcription activator activity, glycosaminoglycan binding and extracellular matrix structural constituent ( Figure 1C ). Interestingly, KEGG pathway enrichment analysis revealed that the DEGs were enriched mainly in the cardiomyopathy and Wnt signaling pathways, and the gene ratios showed that more DEGs were clustered in the neuroactive ligand-receptor interaction pathway, MAPK signaling pathway, and Rap1 and cyclic adenosine monophosphate (cAMP) signaling pathway ( Figure 1D ). GSEA showed that the KRAS pathway was upregulated in SUZ12-Mut cells ( Figure 1E ).

The PPI network of the DEGs was constructed ( Figure 2A ) and the most significant module identified using Cytoscape ( Figure 2B ). The top 10 hub genes were ADCY5, ADCY1, BDKRB1, LPAR5, BDKRB2, CHRM2, S1PR5, S1PR3, ADRA2A, and ADRA2C; all of these were upregulated DEGs. TSSs are one of the most important regions in the genome, and the altered distribution of H3K27me3 in TSS regions regulates gene expression. The distribution of H3K27me3 near TSS regions was significantly higher in SUZ12-WT cells than in SUZ12-Mut cells ( Figure 2C ). 4 of 10 hub genes were detected in the gene microarray dataset containing survival, and the other genes were not expressed. Subsequently, associations between overall survival and the expression levels of the 4 hub genes were analyzed using Cox proportional hazards regression for survival analysis ( Supplementary Figure S1 ). MPNST patients with upregulation of only ADCY1 showed worse overall survival ( Figure 2E ). Importantly, compared with that in SUZ12-WT cells, the distribution of H3K27me3 in the ADCY1 promoter region and across the genome in SUZ12-Mut cells was sparse and significantly decreased ( Figure 2D ). KEGG pathway analysis showed that the Rap1 and cAMP signaling pathways were enriched, and the product of ADCY1 is cAMP. Therefore, it was speculated that SUZ12 deletion may amplify RAS signaling via the ADCY1/cAMP/Rap1/ERK pathway.

Clinical information and the SUZ12 expression data for 8 MPNST patients are shown in Table 1 . Immunohistochemical analysis showed that two of the eight MPNST cases were positive for SUZ12, the remaining six did not express SUZ12, and all recurrent MPNSTs were SUZ12-negative. The H3K27me3 level was extremely low in SUZ12-negative patients, while the rates of ADCY1 and p-ERK positivity were significantly higher in SUZ12-negative patients than in SUZ12-positive patients ( Figures 3A, B ). These results suggested that the expression of ADCY1 and phosphorylation of ERK may be regulated by SUZ12, consistent with the differential gene expression analysis results.

SUZ12 siRNA and pENTER-SUZ12 were transfected into ipNF05.5 and sNF96.2 cells to further clarify the effect of SUZ12, and NC siRNAs and empty vector were used as the corresponding controls. As shown by western blot analysis and RT-PCR, S3 had the best knockdown effect among the three siRNA sequences transfected and was thus used for all subsequent experiments ( Figure 4A ). The expression of SUZ12 was knocked down or upregulated after transfection of the siRNA and overexpression plasmid, respectively. In addition, after transfection, the H3K27me3 level was consistent with the SUZ12 expression level but was inversely related to the ADCY1 level. SUZ12 did not affect the expression level of ERK2, while the level of p-ERK varied, albeit non significantly, with that of ADCY1 in both ipNF05.5 and sNF96.2 cells ( Figure 4B ).

Forskolin, an activator of adenylate cyclase, increases the intracellular level of cAMP (19). Additional elevation of the cAMP level can be achieved through treatment with IBMX, a nonselective phosphodiesterase inhibitor (20). Cells were stimulated with this mixture of forskolin and IBMX (i.e., F/I) for 2, 5, 10 and 20 min, and ERK was found to be activated at approximately 20 min, consistent with the manufacturer’s instructions ( Figure 5A ). F/I was added for 20 min and was then removed to observe the trend in ERK phosphorylation over time. The p-ERK level gradually increased and then decreased at 20 min in the control cells. However, the activation of ERK phosphorylation was accelerated and prolonged after transfection of siRNA, especially in ipNF05.5 cells, starting at 0 min and lasting for 40 min ( Figure 5B ). Interestingly, the level of p-ERK in the SUZ12 si group decreased at 5-10 min. However, the intensity and duration of p-ERK activation were reduced after overexpression of SUZ12. Under the same treatment conditions, SNF96.2 cells exhibited exactly the same pattern as ipNF05.5 cells ( Figure 5C ). Therefore, we confirmed that SUZ12 can affect the phosphorylation level of ERK through ADCY1.

ERK activation is mediated by PKA-independent Ras proteins that cooperate with the Rap1 protein through cAMP (20). To determine whether the effect of SUZ12 is mediated by Rap1, we utilized a pulldown assay to detect the effect of SUZ12 knockdown and PKA inhibitor (H89) treatment on the level of activated Rap1 (Rap1-GTP). Rap1 activation was significantly elevated after interference with SUZ12 expression and was blocked by H89 treatment ( Figure 6A ). More importantly, the effects of SUZ12 knockdown on ERK phosphorylation were eliminated after the addition of H89 to both ipNF05.5 ( Figure 6B ) and sNF96.2 cells ( Figure 6C ) in the SUZ12 si group treated with F/I. Therefore, we confirmed that SUZ12 can amplify RAS signaling through the cAMP/PKA/Rap1/ERK pathway.