SV2B/miR-34a/miR-128 axis as prognostic biomarker in glioblastoma multiforme

Glioblastoma (GBM) is a heterogenous primary brain tumour that is characterised with unfavourable patient prognosis. The identification of biomarkers for managing brain malignancies is of utmost importance. MicroRNAs (miRNAs) are small, non-coding RNAs implicated in cancer development. This study aimed to assess the prognostic significance of miRNAs and their gene targets in GBM. An in silico approach was employed to investigate the differentially expressed miRNAs in GBM. The most dysregulated miRNAs were identified and analysed via Sfold in association with their gene target. The candidate gene was studied via multi-omics approaches, followed by in vitro and in vivo experiments. The in silico analyses revealed that miR-128a and miR-34a were significantly downregulated within GBM. Both miRNAs displayed high binding affinity to the synaptic vesicle glycoprotein 2B (SV2B) 3′ untranslated region (3′UTR). SV2B exhibited upregulation within brain regions with high synaptic activity. Significantly higher SV2B levels were observed in high grade brain malignancies in comparison to their normal counterparts. SV2B expression was observed across the cytoplasm of GBM cells. Our findings underscored the downregulated expression patterns of miR-128a and miR-34a, alongside the upregulation of SV2B in GBM suggesting the importance of the SV2B/miR-34a/miR-128 axis as a potential prognostic approach in GBM management.

Glioblastoma multiforme (GBM) is an aggressive and heterogeneous primary brain malignancy characterised by its invasive nature and resistance to therapy.The tumour can originate in any part of the central nervous system (CNS), but it is frequently found in the frontal or temporal lobes 1 .GBM accounts for nearly 3350 newly diagnosed cases in the UK annually with an overall survival rate between 6-17 months 2 .Currently, aggressive multimodal therapeutic approaches involving maximal surgical resection followed by radiotherapy and chemotherapy with temozolomide have failed to improve overall survival rates and patients' prognosis remains poor 3 .Immunotherapies incorporating treatment with nivolumab and pembrolizumab have shown no safety concerns and neurotoxicity, but no survival improvements have been noticed in recurrent GBM 4,5 .Therefore, there is an urgent demand for innovative therapies for GBM, due to its aggressive nature and the limited effectiveness of current treatment options.
Responsible for approximately 50% of all gliomas and 15% of all intracranial neoplasms, GBM is associated with a complex heterogenic molecular profile 6 .The clinical application of molecular profiles for GBM diagnosis is often limited due to the dynamic microenvironment of GBMs and their ability to transition between different molecular subtypes.Current diagnostic practices for GBM are based on magnetic resonance imaging (MRI).However, relying solely on MRI for a conclusive clinical diagnosis can lead to errors in some instances 7 .Therefore, the need for tissue biopsy in the diagnostic decision-making process is nearly inevitable.Nonetheless, this approach has its limitations, including the invasive nature of the procedure, the lack of serial monitoring of the malignancy, tissue heterogeneity and the potential risk of neurological damage.To address these challenges, we have turned our attention to microRNAs (miRNAs).
MiRNAs are a class of small non-coding, single-stranded RNA molecules widely found in eukaryotic organisms consisting of approximately 22 nucleotides.The process of miRNA biogenesis undergoes intricate regulation at both the transcriptional and post-transcriptional levels to produce fully functional, mature miRNAs, which play a vital role in maintaining cellular homeostasis 8 .However, alternations of miRNAs functional properties do occur occasionally and could potentially lead to a transcriptome imbalance associated with different malignancies 9 .Deregulation of miRNAs in GBM can result from various faulty events during their biogenesis, such as amplifications, deletions, epigenetic modifications, translocations, and silencing.The overexpression of certain miRNAs can lead to the suppression of tumour suppressor genes, while the downregulation of others can be associated with increased expression of oncogenes.Both scenarios could affect cell proliferation, differentiation, and apoptosis, ultimately contributing to the growth and progression of tumours 10 .The detection of miRNA signatures in primary brain tumours has provided valuable insights into the diagnosis, prognosis, and monitoring of patients 11 , but more research is necessitated to unravel their exact role in the progression of those types of tumours.
Herein, we aimed to screen for the differentially expressed miRNAs and their target genes within GBM patients to identify valid prognostic biomarkers or therapeutic targets.Specifically, in silico analysis revealed that miR-34a and miR-128a were significantly downregulated in GBM cells and patients verifying their tumour suppressor properties.Additional bioinformatics analysis demonstrated that a common target gene for both miRNAs was the synaptic vesicle glycoprotein 2B (SV2B) gene which was found upregulated in GBM patients and hence manifesting oncogenic properties.The effect of current miRNA regulation on SV2B protein function was validated using brain tumour patient samples, including GBM.Based on current findings, we are proposing that the SV2B/miR-34a/miR-128 axis could serve as a putative prognostic biomarker for glioblastoma multiforme.

miRNA analyses
Throughout our bioinformatics analysis, illustrated in Table 1, we analysed miRNAs and their gene targets using a formatted miRNA summary report, obtained from the NCBI and miRSystem databases.The summary of the hits between miRNAs and their target genes allowed us to select the four miRNAs with evident relevance within GBM pathogenesis, including miR-21-5p, miR-34a, miR-128a-3p, and miR-221-3p.Of these, miR-34a showed the highest number of hits against GBM oncogenes.Subsequent KEGG pathway analysis aided in depicting the differential expression profiles of these miRNAs (Fig. 1a).miR-34a and miR-128a were shown to be downregulated in GBM, whilst miR-21a and miR-221-3p were seen to be upregulated.Further analysis of the commonly co-expressed genes with SV2B and their miRNA regulatory pathways depicted miR-34a as the most important miRNA involved in overall gene regulation for this hub, whereas miR-128a was seen as the most important brain specific miRNA regarding the regulation of the gene hub and SV2B (Fig. 1b,c).Furthermore, common miRNAs targeting the SV2B gene including miR-128a and miR-34a were assessed by employing four databases: miRSystem, TargetScan, miRWalk, and ENCORI.The findings indicated that there were 26 common miRNA gene targets (1%) that were identified by all four databases.Within this set of shared miRNAs, miR-34a-5p was seen as downregulated and depicted as the main target for the gene (Fig. 1d).We selected the downregulated miR-34a and miR-128a for subsequent Sfold base-pair matching analysis.As illustrated in Fig. 1e, miR-34a featured a well conserved binding site for the 3′UTR of SV2B at site position 3909-3932, with a ΔGhybrid value of − 27.500.The binding type was 7mer-A1.Figure 1f-h revealed the Sfold predicted binding sites between SV2B and miR-128a.The outcome of the 3′UTR-miRNA binding depicted three site positions, including 8986-8999, 9532-9552, and 5671-5685.The hybrid binding interaction at these sites was − 17.900, − 19.300, and − 17.600, respectively.The seed interactions were described as 7mre-A1 for the first two sites, and as 6mer for the latest.

SV2B expression
Figure 2a illustrates the normal expression of the SV2B gene within various human tissues.The results demonstrated that the expression of this gene is highly elevated in the following brain regions: frontal cortex (BA9) (8.9 times), cortex (8.2 times), cerebellar hemisphere (7.2 times), anterior cingulate cortex (BA24) (7.1 times), and cerebellum (5.8 times).Furthermore, the proteomics analysis shown in Fig. 2b revealed that the median protein log 10 normalized iBAQ intensity expression of SV2B was highest within the prefrontal cortex, the retina, and the brain as a whole (6.01, 5.98, and 5.48 iBAQ, respectively).The incorporated KEGG pathway analysis as shown in Fig. 2c depicted that members of the SV2 protein family were classed as proteoglycans that were involved in extracellular matrix-receptor interactions (ECM) with laminins.Specifically, SV2B was seen to interact closely with SYT1 as part of the ECM-receptor interaction signalling.

Immunohistochemistry SV2B
The immunohistochemical analysis of SV2B performed within patient cohorts and normal brain samples revealed higher expression of the protein in all age patient groups in comparison to normal and normal adjacent brain tissue samples.However, only the age groups of 0-20 years and 61-80 years reached a significant difference regarding protein expression when compared to its normal counterparts (Fig. 4a).Figure 4b illustrates the different patterns of SV2B expression within different tumour types.The analysis depicted significant differences of SV2B levels between astrocytomas, IDH-wildtype diffuse astrocytic tumours, and oligodendrogliomas when compared to glioblastomas, the latter expressing higher protein abundance within patient samples.Medulloblastomas also possessed a significantly higher abundance of SV2B in comparison to astrocytomas.Malignant brain tumours evidently expressed higher levels of SV2B when compared to normal and normal adjacent tissue samples, reaching significance.Figure 4c depicted a significant increase of SV2B expression with increasing tumour grade.Significant differences of measured SV2B levels were seen between grades 1 and 2 in comparison to grade 4 brain tumour samples.Furthermore, the differences between grades 2 and 3 we also denoted with significance.Grades 3 and 4 were shown to express much higher levels of SV2B in comparison to lower tumour grades.As illustrated in the panel of microscopic images in Fig. 4d, the intensity of the brown colourised tissues sections increased within high grade tumours.The panels in Fig. 4e showed different patient samples with progressing tumour stages.

Immunofluorescence
The immunofluorescence analyses of stained GBM cells revealed that there was a specific localisation of SV2B within the cytoplasm of GBM cells.Figures 5a illustrated the negative control condition which did not visualise any specific binding.Βeta-actin was used as a positive control and showed similarly uniform staining across both the U87MG and U251MG cell lines (Fig. 5b,c).Figure 5d,e represented that SV2B was enriched around the nucleus, suggesting a possibility of being localised in the endoplasmic reticulum or Golgi.Staining intensity was variable across the two cell lines.The U87MG cell line exhibited more intense SV2B staining (~ 1300-2200  range) compared to the U251MG cell line which exhibited lower intensity of staining for SV2B (~ 900-2000).Despite the higher intensity, there were less SV2B U87MG positive cells and therefore lower signal frequency than in U251MG cells, where the signal intensity was lower, but staining frequency was higher.

Discussion
The limited success in achieving favourable clinical outcomes for primary brain tumours has spurred increased efforts to comprehend the role of miRNAs.Our research attempted to elucidate the role of miR-128a and miR-34a, as emerged through our analyses, and their action upon their target gene SV2B.
The conducted in silico analysis demonstrated that miR-21 and miR-221 were overexpressed in GBM tumours, whereas miR-128a and miR-34a were downregulated.This was further supported by the KEGG pathway analysis, where miR-21 and miR-221 were observed to be upregulated, whilst miR-34a and miR-128a were found to be downregulated.Given the oncogenic nature of miR-21 and miR-221, they were excluded from further analysis in relation to SV2B.
The potential therapeutic use of miR-34a, which might function as a tumour suppressor by targeting gene promoters within the p53 pathway has arisen for patients with GBM 12 .Vaitkiene et al., 2019 also demonstrated lower miR-34a expression which was associated with larger tumour volume, decreased physical functioning, and lower Karnofsky Performance Status (KPS) scores in patients with GBM.However, these findings were described as preliminary, due to the small sample size incorporated in that study.miR-34a was also found significantly downregulated within patients' serums samples highlighting its relevance as a promising non-invasive biomarker for GBM 13 .However, larger sample size validations are further required to strengthen its clinical importance.
The miR-128a is an intronic miRNA primarily located in the brain and has been previously associated with normal brain development 14 .However, miRNA assays, RT-qPCR, and western blot analyses have indicated that miR-128a is expressed at lower levels in aggressive solid brain tumours, including GBM.On the contrary, Roth, et al., 2011 demonstrated that there were elevated levels of miR-128a within blood samples 15 .Our in-silico analysis revealed that miR-128a was downregulated within glioblastoma patient samples suggesting its tumour suppressive properties.These findings were in line with previous bioinformatics analyses which had identified low miR-128 levels in serum and tissue patient samples 16 .The low levels if miR-128a were associated with high pathological grade and low Karnofsky Performance Status score (KPS).This in turn has suggested that miR-128 could act as a serum detectable biomarker in GBM patients.
Our results demonstrated a potential interaction between miRNAs and the SV2B gene.SV2B is involved in synaptic vesicle trafficking and neurotransmitter release 17 .Our findings showed that miR-128a and miR-34a featured conserved binding sites towards SV2B.However, miR-34a stood out as a potential target for SV2B, as detected in all four miRNA-databases.To the best of our knowledge, the potential prognostic value of the miR-34a/mir-128a/SV2B axis has not been reported before.Previous work has demonstrated that miR-330-3p was implicated in SV2B regulation via directly targeting SV2B mRNA and inhibiting its expression in pancreatic β-cells 18 .In glioma tissues and cells, miR-330-3p was observed to exert low expression levels compared to adjacent tissues and normal astrocytes 19 .Our in-silico analyses depicted that miR-330-3p and miR-34a targeted SV2B (Supplementary Table 1).Thus, an association between the two miRNAs and their target can potentially occur leading to the degradation of the gene.According to Morris et al., 2019 mir-134 was found to target SV2B in the brain 20 .In line with these findings, our bioinformatic analysis demonstrated that miR-134 had 167 binding sites in the 3′UTR-seedless region of SV2B (Supplementary Table 2) suggesting its high binding affinity to SV2B.In neuronal cells, miR-134 can downregulate SV2B expression, potentially affecting synaptic transmission and plasticity.The brain regional distribution between SV2B and miR-134 led to the suggestion that other brain specific miRNAs, such as miR-128a could act as targets for SV2B and thus influence its expression.
Expressed as a major form of the SV2 family, SV2B distribution across the brain was previously depicted and the gene was found to be exclusively expressed within glutamatergic neurons 21 .An immunohistochemical study has found the expression of the synaptic protein within the cerebellum, basal ganglia, and hippocampus 22 .This was in support to our findings which demonstrated that SV2B was predominantly expressed within synaptically active brain areas, such as the frontal cortex, cortex, cerebellar hemisphere, anterior cingulate cortex, and cerebellum.Given the primary function of SV2B, the protein product acts as a store of neurotransmitters and releases them upon neuronal excitation 23,24 .This has suggested that genes that align with the spatial location of GBM were associated with neuroactive ligand-receptor interactions and biological processes including the regulation of synapse organization, the modulation of chemical synaptic transmission, and interactions involving neuroactive ligands and receptors.Furthermore, genes associated with synapses, which possess postsynaptic structures and glutamate receptors, such as SV2B, have been observed to exhibit elevated expression levels in various subgroups of malignant gliomas 25 .When exposed to high levels of Ca 2+ glioma cells increase the release of glutamate into the tumour microenvironment (TME).This in turn leads to excessive glutamatergic and calcium ion signalling and promotes the spread of glioma calls 26 .Our results demonstrated that high levels of SV2B led to shorter overall survival in patients with GBM, suggesting its diagnostic value for GBM patient cohorts.The energy conductivity observed in gliomas may indicate an intensified information transmission.Conversely, it is worth noting that this abnormal and prolonged activity can result in a reduction in the physiologically normal conductivity.Despite the heightened activity in gliomas, it is important to acknowledge that their activity may not be solely dependent on SV2B.
Our STRING analysis indicated that SV2B is significantly co-expressed with genes involved in the ECMreceptor interaction, making it an important hub gene associated with signalling transduction.Previous evidence suggested that PAK-1 and SV2B might act as GBM prognostic biomarkers when studied together 27 .As apparent from our KEGG pathway and GeneMania results, SV2B had a very strong interaction and co-expression with Synaptotagmin-1 (SYT1) instead.SYT1 is situated on the vesicular membrane of nerve and endocrine cells.It acts as a primary calcium (Ca 2+ ) sensor in the mechanisms of neurotransmission and hormone secretion, and it plays a crucial role in processes triggered by Ca 2+ for secretion 28 .Given the involvement of SYT1 and its calcium-binding protein product, this gene has appeared in GBM hub oncogene nodes alongside SV2B 29,30 .Moreover, a larger scale analysis of glioblastoma and normal brain tissue samples obtained from the TCGA and GEO databases revealed that SYT1 was one of the core genes associated with GBM progression among the 552 differentially expressed genes in that analysis 31 .Nevertheless, Losada-Pérez et al. 30 identified that SYT1 gathers on the membrane of Drosophila glial cells, which in turn appears to intensify its expression within GBM samples.These statements, along with our findings suggested that the SV2B and SYT1 could act as potential prognostic biomarkers in combination for GBM patients as both oncogenes are associated with shorter overall survival.
Previous bioinformatics studies suggested that SV2B expression may be associated with glioma grade or unfavourable prognosis, suggesting its oncogenic properties 32,33 .Additionally, SV2B was found to be overexpressed within breast cancer tissue when compared to normal tissue 34 .This is in support with our IHC results, which depicted significantly high SV2B protein expression in high grade brain malignancies, such as medulloblastoma and GBM, respectively.To the best of our knowledge, no current immunohistochemical evidence of expression for SV2B within patient cohorts diagnosed with brain tumours exists.One of the possible reasons for the high accumulation of SV2B within high grade brain malignancies could be widely distributed SV2B messenger neurotransmitter within the brain.In turn, that leads to elevated intracellular Ca 2+ levels which stimulate cell proliferation and migration, leading to glioma cells acquiring their tumorigenic characteristics 35 .Additionally, SV2B has been demonstrated to be associated with energy metabolism, as the substantial energy requirements within the tumour may heavily depend on SV2B for regulating transporter protein activity, glucose transport, and other functions 36 .Our IHC findings revealed that the expression of SV2B increased with tumour grade, which suggested that highly metabolically active malignancies might require more of the protein to be able to proliferate and metastasise.GBMs and medulloblastoma represent advanced stage in the development of brain malignancies, where numerous genes are differentially expressed to varying extents.Additionally, half of the newly diagnosed high grade brain tumours are found in patients over 65 years 37 .Our results showed that patients in the age group of 61-80 years presented significantly high expression of SV2B in comparison to their normal counterparts suggesting its prognostic relevance in high grade brain malignancies.The significant overexpression of SV2B in the age groups of 0-20 years also suggested that the protein is highly expressed within paediatric brain malignancies too.Moreover, the elevated levels of SV2B found in highly malignant paediatric brain malignancy cases could be associated with their expression in brain regions characterized by pronounced synaptic plasticity.Furthermore, the observed difference between NAT and high-grade tumours is worth noting.The significant under expression of SV2B within NAT could be a result of the particular characteristics within this type of tissue that makes it a separate entity and distinguishes it from both healthy and tumour tissue.Thus, it could be suggested that SV2B could potentially act as an oncogene hence its high abundance in highly malignant brain tumours.Furthermore, the expression levels SV2B suggested that it can act as a prognostic marker, enhancing its clinical relevance in combating yet incurable highly malignant brain tumours, such as glioblastoma and 207 patients.The slides featured different brain malignancy types at different stages including astrocytoma, oligoastrocytomas (referred to as IDH-wildtype diffuse astrocytic tumours across the manuscript according to the world health organisation (WHO) classification of brain tumours, 2021), ependymomas, oligodendrogliomas, medulloblastomas, and glioblastomas, alongside normal and normal adjacent control tissue (Supplementary Table 3).The term "oligoastrocytoma" occurring in the IHC tissue microarray was used according to the old WHO classification brain tumours.This has been replaced with the term "IDH-wild type astrocytic tumours" according to the 2021 WHO classification of brain tumours which was used across this manuscript 43 .The protocol followed was previously described by Filipe et al. 44 .Post deparaffinized, rehydrated, antigen retrieval, treatment with H 2 O 2 , and blocking, the slides were incubated overnight at 4 °C with the SV2B primary antibody (1:200 dilution) (ProteinTech, Manchester, UK).Following an overnight incubation at 4 °C, the slides were incubated with a secondary anti-rabbit antibody from the Zytochem Plus HRP-DAB Kit (HRP008DAB-RB, Zytomed Systems, UK) for 1 h.Then, streptavidin-HRP conjugate was added, followed by a 45min incubation.Staining and counterstaining with DAB and haematoxylin was performed, respectively.Dehydration was performed using ethanol and HistoClear, and coverslips were applied with DPX mounting medium.Immunoreactivity was assessed by three independent viewers using a light microscope (Zeiss Microscopy, Oberkochen, Germany).The extent of brown staining was scored based on the percentage of positively stained cells.Scoring categories included: 0 (< 5% cells), 1 (5-25% cells), 2 (26-50% cells), 3 (51-75% cells), and 4 (> 75% cells).The average score from the three independent assessments was calculated.

Immunofluorescence
U87MG and U251MG glioblastoma cells were seeded on round coverslips in a 12-well plate at a cell density of 2.5 × 10 5 cells/well.The cells were grown in their respective media type for 24 h.Following that, the cells were fixed with 4% paraformaldehyde (PFA) (Sigma-AldrichTM, Dorset, UK) for 45 min.Two 5-min wash steps with phosphate buffered saline (PBS) (ThermofisherTM, CA, USA) were then performed.Blocking of the slides with 5% bovine serum albumin (BSA) (Sigma-AldrichTM, Dorset, UK) for 1 h was then carried out.The blocking solution was aspirated, and cells were incubated in primary antibody (in 5% BSA) overnight at 4 °C with gentle agitation.The dilutions for the SV2B (ProteinTech, Manchester, UK) and the Beta Actin (Bio-Rad Laboratories, California, US) primary antibodies were 1:100.Next day, the antibody solution was aspirated, and cells were washed three times with PBS for 5 min.For secondary antibody alone controls, cells were incubated in blocks without primary antibody.All cells were subsequently incubated in secondary antibody Alexa Fluor 488 Phalloidin (Applied Biosystems ThermoFisher, Pleasanton, CA) (prepared as for primary) at a dilution of 1:500 for 1 h at room temperature with gentle agitation.Cells were then washed three times with PBS for 5 min.The coverslips were then removed from the 12-well plate and excess liquid was gently dabbed off with tissue paper.A few drops of 4′,6-diamidino-2-phenylindole (DAPI) were applied to each slide, and the coverslip was gently dropped downwards at a 45° angle.The coverslips were the sealed using nail varnish.The slides were protected from light and stored at 4 °C overnight allowing the nail varnish to dry.Fluorescent images were obtained using a Zeiss Axioimager M2 microscope fitted with an Axiocam 503 imaging device using Zeiss ZEN software (Zeiss Microscopy × 40 magnification).Cell fluorescence was visualised using an EC Plan-Neofluar 63× (1.25) oil objective.The following channels were used for various fluorophores-DAPI (Ex.− 353, Em. − 465), and Alexa Fluor 488 (Ex.− 493, Em. − 517).The ZenBlue (Zeiss) software was used for image acquisition and subsequent analysis.

Statistical analysis
GraphPad Prism 9.4.1 software (GraphPad Software, San Diego, USA) was used for statistical analysis and graphical representations.Two-way ANOVA followed by Tukey's multiple comparisons test was used for three or more groups.A p-value of < 0.05 was considered statistically significant.