The Quantitative Proteomics Analysis in Glioblastoma After HULC Silencing


 Background: Glioblastoma multiforme (GBM) is a malignant intracranial tumor threatening patients' survival. The study aimed to find the menchanisms of lncRNA highly up-regulated in liver cancer (HULC) affecting GBM or involved signaling pathways, which may providing theoretical support for targeted therapy.Methods: Two cell lines were constructed: HULC-small interfering RNA (siRNA) and negative control. Then qRT-PCR was operated to detect the transfection efficiency and quantitative proteomics based on mass spectrometry (MS) were conducted. Finally, the differentially expressed proteins were analyzed in gene ontology (GO) classification and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment.Results: The relative expression level of HULC in HULC-siRNA was significantly lower than that in NC detected by qRT-PCR. A total of 136 differentially expressed proteins were identified, including 24 down-regulated and 112 up-regulated. GO classification results illustrated that they were centralized in cellular or single-organism process and biological regulation in biological process, cell and organelle in cellular component, binding and catalytic activity in molecular function. KEGG pathway enrichment analysis indicated that some were sinificantly enriched, including tight junction, metabolic pathways and arachidonic acid metabolism. It was further discovered that PLA2G4A was down regulated obviously after HULC silencing.Conclusion: HULC changed the proteomics characteristics of GBM, including regulating the expression of PLA2G4A. This study provided a new perspective on the menchanisms and potential drug targets of GBM.

The present study detected differentially expressed proteins by taking advantage of highly sensitive quantitative proteomics, combined with bioinformatics analysis to nd meaningful targets or signaling pathways involved in pathogenesis between HULC and GBM, which may provide some ideas or directions for targeted therapy of GBM.

Cell culture
Human glioma cell line U87 was obtained from China Center for Type Culture Collection (Wuhan, China) and cultured with DMEM (BD, Thermo Fisher Scienti c, USA) adding 10% FBS (BD, Thermo Fisher Scienti c, USA) in a incubator (Thermo Fisher Scienti c, USA) containing 5% CO 2 at 37°C.

Construction of transfected stable cell lines
After digested, resuspended and counted, cells were plated on six-well plates at a concentration of QRT-PCR RNA was extracted from two samples with total RNA extraction kit (QIAGEN, Germany). cDNA was synthesized according to the protocol of high-throughput cDNA reverse transcription kit (Thermo Fisher Scienti c, USA). Next, polymerase chain reaction was performed. The relative mRNA levels of target genes were obtained by the 2 -ΔΔCt method and determined in triplicate. Upstream sequence of target gene HULC primer: 5'-TCAACCTCCAGAACTGTGATCC-3', down stream sequence: 5'-TGCTTGATGCTTTGGTCTGTT-3'; Upstream sequence of reference gene ACTB primer :5'-CGTGGACATCCGCAAAGA-3', down stream sequence: 5'-GAAGGTGGACAGCGAGGC-3'.

Protein extraction
Samples were sonicated in 4 times volumes of lysis buffer [8 M urea (Sigma,UAS), 1% protease inhibitor (Calbiochem, Germany) and 2 mM EDTA (Sigma USA)]. The supernatant was collected and the protein concentration was determined with BCA kit (Beyotime, China) according to the manufacture's instructions.

Trypsin digestion
The protein solution was reduced with 5 nM dithiothreitol (Sigma, USA) for 30 minutes at 56°C and incubated in 11 nM iodoacetamide (Sigma, USA) for 15 minutes in the dark at room temperature. Then trypsin (Promega, USA) was added at a ratio of 1:50 trypsin to protein at 37°C overnight and 1:100 for 4 hours.

TMT labeling
The digested peptides were desalted with Strata X C18 SPE column (Phenomenex) and freeze-dried in a vacuum environment. Next they were dissolved in 0.5 M NH 4 HCO 3 (Sigma, USA) and labeled with TMT kit (Thermo Fisher Scienti c, USA) according to the protocol.

LC-MS/MS analysis
Peptides were dissolved with liquid chromatography mobile phase [aqueous solution contaning 0.1% formic acid (Fluka, USA) and 2% acetonitrile] and seperated on an EASY-nLC 1000 ultra-high performance liquid chromatography system. Later they were injected into the NSI ion source for ionization and analysis was performed by Orbitrap Fusion Lumos MS. Both the peptide precursor ion and its secondary fragments were detected and analyzed with high-resolution Orbitrap. According to the ratio of mass to charge, they were seperated at the rst-stage MS scan. Finally, second-stage separation was performed.

Bioinformatics analysis
Data of second-stage MS was searched in Maxquant (v.1.5.2.8), relevant parameters as showed in Table   1. The quantitative values of each sample were obtained from 3 repeated total proteins quantitative detections. And the relative standard deviation (RSD) was calculated to assess the degree of sample dispersion. Taking the average of 3 quantitative values, fold change was de ned as the ratio of 2 averages (HULC-siRNA to NC). The values of log2 fold change were needed to make them consistent with normal distribution and evaluated by two-tailed t-test. Differentially expressed proteins were ltrated based on the following criteria where fold change was more than 1.2 or less than 1/1.2, and P value was less than 0.05 simultaneously. GO annotation of every differentially expressed protein was obtained by searching Uniprot-GOA database or applying Interproscan (v.5.14-53.0), an algorithm software. Moreover, KEGG pathway database (KAAS v.2.0 KEGG Mapper v2.5) was utilized to annotate corresponding pathways. The difference in GO enrichment or KEGG pathway enrichment was further evaluated by twotailed Fisher's exact test (Perl module v.1.31), where corrected P value <0.05 was considered statistically different.

Results
The detection of expression of HULC in two stable cell lines by qRT-PCR The relative expression of HULC in NC was 1.05±0.101, and 0.680±0.0890 in HULC-siRNA (t=4.82 P=0.00860). So it was clear that the relative expression of HULC in HULC-siRNA was signi cantly lower than it in NC (Fig. 1A), which illustrated the successful construction of stable cell lines.

Statistics of differentially expressed proteins
Since the overall RSD values of two sample groups were pretty small, protein samples obtained from LC-MS/MS analysis harbored high repeatability and quality (Fig. 1B). Quantitative values of proteins in HULC-siRNA were compared with those in NC. When P value < 0.05, fold change > 1.2 was regarded as up regulation. Conversely, fold change < 1/1.2 was regarded as down regulation. A total of 112 up-regulated proteins and 24 down-regulated proteins were detected (Table 2). Then the volcano plot was made with log2 fold change as the abscissa, and -log10 P value as the ordinate (Fig. 1C).

GO annotation and enrichment
GO annotation is classi ed into 3 categories (biological process, cellular component, molecular function) to explain the biological role of each protein from different perspectives. Two charts illustrated the distribution of differentially expressed proteins in GO terms level 2 ( Fig. 1D and 1E). Taken together, it was clear that dysregulated proteins were mainly involved in cellular or single-organism process and biological regulation under the classi cation of biological process, making up cell and organelle under the classi cation of cellular component. And as for molecular function, they were concentrated on binding and catalytic activity. Moreover, Fisher's exact test was applied to perform GO enrichment analysis. Directed acyclic graphs re ected not only the enrichment difference of GO classi cation, but also the hierarchical relationship intuitively. It could be concluded that signi cantly different GO classi cations were mostly enriched in binding, metabolism, regulation, and catalytic functions. And they were also enriched at deeper and more detailed levels of grade ( Fig. 2 and 3).

KEGG pathway annotation and enrichment
All differentially expressed proteins were annotated with KEGG pathway. Furthermore, in order to re ect whether there was any signi cant enrichment trend of those proteins in annotated KEGG pathways, the enrichment test was performed with Fisher's exact test and P values were presented as -log10 conversion.
It was exhibited that tight junction was most enriched in up-regulated pathways after the interference of HULC, while down-regulated KEGG pathways preferred metabolic pathway, arachidonic acid metabolism, terpenoid backbone biosynthesis and platelet activation (Fig. 3).

Discussion
GBM is the fourth grade of glioma with the highest malignancy. Increasing evidence demonstrates the dysregulation of many kinds of lncRNAs is involved in a series of biological processes in the occurrence and development of GBM. For example, highly expressed PVT1 could improve the proliferation, invasion and aerobic glycolysis in glioma cell by inhibiting the expression of miR-140-5p [17]. Another lncRNA GAS5-AS1 could bind miR-106b-5p to promote the expression of downstream genes and thus play a role in inhibiting the proliferation, migration and invasion of glioma [18]. Studies are emerging quickly on the mechanisms of how lncRNA in uence other tumors [19,20], aimed to search for highly speci c and sensitive biomarkers to help make early diagnosis, predict prognosis, and provide potential therapeutic targets of cancers.
LncRNA HULC has been proved to harbor a signi cantly higher level in GBM cells than normal cells, and promote the proliferation of GBM in vitro [11]. Yu Zhu et al. have found that HULC silencing could inhibit angiogenesis of glioma through ESM-1 mediated PI3K/AKT/mTOR signaling pathway, resulting in the suppression of GBM growth [12].
Proteomics research is currently in full swing. Islam Farhadul et al. studied and analyzed the differences of total proteome between esophageal squamous cell carcinoma and non-tumor cell through quantitative proteomics based on MS [13]. Qingchuan zhao et al. screened tumor-speci c antigens for high-grade serous ovarian cancer with MS, which might become suitable targets for ovarian cancer immunotherapy [14]. In this study, we have obtained the proteome of HULC-siRNA and NC through TMT labeling, HPLC fractionation technology and LC-MS/MS analysis, hoping excavate some promising biomarkers.
Among all differentially expressed proteins, the top 5 of up regulation or down regulation were analyzed in detail respectively. In PubMed database, none of the above 10 proteins were reported to be related to HULC. However, there have been some studies between CRYAB or SASH1 and glioma [21,22]. No related reports have been found on the remaining 8 proteins associated with glioma till now. And other proteins were not analyzed this time. Undoubtedly, deeper learning of these proteins is helpful for further study in molecular mechanisms.
GO database is applied to describe various properties of genes and gene products. GO annotation demonstrated that differentially expressed proteins were primarily involved in cellular biological processes and constituting cell and organelle structures. More importantly, they played a role in molecular binding and catalytic activities. The above could make some in uence to the proliferation, migration of GBM, which undoubtedly re ected the crucial in uence of HULC on GBM. The trend of GO enrichment was consistent with the results of secondary GO classi cation. In addition, it explained some crucial processes of how HULC microscopically regulated GBM in depth, such as calcium-dependent phospholipid binding, methylase activity, regulation of angiogenesis and so on. As we all know, DNA methylation is an early event of tumorigeneisis. For example, the methylation level of the MGMT promoter in gliomas might be associated with prognosis and recurrence [23]. Moreover, the methylation of SASH1 was conducive to weakening cell adhesion [22]. Thus it was considered that the methylation of HULC could be put into following study.
KEGG is an information network describing the interactions between known molecules, such as information transmission, protein-protein interaction, and biochemical reactions. KEGG pathway mostly includes metabolism, cell growth, genetic or environmental information processing, human diseases, drug development and so on. In the present study, it was revealed that tight junction had a strongest enrichment under the up-regulated pathways, which further implied that the interference of HULC could improve cell adhesion and suppress migration of GBM. At the same time, terpenoid backbone biosynthesis presented obvious difference in down-regulated pathways, indicating that HULC silencing might act on inhibiting cell proliferation. In addition, down-regulated pathway of platelet activation prompted that HULC was related to the complication of GBM like thrombosis to some extent.
Comparing some down-regulated pathways (arachidonic acid metabolism, platelet activation), we have found that cPLA2α was one of the common proteins and might act as a crucial role. cPLA2α, encoded by PLA2G4A gene, is the most abundant subtype in the family of phospholipase A2. Phospholipase can hydrolyze membrane phospholipids to arachidonic acid, which is further involved in many Pathological and physiological processes, including inducing in ammation, transmitting signals, and promoting cell growth, etc. Although one study has shown that down expression of PLA2G4A can promote the migration and invasion of esophageal squamous cell carcinoma [24], most researchers considered that PLA2G4A was an oncogene [25][26][27], which was consistent with our results. The data showed that the expression level of PLA2G4A notably decrease after HULC silencing. Thus, it was suspected that PLA2G4A was likely to be one of the key point of how HULC acted on GBM, which also echoed arachidonic acid metabolism, the most signi cant pathway in down-regulated category. Since the concept of tumor-promoting in ammation was proposed in 2011 [28], tumor-associated chronic in ammation is a key trigger point for cancer progression, we can assume that the in ammation induced by arachidonic acid metabolism pathway is an important participant in the development of GBM. Undoubtedly, more speci c mechanisms need further veri cation and research.

Conclusions
We concluded that HULC had changed the proteomics characteristics of GBM. And PLA2G4A was prominantly regulated by HULC in GBM cells. This study provided a new perspective on the pathogenesis of GBM, and a starting point of GBM targeted therapy. Nevertheless, the number of differentially expressed proteins detected this time was relatively small, which stimulated more research for quantitative proteome assay on cell models or tissues of over-expressive HULC.  The GO enrichment of up-regulated proteins were analyzed by Fisher's exact test, which presented the obvious enrichment in binding, metabolism, regulation, and catalytic functions.