Nuclear Factor I A Promotes Temozolomide Resistance in Glioblastoma via Transcriptional Regulation of Nuclear Factor κB Pathway

Background: Glioma is one of the most common primary brain tumors in human with severe mortality based on its therapy resistance and recurrence. Many molecular pathways and regulation factors have been proved to be required for GBM growth and therapy resistance, however, the underlying molecular mechanisms still remains unclear. Methods: Nuclear factor I-A (NFIA) was identified as a key candidate kinase encoding gene in chemoresistance regulation by using kinome-wide bioinformatic analysis. Afterwards, the potential biological functions of NFIA in oncogenesis and chemoresistance were clarified by qRT-PCR, western blotting and in vivo xenograft models followed by temozolomide (TMZ) resistant U87 cell induction. Additionally, immunohistochemistry (IHC) assays were performed to explore the clinical significance of AURKB in glioma patients. At last, lentiviral silencing of NFIA was used to explore the potential downstream targets for NFIA in acquired TMZ resistance in GBM.. Results: We identified NFIA was the most correlated gene for TMZ resistance in GBM. Clinically, elevated NFIA expression was significantly correlated to adverse outcomes of glioma patients especially in GBM patients. Moreover, NFIA was functionally required for TMZ resistance of U87 cells while suppression of NFIA via lentivirus infection reduced cell proliferation, tumorigenesis as well as resistance to TMZ in GBM cells. Lastly, NFIA promoted acquired TMZ resistance in GBM via transcription activity thus regulated the expression of nuclear factor κB (NF-kB). Conclusions: Altogether, our study suggests that NFIA-dependent transcriptional regulation of NF-kB contributes to the acquired TMZ resistance in GBM, indicating that NFIA-NF-κB axis could be a new therapeutic target for TMZ resistant GBM.

to either chemotherapeutics or radiation, which leads to more frequent recurrence compared to low grade glioma [3]. A wide range of molecular pathways and regulation factors have been proved to be required for GBM growth and therapy resistance, which might be potential therapeutic targets for GBM treatment [4]. Therefore, it is meaningful for researchers to deeply investigate the mechanism included in the biological behaviors of GBM. GBM was often chartered by aberrant proliferation and differentiation, indicating a deregulation of the neurodevelopmental process [5]. Accumulating evidence shows that glial fate regulator, nuclear factor I-A (NFIA), is essential for both embryonic development and tumorigenesis for nervous system [6].
NFIA has been proved to be functionally required for glial lineage specification, glial progenitors and astrocyte terminal differentiation, as well as the tumorigenesis of GBM [6]. Moreover, an increased NFIA expression could be found in astrocytoma and GBM and functions as a critical component of the oncogenic network in glioma [7]. Lee et al [7] also reports that NFIA promotes cell growth and inhibits cell apoptosis of glioma through a negative regulation of multiple tumor suppressors such as p65 and p21. However, the molecular mechanisms of NFIA-induced proliferation and drug resistance in GBM are still not clearly understood.
Nuclear Factor κB (NF-κB) pathway is well known for its variety of functions on cellular responses and disease development. Abnormal activation of NF-κB pathway mediates a wide range of cellular processes related to tumorigenesis, including reduction of cell apoptosis, oncogene mutations and immune stimulation in multiple types of human cancers [8][9][10]. Activation of NF-κB in cancers is most likely though either inflammatory stimulation such as tumor necrosis factor (TNF) or upstream regulator such as NF-kB-inducing kinase (NIK), respectively [11,12]. As an important regulator in GBM, elevated NF-κB is related to poor prognosis and enhanced resistance to chemo-therapy or radiation [13]. Kim et al [14] has reported that MLK4-dependent activation of IkB kinase-α (IKKα) enhances NF-kB activity in GBM and induces the subtype transition from proneural to mesenchymal thus promotes the cell proliferation and radio-resistance of GBM. Moreover, NFIA was identified as a upstream transcripton for NF-kB via transcriptional regulation of NF-kB p65 promoter activity[6].
Herein, this study was performed to investigate the mechanism of NFIA-dependent transcriptional cells were dissociated with accutase before seeded. The number of cells was measured by using cell counter with trypan blue and was seeded with a density of 10 6 cells/10 mL.
Inducing TMZ resistance in GBM cells U87 cells were cultured in 6-well plates with DMEM-F12 containing 10% FBS at 37˚C with 5% CO 2 overnight. Cells were treated with TMZ at a starting dose of 100 μM. Medium containing TMZ was replaced every 24 h for the first 5 days continuously. TMZ concentration was added every 2 weeks for 3 months and the maintenance dose was 500μM.
In vitro cell proliferation assay Adhered GBM cells were dissociated into single cell suspension with accutase before using. Cell number was measured by using cell counter with trypan blue. Cells were seeded into 96 wells plate at a density of 1000 cells per well with 100μL fresh medium. Cell number was calculated by alamarBlue according to the manufacturer's protocol at day 0, 2, 4, 6 and 8 after seeding.

In vitro cell viability assay
Single cell suspension was seed into 96 wells plate at a density of 2000 cells/100μL per well and cultured for 12 h at 37˚C with 5% CO 2 then added 100μL of fresh medium containing TMZ at different amount. The cell number was measured by using alamarBlue according to the manufacturer's instructions. IC50 was calculated with SPSS 19.0.
Quantitative RT-PCR (qRT-PCR) Total RNA was prepared by using the RNeasy mini kit according to the manufacturer's instructions.

Western blotting
Western blotting analysis was performed as described previously [4]. Cell lysates were prepared with RIPA buffer containing protease and phosphatase inhibitor cocktail on ice then then concentrations of protein were measured by using the Bradford method. 10ug/lane of protein were fractionated on NuPAGE Novex 4-12% Bis-Tris Protein gel (Invitrogen) and then transferred to PVDF membrane (Invitrogen). Membranes were blocked with 5% skimmed milk for 1h then incubated with the primary antibody overnight at 4°C then incubated with the secondary antibody at room temperature for 1h.
Protein expression was visualized by using ECL methods according to the manufacturer's instructions (GE Healthcare Life Sciences). β-actin served as a control.

Lentivirus production and transduction
Lentivirus production and transduction was performed as mentioned in the previous study [4]. The lentivirus for shNFIA and NFIA overexpression were purchased from Genechem (Shanghai, China). The lentivirus infection was performed according to the manufacturer's protocol.

Luciferase reporter assay
After lentivirus infection, pre-treated 293T or U87 cells were seeded at a concentration of 10 6 cells per well in six-well plates. NF-κB p65 activity was determined by using the NF-κB Reporter kit (BPS Bioscience, Cat log no. 60614, San Diego, CA, USA). The attached cells were transfected with NF-κB reporter and negative control reporter for 24 h following the manufacturer's protocol. Normalized luciferase activity for NF-κB p65 reporter was measured as a ratio of firefly luminescence to Renilla luminescence. 5 replicates were used for each sample and the results were represented as mean ± SD.

Flow cytometry
Flow cytometry was performed as previously described [4]. The Alexa Fluor® 488 Annexin V/Dead Cell Apoptosis kit was used to measure U87 cell apoptosis according to the manufacturer's protocol.

In vivo intracranial xenograft tumor model
All the usage of experimental animals in this study was adheres to the Animal Research: Reporting In Vivo Experiments (ARRIVE) guidelines. 6-week-old female nude mice cultured under specific pathogen free condition (provided by the Experimental Animal Centre of Xi'an Jiaotong University) were used for in vivo xenograft of GBM cells. Prepared GBM cell suspension (pre-transduced with shNT or shNFIA lentivirus) was diluted to the density of 1 × 10 5 cells in 2 μL PBS then implanted into the mice brains as previously described [4,17]. 5 mice were used for each group. TMZ (50mg/kg/d) or DMSO was taken by tail vein injection after 7 days injection of glioma cells. Mice were monitored once a day until the following symptoms appeared: arched back, leg paralysis, unsteady gait or bodyweight loss for more than10%. When neuropathological symptoms developed, the mice were anaesthetized and sacrificed with overdose Ketamine (80mg/kg) and Xylazine (20mg/kg).

Statistical analysis
All the results in this study are presented as mean ± SD. Number of replicates is mentioned in the related figure legends. Statistical differences between 2 groups were evaluated by using 2-tailed t tests. Multiple groups were compared with one-way ANOVA followed by Dunnett's posttest. Kaplan-Meier survival analysis was compared by log-rank analysis. All statistical analysis was performed with SPSS 19.0 or GraphPad Prism 6 software. Statistically significance was considered when P value was less than 0.05.

Enriched NFIA expression could be related to TMZ resistance in GBM
To deeply explore the key regulator and related molecular mechanism of acquired TMZ resistance in glioma, hierarchical bi-clustering based on the previously published microarray databases (Tso et al [15], GSE 68029, 2015 and Mao et al [16], GSE67089, 2015) then the genes were ranked according to their fold changes ( Figure 1A and Figure 1B). Genes with more than 5 fold up-regulation were picked up from these 2 databases. Finally 105 genes were identified for TMZ resistance characters as well as 452 genes for GBM compared to astrocytes. Then we merged the 2 gene lists together and found that NFIA was the only gene which was significantly enriched in both 2 phenotypes ( Figure 1C).
Additionally, we confirmed that NFIA was significantly increased in TMZ resistant glioma cells compared to their naïve control lines ( Figure 1D), demonstrating that NFIA could be essential for glioma cells to gain TMZ resistance. Moreover, an elevated NFIA expression could be observed in GBM, which was considered as the most lethal type of glioma ( Figure 1E). Taken together, these results indicated that NFIA was elevated in GBM and might be essential for therapy resistance and tumor recurrence.

Increased NFIA expression was associated with poor outcomes in GBM patients
Given the results from the previous data, it raised up the question whether NFIA could be a prognostic marker for glioma. To this end, IHC was performed to examine NFIA expression in 68 glioma tumor tissues which were collected from patients underwent surgical therapy from 2008 to 2017 in the First Affiliated Hospital of Xi'an Jiaotong University. As a result, NFIA was found to be expressed in the nucleus of glioma cells (Figure 2A). GIS was used to determine the expression level and NFIA was found to be markedly enriched in GBM, contrarily to low grade glioma samples (Figure 2A and Figure   2B). We next assessed the expression of NFIA in GBM by analyze TCGA database and Rembrandt database. The results demonstrated that NFIA was elevated in GBM compared to normal brain tissue, which were similar with our study ( Figure 2C and Figure 2D). When look into the survival for those patients, patients with lower NFIA expressed glioma represented longer overall survival compared to those with higher NFIA expression ( Figure 2E). Similar results were achieved when we specifically focused on GBM patients ( Figure 2F). Moreover, an analysis of overall survival was performed among 541 glioma patients from Rembrandt database and the results showed that the post-surgical survival for the patients with low NFIA expression was significantly prolonged than that in the patients with increased NFIA expression ( Figure 2G). Altogether, the results showed the possibility that NFIA is supposed to be a specific clinical relevant oncogene for glioma and GBM.

NFIA was functionally required for TMZ resistance in GBM
To investigate the function of NFIA in the acquirement of TMZ resistance in GBM, an exogenous overexpression of NFIA was performed in U87 cells by using lentivirus infection. qRT-PCR and western blotting analysis confirmed that NFIA expression was markedly increased in NFIA overexpressed U87 cells ( Figure 3A and 3B). In vitro cell growth assay indicated that the proliferation and TMZ resistance was obviously enhanced after NFIA overexpression ( Figure 3C). Moreover, U87 cells with or without NFIA overexpression were treated with TMZ thus the apoptosis analysis was performed. The results showed increased TMZ resistance after NFIA overexpression ( Figure 3D), demonstrating that NFIA was functional required for acquired TMZ resistance in GBM.

Suppression of NFIA enhanced the TMZ sensitivity of TMZ-resistant GBM
To thoroughly study the functional role of NFIA in acquired resistance to TMZ in GBM, we established in vitro TMZ resistant U87 cell line according to the previous publications [18,19]. After 3 months culturing with TMZ-contained medium, U87 cells gained stable resistance to TMZ ( Figure 4A). qRT-PCR analysis was performed among the TMZ resistant GBM cell lines and their naïve control lines. The results indicated dramatically increased NFIA expression in TMZ resistant U87 cells ( Figure 4B). Moreover, western blotting results also showed up-regulated NFIA in TMZ resistant population of U87 cells ( Figure 4C).
For further assessment of the functional role of NFIA, TMZ resistant U87 was transduced with a lentiviral shRNA clone for NFIA (shNFIA) or a non-targeting lentivirus (shNT) as a negative control.
Both qRT-PCR and western blotting analysis showed dramatic down-regulation of NFIA at mRNA level ( Figure 4D and 4E). Additionally, to test the function of NFIA on TMZ resistance, U87 TMZ resistant cells treated with or without NFIA suppression then exposed to TMZ treatment at 300μM. In vitro cell growth assay exhibited decreased cell proliferation and increased TMZ sensitivity in U87 TMZ resistant cells transduced with shNFIA lentivirus ( Figure 4F). Similarly, flow cytometry assays for apoptosis were performed with shNFIA or shNT pre-transduced U87 TMZ resistant cells followed with or without TMZ treatment (300μM). The proportions of cells that undergoing early and late apoptosis were both dramatically increased when cells received combined treatment of TMZ and NFIA silencing compared with TMZ alone ( Figure 4G). Next, we investigated the function of NFIA knock-down on in vivo tumorigenesis by using mouse intracranial tumor models. The results indicated that the control mice with xenografts of shNT-transduced U87 TMZ resistant cells rapidly represented tumor-related symptoms compared with those transplanted with shNFIA-transduced U87 TMZ resistant cells combined with TMZ treatment (Figure 4G), highlighting a potent anti-TMZ resistance effects of NFIA knock-down in TMZ resistant GBMs.

NFIA-dependent transcriptional regulation of NF-kB contributes to the acquired TMZ resistance in GBM
Previous study showed NFIA contributes to tumor progression through regulating NF-κB expression. In our study, we found that NF-κB expression was significantly increased in TMZ resistant U87 cells compared to the original non-treated U87 ( Figure 5A and Figure 5B). Furthermore, inhibition of NFIA induced reduction of NF-κB in TMZ resistant U87 ( Figure 5C and Figure 5D). Altogether, these data suggested suggests the presence of a tumor-promoting regulation between NFIA and NF-κB in GBM.
We then performed a luciferase reporter assay with constructs driven by a human NF-κB promoter. As expected, overexpression of NFIA leaded to a significant increase of NF-κB promoter activity in both 293T and U87 cells ( Figure 5E). Contrarily, shRNA-mediated-knockdown of NFIA resulted in a marked decrease in transcription activity of NF-κB promoter region in U87 cells, especially in TMZ resistant U87 cells which exhibited a higher expression of NF-κB ( Figure 5F). These data suggest that the NFIAdependent transcriptional regulation of NF-kB contributes to the acquired TMZ resistance in GBM.

Discussion
Accumulating data demonstrates that acquired resistance to radio therapy and chemotherapy is essential for the recurrence and lethal mortality of GBM [20]. Multiple mechanisms have been identified to be functionally for GBM cells to gain therapy resistance [4,21]. Our findings here indicates that NFIA promotes TMZ resistant in GBM via a transcriptional regulation of NF-κB thus contributes to poor prognosis and recurrence for GBM patients, suggesting that NFIA-NF-κB axis could be a new therapeutic target for TMZ resistant GBM.
The NFI family of site specific DNA-binding proteins, which includes NFIA, NFIB, NFIC, and NFIX, was Our study identified that NFIA was highly expressed in GBM compared to normal brain, moreover, was significantly enriched in TMZ resistant GBM. Additionally, high expression of NFIA implied poor prognosis of glioma patients. Similar results could be observed when we look into databases such as TCGA and Rembrandt in which more patients are included, providing a potential therapeutic target for GBM treatment. Due to the limitation of the in vitro research and small amount of data, it still remains to be seen whether these findings could be extended to the more complex in vivo situations. Also, an analysis of a larger cohort is planned to strengthen the conclusion.
NF-κB p65 is a well-recognized anti-apoptotic transcription factor and abnormal expression or activation of NF-κB p65 has been found in multiple types of malignant cancers such as cervical cancer, esophageal squamous cancer, ovarian cancer, breast cancer and glioma et al [13,[26][27][28][29]