Background
Glioma originated from the neuroectoderm and accounts for approximately 81% of primary malignant brain tumors [
1]. Glioma is graded I-IV according to the 2016 World Health Organization classification of central nervous system tumors. Glioblastoma is the most common and malignant subtype and the dismal 2-year and 5-year survival rate are < 40 and < 10%,respectively [
2]. Low grade gliomas (WHO I-II) usually progress to higher grade and eventually have poor outcomes despite treating with standard care (surgical resection combined with postoperative radiotherapy and chemotherapy) [
3]. Therefore, it is urgent to understand the key molecular mechanisms in the malignancy progression of glioma to develop more effective treatments.
Hypoxia is a common pathological feature in glioma. Increasing grade of gliomas correlate with an increase in absence of oxygen [
4]. Chronic hypoxia often leads to necrosis in tumor tissues, which is one of the most distinct characteristics of glioblastoma. Hypoxia microenvironment promotes glioma cells aggressive phenotype by upregulating hypoxia-inducible factor (HIF) family [
5‐
7].HIF1A is highly expressed in glioblastoma and significantly correlated with IDH1/2 mutation [
8]. Moreover, previous studies have demonstrated that hypoxia environment could induce the epithelial-mesenchymal transition (EMT) during the progression of glioma by regulating several pathways, such as Wnt/β-catenin [
9], transforming growth factor β (TGF-β) [
10] and Sonic Hedgehog (SHH) pathway [
11]. This phenotype transition facilitated glioma cells easier to infiltrate the adjacent brain tissues and more resistance to chemo/radiotherapy [
12]. However, the specific mechanism underlying which hypoxia promoting glioma malignancy remains to be further illustrated.
Ferritin was consisted of 24 units of heavy chain (FTH) and light chain (FTL). FTL had been widely recognized as one of the iron metabolism regulators for a long time. While in recent years, a growing number of studies have revealed the close relationship between FTL and tumor malignancies [
13‐
16]. It was revealed that FTL could be used as a biomarker to discriminate benign and malignant tumors, and to predict the prognosis of patients with tumors [
17,
18]. Besides, FTL was found to be overexpressed in various malignant tumors, and played a crucial role in regulating malignancy progress of cancers [
19,
20]. Recent studies revealed that FTL could be upregulated on the post-transcriptional level by hypoxic conditions. Alveolar macrophages had 2.5 folds content of FTL when cells were exposed to hypoxia. Similar results were also observed in another study conducted by Sammarco et. HEK 293 cells were cultured in 1% oxygen and the results showed that FTL and FTH were differentially upregulated [
21,
22]. These results suggested that FTL might be regulated under the hypoxia environment. While the effect of hypoxic environment on FTL expression and its regulation in process of glioma malignancy have not been well investigated so far.
In this study, we found that FTL was higher in HGG than in Low grade glioma (LGG). High FTL expression closely associated with wildtype IDH 1/2 and poor prognosis. We showed, for the first time that FTL was a hypoxia-responsive gene that significantly elevated under hypoxia in a time-dependent manner in U87 and U251 cells. Further analysis revealed that HIF-1A regulates FTL expression by directly binding to HRE-3 in the FTL promoter region. Functionally, we showed that FTL induced the EMT and promoted migration, invasion and chemo-resistance of glioma both in vitro and in vivo. Mechanistically, oncogenic role of FTL was functioned by regulating AKT/GSK3β/ β-catenin signaling.
Materials and methods
Clinical samples
Glioma tissues were obtained from the department of neurosurgery in Renmin hospital of Wuhan University from July 2015 to July 2018.A total of 142 paraffin-embedded glioma tissue were used for immunohistochemical staining. For western blot,28 glioma frozen tissues (stored at − 80 °C) of different grades were evaluated. Details of clinical information for all patients was presented in Table S
1. None patients received any chemo- or radiotherapy before surgery. All patients signed informed consents and this study received the approval of the Ethics Committee of Renmin Hospital of Wuhan University (approved number: 2012LKSZ (010) H).
Immunohistochemical (IHC) staining
The paraformaldehyde-fixed paraffin tissue microarray that contained 142 glioma tissues was used. The microarray was incubated with a primary anti-FTL monoclonal antibody (Abcam, USA: Ab109019) overnight at 4 °C. Images were captured using an Olympus BX40 microscope (Tokyo, Japan). Two individuals were separately responsible for the assessment of the results. The result was primarily based on the strength of staining and the number of positive cells.10 high magnification fields were randomly selected for observation. Positive staining rate was scored as: 0 points for less than 5%, 1 point for 5–25%, 2 points for 26–50%, 3 points for 51–75% and 4 points for 75%. Besides, the classification of the strength of staining was followed: non-staining is 0 points, light yellow is 1 point, brown yellow is 2 points, and brown is 3 points. Finally, multiply the two scores to get the final score which would be graded into 4 grades: negative (0 point), weakly positive (1–4 points), positive(5–8 points) and strongly positive(9–12 points). Final score less than 5 was defined as low expression and IHC score 5–12 was considered as high expression.
Immunofluorescence staining
Cells were fixed with 4% paraformaldehyde for 15 min and penetrated with 0.5% Triton X-100 (made in PBS) at room temperature for 10 min. Then slides were washed with PBS 3 times. 1%BSA was added dropwise on the slides at room temperature for 30 min. Then we added a sufficient amount of the diluted primary antibody and placed it in a wet box, incubate at 4 °C overnight. Incubation with secondary antibody (Antgene, Wuhan, China) was performed in a wet box at 37 °C for 1 h under dark conditions. DAPI (ANT046, Antgene) was added in the dark for 5 min. Finally, slides were observed under a fluorescence microscope (Olympus BX51, Japan) to acquiring images.
Cells, cell culture and transfection
Two human glioblastoma-derived cancer lines, U251 and U87 were purchased from the Cell Bank Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). Cell lines were identified by Procell Life Science&Technology Co.,Ltd. (Wuhan, China). Cell lines were all cultured at 37 °C under a humidified atmosphere of 5% CO2 by using Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Gibco, Invitrogen, Carlsbad, CA, USA). All cell lines were cultured without antibiotics. Short hairpin RNA (shRNA) targeted FTL and a scramble shRNA were purchased from Genechem Co., Ltd. (Shanghai, China). The target sequences against human FTL(5′-GGCGA GTATCTCTTCGAAA-3′) and scrambled shRNA(5′-TTTCGAAGA GATACTCGCC-3′) were cloned into the GV248 lentiviral vector. U251 and U87 cells were transfected with Lentivirus for 72 h and treated with puromycin(4 μg/ml) for 7 days.The specific small interfering RNA (siRNA) for FTL (siG143101050 18–1-5),HIF1A(siG0811494537–1-5), CTNNB1(siB08220115 751-1-5), HIF2A (siG170217101514–1-5) and scramble siRNAs were purchased from RiboBio (Guangzhou, China). Cells were cultured in a 6-well plate and transfected with lip2000(Invitrogen, Carlsbad, CA, USA) following the instructions of the manufacturer. After 48 h of transfection, the cells were used for subsequent experiments. FTL overexpression plasmid and a blank pcDNA3.1 vector were constructed.2 × 105 cells were cultured in a 6-well plate and transfected with 2 ng plasmid using lip3000(Invitrogen, Carlsbad, CA, USA) following the instructions of the manufacturer. The cells were harvested after transfected for 48 h for further experiments. For in vitro hypoxia experiments, cells were cultured in a consistent 1% O 2 hypoxic condition. The hypoxia mimetic cobalt chloride (CoCl2) (Sigma, NO.232696) was dissolved in sterile PBS and the final concentration of CoCl2 in the medium was 200 uM. To monitor resistance to temozolomide (TMZ), the U87 and U251 cells were treated with TMZ (Selleck, NO.S1237) at various concentrations for 24 h.
To clarify the expression and prognostic role of FTL in gliomas, we used the Gliovis database (
http://gliovis.bioinfo.cnio.es/) and the UCSC Xena platform (
http://xena.ucsc.edu/). Normalized RSEM gene-level RNAseq and corresponding clinical data of The Cancer Genome Atlas (TCGA),Rembrandt and IVY dataset were downloaded from Gliovis. Specific information on postoperative treatments (chemo/ radiotherapy) of glioma patients was downloaded from UCSC Xena platform. Besides, normalized mRNA expression (mRNA-array_693, (batch 1)) and clinical data were downloaded from Chinese Glioma Genome Atlas (CGGA). Low grade glioma was defined as WHO grade I-II and High grade glioma was defined as WHO III-IV according to the 2016 World Health Organization classification of central nervous system tumors [
23].
RNA isolation and RT-PCR
Total RNA of U251 and U87 cell lines were extracted using TRIzol regent (Invitrogen). We used PrimeScript RT reagent kit with gDNA Eraser (Takara, Tokyo, Japan) to prepared for cDNA and real-time PCR was performed by using SYBR Green II Mixture (TaKaRa) according to the manufacturer’s protocol. GAPDH was used for normalization and the comparative Ct method (ΔΔCt) was used to evaluate mRNA expression. The specific primer pairs were as follows: GAPDH (internal control gene) primer (forward primer, 5′-ACAACTTTGGTATCGTGGA AGG-3′; reverse primer, 5′-GCCATCACGCCACAGTTTC-3′); FTL primer (forward primer, 5′-CAGCCTGGTCAATTTGTACCT-3′; reverse primer, 5′-GCCAATTCG CGGAAGAAGTG-3′).
Western blot
U251 and U87 were lysed in a modified RIPA buffer
(NO.P0013B, Beyotime Biotechnology, China) on ice for about 30 min, then centrifuged at 12,000 rpm for 15 min. For frozen glioma tissues, we added 1 ml of RIPA lysate per 100 mg of tissues. The concentration of the sample was quantitatively determined by BCA protein assay. The lysate was mixed with loading buffer after heated at 100 °C for 5mins.In brief, equal protein amount was loaded on 8–12% SDS-PAGE and then transferred to a nitrocellulose membrane. Next PVDF membrane was blocked in 5% non-fat milk for 1 h and incubated with primary antibody at 4 °C overnight. Secondary antibodies (Antgene,Wuhan,Chian,1:10000) were used to incubate the membrane in shade environment at room temperature for 1 h.The membranes were visualized with Odyssey (LI-COR biosciences, USA). Primary antibodies used were presented in Table S
2.Western blot analysis were repeated three times.
Wound healing and transwell assay
Cells were seeded in a 6-well plate and cultured for a certain time to reach a > 90% confluence. The sterile pipette tip was used to scratch a linear wound and serum free DMEM was added for further culturing. Wound healing images were captured using an inverted microscope (Olympus BX51, Japan) and ImageJ software was used to analyze relative area of wound closed. For transwell assay, appropriated glioma cells were seeded into the upper well (Corning, USA) precoated with Matrigel (R&D, USA). The lower chamber was filled with 600 μl of DMEM containing 10% FBS. Transwell chambers were placed in an incubator (37 °C,5% CO2) for 24 h. Cells in the upper chamber were fixed with 4% paraformaldehyde for 15mins, stained with 0.1% crystal violet for 15mins and counted under an inverted microscope (Olympus BX51, Japan). We randomly selected 6 fields to count the number of invading cells in each set of experiments. All assays were repeated 3 times.
1000 glioma cells were counted and seeded in 6-well plates. Cells were cultured with DMEM supplemented with 10% FBS and then Temozolomide was used to treat cells for 24 h.Cells were continued to be cultured in complete medium for about 10 days. Clones that contained more than 50 cells were scored. The clone formation rate was defined as the number of scored clones divided by the total cells seeded. For CCK8 assay,3000 cells were resuspended in 100 μl DMEM supplemented with 10% FBS and then added to a 96 well plate. Various concentrations of temozolomide were added. Cell proliferation was investigated using CCK8(Dojindo Molecular Technologies, USA) according to the manufacturer’s instruction.
Flow cytometric analysis
Cells were seeded in a 6-well plate and treated with temozolomide(400 μM) for 72 h. Annexin V-PE/7- ADD kit (Becton Dickinson, USA) were used to measure the apoptosis of glioma cells. All operations were carried out according to the manufacturer’s instruction. In briefly, cells were harvested and washed three times with PBS. Then cells were stained with Annexin V-PE/7- ADD for 10 min under dark conditions. The apoptosis of samples was measured by FACS Calibur flow cytometer (BD Biosciences, USA). Early apoptosis and late apoptosis were summed and the total apoptosis rate was calculated.
TUNEL assay
In Situ Cell Death Detection Kit was used to detected DNA fragmentation in apoptotic cells in xenografts according to the manufacturer’s protocol (Roche). In short, sections were deparaffinised at 60 °C for 20 min on a heat block and then incubated in xylene (3 × 5 min). Tissue was then rehydrated by washing in graded alcohol, 3 min for each, after which they were rinsed in PBS three times. After treated with 0.1%Triton X-100 and Proteinase K, the sections were incubated with TUNEL reaction mixture and incubated with converter-POD. Subsequently, DAB was used to stain slides. An Olympus BX51 microscope (Olympus) was used for image acquisition.
Luciferase assays
To investigate the role of interaction between HIF1A and FTL, we constructed FTL promoter-driven luciferase reporter plasmids and transfected into U87 and U251 cells. Cells were pretreated with si-HIF1A or scramble for 48 h and then exposed to 1% O2 for 24 h.To further determine the direct binding between HIF1A and HREs in FTL promoter, we constructed mutant luciferase plasmid with ablation of HREs on FTL promoter by changing 5′–GCGTG-3′ to 5′–GCTCT-3′ and then co-transfected with plasmid containing the Renilla luciferase gene. Firefly luciferase activities were normalized using Renilla luciferase. Besides, TCF/LEF luciferase reporter (No.11542ES03) assay was performed using the Dual Luciferase Reporter Gene Assay Kit (Yeasen Biotech Co., Ltd. Shanghai, China) according to the manufacturer’s protocol. All experiments were repeated three times.
Chromatin immunoprecipitation (ChIP)
ChIP was performed to explore the potential binding between the promoter region of FTL and HIF1A in glioma cells. U87 cells were incubated under the hypoxic condition for 24 h.Antibody against HIF1A was purchased from cell signaling technology (Danvers, USA). Precipitated DNAs that contained FTL fragments were then amplified using quantitative PCR. The sequences of three primers used to detect HREs in FTL promoter were as follows:1# (F): 5′-CGCAGGGCTTCTCTTTGTGG-3′, (R): 5′- TGAACAGTGTCTCTGAAGTTGCC -3′; 2# (F): 5′- CCACAACGCAGGGCTTCT C-3′, (R): 5′-TTTGGAGACAACTCACAGACTTCG -3′; 3# (F): 5′- CGCAGG GCTTCTCTTTGTGG -3′, (R): 5′- GGAGTGGAAATGGGGAGGAATG − 3′.
In vivo experiments
All nude mice were purchased from Shulaibao (Wuhan,China) Biotechnology Co., Ltd. Animal feeding and experimental operations were in line with the guidelines of the Animal Ethics Committee of Wuhan University People’s Hospital. Stably transfected U87 cells that were growing in the logarithmic phase were prepared. Cells were resuspended in PBS at a concentration of 5 × 106 cells/100 μL and then subcutaneously injected into the armpits of 5-week-old Balb/c nude mice. For in vivo temozolomide (TMZ) treatment, Six nude mice were randomly divided into two groups. Starting on day 5, the mice were injected intraperitoneally with TMZ (50 mg/kg) for 5 days. After subcutaneous implantation, the condition of the nude mice was observed daily. Recorded dynamic changes in the size of subcutaneous xenografts (longest diameter * shortest diameter 2/ 2). All nude mice were sacrificed on the 29th day after transplantation and the tumors were weighted.
Statistical analysis
Data were presented as mean values ± standard deviation (SD) from at least three experiments. Student’s t-test was used to analyze the differences between two groups. One-way analysis of variance (ANOVA) was used for the comparison among three or more groups and the student-Newman-Keuls (SNK) method was used for post-analysis. Patients were divided into high and low groups according to the 50% cutoff point of FTL expression and Kaplan–Meier survival analysis was used to analyzed significance between groups. Univariate and multivariates Cox regression analysis was assessed by SPSS.21 (IBM, New York) software. Graphs production were performed by GraphPad Prism 5.0 software (GraphPad Software, Inc., La Jolla, USA). A p value of less than 0.05 was considered as statistical significance.
Discussion
Hypoxia environment is a fundamental characteristic of malignancy of glioma. The aggressive clinical behavior of glioma are closely related to the tension of oxygen in tumor microenvironment [
25]. Hypoxia is considered to be a major driver of malignancy progression and treatment resistance of glioma. Hypoxia inducible factor 1(HIF1) had long been recognized as regulator of the mechanism of hypoxia-promoted progression in glioma [
26].. Once cells are under hypoxic condition, HIF-1α will be gradually accumulated and then translocate to the nucleus. Subsequently, HIF-1α binds with HIF-1β to form a stable complex which can bind with the hypoxia responsive elements (HREs) in the promoter region of target gene [
27]. The promoter of the ferritin gene contains regions of HREs that may interact with HIF-1α.Mitochondrial Ferritin (FtMt) is a form of ferritin distribution in mitochondria. FtMt was proved to be an potential target gene of HIF1a,as well as stabilized HIF1a in a hypoxic environment by binding to the HREs that located in the promoter regions of human FTMT gene [
20]. In this study, we found a novel hypoxia response gene, FTL, which obviously increased under hypoxia in a time-dependent manner. Besides, there were three HREs in the region of FTL promoter and HIF-1α, not HIF2α regulates FTL expression by directly binding to HRE-3 in FTL promoter. FTL expression significantly correlated with glioma grade. As it was described before that increasing grade of gliomas correlates with an increase absence of oxygen [
4]. So it was reasonable to infer that the high expression of FTL in HGG might be caused by the ubiquitous hypoxic-microenvironment in glioma.
Ferritinwas reported to play crucial role in regulating several solid tumors, including glioma [
16,
28,
29]. Glioma cells were considered to undergo EMT during tumorigenesis and progression to higher grade. Recently, several studies uncover closely associated between ferritin and epithelial- mesenchymal transition (EMT). Previous study found that FTL was downregulated in osteosacrcoma (OS) and overexpression FTL in MG-63 cells enhanced the invasion and altered the expression of multiply EMT-related markers [
30]. On the contrary, ferritin was evaluated by means of western blot in breast cancer cell lines, and results showed that FTL level was significantly elevated in mesenchymal phenotype cell lines compared with epithelial phenotype [
17]. These findings indicate importance of FTL in process of EMT. However, the specific molecular mechanism of how FTL regulates EMT remains unclear. Moreover, authors of another study demonstrated that FTL was a marker of breast tumors with an aggressive phenotype [
31]. Consistent with this finding, we found that FTL expression was mainly enriched in mesenchymal subtype and correlated with multiply EMT-related markers in TCGA and CGGA datasets. Also, knocking down FTL dramatically altered glioma cell to blunt morphology and reduced the migration and invasion of glioma cells, as well as alter expression of snail and E-cadherin. The reversal of the EMT process is accompanied by a decrease in the expression of EMT transcription factors and remarkable decline in cell invasiveness [
31]. Our study provided solid evidence FTL might be a novel regulator of EMT in glioma. Targeting crucial control mechanisms of EMT could prevent the transformation from epithelial to mesenchymal subtype, which inhibited progression and enhance therapeutic effect of glioma [
32].
Accumulating evidence shows that nuclear accumulation of β-catenin play crucial role in regulating EMT [
33,
34]. Recently, Dong Xiao et al. found that SPHK2, a direct target of miR-708, triggered a cascade of signals leading to the activation of Akt pathway and the phosphorylation of GSK-3β and finally to the nuclear translocation of β-catenin to regulate EMT in glioma cells [
35]. The nuclear accumulation of β-catenin correlated with WHO grades and cytoplasmic- nuclear β-catenin was an independent prognostic factors in glioma [
36].FTL positively correlated with β-catenin and the nuclear accumulation of β-catenin was more common in glioma tissues with high FTL expression. Therefore it was reasonable to infer that FTL may have a certain relationship with β-catenin. We found that knocking down FTL in glioma cells dramatically reduced nucleus accumulation of β-catenin and dramatical decrease of activity of β-catenin signaling detected by Luciferase reporter system. The nuclear translocation of β-catenin bond with TCF/LEF transcription factors to induce the expression of vimentin and snail1,and subsequently activated the EMT process [
37]. In our study, inhibition of FTL inactivated AKT by phosphorylation (ser473) and decrease of the phosphorylation level of GSK3β (ser9). Generally, inactive GSK3β, together with Axin, adenomatous polyposis coli (APC), casein kinase 1 (CK1) stabilized β-catenin which subsequently resulted in nucleus translocation of β-catenin. Therefore, we inferred that FTL was a regulator of AKT/GSK3β/ β-catenin signaling. Using IM-12 or CTNNB1 plasmid significant reversed the oncogenic function of FTL in mediating EMT in glioma, which strongly indicated that FTL promoted EMT by regulating AKT/GSK3β/ β-catenin signaling.
Interaction between hypoxia and EMT was mediated by multiply genes and pathways. HIF1A might directly induced the expression of TWIST which promoted EMT by interacting with Ring1B and EZH2. However, multiply genes might as crucial mediators by which hypoxia induced EMT. Previous study revealed that FAT1 positively correlated with multiply hypoxia related genes and it was a potent regulator of EMT both via or independent of HIF1a in glioblastoma [
38]. We found that hypoxia enhanced the invasion of glioma cells, while inhibition FTL in glioma cells could mostly eliminate hypoxia-promoted invasion. Consistent with these findings, expressions of EMT-related markers were also obviously altered. Overall, our study revealed that HIF1A directly bond with HRE-3 in the region of FTL promoter to enhance its expression and FTL might act as a crucial gene that regulated EMT process of glioma. Targeting FTL in glioma cells could dramatically inhibit EMT induced by hypoxia, which indicated that FTL could be a potential target for therapy.
As a consequence of EMT process, glioma cells gradually become more invasive, and the adhesion between cells and cells is reduced, which makes it easier to infiltrate the adjacent brain tissues or escape from chemo/radiotherapy [
39,
40]. In addition, EMT alter the stem cell characteristics of tumor cells and express more stem cell markers, which facilitates glioma resistance to chemotherapy and more likely to relapse. Hongbo Guo et al. found that miR-203 expressed low in imatinib-resistant GBM cells(U87AR,U251AR), and ectopic expression of miR-203 obviously reversed EMT by directly targeting SNAI2,which sensitized glioma cells to chemotherapy [
41]. Also the results of another study revealed that miR-140 that targeted CTSB signaling suppressed the mesenchymal transition of GBM and enhanced TMZ cytotoxicity [
42]. Previous studies demonstrated that FTL was participate in chemo-resistance of human breast cancer cells and colorectal cancer and inhibition of FTL induced sensitivity of cells to chemotherapy agents [
20]. Consistent with these findings, we found thatknocking down FTL significantly inhibited the proliferation and increased apoptosis of glioma cells treated with TMZ(400 μM). Both in vitro and in vivo showed that cells transfected with sh-FTL were more sensitive to TMZ,which resulted in more apoptotic cells. Together, FTL could enhanced TMZ resistance and decreased the cytotoxic effect of TMZ therapy on glioma cells. The possible mechanistic explanations of FTL-mediated TMZ resistance are that FTL may be an important upstream regulatory protein in the process of MGMT methylation. Besides, HIF1A can activate autophagy. It’s possible that hypoxia induced FTL may also affect the autophagic degradation of proteins to affect the resistance of TMZ.
The significance of FTL expression was also demonstrated by its correlation with the clinical prognosis of glioma patients. Upregulation of FTL expression in glioma had been found in several studies, but the relationship of FTL expression and prognosis of glioma has not been well documented. Through bioinformatics analysis and in-house cohort validation, we have identified FTL as a novel biomarker of prognosis, as well as response to TMZ in glioma. Considering the correlation between FTL expression and IDH1/2 or subtypes, use of combination molecular analysis containing FTL might provide a more effective method for predicting prognosis of glioma. Moreover, FTL can be secreted into blood by glioma cells, detecting the level of FTL in plasma may predict the prognosis of glioma. This gives us a hint that FTL may become an important indicator in glioma liquid biopsy. Certainly, more clinical research is needed to clarify these issues regarding biomarker of plasma FTL in glioma.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.