Coiled-coil domain containing 109B is a HIF1α-regulated gene critical for progression of human gliomas

Human brain tumor (n = 68; WHO grade II–IV) and non-neoplastic tissue (n = 4) samples were obtained from surgeries performed at the Department of Neurosurgery at Qilu Hospital (Shandong, China). Written informed consent was obtained from all patients, and approval for experiments was obtained from Ethics Committee of the Qilu Hospital. All surgeries and post-operative animal care were approved by the Institutional Animal Care and Use Committee (IACUC) of Shandong University (Shandong, China). Our research complies with the commonly-accepted ‘3Rs’: replacement of animals by alternatives wherever possible, reduction in the number of animals used, and refinement of experimental conditions and procedures to minimize harm to animals.

Human glioma cell lines, U87MG, U251 and T98 were obtained from the Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The normal human astrocytes (NHA) cell line was a kind gift from the Department of Biomedicine at the University of Bergen (Bergen, Norway). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific) and maintained at 37 °C in a humidified chamber containing 5% CO₂. For hypoxic treatment, cells were placed in a modulator incubator (HERAcell 150i, Thermo Fisher Scientific) in 94% N₂, 5% CO₂, and 1% O₂. For stable CCDC109B-knockdown, U87MG and U251 cells were infected with lentivirus expressing short hairpin RNA (shRNA) (sh-CCDC109B-1). After 48 h, U87MG or U251 cells were exposed to 0.5 or 2 µg/mL puromycin (A1113802, Thermo Fisher Scientific), respectively, in complete DMEM for an additional 2 weeks. Cells were subsequently treated with PX478 (S7612, Selleck Chemicals; Shanghai, China) and HIF1α siRNA to inhibit HIF1α expression and harvested after 48 h. Sequences of synthesized shRNAs (Genepharma; Shanghai, China) were the following: sh-Negative Control (sh-NC) 5′-TTCTCCGAACGTGTCACGTtt-3′; sh-CCDC109B-1 5′-CAGTCACACCATTATAGTAtt-3′; sh-CCDC109B-2 5′-CTCGACAGGATTATACTTAtt-3′; sh-CCDC109B-3 5′-GCAAGTAGAAGAACTCAATtt-3′. Sequences of synthesized siRNAs (Genepharma) were the following: si-NC 5′-TTCTCCGAAGGTGTCACGG-3′; si-HIF1α-1 5′TACGTTGTGAGTGGTATTATT-3′; si-HIF1α-2 5′-CTGATGACCAGCAACTTGA-3′.

Samples were fixed in 4% formalin, paraffin-embedded, and sectioned (4 µm). After de-waxing and rehydration, the sections were incubated with 0.01 M citrate buffer for 20 min at 95 °C for antigen retrieval. Endogenous peroxidase activity and non-specific antigens were blocked with 3% hydrogen peroxide (ZSGB-Bio; Beijing, China) and 10% normal goat serum (ZSGB-Bio) respectively, followed by incubation with primary antibody at 4 °C overnight. Sections were rinsed with phosphate buffered saline (PBS), treated with goat anti-rabbit secondary antibody (ZSGB-Bio), visualized using 3, 3′-diaminobenzidine (DAB, ZSGB-Bio) as substrate, and counterstained with hematoxylin (Beyotime; Haimen, China). Normal mouse serum was used as the negative control. Staining of cancer cells was scored as follows: 0, no staining; 1, weak staining in <50% cells; 2, weak staining in ≥50% cells; 3, strong staining in <50% cells; and 4, strong staining in ≥50% cells. The following primary antibodies (Abcam, Cambridge, UK) were used at the dilutions indicated: CCDC109B (1:200), HIF1α (1:200), Ki-67 (1:500), MMP2 (1:100) and MMP9 (1:200).

Cells and tissues were incubated 30 min in RIPA buffer containing protein inhibitor cocktail for lysis (Thermo Fisher Scientific). After centrifugation and denaturation, protein (20 μg) was separated by 10% polyacrylamide gel electrophoresis and electrophoretically transferred to polyvinylidene difluoride (PVDF) membranes (Merck Millipore; Shanghai, China). Membranes were blocked with Tris Buffered Saline with Tween 20 (TBST, 10 mM Tris, 150 mM NaCl, 0.1% Tween 20) containing 5% bovine serum albumin (BSA, Thermo Fisher Scientific),and incubated overnight at 4 °C with the following primary antibodies against CCDC109B (1:500), HIF1α (1:1000), MMP2 (1:1000), MMP9 (1:1000) and β-Tubulin (1:1000; Cell Signaling Technology; Danvers, MA, USA). Membranes were incubated the next day with secondary antibody (1:5000; Santa Cruz; Dallas, TX, USA) conjugated to horseradish peroxidase (HRP) for 1 h at room temperature. Proteins were quantified using a system for detecting chemiluminescence (Bio-Rad; Irvine, CA, USA), according to the manufacturer’s protocol. Representative images and data were obtained from at least three independent biological replicate experiments.

Cell migration and invasion assays were performed in uncoated and matrigel-coated (BD Biosciences; San Jose, CA, USA) Transwell chambers (8 μm pores; Corning Costar; Corning, NY, USA). Cells (2 × 10⁴) in medium (200 µL) with 1% FBS were seeded in the top chamber. The lower chamber was filled with medium (600 µL) containing 30% FBS. Chambers were incubated for 24 h under normoxic or hypoxic conditions. Cells that migrated to or invaded into the lower surface were fixed with 4% paraformaldehyde (Solarbio; Beijing, China), stained with crystal violet (Solarbio) for 15 min and counted under bright field microscopy. Images were acquired from 5 random fields in each well, and cell numbers were determined using Kodak MI software. Each experiment was repeated three times in triplicate.

To assess the distribution and expression levels of CCDC109B, NHA and glioma cells were seeded onto glass slides. The cells were then washed twice with PBS and fixed with 4% paraformaldehyde for 20 min at room temperature. Cells were rinsed with PBS, permeabilized with 0.5% Triton X-100 (Solarbio) for 15 min, and blocked with 10% normal goat serum for 60 min at room temperature. Cells were stained with primary antibody against CCDC109B (1:100) at 4 °C overnight, followed by incubation with Alexa Fluor 594 goat anti-rabbit IgG (Abcam, UK; 1:800) for 1 h at room temperature. Cell nuclei were stained with DAPI (Sigma-Aldrich, Germany) at 37 °C for 10 min, and images were obtained with confocal microscopy (LSM780, Zeiss).

Cell proliferation was measured using the EdU Apollo 567 Cell Tracking Kit (Ribo-bio; Guangzhou, China). Cells (2 × 10⁴) under different treatments were seeded onto 24-well plates, exposed to 200 μM of 5-ethynyl-20-deoxyuridine for 2 h at 37 °C, fixed with 4% paraformaldehyde for 20 min, and treated with 0.5% Triton X-100 for 10 min. Cells were rinsed with PBS three times, and incubated with 100 μL of Apollo reagent for 30 min. Nuclie were stained with Hochest33342. The percentages of EdU-positive cells were determined from 500 cells and three independent experiments were performed.

NC and sh-CCDC109B-1 glioma cells were seeded onto six-well plates (120 cells per well) and cultured for 2 weeks in medium that was changed twice each week. Colonies of more than 50 cells were counted after fixation and staining with 100% methanol and 5% crystal violet. Data reported represent the average of three independent experiments.

Total RNA was isolated from cells using Trizol reagent (Takara; Tokyo, Japan) according to the manufacturer’s protocol. Total RNA was reverse-transcribed, and the resulting cDNA was used as template in real-time quantitative PCR performed with the standard SYBR premix Ex Taq (Takara) on the Real Time PCR Detection System (480II, Roche; Pleasanton, CA, USA). GAPDH served as an internal control, and independent experiments were conducted in triplicate. The following primers were used: GAPDH, forward, 5′-AATGAAGGGGTCATTGATGG-3′, reverse, 5′-AAGGTGAAGGTCGGAGTCAA-3′; HIF1α, forward, 5′-TGGCAGCAACGACACAGAAA-3′, reverse, 5′-TGCAGGGTCAGCACTACTTC-3′; CCDC109B, forward, 5′-ACACTGCTGAGATGGAACACAT-3′, reverse, 5′- TTGGCTTCCGAATGAGCTTCTA-3′.

For generation of the subcutaneous GBM model, female 4-week-old nude mice (SLAC laboratory animal Center; Shanghai, China) were maintained in a barrier facility on high-efficiency particulate air (HEPA)-filtered racks. Digoxin and saline were purchased from Qilu Hospital, Shandong University. Nude mice (n = 16) were divided into two groups (U87MG + saline, U87MG + digoxin, 8 mice per group). Cells were harvested by trypsinization, resuspended at 10⁷ cells/mL in a 1:1 solution of PBS/Matrigel (BD Biosciences, USA), and injected subcutaneously into the right shoulder of the mouse. The tumor tissues were isolated 37 days after injection, and then used for protein extraction.

For orthotopic xenografts, 4-week-old female nude mice (n = 16) were divided into two groups (sh-CCDC109B-1 and NC group), and U87MG or U87MG modified cells (1 × 10⁶) were implanted into the brain using a stereotactic apparatus (KDS310, KD Scientific; Holliston, MA, USA). Animals which displayed symptoms such as severe hunchback posture, apathy, decreased motion or activity, dragging legs, or drastic loss of body weight were euthanized by cervical dislocation. Excised tumor tissues were formalin-fixed, paraffin-embedded, and sectioned for Hematoxylin–Eosin (HE) staining and IHC.

All data are presented as a mean ± the standard error of the mean (S.E.M). The Student’s t test was used when only two groups were being compared. Analysis of variance (ANOVA) was used in cases where there were more than two groups being compared. Survival curves were estimated by the Kaplan–Meier method and compared using the log-rank test. For multivariate analysis, independent prognostic factors were determined using the Cox’ proportional hazards model. Variables that might be dependent on other variables were excluded from the model. A two-tailed χ² test was used to determine the association between CCDC109B and HIF1α. GraphPad Prism version 7.00 software program (GraphPad; La Jolla, CA, USA) was used to analyze in vitro and in vivo experiments. Differences were considered to be statistically significant when P < 0.05.

Immunofluorescence staining were used to detect localization and expression level of CCDC109B in NHA cell line and human glioma cell lines in vitro. The results revealed cytoplasmic localization and increased expression levels of CCDC109B protein in U87MG, U251 and T98 glioma cells compared to NHA (Fig. 1a). Western blot analysis confirmed the cell staining. Expression levels of CCDC109B protein was increased in glioma cell lines relative to NHA in vitro (Fig. 1b). To further confirm the level of CCDC109B in normal brain tissue samples and different grades glioma tissues, we searched publicly available databases, Rembrandt, TCGA, Chinese Glioma Genome Atlas (CGGA) and found a relatively higher mRNA level of CCDC109B in HGG in contrast to low grade gliomas (LGG; WHOI-II) and normal brain tissues (P < 0.001, Fig. 1c). Expression levels of CCDC109B were also stratified on the basis of the molecular subtypes of human glioma (mesenchymal, classical, neural, and proneural) in TCGA, CGGA and Gene Expression Omnibus (GSE4271) databases. Intriguingly, CCDC109B was increased in the mesenchymal glioma molecular subtype compared to other subtypes (P < 0.001, Fig. 1c), which indicates a potential role of CCDC109B expression in glioma migration and invasion. We validated the results of our molecular analysis in a cohort of glioma and non-neoplastic brain tissue samples from our own institution using IHC and western blot analysis. CCDC109B protein was highly expressed (scores ≥ 3) in majority of HGG (29/49, 59.2%) and very few LGG (2/19, 10.5%), with almost no expression in normal brain tissue samples (n = 4; Fig. 1d, e). The difference in expression levels between these groups was statistically significant (P < 0.001, Table 1), with high CCDC109B expression correlating with increased tumor grade (P < 0.001, Fig. 1e). Expression by western blot corroborated these results. CCDC109B protein levels were increased in HGG cases (n = 5) relative to normal brain tissues (n = 3) and LGG (n = 4; Fig. 1f). These results all together indicated that CCDC109B levels were elevated in HGG compared to LGG and non-neoplastic brain tissue samples.

The difference in expression levels of CCDC109B between HGG and LGG drove us to further investigate whether CCDC109B could serve as a prognostic marker in glioma patients. We analyzed the relationship between CCDC109B level and overall survival (OS) of glioma patients in TCGA, Rembrandt and CGGA databases based on tumor grade. LGG patients with a high or low expression of CCDC109B displayed a considerably different median OS in all three databases (all P < 0.001, Figs. 2a–c). Furthermore, levels of CCDC109B also exhibited a significant inverse relationship with median survival time of GBM patients in TCGA (P < 0.01, Fig. 2d) and Rembrandt (P < 0.001, Fig. 2e) databases. This correlation however was not significant in GBM patients from the CGGA database (P = 0.426, Fig. 2f).

To further confirm the prognostic value of CCDC109B in glioma, univariate Cox analysis was performed with clinical and molecular data of glioma patients in TCGA. The results demonstrated that age (HR = 1.075, P < 0.001), WHO grade (HR = 9.560, P < 0.001), CCDC109B expression (HR = 1.861, P < 0.001), and mutation status of isocitrate dehydrogenase 1 (IDH1, HR = 0.244, P < 0.001), were all prognostic indicators for glioma patients (Table 2).

To determine whether the protein has a biological role in glioma, we designed lentiviral constructs expressing shRNAs targeted against CCDC109B for stably knockdown of expression. Compared to NC constructs, the mRNA levels of CCDC109B in U87MG and U251 cells were significantly down-regulated after infection with three different shRNAs targeting CCDC109B (sh-CCDC109B-1; sh-CCDC109B-2; sh-CCDC109B-3; P < 0.001, Fig. 3a). Protein was nearly undetectable in cells infected with sh-CCDC109B-1 (Fig. 3b). Therefore, this shRNA was used in subsequent functional assays.

We evaluated the effects of CCDC109B knockdown on glioma cell proliferation using EdU (Fig. 3c) and plate colony forming assays (Fig. 3e). Loss of CCDC109B led to significant decreases in the percentage of EdU positive cells (all P < 0.05, Fig. 3d) and colony forming ability (all P < 0.05, Fig. 3f) in both U87MG and U251 cells.

In Transwell migration and invasion assays (Fig. 3g), CCDC109B knockdown attenuated the number of U87MG and U251 cells that had migrated/invaded after a 24-h incubation (all P < 0.05, Fig. 3h). Western blot analysis revealed that MMP2 and MMP9, two metalloproteinases which play critical roles in tumor invasion and migration [21, 22], were also reduced after CCDC109B knockdown (Fig. 3i). Taken together, these functional assays indicated that expression levels of CCDC109B potentially promoted glioma cell proliferation, migration and invasion in vitro.

We next established orthotopic tumor models by implanting U87MG-NC cells or U87MG-sh-CCDC109B-1 cells intracranially in nude mice to investigate whether CCDC109B mediated proliferation and invasion of glioma cells in vivo. Tumor volume was decreased with CCDC109B knockdown (Fig. 4a) and OS was prolonged in mice when compared to controls (P < 0.05, Fig. 4b). IHC staining for CCDC109B, and markers for proliferation (Ki-67), and invasion (MMP2 and MMP9) performed on sections from xenografts further established a potential role for CCDC109B in regulating these pathways (Fig. 4c). Lower levels of all three markers, Ki-67, MMP2, and MMP9, were observed in xenografts of U87MG-sh-CCDC109B-1 compared to controls (all P < 0.01, Fig. 4d).

One of the unexpected findings from IHC performed on primary GBM samples was the high expression of CCDC109B localized in areas bordering necrosis. Increased expression of HIF1α, a transcriptional regulator typically induced by hypoxia, was also increased in these areas (Fig. 5a). IHC staining was used to further examine the relationship between HIF1α and CCDC109B in a cohort of GBM specimens (n = 32; Fig. 5b; Additional file 1: Table S1; P = 0.020).

We next wanted to establish whether HIF1α might induce CCDC109B under hypoxia. We selected glioma cell lines, U87MG and U251, to further examine the relationship between these two proteins, as they express higher levels of HIF1α protein than T98 or NHA (Fig. 5c). We cultured U87MG and U251 cells under hypoxia (1% O₂) for 6, 12, 24 and 48 h. mRNAs levels of CCDC109B were increased by ~twofold under hypoxia (P < 0.001, Fig. 5d), and coordinate increases in CCDC109B and HIF1α at the protein level were confirmed by western blot (Fig. 5e). U87MG and U251 cells were treated with siRNAs targeting HIF1α (si-HIF1α and si-HIF1α-2) or an inhibitor of HIF1α (PX478) [23‐25] to test whether HIF1α is involved in regulating CCDC109B expression. Down-regulation of HIF1α reduced mRNA levels of CCDC109B (Additional file 2: Figure S1A, B) and led to moderate decreases in CCDC109B protein (Fig. 5f, g).

To verify these results in vivo, we implanted U87MG into the right shoulder of nude mice to establish subcutaneously xenografts. Digoxin, a drug widely used to inhibit HIF1α activity [26‐28], was subsequently injected into implanted animals to investigate whether HIF1α induced CCDC109B in vivo. Mice were injected one week after implantation with saline or digoxin (2 mg/kg) intraperitoneally every day for 30 days. Tumor size was significantly larger in the saline than the digoxin treated animals (Fig. 5h). We next measured protein levels of HIF1α and CCDC109B in treated and untreated xenografts by western blot. CCDC109B expression was decreased in digoxin relative to saline treated animals (Fig. 5i). Taken together, these results demonstrated that hypoxia enhanced CCDC109B expression and that HIF1α potentially induced expression of CCDC109B.

We next investigated whether CCDC109B knockdown altered hypoxia-induced migration and invasion of U87MG and U251 cells. Knockdown of CCDC109B in glioma cells under hypoxia was confirmed by qRT-PCR and western blot analysis (Fig. 6a, b). In Transwell invasion and migration assays, hypoxia significantly enhanced invasion and migration of U87MG and U251 cells (Fig. 6c, d). In contrast, glioma cell migration and invasion was significantly attenuated in U87MG- and U251-sh-CCDC109B-1 cells (all P < 0.01, Fig. 6c, d). These results indicated that CCDC109B promoted hypoxia-induced invasion and migration in human glioma cell lines U87MG and U251 in vitro.