※ 본지는 의학계 발전을 위해 고군분투하는 의학자들의 연구의지를 고취하고 우수한 연구성과를 공유해 보다 많은 의료진과 의학계 관계자들에게 질환으로 고통받고 있는 환자의 진단과 치료는 물론 예방에 이르기까지 긍정적이고 바람직한 영향을 미칠수 있도록 매월 최근 연구논문 중 이 같은 목적에 부합하는 의미를 갖는 논문을 선정해 커버스토리로 논문전문과 저자의 연구 배경. 연구 과정과 성과를 게재한다.
[헬스앤라이프 윤혜진 기자] 헬스앤라이프저널이 선정한 2019년 4월호 <이달의 논문>은 박병규 국립암센터 국제암대학원대학교 교수가 국제학술지 < BBA - Molecular Cell Research >에 최근 게재한 'Upregulation of transforming growth factor-beta type I receptor by interferon consensus sequence-binding protein in osteosarcoma cells'이다.
전이성 골육종은 지난 30년간 5년 생존율이 제자리에 머물고 있다. 현재까지 골육종 전이 억제를 입증하는 연구는 드문데다 임상 단계에서 골육종 전이 억제 기전을 규명한 연구는 아예 부재하다. 이런 현실 속에서 임상과 연구 현장을 오가며 골육종 치료 성적 향상을 위해 매진하고 있는 박병규 국립암센터 임상의학연구부 교수(소아청소년과 전문의)는 최근 골육종 전이 기전을 규명하는 데 성공했다.
Interferon consensus sequence-binding protein
Transforming growth factor-beta type I
Transforming growth factor-beta (TGF-β) is a known tumor suppressor, which also exerts a tumor promoting activity at an advanced stage of cancer. Previously, we reported that expression of interferon consensus sequence-binding protein (ICSBP), also known as interferon regulatory factor-8, is positively correlated with TGF-β type I receptor (TGF-β RI) expression in osteosarcoma patient tissues. In this study, we demonstrated that ICSBP upregulated TGF-β RI and induced epithelial-to-mesenchymal transition-like phenomena in human osteo-sarcoma cell lines. As determined by soft agar growth of osteosarcoma cells and xenografted mouse models, ICSBP increased tumorigenicity, which was reversed by ICSBP knock-down or a TGF-β RI inhibitor. To test whether ICSBP directly regulates the promoter activity of TGF-β RI, we performed a TGF-β RI promoter assay, an electro mobility shift assay, and a chromatin immunoprecipitation assay. We observed that TGF-β RI promoter was activated in ICSBP-overexpressing osteosarcoma cells. Exploiting serial deletions and mutations of the TGF-β RI promoter, we found a putative ICSBP-binding site at nucleotides −216/−211 (GGXXTC) in the TGF-β RI promoter. Our data suggest that ICSBP upregulates TGF-β RI expression by binding to this site, causing ICSBP-mediated tumor progression in osteosarcoma cells. In addition, we found a positive correlation between ICSBP and TGF-β RI expression in several types of tumors using the cBioportal database.
Summary: We demonstrated that interferon consensus sequence-binding protein upregulates transforming growth factor-beta type I receptor (TGF-β RI) expression by binding to nucleotides −216/−211 (GGXXTC) in the TGF-β RI promoter, which resulted in increased tumorigenicity and tumor progression in human osteo-sarcoma cells.
Interferon consensus sequence-binding protein (ICSBP), also known as interferon regulatory factor-8 (IRF8), is a member of the interferon(IFN)-γ inducible family of transcription factors expressed in hemato-poietic cells . ICSBP forms heterodimers with other transcription factors and binds to specific DNA sequences; thereby, ICSBP is capable of either activating or repressing gene transcription. For instance, ICSBP associates with Ets transcription factors, including PU.1, to activate transcription driven by the Ets/IRF composite element (EICE) . On the other hand, ICSBP associates with other IRFs, such as IRF1 and IRF2 on the IFN-stimulated response element (ISRE), to repress the transcription of ISRE-driven genes [3,4]. It has been shown that ICSBP is essential for immune cell development, including development of monocyte and dendritic cells. ICSBP-deficient mice have been shown to develop immunodeficiency and chronic myeloid leukemia (CML)-like syndrome . In addition, downregulation of ICSBP is observed in various non-hematopoietic cancers, such as nasopharyngeal, esophageal, and multiple other carcinomas . Accordingly, ICSBP is generally considered as a tumor suppressor. However, a recent study has suggested that ICSBP also has a role as a tumor promoter via transforming growth factor-beta (TGF-β) receptor upregulation and activation .
TGF-β is a tumor suppressor at an early-stage tumorigenesis; however, the cytokine is also known to act as a tumor promoter during tumor progression . The signaling pathway begins with TGF-β binding to TGF-β type II receptor (TGF-β RII), which in turn forms a heterodimer with TGF-β type I receptor (TGF-β RI), leading to receptor kinase activation. Activated TGF-β RI phosphorylates Smad2 and Smad3, which then bind to Smad4 and translocate to the nucleus to regulate gene expression. TGF-β also signals through non-Smad pathways, including p38, mitogen-activated protein kinase, Src, Akt, and mechanistic target of rapamycin pathways . Notably, cell growth and metastasis are controlled by TGF-β-induced activation of both Smad and non-Smad pathways. Further, TGF-β induces epithelial-to-mesenchymal transition (EMT), which is a critical phenomenon for cancer cells to acquire migratory and invasive properties during metastasis. Recently, we demonstrated that ICSBP overexpression enhances cell proliferation via TGF-β receptor/TGF-β-activated kinase(TAK) signaling pathways in human leukemic HL-60 cells . We further reported that ICSBP overexpression enhances TGF-β receptor and Snail expression, and induces EMT-like phenomena (ELP) in human osteosarcoma cells . Although these studies indicate that ICSBP does not function entirely as a tumor suppressor, but serves as a tumor promoter depending on tumor types, additional efforts are required to delineate the association between ICSBP and tumor development, including in vivo experiments.
In the present study, we demonstrate that ICSBP expression induces ELP and enhances osteosarcoma growth both in vitro and in vivo through upregulation of TGF-β RI. Further, we identified a novel putative ICSBP-binding site in the TGF-β RI promoter.
2. Materials and methods
2.1. Cell lines and culture
Human osteosarcoma cell lines Saos-2, U2OS, and 143B were obtained from the American Type Culture Collection (Rockville, MD, USA). Cells were maintained in RPMI 1640 medium(Life Technologies, Grand Island, NY, USA)supplemented with 10% fetal bovine serum(FBS; Life Technologies), 100 units/ml penicillin, and 100 μg/ml streptomycin (Life Technologies) at 37 °C in a humidified incubator with an atmosphere containing 5% CO2.
2.2. Establishment of ICSBP-expressing stable cell lines
ICSBP PCR product was cloned into the HindIII and XhoI sites of pcDNA3.1/V5-HisA vector (Invitrogen, Waltham, MA, USA). Saos-2, U2OS, and 143B cells were transfected with ICSBP construct or empty vector (Mock) using Lipofectamine 2000 (Life Technologies) according to the manufacturer's instructions. Stable cell lines, including Saos-2-Mock, Saos-2-ICSBP, U2OS-Mock, U2OS-ICSBP, 143B-Mock, and 143B-ICSBP cells were established by selection with 500 μg/ml of geneticin(G418, Calbiochem, La Jolla, CA, USA) for 4 weeks.
2.3. Antibodies and reagents
Polyclonal antibodies against TGF-β RI, phospho-Smad2, Smad2, phospho-Smad3, Smad3, Ki67, PARP, caspase3, GAPDH and β-actin were purchased from Cell Signaling Technology (Beverly, MA, USA). The V5 antibody was obtained from Invitrogen. The ICSBP antibodies(anti-goat and anti-mouse) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). TGF-β RI inhibitor, SB431542, was purchased from LC Labs (Woburn, MA, USA).
2.4. RNA isolation and reverse transcription-PCR (RT-PCR)
Total RNA was extracted from cells using Trizol reagent (Invitrogen) according to the manufacturer's instructions. cDNA was synthesized using the Maximum RT preMix Kit (iNtRON Biotechnology, Seoul, Korea). Each reaction was performed in a thermal cycler for 60 min at 45 °C. cDNA fragments were amplified using the following primer pairs: human ICSBP 5′-AAT GGT GGT CGG CGG CTT CG-3′ (sense), 5′-TGC TGA ATG GTG CGC GTC GT-3′ (anti-sense), human TGF-β RI 5′-CAC CAT TTC GAG CAA CC-3′ (sense), 5′-TCT GGA GGA CCT GGT AGA GG-3′ (anti-sense), human GAPDH 5′-GAC CCC TTC ATT GAC CTC AAC-3′(sense), 5′-CTT CTC CAT GGT GGT GAA GA-3′ (anti-sense). PCR was performed in the Unit Block Assembly for PTC DNA Engine system (AlphaTM, Gottingen, Germany) with a program consisting of 30 cycles of 95 °C for 40 s, 58 °C for 1 min, and 72 °C for 1 min. PCR products were resolved on a 1.5% agarose gel.
2.5. RNA interference
siRNAs specific for human ICSBP (si-ICSBP) and non-specific siRNA(si-Cont) were purchased from Dharmacon (Lafayette, CO, USA). Cells were transiently transfected with 100 nM siRNAs using Lipofectamine 2000 as described by the manufacturer.
2.6. Lentivirus production and establishment of luciferase-expressing cell lines
The luciferase-expressing lentivirus construct (pLECE3-Luc) was a kind gift from Dr. Sang Jin Lee (National Cancer Center, Goyang, Korea). Lentivirus production was performed as previously described . Briefly, we cotransfected lentiviral vector with the Lentiviral Packaging Mix (Invitrogen) into 293 FT cells. After 48 h, viral supernatants were harvested, spun to remove cell debris, filtered through a Millex-HV syringe filter (0.45 μm; Millipore, Bedford, MA, USA), supplemented with 4 μg/ml of polybrene, and stored at −80 °C. 143B-Mock and 143B-ICSBP cells were infected with lentiviral supernatants. After 48 h, cells were rinsed and plated in selective media containing 1 μg/ml puromycin (Sigma, St. Louis, MO, USA) for 14 days to generatestable cell lines.
2.7. Measurement of cell proliferation
Cells were plated at 1 × 104 /100 μl/well in a 12-well plate. Viable cells were counted at the indicated times using a hemocytometer after trypan blue staining.
2.8. Soft agar colony formation assay
Cells were transfected with siRNAs for 24 h. Soft agar assays were performed in six-well plates by placing 1 × 103 cells in 1 ml of 0.3% agar onto a 2 ml of 0.8% agar base layer. Plates were covered with 1 ml fresh RPMI medium containing 10% FBS and incubated in a 5% CO2 atmosphere at 37 °C for 3 weeks. Cell growth medium was changed every third day. Colonies were stained with iodonitro tetrazolium violet(INT) solution (0.5 mg/ml, Sigma) and images were taken with Kodak Image Station 2000R (Eastman Kodak Company, New Haven, CT, USA). Colonies over 100 μm in diameter were counted.
2.9. Cloning and mutagenesis for reporter assay
TGF-β RI promoter reporter (TRIP) was generated by inserting the nt (−1966/+50) region of the human TGF-β RI promoter into pGL3-basic firefly luciferase vector (Promega, Madison, WI, USA) between XhoI and HindIII sites. Deleted forms of the TGF-β RI promoter were generated by PCR amplification: 5′ deletion mutants were produced by PCR amplification of TGF-β RI promoter region spanning from designated 5′ starting points to nt (+50); 3′ truncated mutants by PCR amplification of TGF-β RI promoter region spanning from nt (−1966) to designated 3′ end points. Amplified DNA fragments were then cloned into pGL3-basic vector. Sequences of all mutants were verified. Two-base substitution mutants of the nt (−220/−210) region of TRIP were generated using the QuikChange site-directed mutagenesis kit(Stratagene, La Jolla, CA, USA). U2OS-Mock, U2OS-ICSBP, 143B-Mock, and 143B-ICSBP cells were grown in 12-well plates and co-transfected with reporter vectors and internal control, Renilla luciferase plasmid pRL-TK vectors (Promega), using Lipofectamine 2000. After 24 h, cells were harvested. Renilla and firefly luciferase activities were determined using the Dual-luciferase assay kit (Promega) as described by the manufacturer, with a luminescence plate reader (VICTORTM X, PerkinElmer, Waltham, MA, USA). The firefly luciferase activity was normalized for transfection efficiency with the Renilla luciferase activity as an internal control. Data were expressed as relative luciferase activity. U2OS-ICSBP cells transfected with pGL3-basic vector were used as negative control.
2.10. Electrophoretic mobility shift assay
Nuclear proteins were extracted from U2OS-ICSBP cells as previously described . Nuclear extracts (2 μg) were incubated in binding buffer (100 mM Tris pH 7.5, 500 mM KCl, 10 mM dithiothreitol [DTT]) with 4 pmol biotinlabeled, double-stranded oligonucleotide probes: −230/−210 (5′-CGGGCCCGCAGGCGGGGCTC-3′) and −225/−206 (5′-CCGCAGGCGGGGCTCCCGGC-3′), or mutated oligonucleo-tide −225/−206(5′-CCGCAGGCGATGCCTCCGGC-3′). Unlabeled, competing oligonucleotides were added to the binding reactions at 100-fold molar excess. For supershift assay, complete reaction mixture was incubated with control IgG or anti-V5 antibodies. DNA/protein complexes were separated by 6% native polyacrylamide gel electrophoresis(PAGE) and visualized by chemiluminescence imaging.
2.11. Establishment of xenograft and in vivo treatment of mice with either siRNA or TGF-β RI inhibitor
All animal experiments in this study were performed in accordance with the Guideline for the Care and Use of Laboratory Animals of the National Cancer Center, Korea. Luciferase-expressing 143B-Mock or 143B-ICSBP cells (1 × 106) were injected subcutaneously into male Balb/c nude mice. The greatest longitudinal diameter (length) and the greatest transverse diameter (width) of each tumor were determined once a week using a caliper. Each tumor volume was calculated by the modified ellipsoidal formula (tumor volume = 1/2 [length × width2]). When tumor size reached 30 mm3, the mice were injected intraperitoneally with either dimethyl sulfoxide (DMSO) or SB431542(20 mg/kg). Injection was performed twice a week for 3 weeks. In addition, when tumor size reached 50 mm3, 250 pmol siRNA solution in phosphate-buffered saline (PBS) was injected into the tumor by electroporation according to the manufacturer's instructions (NEPA21, Nepa gene Co., Chiba, Japan). We repeated electroporation every 7 days and monitored tumor size for up to 3 weeks.
2.12. In vivo measurement of luciferase activity
Mice were injected intraperitoneally with 50 μl D-luciferin(PerkinElmer) dissolved as 30 mg/ml in PBS and anaesthetized with isoflurane (Choongwae, Seoul, Korea). After 10 min of luciferin injection, bioluminescence was detected for 1 min with Xenogen IVIS XR(Caliper Life Science, Mountain View, CA, USA).
2.13. Immunohistochemical staining
Tumor tissues were fixed with 10% neutral buffered formalin. Formaldehyde-fixed specimens were paraffinembedded and cut to a thickness of 4 μm. Sections were dried at 56 °C for 1 h, and immunohistochemical staining was performed with Discovery XT(Ventana Medical Systems, Tucson, Arizona, USA) as follows. Sections were deparaffinized, rehydrated with EZ prep (Ventana Medical Systems), and washed with reaction buffer. Antigens were retrieved with heat treatment in Tris-ethylenediaminetetraacetic acid (EDTA) pH 8.0 buffer (CC1, Ventana Medical Systems) at 90 °C for 30 min. Then, they were allowed to bind specific antibodies, including anti-ICSBP (1:100 dilutions; Santa Cruz), anti-TGF-β RI (1:100 dilution; Abcam, Cambridge, UK), anti-phospho-Smad2 (1:1000 dilution; Cell Signaling Technology), anti-Smad2 (1:200 dilution; Cell Signaling Technology), and anti-Ki67 antibodies (1:200 dilution; Cell Signaling Technology). Parallel sections incubated with normal IgGs instead of primary antibodies were used as negative controls.
2.14. Immunoblot analyses
After washing with ice-cold PBS (10 mM Na2HPO4 pH 7.4, 145 mM NaCl, and 2.7 mM KCl), cells were lysed with 2 × sodium docecyl sulfate (SDS)-PAGE sample buffer (20 mM Tris pH 8.0, 2% SDS, 2 mM DTT, 1 mM Na3VO4, 2 mM EDTA, and 20% glycerol) and boiled for 5 min. The protein concentration of each sample was determined using the Micro-Bicinchoninic Acid protein assay reagent as described by the manufacturer (Thermo Scientific, Rockford, IL, USA). Total cellular protein (30 μg/lane) was separated by 10% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes. Membranes were blocked overnight at 4 °C in TBST (20 mM Tris pH 8.0, 150 mM NaCl, and 0.05% Tween 20) containing 5% non-fat milk. Membranes were then incubated overnight at 4 °C with primary antibody, washed three times with TBST, incubated with horseradish peroxidase (HRP)-con-jugated goat anti-rabbit IgG secondary antibody for 1 h at room temperature, and washed three times with TBST. Proteins were visualized using an enhanced chemiluminescence reagent (Millipore).
2.15. Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) assay was performed using EZ-ChIP (Millipore) according to the manufacturer's protocol. Crosslinking was done with 1% formaldehyde at RT for 10 min and stopped by the addition of glycine to make a final concentration of 0.125 M. Cells were collected and rinsed with cold PBS. Cell pellets were resuspended in hypotonic buffer (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA) containing 0.5 mM PMSF, protease inhibitor cocktail, and incubated on ice for 15 min. The nuclei were collected by micro-centrifugation, resuspended in SDS lysis buffer, and sonicated to an average length of 300–600 bp with S220 Focused ultrasonicator (Covaris, MA, USA). ChIP was conducted with monoclonal anti-ICSBP antibody (Santacruz Biotechnology) and protein A and Gagarose. DNA-protein complexes were eluted from beads, and DNA was recovered by reverse crosslinking using 0.2 M NaCl containing buffer and purified with spin columns. The ICSBP-binding site in the TGF-β RI promoter was analyzed by PCR using sense, 5′-ACTCACACAGACACA CCCAT-3′ and antisense, 5′-CCTCACCCCAGCAAACCT-3′ primers. DNA densities measured with densitometer were normalized to input DNA.
2.16. Statistical analysis
Comparison between two groups was performed using Student's t-test. Statistical significance was defined as P < 0.05. Data represent average values and standard deviation (error bars) obtained from two or three independent experiments.
Fig. 1. Effects of ICSBP overexpression on cell morphology, growth, and colony formation in human osteosarcoma cell lines. (A) Saos-2, U2OS, and 143B cells were transfected with either empty vector (Mock) or ICSBP-expression vector (ICSBP) and stable cells lines were established as described in the “Materials and methods”. Cell lysates were analyzed by immunoblotting using anti-ICSBP and anti-β-actin antibodies (upper panel). Representative cell images were taken using phase contrast microscope. Scale bar, 100 μm. (lower panel) (B) 143B-Mock (Mock) and 143B-ICSBP (ICSBP) cells (1 × 104) were plated in a 12-well plate and cell numbers were counted after trypan blue staining at the indicated times. The data represent the mean from three independent experiments with error bars representing the standard deviation. (C) Mock and ICSBP cells were seeded (1 × 103 cells/35 mm culture plate) and cultured in soft agar medium for 21 days to allow colony formation, followed by INT staining and image capturing. Colony sizes were measured and those bigger than 100 μm in diameter were counted (*P < 0.05). (D) ICSBP cells were transfected either with control si-RNA (si-Cont) or si-RNA specific for ICSBP (si-ICSBP), and then processed either for immunoblotting analysis or soft agar colony forming assay as in panel C. Similar results were observed in two independent experiments.
3.1. ICSBP overexpression induces ELP, enhances cell growth, and upregulates TGF-β receptor in human osteosarcoma cells
Previously, we reported that ICSBP overexpression induces ELP in human osteosarcoma cell lines . Here, we generated stable osteo-sarcoma cell lines (Saos-2-ICSBP, U2OS-ICSBP, and 143B-ICSBP) that overexpress ICSBP(Fig. 1A). Because morphologic change following ICSBP overexpression was most notable in 143B cells among these three(Fig. 1A) and 143B is known as highly tumorigenic and metastatic cells, we chose 143B cells for animal experiments. To quantitatively evaluate ELP induction, we measured the expression level of EMT maker using qPCR, and demonstrated that decreased E-cadherin and increased N-cadherin mRNA in 143B-ICSBP cells compared to 143B-Mock cells(Fig.S1). Although ICSBP overexpression increased 143B cell growth slightly(Fig. 1B), it enhanced anchorage-independent cell growth in soft agar(Fig. 1C), indicating ICSBP's role for tumorigenicity. Transfection of 143B cells with small interfering RNAs (siRNAs) specific for ICSBP(si-ICSBP) inhibited colony formation in soft agar (Fig. 1D). Taken together, the data indicate that ICSBP induces ELP and increases anchorage-independent cell growth.
Fig. 2. Effects of TGF-β RI upregulated by ICSBP on ELP and soft agar colony formation. (A) Cell lysates from 143B-Mock (Mock) and 143B-ICSBP (ICSBP) cells were analyzed by immunoblotting using the indicated antibodies. The levels of TGFβ RI were quantified by densitometry and normalized to β-actin. (B) Total mRNAs were extracted from Mock and ICSBP cells and levels of ICSBP and TGF-β RI mRNAs were determined by RT-PCR. GAPDH was used as a control. (C) ICSBP cells were transfected with either si-Cont or si-ICSBP for 24 h and cell lysates were processed for immunoblot analysis with the indicated antibodies. (D) ICSBP cells were transfected with either si-Cont or si-ICSBP for 24 h and representative cell images were taken using phase contrast microscope. Scale bar, 100 μm. (E) ICSBP cells were treated without or with 50 μM SB431542 (TGF-β RI inhibitor) for 24 h and representative cell images were taken using phase contrast microscope. Scale bar, 100 μm. (F) ICSBP cells were treated with varying concentrations of SB431542 and counted after trypan blue staining at the indicated times. Cell lysates from ICSBP cells treated with 50 μM SB431542 were analyzed by immunoblotting using the indicated antibodies (G) ICSBP cells were pre-treated with 50 μM SB431542 for 30 min and subjected to colony formation assay as in Fig. 1C. (*P < 0.05). Similar results were observed in two independent experiments.
TGF-β receptor signaling is associated with EMT (ELP for mesenchymal cells) [9,10]. ICSBP overexpression induced TGF-β RI upregulation at both protein and mRNA levels in 143B-ICSBP cells(Fig. 2A and B). Conversely, knockdown of ICSBP decreased TGF-β RI levels and attenuated basal activation of TGF-β signaling as shown by reduced phosphorylation of Smad2 and Smad3 (Fig. 2C). Additionally, ICSBP (IFN-γ induced transcription factor) and TGF-β RI mRNAs were induced by IFN-γ treatment (200 IU/ml for 24 h) in parental osteosarcoma cells (Fig. S2). As for non-Smad pathway molecules, phosphorylation of AKT and ERK was minimally increased in 143B-ICSBP cells as compared with 143B-Mock cells, but the activation of p38 was more evident in the former cells (Fig. S3). Transfection of si-ICSBP decreased levels of vimentin, a marker for EMT (Fig. 2C) and reversed ELP morphology in 143B-ICSBP cells (Fig. 2D). In agreement with this finding, transfection of si-ICSBP into 143B-ICSBP cells resulted in increased E-cadherin level, but decreased N-cadherin level (Fig. S1). Similarly, SB431542, a selective inhibitor of TGF-β RI , reversed ELP morphology (Fig. 2E). These data indicate that ICSBP-induced ELP is associated with TGF-β RI upregulation in 143B cells. Consistent with this, SB43152 decreased the growth of 143B-ICSBP cells, without causing apoptosis, as determined by cell counting and colony formation in soft agar (Fig. 2F and G). These results suggest that ICSBP upregulates TGF-β RI, which causes cell growth and ELP.
Fig. 3. Enhanced tumorigenicity by ICSBP in mouse xenograft model. (A) Mock and ICSBP cells (1 × 106) were injected into nude mice (n = 10) subcutaneously and tumor growth was monitored using IVIS. Tumor size was measured with caliper at the indicated times (left panel). After sacrificing the mice, tumor tissues were taken and representative images of tumors are obtained (right panel). (B) Tumor tissues were stained using the indicated antibodies for immunochemistry analysis and control IgG was used as a negative staining control. Scale bar, 100 μm. (C) ICSBP cells were transfected with either si-Cont or si-ICSBP and injected subcutaneously into nude mice (n = 10), and tumor growth was monitored using IVIS (left panel). Lysates from injected cells were processed for immunoblot analysis with the indicated antibodies (right panel). (D) Tumor size was measured with caliper at the indicated times. After sacrificing the mice, tumor tissues were taken and representative images of tumors were captured (left panel). Lysates from injected cells were processed for immunoblot analysis with the indicated antibodies (right panel). Similar results were observed in two independent experiments.
3.2. ICSBP promotes osteosarcoma growth, which is suppressed by either ICSBP knockdown or TGF-β RI inhibition in xenografted mouse models
Because ICSBP increased cell growth and colony formation in soft agar, we examined whether ICSBP enhances tumor growth in xeno-grafted Balb/c nude mice. Tumors were generated by injecting the same number of luciferase-expressing 143B-Mock and 143B-ICSBP cells subcutaneously into the left and right back, respectively. Subsequently, changes in tumor volumes and luciferase activities were monitored. As shown in Fig. 3A and B, tumors composed of ICSBP cells grew faster than tumors of Mock cells. Then, the tumors were removed from the sacrificed mice and processed for immunocytochemistry. TGF-β RI expression was higher in tumor specimens of 143B-ICSBP cells than in ones of 143B-Mock cells (Figs. 3B and S4), which agreed with Fig. 2. Moreover, expression levels of Ki67, a proliferation marker, showed similar pattern (Fig. 3B). Next, we tested whether knockdown of ICSBP affects tumor growth in vivo. ICSBP cells transfected with either nonspecific siRNA (si-Cont) or si-ICSBP were injected subcutaneously into the left and right back of Balb/c mice, respectively. Knockdown of ICSBP resulted in decreased tumor growth compared with controls(Fig. 3C and D). Downregulation of ICSBP in si-ICSBP transfected cells was verified by immunoblotting analysis (Fig. 3C and D). To investigate whether ICSBP knockdown has antitumor effects even in pre-established tumors, we generated ICSBP xenograft in nude mice and later injected si-ICSBP into the tumor mass using in vivo siRNA delivery system when the tumor size reached 50 mm3. Interestingly, si-ICSBP treatment suppressed the tumor growth (Fig. 4A). Down-regulation of ICSBP was verified in the removed tumor tissues by immunoblotting and immunocytochemistry analyses (Fig. 4B and C). Decreased expression of TGF-β RI and Ki67 was also demonstrated in the tumors treated with si-ICSBP (Fig. 4C). Next, we tested whether SB431542 shows an antitumor effect. ICSBP xenograft was prepared in nude mice, and treated with SB431542 when the tumor size reached 50 mm3. As expected, tumor growth was retarded in SB431542-treated mice compared with control mice (Fig. 4D). Protein levels of TGF-β RI, Smad2, pSmad2, and Ki67 were lower in the tumor specimens taken from SB431542-treated mice than controls (Fig. 4E and F).
3.3. ICSBP binds to the promoter of TGF-β RI and activates it
Because ICSBP increased both mRNA and protein levels of TGF-β RI(Fig. 2), we explored whether ICSBP directly binds and activates the TGF-β RI promoter. Human TGF-β RI promoter/luciferase reporter constructs, denoted by TRIP, were generated which encompassed nt(−1966/+50) in the promoter region of TGF-β RI (Fig. 5A). TRIP constructs were transiently transfected into U2OS-Mock and U2OS-ICSBP cells and luciferase reporter assay was conducted. Luciferase activities were increased in ICSBP cells compared with Mock cells(Fig. 5B), indicating that activation of the TGF-β RI promoter is dependent on ICSBP expression. To identify the regulatory elements critical for promoter activation by ICSBP, a series of 5′-deletions or 3′-truncations of TGF-β RI promoter constructs were prepared as in Fig. 5. Constructs were transfected into ICSBP cells for promoter activity analysis. Unexpectedly, only negligible luciferase activity was noted with the DM1 construct (−1200/+50), despite that it contains a putative ICSBP-binding site, EICE, located at nt (−989) of the TGF-β RI promoter . Further, with the DM3 (−600/+50) reporter, which lacks an EICE, even a higher luciferase activity than with the TRIP reporter was measured (Fig. 5B and C). Very weak luciferase activities with various constructs lacking the −600/+50 region (ΔDM1, ΔDM2, and ΔDM3 in Fig. 5C) suggest that the −600/+50 region is important for promoter activity. We further sought for positive regulatory elements by testing serial deletion mutants of the DM3 promoter region. DM3-1 (−300/+50) elicited higher promoter activity than DM3 did, whereas DM3-2 (−200/+50) did not (Fig. 5D); DM3-4 (−600/−200), which contains the −300/−200 region, elicited far greater luciferase activity than DM3-3 (−600/−300) did. These findings indicate that the −300/−200 region is crucial for promoter activity. As a next step, we generated progressive 5′-deletions from the −320/+50 region(Fig. 5E). Only negligible luciferase activity with the DM3-9 (−200/+50) construct lacking the −240/−200 region was demonstrated, indicating that the −240/−200 region is crucial for ICSBP-mediated TGF-β RI promoter activity (Fig. 5E). We further dissected the −270/+50 region by making successive deletions of 10 bases and found that deletion of the −220/−210 region nearly abrogated promoter activity(Fig. 5F). To test whether similar results are obtained in another osteosarcoma cell, we performed reporter assay using 143B cells. Consistent with the data above (Fig. 5B and F), DM3 and DM3-15 elicited higher luciferase activities in 143B-ICSBP cells than in 143B-Mock cells(Fig. 5G). Collectively, these data indicate that activation of the TGF-β RI promoter is dependent on ICSBP, and the −220/−210 region is critical for ICSBP-mediated promoter activity.
To verify that the −220/−210 region is critical for ICSBP binding, we carried out an electrophoretic mobility shift assay (EMSA) using a biotin-labeled −230/−210 oligonucleotide probe. With the addition of oligonucleotide probe to U2OS-ICSBP nuclear extracts, a band formed by association of ICSBP and the oligonucleotide probe was visualized(Fig. 6A). The band intensity was diminished greatly by adding excess unlabeled oligonucleotide to reaction mixtures. Same experiment using nuclear extracts obtained from si-ICSBP transfected Mock and ICSBP cells resulted in diminished intensity of the ICSBP-oligonucleotide band(Fig. S5). Furthermore, the ICSBP-oligonucleotide band was supershifted with the addition of anti-ICSBP antibody but not with the addition of control isotype IgG (Fig. 6B), indicating that ICSBP binds to the −230/−210 oligonucleotide specifically. To determine the specific ICSBP binding site within the −220/−210 region, we generated series of two-base substitution mutants in the region (Fig. 6C). Mutations at nt(−211/−212; M212) or nt (−215/−216; M216) resulted in decreased luciferase activities, although nt (−211/−212) appeared to be more critical for TGF-β RI promoter activation. To verify the binding specificity, we performed EMSA using either wild type (WT) oligonucleotide(−225/−206) or mutated oligonucleotide at nt (−216, −215, −212, −211) (Fig. 6D). A distinct ICSBP/WT oligonucleotide band was visualized, which disappeared by adding excess unlabeled oligonucleotide. However, ICSBP bound to the mutated oligonucleotide much weaker than to the WT oligonucleotide. Finally, to further verify the binding of native promoter region encompassing nt (−220/−210) by ICSBP, ChIP assay was conducted. DNA band (382 bp) released from ICSBP binding was clearly visible in ICSBP cells but barely visible in parental or Mock cells (Fig. 6E). Binding of native promoter region by ICSBP was also demonstrated in IFN-γ treated parental cells (Fig. S6). Taken together, these data indicate that ICSBP specifically binds to the TGF-β RI promoter region at −220/−210, and the sequence of binding motif is GGXXTC (−216/−211) (Fig. 6F).
3.4. ICSBP expression is correlated with TGF-β RI mRNA expression in brain glioma cells
We examined the cBioportal database to see if there is any tumor having correlation between its ICSBP and TGF-β RI expression. Eleven studies revealed positive correlation at mRNA levels of these two genes(Table 1). Of 11tumor types, highest correlation was demonstrated in brain lower grade glioma and glioblastoma multiforme. The odds of co-occurrence in these tumor types were significant (P < 0.05; log of odds > 3). Positive correlation was demonstrated in both Pearson's and Spearman's test (Fig. S6A and B). In support of this correlation, si-ICSBP transfected U373 human glioma cells resulted in decreased TGF-β RI protein expression (Fig. S6C).
Fig. 4. Enhanced tumorigenicity caused by ICSBP-induced TGF-β RI in mouse xenograft model. (A, B, C) 143B-ICSBP cells (1 × 106) were injected subcutaneously into nude mice. When tumor size reached 50 mm3, tumors were treated with either si-Cont or si-ICSBP by electroporation once a week for 3 weeks. At the indicated times tumor volumes were measured with caliper. After sacrificing the mice, tumors were taken from the mice and processed for tumor imaging (A), immunoblot analysis (B), and immunohistochemistry analysis using the indicated antibodies or control IgG as a negative staining control. Scale bar 100 μm (C). (D–F) ICSBP cells(1 × 106) were injected subcutaneously into nude mice (n = 10). When tumor size reached 30 mm3, tumor bearing mice received intraperitoneal injection with either vehicle DMSO or 20 mg/kg of SB431542 twice a week for three weeks. At the indicated times tumor volumes were measured with caliper. After sacrificing the mice, tumors were taken and processed for tumor imaging (D), immunoblot analysis (E), and immunohistochemistry analysis using the indicated antibodies (F). Similar results were observed in two independent experiments.
ICSBP, a transcription factor of the IFN regulatory factor family, is important for myeloid ell differentiation, such as macrophages and dendritic cells [14,15]. It has been demonstrated that ICSBP expression is high in normal hematopoietic cells but downregulated in myeloid leukemia, nasopharyngeal carcinoma, and esophageal carcinoma . In addition, ICSBP overexpression increases Fas ligand-induced apoptosis and decreases B-cell lymphoma-2 expression in mouse leukemic cells . Accordingly, ICSBP has been described as a tumor suppressor. In contrast to these reports, we previously demonstrated that ICSBP overexpression is associated with tumor progression via TGF-β RI upregulation in leukemia and osteosarcoma cell lines [7,10]. Moreover, we showed that ICSBP expression is associated with TGF-β RI expression using osteosarcoma patient tissues . Collectively, our previous works indicate a new function of ICSBP in tumor development, which drove us to investigate the mechanism underlying ICSBP-mediated TGF-β RI induction. In this study, we also explored the role of ICSBP in regulating tumor progression via TGF-β receptor signaling in vivo with xenografted mouse models. Finally, we could identify a novel binding site in the TGF-β RI promoter.
We showed that ICSBP overexpression augmented osteosarcoma cell growth in soft agar (Fig. 1), which is a characteristic of tumorigenicity. ICSBP knockdown reversed it. In addition, ICSBP overexpression significantly enhanced tumor growth in xenografted mouse models, whereas si-ICSBP treatment suppressed it (Figs. 3 and 4), indicating that ICSBP promotes tumor progression in 143B osteosarcoma cells. Intriguingly, ICSBP increased TGF-β RI levels, which was responsible for ICSBP-induced cell growth and ELP in osteosarcoma cells (Figs. 1, 2). Knockdown of ICSBP decreased TGF-β RI levels with concomitant dephosphorylation of Smad2 and Smad3, indicating that canonical TGF-β signaling pathway was suppressed (Fig. 2C). As for non-canonical pathway molecules, increased phosphorylation of p38 was evident in 143B-ICSBP cells, but that of AKT and ERK was minimal (Fig. S3). The data agree with our previous studies [7,10], in that the activation both Smad and non-Smad pathways occurred not only in U2OS osteosarcoma cells but in hematopoietic HL-60 cells as well. U2OS-ICSBP and HL60-ICSBP cells showed greater RI up-regulation than 143B-ICSBP cells, and this might underlie the finding that the enhanced phosphorylation of AKT, ERK, and p38 cells was less in 143B-ICSBP cells. Treatment with TGF-β RI inhibitor, SB431542, resulted in decreased cell growth and ELP (Fig. 2). Moreover, SB431542 treatment suppressed ICSBP-induced tumor growth in xenografted mouse models (Fig. 4D). Furthermore, transfection of si-ICSBP onto mouse tumors resulted in downregulation of TGF-β RI (Fig. 4B and C). Taken together, these findings strongly suggest that ICSBP exerts pro-tumorigenic activity in osteosarcoma cells via TGF-β RI signaling.
In an early stage of tumor development, TGF-β plays an inhibitory role for tumor growth; however, it also plays a critical role in tumor malignancy, such as tumor expansion and metastasis, at an advanced stage of cancer . EMT is an important process for tumor malignancy, such as tumor metastasis and drug resistance, which is mainly regulated by TGF-β signaling . Our previous reports and the current study revealed that ICSBP-mediated ELP and tumor growth are associated with TGF-β signaling. We reported no difference in the secreted amounts of TGF-β between U2OS-Mock and U2OS-ICSBP cells, indicating that the activation of TGF-β signaling by ICSBP is probably mediated by TGF-β receptor upregulation rather than the consequence of TGF-β secretion . Apart from TGF-β binding, TGF-β receptor levels are important for TGF-β-mediated tumor growth regulation. For example, TGF-β receptor downregulation, which in turn leads to a loss of growth inhibition by TGF-β, has been implicated in tumor progression in cancers such as prostate and colorectal cancers [18,19]. On the other hand, TGF-β RI expression is higher in non-small cell lung carcinoma than in controls, while TGF-β RII levels are not different between tumors and controls . In breast cancer, TGF-β RI expression is also correlated with tumor progression . Finally, we reported upregulation of TGF-β RI levels in osteosarcoma patient tissues. Overall, these findings suggest that TGF-β RI levels are linked to tumor progression.
In addition to stimulating ICSBP production, INF-γ also induces high-affinity TGF-β receptors in human corneal fibroblasts . In agreement with the report, treatment with INF-γ resulted in increased TGF-β RI mRNA expression in parental U2OS and 143B cells (Fig. S2). Because ICSBP, a transcription factor, increased TGF-β RI levels (Fig. 2), we sought for ICSBP-binding sites in the promoter of TGF-β RI. Although there were known ICSBP-binding sites, EICEs, located at nt−989 of the TGF-β RI promoter , our data revealed that the −220/−210 region, is critical for the ICSBP-mediated TGF-β RI promoter activation (Figs. 5 and 6). In addition, we identified the ICSBP-binding site with a GGXXTC sequence (−216/−211) using site-directed mutagenesis analysis. Similarly, this binding was also demonstrated in parental U2OS cells treated with IFN-γ (Fig. S6). Finally, binding of ICSBP to the native promoter region encompassing nt (−220/−210) was verified using ChIP assay (Fig. 6E). We investigated whether the association of ICSBP and TGF-β RI was documented in tumor types other than osteosarcoma, and found correlation between mRNA levels of the two genes in tumors, such as brain lower grade glioma and glioblastoma multiforme through cBioportal database searching (Fig. S6). It is interesting that positive correlation in these cancer types was demonstrated in both parametric Pearson's test and non-parametric Spearman's test. The former test assumes linear relations of normally distributed data and the latter test assumes biases of distribution with outliers. This suggests that the expression of ICSBP and TGF-β RI is correlated both in terms of their expression levels and their ranks.
Fig. 5. Activation of TGF-β RI promoter by ICSBP. (A) Schematic representation of the pGL3-lucifease vector (pGL3) and the human TGF-β RI gene promoter (−1966/+50) cloned into the pGL3 (TRIP). (B–G) pGL3, TRIP or deleted/truncated derivatives of TRIP were transfected into either U2OS-Mock or U2OS-ICSBP cells (B–F) or into 143B-Mock or 143B-ICSBP cells (G) for 24 h, and then luciferase activities were measured. Schematic illustration of constructs is shown in the left panels and results of luciferase assay in the right panels (*P < 0.05). Similar results were observed in three independent experiments.
Fig. 6. GGXXTC motif, at −216/−211 of the TGF-β RI promoter, specifically bound by ICSBP. (A) EMSA was performed with nuclear extract from U2OS-ICSBP cells and biotin-labeled oligonucleotide probe containing the −230/−210 region (labeled) of the TGF-β RI promoter in the absence or presence of 100-fold molar excess of unlabeled oligonucleotide (unlabeled). Arrowhead indicates ICSBP bound to the probe. (B) Nuclear extracts from ICSBP cells were incubated with biotin-labeled probe (labeled) and ICSBP antibody for supershift assay. Non-specific IgG was used as a negative control. Filled arrowhead indicates ICSBP bound to the probe and empty one supershifted ICSBP by ICSBP antibody. (C) ICSBP cells were transfected with either pGL3 or TRIP mutants containing successive 2-base substitutions(shown in bold characters) within the −220/−211 region, followed by luciferase assay. (D) Nuclear extracts from U2OS-ICSBP cells were incubated either with biotin-labeled WT oligonucleotide probe (−225/−206) (oligo) or with mutant one (mutant oligo) in the absence or presence of 100-fold molar excess of unlabeled WT oligo or unlabeled mutant oligo, respectively, followed by EMSA. Substituted sites of mutant probe are shown in bold characters at nt (−216, −215, −212, and −211). (E) Soluble chromatins from U2OS-parental, -Mock, and -ICSBP cells were immunoprecipitated with either anti-ICSBP antibody or non-specific control IgG described in the “Materials and methods”. Amounts equivalent to 1% of precleared lysates were saved for input control prior to incubation with anti-ICSBP antibody. ChIP assay using anti-ICSBP antibody or mouse IgG was conducted. DNA fragments in ChIP were purified and subjected to PCR using primers encompassing ICSBP binding sites within the TGF-β RI promoter. PCR products were quantified by densitometry and normalized as bound/input ratios. Similar results were observed in two independent experiments. (F) Schematic illustration of the binding motif of ICSBP in the TGF-β RI promoter.
We demonstrated that ICSBP exerts its tumor promoting activity via upregulation of TGF-β RI in osteosarcoma cells. Upregulation of TGF-β RI was driven by its promoter binding by ICSBP. We identified the putative binding site for ICSBP, which was located at nt (−216/−211) in the TGF-β RI promoter with GGXXTC sequence. In addition, tumor promoting activity of ICSBP in osteosarcoma cells was demonstrated by increased tumorigenicity and accelerated tumor growth in xenografted mouse models. Finally, positive correlation between ICSBP and TGF-β
RI expression was noted in several types of tumors, including brain glial tumors in the cBioportal database.
Downloaded from www.elsevisr.com/locate/bbamcr
#헬스앤라이프 #이달의논문 #4월호 #박병규 #국립암센터 #국립암대학원대학교 #골육종 #전이기전 #규명 #BBA # MolecularCellResearch #Upregulation of transforming growth factor-beta type I receptor by interferon consensus sequence-binding protein in osteosarcoma cells