Novel lymphoid enhancer-binding factor 1-cytoglobin axis promotes extravasation of osteosarcoma cells into the lungs.

Lung metastasis is a major cause of mortality in patients with osteosarcoma (OS). A better understanding of the molecular mechanism of OS lung metastasis may facilitate development of new therapeutic strategies to prevent the metastasis. We have established high- and low-metastatic sublines (LM8-H and LM8-L, respectively) from Dunn OS cell line LM8 by using in vivo image-guided screening. Among the genes whose expression was significantly increased in LM8-H compared to LM8-L, the transcription factor lymphoid enhancer-binding factor 1 (LEF1) was identified as a factor that promotes LM8-H cell extravasation into the lungs. To identify downstream effectors of LEF1 that are involved in OS lung metastasis, 13 genes were selected based on LM8 microarray data and genomewide meta-analysis of a public database for OS patients. Among them, the cytoglobin (Cygb) gene was identified as a key effector in promoting OS extravasation into the lungs. CYGB overexpression increased the extravasation ability of LM8-L cells, whereas knocking out the Cygb gene in LM8-H cells reduced this ability. Our results showed a novel LEF1-CYGB axis in OS lung metastasis and may provide a new way of developing therapeutic strategies to prevent OS lung metastasis.

INTRODUCTION

Osteosarcoma (OS) is the most common malignant bone tumor in both children and adults.1, 2 Lung metastasis is a major cause of poor prognosis in OS patients. Although the 5‐year survival rate of patients with localized disease is approximately 70%,3, 4 that of patients with lung metastasis is less than 40%.4 Because no effective treatment is currently available to improve outcomes, the development of new therapeutic strategies based on the molecular mechanisms of OS lung metastasis is urgently required.

Metastasis is a multistep processes involving epithelial‐mesenchymal transition (EMT), invasion, intravasation, circulation, extravasation, seeding, and outgrowth in secondary organs.5, 6 Because many genes participate in each step, identification of a gene whose function is crucial for a particular step is important for developing a strategy to prevent metastasis. Extravasation occurs when cancer cells migrate to distant organs after entering the circulation. The process involves the adhesion of cancer cells to the endothelium, modulation of the endothelial barrier, and transmigration of cancer cells to reach the underlying tissue. First, adhesion between cancer cells and vascular endothelial cells occurs through cell‐adhesion molecules with ligand and receptor functions after cancer cell arrest in small capillaries.7, 8 Then, secreted molecules from cancer cells induce alterations in the endothelial cell‐cell junction architecture.8 In addition, cancer cells can interact with immune cells and platelets to induce cytokine production from those cells, facilitating the alterations.7, 8 After endothelial barriers are disorganized, endothelial leakiness and cancer cell transendothelial migration occurs. Identification of genes controlling OS extravasation may be able to prevent OS lung metastasis, but such genes have not been identified.

Wnt signaling and lymphoid enhancer‐binding factor 1 (LEF1) advocate metastasis‐promoting pathways in OS and other cancers.9, 10, 11 Inhibition of Wnt or silencing of Lef1 expression significantly suppressed metastasis in vivo.9, 10 LEF1, a member of the T‐cell factor (TCF)/LEF family of high‐mobility group transcription factors, is primarily involved in the canonical Wnt/β‐catenin signaling pathway.11, 12 Although LEF1 is implicated in many steps of metastasis,11 the underlying mechanism whereby LEF1 enhances lung metastasis in OS is still unclear.

Cytoglobin (CYGB) is a member of the globin family of proteins, which include hemoglobin and myoglobin.13, 14 Cygb was first identified as an inflammatory‐ and fibrosis‐related gene in the liver.15 In addition, Cygb is also known to function as a tumor suppressor gene16, 17, 18 and is involved in protective mechanisms against cellular stresses such as cell injury, DNA damage, and hypoxia.13, 16, 19, 20, 21, 22 CYGB is induced by hypoxia‐inducible factor‐1α (HIF‐1α), nuclear factor kappa‐light‐chain enhancer of activated B cells (NF‐κB), and other inflammation‐related transcription factors.23 Overexpression (OE) of CYGB in lung cancer cells impaired transmigration and anchorage‐independent growth under normoxic conditions but promoted these abilities under hypoxic conditions.19 In the present study, we isolated LM8 sublines with differential abilities to metastasize to the lungs, and molecular genetic analyses of these sublines showed that LEF1‐induced CYGB plays a crucial role in the extravasation step during lung metastasis. Our results indicate that a novel LEF1‐CYGB axis can potentially serve as a therapeutic target for preventing the lung metastasis of OS.

MATERIALS AND METHODS

Cell culture

Murine OS LM8 cell line24 was gifted by Dr Hideki Yoshikawa (Osaka University, Osaka, Japan). All LM8 sublines were cultured in DMEM supplemented with 5% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL at 37°C, 5% CO2). Murine vascular endothelia bEnd.3 cells were purchased from the ATCC (Manassas, VA, USA). bEnd.3 cells were cultured in DMEM supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL) at 37°C, 5% CO2.

Mice

Male BALB/c nu/nu, SCID, and C3H mice were obtained from Charles River Laboratory, Japan (Yokohama, Japan). All mice used were 6‐8 weeks of age and were housed in the animal facilities at Tokyo Institute of Technology. All experimental procedures involving mice were approved by the Animal Experiment Committees of Tokyo Institute of Technology (authorization numbers 2010006 and 2014005) and carried out in accordance with relevant national and international guidelines.

In vivo and ex vivo bioluminescence imaging

Bioluminescence (BL) images of mice were acquired using the IVIS® Spectrum system (PerkinElmer, Waltham, MA, USA) 15 minutes after i.p. injection with d‐luciferin (50 mg/kg) (Promega, Madison, WI, USA). Ex vivo imaging was immediately carried out after the last in vivo image was taken. The following conditions were used for image acquisition: open emission filter, exposure time = 60 seconds, binning = medium 8, field of view = 12.9 × 12.9 cm, and f/stop = 1. BL images were analyzed using Living Image 4.3 software (PerkinElmer).

Establishment of LM8‐L and LM8‐H

The LM8/luc cell line was established by stable transfection with a firefly luciferase gene as described previously.25 To establish LM8‐L cells, which have lost the ability to metastasize to the lungs, LM8/luc cells were intracardially injected into BALB/c nude mice, and LM8/luc cells that metastasized to the bone were isolated with a BL image‐guided approach. The isolated cells were cultured and reinjected into nude mice. LM8‐L was established after 4 rounds of the image‐guided in vivo screening process. LM8‐H was selected based on metastatic ability to the lung in C3H mice from LM8‐L sublines that were isolated from lung metastases generated after injection of LM8‐L into the tibia of SCID mice.

Lung metastasis assay

C3H mice were i.v. injected with LM8 sublines (106 cells/100 μL PBS: 137 mmol/L NaCl; 2.7 mmol/L KCl; 4.3 mmol/L Na2HPO4; 1.47 mmol/L KH2PO4). BL signals from the lungs were monitored through in vivo BL imaging on indicated days.

Histology analysis

Isolated lungs were embedded in optimal cutting temperature (OCT) compound (Sakura Fine Tech, Tokyo, Japan) and stored at −80°C. Fixed lung cryosections of the lung (10‐μm thick) were then stained with HE.

Tumor formation ability assay

For s.c. transplantation, cell suspensions (106/20 μL PBS) were mixed with 20 μL Geltrex® (Thermo Fisher Scientific, Waltham, MA, USA) after which the mixture was s.c. injected into the hind limb of mice anesthetized with pentobarbital sodium (Somnopentyl; Kyoritsu Seiyaku Corp., Tokyo, Japan). Thirty days after transplantation, tumors were excised and their weights were measured.

Western blotting

Cell lysates were prepared with RIPA buffer and subjected to western blot analysis using a rabbit anti‐LEF1 polyclonal antibody (Ab; Cell Signaling Technology, Danvers, MA, USA), a rabbit anti‐CYGB polyclonal Ab (Thermo Fisher Scientific), and a mouse anti‐β‐actin monoclonal Ab (Sigma Aldrich, St Louis, MO, USA).

Cell‐proliferation assay

Cell proliferation was evaluated with the water‐soluble tetrazolium salt 1 (WST‐1) reagent (Roche Diagnostics, Basel, Switzerland) according to the manufacturer's instructions. Briefly, cells (103 cells/100 μL culture medium) were seeded in a 96‐well plate. After culturing for 24, 48, or 72 hours, the medium was removed and 100 μL WST‐1 (10‐fold dilution with culture medium) was added to each well. The cells were further incubated for 3 hours and, then, the absorbance of each well was measured at 450 nm with a reference wavelength of 750 nm, after shaking the plate for 1 minute with a 680XR microplate reader Model (Bio‐Rad, Hercules, CA, USA).

Extravasation assay

Cells were labeled with 25 μmol/L CellTracker® Green and i.v. injected into C3H mice (106 cells/100 μL PBS). DyLight® 594‐labeled isolectin B4 (6 mg/kg) (Vector Laboratories, Burlingame, CA, USA) was i.v. injected to stain endothelial cells 5 minutes before dissecting mice. The lungs were removed and observed under a confocal fluorescence microscope (Carl Zeiss, Jena, Germany) 48 hours after LM8 injection. Average fluorescence intensity of 3 fields/sample was quantitatively analyzed using ImageJ software.26

Cell‐adhesion assay

bEnd.3 cells were seeded on a 24‐well plate (105 cells/well) and cultured for 3 days. LM8 sublines were labeled with 25 μmol/L CellTracker® Green for 30 minutes. After washing with PBS, LM8 cells (5 × 104) were seeded into 24‐well plates with bEnd3 monolayers. After a 1‐hour incubation, each well was washed 3 times with PBS. Number of LM8 cells attached to the bEnd3 monolayer was observed by fluorescence microscopy (4 fields/well) and quantitatively analyzed using ImageJ software.26 Each sample was analyzed in triplicate.

Transmigration assay

bEnd.3 cells (105) were seeded in the top filters with 8‐μm‐pore Transwell® plate (Corning, Corning, NY, USA) and grown for 3 days. LM8 sublines were labeled with 25 μmol/L CellTracker® Green for 30 minutes. After washing with PBS, the cells (5 × 104) were seeded on bEnd.3 monolayers. After a 24‐hour incubation, the unmigrated cells were wiped off with a cotton swab and, then, the filter was fixed with 4% paraformaldehyde for 20 minutes. The migrated cells on the filter were then observed under a fluorescence microscope (4 fields/filter) and the number of migrated cells was analyzed using ImageJ software.26 Results are shown as the average number of cells per field. Each sample was analyzed in triplicate.

Genomewide meta‐analysis

Microarray data sets were downloaded from the public Gene Expression Omnibus (GEO) repository (http://www.ncbi.nlm.nih.gov/gds). The first data set included low‐ and high‐metastatic clinical OS samples (GSE21257). The second data set included low‐ and high‐metastatic human OS cell lines (GSE49003). The genes commonly upregulated in highly metastatic OS in both data sets were analyzed using GEO2R (https://www.ncbi.nlm.nih.gov/gds), and genes whose P‐values were <.05 were selected. Gene IDs of selected human genes were converted to murine gene IDs using BioMart web software.27

Reverse transcription‐PCR and qPCR

Total RNA was extracted from cell pellets using the RNeasy® Mini Kit (Qiagen, Valencia, CA, USA), as recommended by the manufacturer. Total RNA (1 μg) was reverse‐transcribed with Oligo(dT)20 Primer (Toyobo Co., Osaka, Japan) and ReverTra Ace (Toyobo Co.). qPCR and RT‐PCR were carried out using Thunderbird® SYBR qPCR Mix (Toyobo) and EmeraldAmp® GT PCR Master Mix (Takara, Tokyo, Japan), respectively. The primer sets used for qPCR and RT‐PCR are shown in Table S1.

Gene KO using the CRISPR‐Cas9 system

Two guide RNAs (gRNA1 and gRNA2) used for editing Lef1 and Cygb were constructed using CRISPR design software.28 The sequences of gRNA1 and gRNA2 used for targeting Lef1 are 5′‐TTGTTGTACAGGCCTCCGTC‐3′ and 5′‐GTACGGGTCGCTGTTCATAT‐3′, respectively. The sequences of gRNA1 and gRNA2 used for targeting Cygb are 5′‐GAAGGCGGTTCAGGCTACGT‐3′ and 5′‐TGAAGTACTGCTTGGCCGAA‐3′, respectively. The Lef1 and Cygb gRNAs were inserted into a unique BbsI site of the pX330 plasmid (42230; Addgene, Cambridge, MA, USA). We used a fluorescence indicator system using the pCAG/EGxxFP plasmid29 provided by Dr Ikawa (Osaka University, Osaka, Japan) to select cells whose genomes were correctly edited using the CRISPR‐Cas9 system. GFP‐positive colonies were selected, and 2 independent LM8‐H/Lef1‐KO and LM8‐H/Cygb‐KO clones each were established from the gRNA1‐ and gRNA2‐mediated KO cells.

Vector construction

The coding sequence of Cygb (NM_030206) was amplified using the KOD® FX Kit (Toyobo) with the following primer set: forward, 5′‐TCATGGAGAAA‐GTGCCGGGCG‐3′ and reverse, 5′‐CCCAAAGTGCTGCCAGGGAGG‐3′. The PCR product was purified by Gelase® (Epicentre, Madison, WI, USA) and ligated into an EcoRV‐digested pcDNA3.1‐myc‐His plasmid (Invitrogen, Carlsbad, CA, USA) using the Quick Ligation Kit (New England BioLabs, Cambridge, MA, USA) to construct the pcDNA3.1/Cygb. The pT2‐MCS‐SVNeo vector containing multicloning sites (MCS) was constructed using the pT2‐SVNeo vector (26553; Addgene), as described previously.30 The fragment containing the Cygb cording sequence was obtained by digesting pcDNA3.1/Cygb with EcoRI and NotI and inserting the liberated fragment into an EcoRI‐ and NotI‐digested pT2‐MCS‐SVNeo plasmid.

Establishment of an LM8 cell line with stable CYGB overexpression

To establish stable cell lines overexpressing Cygb, the sleeping beauty transposon system was used.30, 31, 32 The pT2/Cygb or pT2‐MCS‐SV Neo and pCMV(CAT)T7‐SB100 plasmids (34879; Addgene) were cotransfected into LM8‐L cells using an electroporator (Nepa Gene, Chiba, Japan) according to the manufacturer's instructions. After 48 hours, the culture medium was changed to selection medium containing 1 μg/mL G418 (Roche Life Sciences). The cells were further cultured in selection medium for 14 days, and single colonies were isolated to establish CYGB‐overexpressing LM8‐L (CYGB‐OE) cells.

Statistical analysis

Data are presented as the mean ± SE and were statistically analyzed with a 2‐sided Student's t test. P‐values <.05 were considered statistically significant.

RESULTS

Establishment of murine OS sublines with differential lung‐metastatic abilities

The LM8/luc cell line was previously established25 by stable transfection with a constitutive firefly luciferase reporter gene in the highly lung‐metastatic murine OS cell line, LM8.24 LM8‐L cells, with a low lung‐metastatic ability, were established after 4 rounds of image‐guided in vivo screening of LM8/luc cells, and LM8‐H cells, with a high lung‐metastatic ability, were isolated from cells in the LM8‐L subline that regained lung‐metastatic ability (Figure S1). Although LM8‐L and LM8‐H showed different cellular morphologies (Figure 1A), the expression levels of genes related to EMT therein did not show typical expression patterns of epithelial or mesenchymal cells (Table S2). Their differential lung‐metastatic abilities were confirmed by i.v. injection into syngeneic C3H mice: significantly higher BL signals were observed in the lungs of mice injected with LM8‐H cells compared to mice injected with LM8‐L (Figure 1B). Histological analysis of the lungs further confirmed the higher lung‐metastatic ability of LM8‐H cells (Figure 1C). However, the in vitro proliferation rate of LM8‐H cells was significantly lower than that of LM8‐L cells (Figure S2), and the growth of s.c. LM8‐H tumors was significantly slower than that of s.c. LM8‐L tumors (Figure 1D). These results are consistent with the previous reports33, 34 and suggest that the larger metastatic foci observed with LM8‐H cells were not as a result of its higher degree of outgrowth in the lungs.

Figure 1

Lm8 cells with different lung‐metastatic potential. A, Microphotographs of the high‐metastatic (LM8‐H) and low‐metastatic (LM8‐L) cells. B, Lung‐metastatic abilities of LM8‐L and LM8‐H cells. Representative in vivo bioluminescence (BL) images on day 12 (left) and quantitative analysis (right) are shown. BL signals from the lungs were normalized to those on day 0. n = 5, *P < .05. C, Representative HE‐stained images of lung tissues at 14 days after injection of the indicated LM8 sublines. D, Subcutaneous tumor formation. LM8 sublines were injected s.c. into C3H mice. Tumors were excised on day 30 (bottom photos). Their weights were measured, and average values are shown. n = 6. *P < .05

Lymphoid enhancer‐binding factor 1 regulated OS lung metastasis

To identify genes responsible for their differential lung‐metastatic abilities of LM8‐L and LM8‐H cells, their gene‐expression profiles were obtained through DNA microarray analysis (Figure 2A). Gene set‐enrichment analysis showed that genes associated with LEF1 and several pathways such as AKT and MAPK were enriched in LM8‐H cells (Figure 2B; Table S3). Among them, LEF1‐related genes were prominent and highly expressed in LM8‐H. Differential LEF1 protein‐expression levels in these LM8 sublines were confirmed by western blotting (Figure 2C). To examine the significance of LEF1 expression in lung‐metastatic ability of LM8‐H cells, Lef1 KO cell lines were generated using the clustered regularly interspaced short palindromic repeats (CRISPR)‐CRISPR‐associated protein 9 (Cas9) system, and 2 independent Lef1 KO clones (LM8‐H/Lef1‐KO1 and LM8‐H/Lef1‐KO2) were established using different gRNA (Figure 2D). When the Lef1 KO clones were i.v. injected into C3H mice, it was found that KO of Lef1 significantly suppressed the lung metastasis of LM8‐H cells (Figure 2E,F).

Figure 2

Lymphoid enhancer‐binding factor 1 (LEF1) expression level correlated with the lung‐metastatic potential of LM8 cells. A, Heat map analysis of microarray data of LM8 sublines. The heat map shows 500 genes with the greatest differential expression (fold‐change >2) between the LM8‐H and LM8‐L cells. B, Gene set‐enrichment analysis of Lef1 in the LM8 sublines. C, LEF1 protein‐expression levels in LM8‐H (H) and LM8‐L (L) cells. D, Establishment of Lef1‐KO sublines from LM8‐H (H) cells. LEF1 expression in LM8‐H/Lef1‐KO1 (KO1) and LM8‐H/Lef1‐KO2 (KO2) cells was examined by western blotting. E, Lung‐metastatic ability of the LM8‐H and LM8‐H/Lef1‐KO sublines. Representative in vivo bioluminescence images on day 15 (top) and quantitative analysis of BL signals (bottom) are shown. The inset graph shows an enlarged view from days 3 and 6. BL signals from the lungs were normalized by those on day 0. n = 3, *P < .05 (LM8‐H vs LM8‐H/Lef1‐KO1 or LM8‐H/Lef1‐KO2 cells). F, Representative HE staining of the lungs at 15 days after i.v. injection of the LM8 sublines

Lymphoid enhancer‐binding factor 1 regulated the extravasation step of LM8 sublines

Considering that the results described thus far were obtained using i.v. injection of LM8 sublines, promotion of metastasis by LEF1 observed in LM8‐H cells occurred at the step after the cells enter the circulation. To identify the step in which LEF1 plays a crucial role in LM8‐H, we first observed lung tissues at 48 hours after tail vein injection with LM8 sublines by fluorescent confocal microcopy, Strikingly, LM8‐H cells successfully extravasated into the lung parenchyma and proliferated, whereas all LM8‐L and LM8‐H/Lef1‐KO cells remained in the blood vessels (Figure 3A; Figure S3), suggesting that LEF1 functions in the extravasation step. To assess the extravasation abilities, which are the abilities of the LM8 sublines in vitro to adhere to and transmigrate through the blood vessels (adhesion and transmigration abilities), adhesion and transmigration assays were carried out with cultured monolayers of murine endothelial cells. Consistent with the observations in lung tissues, LM8‐L and LM8‐H/Lef1‐KO1 cells showed significantly reduced adhesion and transmigration abilities compared to LM8‐H cells (Figure 3B,C). These results strongly suggest that the extravasation step could be responsible for differential lung metastasis abilities of LM8‐L and LM8‐H cells and that LEF1 function is indispensable in the extravasation of LM8‐H cells to the lungs.

Figure 3

Extravasation abilities of LM8 sublines. A, Quantitative analysis (left) and representative images (right) of the lung 48 h after i.v. injection of the green‐fluorescently labeled LM8 sublines. Endothelial cells stained red. n = 9, *P < .05. B, Adhesion abilities. Number of fluorescently labeled cells (LM8 sublines) attached to the endothelial monolayer was counted. n = 3, *P < .05. C, Transmigration ability. Number of fluorescently labeled cells that migrated through the vascular endothelial monolayer was counted. n = 3, *P < .05

Identification of CYGB as a downstream effector of LEF1

Lymphoid enhancer‐binding factor 1 is a well‐known transcriptional factor that mediates nuclear responses to Wnt signaling.11 To identify genes that are regulated by LEF1 and are responsible for the differential extravasation abilities among the LM8 sublines, prometastatic genes downstream of Lef1 were first selected by genomewide meta‐analysis and DNA microarray data for the LM8 sublines (Figure 4A). By carrying out genomewide meta‐analysis with public data sets for human OS cells, 1912 genes were selected as upregulated genes detected in patients with highly metastatic OS. From the DNA microarray data for the LM8‐L, LM8‐H, and LM8‐H/Lef1‐KO sublines, 737 genes were selected as upregulated genes in correlation with the metastatic phenotype of the LM8 sublines. The genes selected by the genomewide meta‐analysis and the DNA microarray data were compared and 21 overlapping genes were extracted. Then, 13 out of 21 genes were selected based on their correlation with patient prognosis using the PROGgene database35 (Table S4; Figure S4). Expression levels of the 13 candidate genes were analyzed by RT‐PCR to examine their correlations with the metastatic phenotype of the LM8 sublines (Figure S5). Among them, the expression levels of Cygb and Ddx58 were well correlated with the metastatic phenotype of the LM8 sublines: their expression levels were high in LM8‐H cells and low in LM8‐L and LM8‐H/lef1KO cells. Their expression levels were confirmed by qualitative PCR (qPCR) (Figure 4B; Figure S6). Protein‐expression level of CYGB was sufficiently correlated with the metastatic phenotype of LM8 sublines: CYGB was highly expressed in LM8‐H cells but not in LM8‐L and LM8‐H/lef1KO cells (Figure 4C). For subsequent studies, we analyzed Cygb because the function of Cygb in metastasis has not yet been described, whereas Ddx58 has been reported to promote lung metastasis in several types of cancer.36, 37

Figure 4

Identification of CYGB as downstream effectors of lymphoid enhancer‐binding factor 1 (LEF1). A, Diagram of the process used to narrow down candidate genes by genomewide meta‐analysis. Number of genes indicated for each step is the number of selected genes. B, Relative Cygb mRNA expression levels in the LM8 sublines to LM8‐L analyzed by qRT‐PCR. C, Protein‐expression levels of CYGB in the LM8 sublines. CYGB, cytoglobin; OS, osteosarcoma

Function of CYGB in the extravasation of LM8 sublines

To assess the correlation between CYGB‐expression level and extravasation ability of the LM8 sublines, we established stable cell clones, including a CYGB‐overexpressing LM8‐L (LM8‐L/CYGB‐OE) clone and 2 independent Cygb‐KO LM8‐H (LM8‐H/Cygb‐KO1 and LM8‐H/Cygb‐KO2) clones using 2 gRNAs (Figure 5A). KO of Cygb increased the cell‐proliferation rate compared to LM8‐H cells and CYGB‐OE decreased it (Figure 5B). These results are consistent with our results showing that Lef1‐KO cells increased the proliferation rate of LM8‐H cells (Figure S2), suggesting that CYGB is a downstream effector of LEF1. Adhesion and transmigration abilities were significantly higher in LM8‐L/CYGB‐OE cells compared to LM8‐L cells (Figure 5C). In contrast, KO of Cygb significantly decreased the adhesion and transmigration abilities of LM8‐H cells (Figure 5D). These results strongly suggest that CYGB promoted extravasation with the LM8 sublines by increasing their adhesion and transmigration abilities.

Figure 5

Function of CYGB in the extravasation of LM8 sublines. A, Protein‐expression levels of CYGB in LM8‐H/Cygb‐KO and LM8‐L/CYGB‐OE cells. B, Cell‐proliferation rates of the LM8 sublines. C,D, Ability of the LM8 sublines to adhere to and transmigrate to the vascular endothelial monolayer. Adhesion and transmigration abilities of LM8‐L/CYGB‐OE (C) and LM8‐H/Cygb‐KO (D) cells were quantitatively analyzed (graphs) by determining the number of fluorescently labeled cells attached to (upper) and transmigrating through (lower) a vascular endothelial monolayer, respectively. Representative images are shown depicting the fluorescently labeled cells that were counted (photos). n = 3, *P < .05. Bars, 100 μm. CYGB, cytoglobin

Knockout of Cygb suppressed the lung‐metastatic ability of LM8‐H cells in vivo

To assess the function of CYGB in lung extravasation, the metastatic ability of Cygb‐KO LM8‐H (LM8‐H/Cygb‐KO) cells was examined in vivo. LM8 sublines were i.v. injected into C3H mice and BL signals in the lungs were monitored for 16 days. Strikingly, LM8‐H/Cygb‐KO cells showed significantly decreased lung metastasis compared to LM8‐H cells (Figure 6A). HE staining of the lungs from mice 16 days after injection of the LM8 sublines confirmed the regulatory function of CYGB in lung metastasis: KO of Cygb in LM8‐H cells significantly reduced the number of lung‐metastatic foci (Figure 6B). The size of foci in the lungs was significantly smaller in the LM8‐H/Cygb‐KO group compared to the LM8‐H group (Figure 6C). LM8‐H/Cygb‐KO cells failed to extravasate into the lung parenchyma and remained and grew in the lung blood vessels (Figure 6D; Figure S3). Together, these results support the function of CYGB in extravasation in lung metastasis of the LM8 sublines.

Figure 6

Knockout of Cygb suppressed the lung‐metastatic ability of LM8‐H cells in vivo. A, Bioluminescence (BL) images of mice injected with the indicated cells on day 16. Relative BL signals to day 0 are shown in the right box plot. B, Representative HE staining of the lung at 16 days after i.v. injection of LM8 sublines (left). Number of foci in the lung on day 16 (right). n = 4 *P < .05. C, Representative HE‐stained lung (left) and size of foci in the lung (right) at 16 days after i.v. injection. n = 32 *P < .05. D, Enlarged representative images of metastatic foci of the LM8 sublines on day 16. Cygb, cytoglobin gene

DISCUSSION

In the present study, we identified CYGB as an important regulator of OS extravasation and highlighted the importance of the LEF1‐CYGB regulatory axis in OS metastasis to the lungs. To our knowledge, this is the first report showing a functional connection between LEF1 and CYGB.

To explore the possibility that LEF1 directly regulates the expression of CYGB, genome analysis of transcription factor binding sites was carried out using several databases38, 39, 40, 41, 42 and showed that Cygb has potential LEF1 binding sites in the promoter region (Figure S7). LEF1 may directly regulate Cygb expression. As CYGB is known to be induced by HIF‐1α, NF‐κB, and other inflammation‐related transcription factors,23 the interaction of LEF1 with these transcription factors may be important for regulating Cygb expression. Among the 737 genes that were differentially expressed between LM8‐L and LM8‐H cells, 21 genes were selected as candidate genes responsible for lung metastasis of LM8‐H by genomewide meta‐analysis, using public data sets from patients with high‐ and low‐metastatic OS. Expression of 13 out of 21 genes correlated well with cancer prognosis in patients with OS (Figure S4). The gene products of the 13 selected genes function as cell receptors/transporters, cytokines, GTPase‐activating proteins and in cell movement (Table S3) and most of the candidate genes have previously been associated with malignant progression. Cygb is defined as a tumor suppressor gene, and its expression is suppressed by the methylation of nucleotides in the promoter region of many types of cancer.19 To our knowledge, the function of CYGB in metastasis, especially extravasation and cell‐cell interactions in cancer, has not yet been reported.

The KO of Cygb in LM8‐H cells clearly demonstrated the significance of CYGB in lung metastasis. The number and size of metastatic foci in LM8‐H/Cygb‐KO cells were strongly suppressed, even though the proliferation rate of LM8‐H/Cygb‐KO cells was higher than that of LM8‐H cells (Figure 5B). Moreover, LM8‐L and LM8‐H/Cygb‐KO cells showed a higher frequency of remaining in pulmonary blood vessels than LM8‐H (Figure S7), supporting lower extravasation ability of these cells. However, the number of metastatic foci of LM8‐H/Cygb‐KO cells was higher than that of LM8‐L cells, suggesting that other downstream effectors of LEF1 also contribute to the lung‐metastatic phenotype of LM8‐H cells. Furthermore, the ability of adhesion and transmigration (Figure 3B,C) and the level of Cygb expression (Figure 4B) were higher in LM8‐H/Lef1‐KO1 cells than in LM8‐L cells. These results suggest that the expression of Cygb could also be regulated by a LEF1‐independent mechanism.

We conducted 2 experiments to elucidate the molecular mechanism by which CYGB functions directly in promoting extravasation based on the proposed functions for CYGB. First, as CYGB has been suggested to act as an NO dioxygenase43 and as NO prevents endothelial activation by inhibiting the expression of adhesion molecules such as vascular cell adhesion molecule 1 (VCAM‐1) and intercellular adhesion molecule 1 (ICAM‐1),44, 45 we investigated whether CYGB reduces NO levels in the lung blood vessels, in turn, reducing the expression of adhesion molecules. We found only slight differences in the total NO species levels and little influence of inhibitors of NO species among the LM8 sublines (Figure S9). Although NO reduction by CYGB may contribute to metastasis, the subtle differences in NO levels among the LM8 sublines cannot explain the significant differences in lung‐metastatic potential found among them, suggesting that NO is not a major player in CYGB‐mediated extravasation. Second, as arachidonic acid (AA) increases the migratory activity of cancer cells through activation of RhoA and RhoC,46 and as recent findings imply that CYGB functions in lipid oxidation,47 thereby modifying several types of phospholipids such as phosphatidylinositol (3,4,5)‐trisphosphate (an important signal mediator of the AKT signaling pathway) to form arachidonyl‐containing lipid, we examined the content of AA in the LM8 sublines. No difference was found in the AA contents among the LM8 sublines (Figure S10), suggesting that AA did not function as an effector molecule of CYGB to promote lung metastasis in the LM8 sublines. These results suggest that the function of CYGB which has not yet been identified contributes to promotion of the extravasation.

As we used an i.v. injection model to examine the lung‐metastatic abilities of the LM8 sublines, our evaluation was limited to the postcirculation steps of metastasis. To explore the differences between LM8‐H and LM8‐L cells in precirculation steps, we analyzed the EMT and invasion steps. Although the morphologies of LM8‐H and LM8‐L cells differ, the expression levels of EMT‐related genes (based on microarray data) and the migration and invasion abilities (assessed by wound‐healing and invasion assays, respectively) were not significantly different between either cell line (Table S4; Figure S11). These findings suggest that their metastatic abilities may not be so very different before the intravasation steps. Because in vitro analyses using cell‐adhesion and transmigration assays can also assess their intravasation capacities, the results suggest that LM8‐L also has low intravasation ability. Therefore, LM8‐L cells may further reduce the frequency of metastasis from the primary tumor site to the lungs when assessed in a model of the total metastasis process. Further studies elucidating the entire LEF1‐CYGB axis should contribute to the development of new therapeutic approaches for preventing OS metastasis to the lungs.

CONFLICTS OF INTEREST

Authors declare no conflicts of interest for this article.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.