IRF2 Destabilizes Oncogenic KPNA2 to Modulate the Tumorigenesis of Osteosarcoma via Regulating NF-κB/p65

Background: Osteosarcomas (OS) are the most frequent primary malignant bone tumor. Emerging evidence revealed that karyopherin alpha 2 (KPNA2) was strongly associated with the tumorigenesis and development of numerous human cancers. The aim of the present study was to investigate the expression pattern, biological functions and underlying mechanism of KPNA2 in OS. Methods: Bioinformatics TFBIND online was applied to forecast the transcription factor (TF) binding sites in the promoter region of KPNA2. The expression prole of KPNA2 in OS tissues were rstly assessed using TARGET dataset. The expression of KPNA2 in clinical OS samples and normal human adjacent samples were analyzed by RT-qPCR and western blot. CCK8, colony formation, wound-healing, and Transwell assays were used to assess cell viability, proliferation and migration in vitro, and in vivo experiments were performed to explore the effects of KPNA2 and interferon regulatory factor-2 (IRF2) on tumor growth. In addition, the correlation between IRF2 and KPNA2, and their roles on the NF-κB/p65 was investigated using chromatin immunoprecipitation (ChIP), RT-qPCR, western blot and dual-luciferase assays. Results: KPNA2 was obviously upregulated while IRF2 was signicantly decreased in OS tissues and cell lines, as well as they were negatively correlated with each other. KPNA2 knockdown remarkably suppressed OS cell growth, migration, invasion in vitro and tumor growth in vivo, while IRF2 knockdown exerts an opposing effect. IRF2 binds to KPNA2 promoter to modulate the tumorigenic malignant phenotypes of OS via regulating NF-κB/p65 signaling. Conclusion: The present study demonstrated that KPNA2 performed the oncogenic function, possibly regulating tumorigenesis through NF-κB/p65 signaling pathway. Importantly, IRF2 was conrmed to serve a potential upstream TF of KPNA2 involving in the regulation of NF-κB/p65 pathway in OS.


Introduction
Osteosarcoma (OS) is the frequent primary malignant bone tumor, affecting mainly pediatric and adolescents (1), which is composed of malignant mesenchymal cells producing osteoid and/or immature bone (2). It typically formed in the metaphysis of long bones, speci cally the proximal tibia, distal femur, and proximal humerus accompanied by bone swelling and pain (3). It is common that metastasis to the lungs of OS (4). Before the use of neoadjuvant and adjuvant chemotherapy, approximately 90% of OS patients died due to pulmonary metastases (5). OS is characterized by high levels of genomic instability.
However, the molecular basis involved in OS remains unclear, and continuing to seek new treatments is urgently needed.
Karyopherin alpha 2 (KPNA2, 58 kDa) is one of seven members of the karyopherin α-family (6, 7); Dysregulation of KPNA2 had been reported serving as as a potential biomarker in several malignancies, including breast cancer (8), gastric cancer (9), lung cancer (10) and glioma (11). KPNA2 served as the adaptor to transfer p65 to the nucleus to identify the classic nuclear localization signal (7,12). A Previous study had shown that KPNA2 interacts with p65, and facilitated the NF-κB p65/p50 nuclear transportation in TNF-α-stimulated lung cancer cells (13). As is well-known, NF-κB/p65 signaling is rest and concentrated in the cytoplasm until being activated and migrating to the nucleus in response to stress. NF-κB is involved in the regulation of cell growth and survival, and its abnormal activation is associated with the malignant progression of various cancers (14,15), includes OS (16). Although, a recent study revealed that KPNA2 promoted NF-κB activation and thus subsequently accelerate the development of osteoarthritis (17). However, the molecular pathways regulated by KPNA2-mediated NF-κB/p65 in OS are yet to be elucidated.
To better understand the molecular mechanism of KPNA2 in OS, we rst applied an online database of TARGET to seek KPNA2-related factors and ultimately found that the interferon regulatory factor-1 (IRF2) expression was positively correlation with KPNA2 in OS tissues. Besides, transcription factors (TFs) are a kind of proteins that speci cally recognized DNA through consensus sequences, thus to control chromatin and transcription, guiding expression of the genome (18). To identify the TF of KPNA2 transcription, the online software of TFBIND (http://tfbind.hgc.jp/) was employed to identify putative binding sites in the promoter region of KPNA2. Speci cally, several canonical IRF2-binding sites on the promoter region of KPNA2 were observed and IRF2 was then extracted from 147 candidate genes. IRF2, belonging to the IRF family, was widely expressed in various tissues (19). Among several cancer types, belong to IRF family, IRF1 signaling pathways may directly induced p21-dependent G0/G1 cell cycle arrest and p21-independent modulation of surviving (20). IRF2 was proved to serves as an important regulator in acute myeloid leukemia by targeting INPP4B (21). Recent study has indicated that IRF2 was able to suppress the strengthen of cell migration and invasion in OS, which were mediated by miR-18a-5p (22). Similarly, IRF2 was not expressed or at a low level in OS tissues (23). Intriguingly, IRF was already identi ed as a functional TF in non-small-cell lung cancer (NSCLC) that suppressed KPNA2 expression (24). Therefore, we speculate that IRF2 may negatively regulated KPNA2 as its upstream TF to modulate OS progression by a p65-dependent signaling.
This study aimed to investigate the role of IRF2 in modulating KPNA2 expression, which may serve an important role in p65 nuclear importation in the progression of OS. Here, we found that KPNA2 de ciency suppressed the malignant behaviors of OS cells, and that the underlying mechanisms involved were regulated by IRF2 and associated with NF-κB signaling pathway by p65 nucleus translocation. Taken together, our ndings demonstrated that KPNA2 may serve as a new potential prognostic indicator and therapeutic target for OS.

Cell culture
Four human OS cell lines (Saos-2, HOS, U2OS and MG-63) and human normal osteoblastic cell line (hFOB 1.19) were obtained from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The cells were grown in Dulbecco's modi ed Eagle's medium (DMEM; Gibco) harboring with 10% fetal bovine serum (FBS) (Gibco). All OS cell lines were cultured in a humidi ed incubator under 5% CO 2 at 37 °C, while hFOB1.19 cells were grown at 34 °C.

Patients and tissue specimens
Twenty-ve paired tumor samples and their adjacent non-tumor tissues from patients who had undergone surgery were obtained from Zhongshan Hospital, Fudan University. This study was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (Y2014-185) according to the Declaration of Helsinki, and written informed consents were obtained from all the patients. No patients had undergone chemotherapy, radiation therapy or other related targeted therapy before surgery. The diagnosis of OS was con rmed by at least two pathologists. All surgical tissue samples used in our study were immediately placed in liquid nitrogen and then stored at -80 °C until use.

Cell proliferation assay
The cell proliferation assay was performed using a Cell Counting Kit (CCK-8, Dojindo, Japan) following the manufacturer's instructions. Cells at a density of 5 x10 3 were added into the 96-well plate and 10 microliters of CCK-8 solution was added to each well at 1, 2, 3, 4 and 5 days at 37°C. An additional 1 h later, the absorbance at wavelength of 450 nm was then measured under a microplate reader.

Colony formation assay
For the colony formation assay, 500 cells were plated into each well of a 6-well culture plate. The plates containing DMEM were incubated at 37°C for 2 weeks. After washed with PBS for three time, cells were xed by 4% paraformaldehyde for 10 min at room temperature, followed by staining with 0.5% crystal violet solution for another 20 min. Lastly, the visible colonies more than 50 cells were manually counted and imaged under a microscope.

Transwell invasion assay
Cells at the density of 1×10 4 were seeded into a diameter Transwell palte with 8-μm pores (Sigma-Aldrich). The upper chamber of the plate was added with 50 µl of Matrigel collagen, and 600 μL of complete DMEM was added to the lower chambers, and then the cells were incubated for 24 h. The cells on the upper layer were removed and the invasive cells were xed with 4% formaldehyde for 20 min, and then stained with crystal violet for 15 min. Cells that had invaded the bottom surface of the lter were counted to assess the invasive ability. Invaded cells were quanti ed by at least ve elds of view under a light microscopy (Leica) to obtain the representative images.
Wound healing assay Cells were cultured in six-well plates. After reaching 90% con uence, a 200 µl pipette tip was used to create the scratch wounds in the cell monolayer. Representative images of cell migration were photographed under a light microscopy (Leica) at 0 and 24 h after wounding. The migration ability was assessed by measuring changes in the size of the wound width or area with Image J software.

Western blot assay
The protein was lysed from tissues and cells with RIPA buffer (Thermo Fisher Scienti c). Protein's concentration was assessed by the bicinchoninic acid (BCA) assay kit (Thermo Fisher Scienti c). Equal amounts of protein samples were separated by 12% SDS-PAGE gels and transferred onto polyvinylidene di uoride (PVDF) membranes. After incubating with 5% not-fat milk for 2 h at room temperature, to hatch the blots with the primary antibodies including anti-KPNA2, IRF2, NF-kB p65, p-NF-kB p65, GAPDH overnight at 4°C. Hatching the horseradish peroxidase-conjugated secondary antibodies at romm temperature. The endogenous GAPDH is the internal reference protein. The protein band signals were visualized on an ECL detector (Pierce) and quanti ed by scanning the densitometry using image J software.
Chromatin immunoprecipitation (ChIP) assay ChIP assays were performed using a kit (Sigma-Aldrich) following the protocol provided by manufacturer.
To hatch the diluted DNA-protein complex, the antibodies of anti-IRF2 and mouse IgG (Sigma-Aldrich) were added in the presence of protein A/G beads and incubated at 4°C overnight. RT-qPCR assay was applied to examine the ChIP DNA samples. IgG was the negative control.

Dual luciferase test
The wild (WT) and -mutant (MUT) of KPNA2 were inserted into the pGL3 promoter vector, which was transfected into U2OS and MG-63 cells using Lipofectamine 2000 (Invitrogen) along with plasmid of IRF2 overexpression or empty plasmid (NC). 48 h later, luciferase activity was measured using a dualluciferase reporter assay system (Promega) following the manufacture's protocols.
Tumor xenograft assay

Histological analysis and immunohistochemistry
Xenotransplant tumor samples were isolated and xed in 4% paraformaldehyde. Para n was used to embed the tumor tissues and then sectioned into a thickness of 5-μm. The para n sections were dewaxed, hydrated, and then stained with hematoxylin and counterstained with eosin. Antigen were retrieved by citrate buffer and blocked with 3% H2O2, immunohistochemistry (IHC) was performed with diluted primary KI67 antibody overnight at 4°C, followed by incubation with secondary antibody at room temperature. The slides were developed by diaminobenzidine (DAB) and stained with hematoxylin. IHC staining pictures were obtained under a light microscope.

Statistical analysis
All obtained data are expressed as the mean ± standard deviation (SD). Differences Student's t-test or one-way ANOVA followed by a Tukey's post hoc test were used to compare data between two group or among multiple groups. Statistical difference was analyzed using GraphPad Prism 8.0 software. followed by a Tukey's post hoc test.

Overexpression of KPNA2 in the osteosarcoma tissues
Based on the TARGET database, we previously revealed KPNA2 expression was notably upregulated in the OS tissues when compared with the non-tumor tissues ( Figure 1A). To examine whether KPNA2 altered in clinical OS, qRT-PCR assay was performed to examine KPNA2 levels in 25 paired of cancerous OS samples and their adjacent normal samples. Figure 1B displayed that the mRNA levels of KPNA2 was obviously up-regulated in OS samples compared with that normal samples. Representative IHC images showed similar results ( Figure 1C). Furthermore, the protein levels of KPNA2 in 8 cases of OS samples were obviously elevated as well ( Figure 1D). Besides, compared to hFOB1.19 cells, a higher level of KPNA2 in four OS cell lines, including U2OS, HOS, Saos2 and MG-63, was able to been observed ( Figure  1E). KPNA2 expression was the highest in U2OS and MG-63, which were chosen to use for subsequent analyzes. These results suggest that KPNA2 might play a vital role in the progression of OS.

Transcription factor IRF2 speci cally regulates KPNA2 expression in osteosarcomas
Considering that the regulation of KPNA2 in OS focused on the transcriptional level, we performed bioinformatic analysis for seeking a transcription factor (TF) in the KPNA2 promoter region. Then we extracted 147 TFs of KPNA2 from online dataset. Data from TARGET datasets showed total of 9792 genes were negatively associated with KPNA2 in OS (R>0.2, FDR<0.05). We further found 5950 genes were overlapped in down-regulated gens from GSE157322 dataset, and 31 genes were overlapped in 147 TFs of KPNA2. Therefore, a total of 26 genes were at the intersection of three (Figure 2A). According to the correlation index of these 26 genes with KPNA2, we chosen the top 5 TFs of KPNA2, includes RFX1, STAT3, IRF2, PPARA and MZF1. To assess whether these ve TFs could alter KPNA2 expression, we overexpressed these TFs in two OS cell lines. As shown in Figure 2B, these results demonstrated that only overexpression of IRF2 signi cantly suppressed KPNA2 expression in two OS cell lines, while other four TFs had no signi cant effect on the KPNA2 expression. Meanwhile, due to KPNA2 was overexpressed in four OS cell lines, we also determined the change of these ve TFs after KPNA2 knockdown using RT-qPCR. Silence of KPNA2 markedly upregulated the mRNA levels of IRF2 while other four TFs had not obviously changed response to KPNA2 knockdown ( Figure 2C). Therefore, we decided to use IRF2 for subsequent experiments.
Through the TARGET dataset, IRF2 expression was mildly downregulated in OS tissues without signi cant difference ( Figure 2D), while IRF2 expression was strongly negatively correlated with KPNA2 ( Figure 2E). Consistently, the mRNA level of IRF2 was signi cantly downregulated in 25 cases of clinical OS tissue samples when compared with adjacent normal tissues ( Figure 2F). Besides, a ChIP test was conducted in two OS cell lines and hFOB1.19 to evaluate the binding relationship of IRF2 with KPNA2, and we found that the enrichment of IRF2 binding was prominently decreased in two OS cell lines comparing to the hFOB1.19 cells ( Figure 2G). The present ndings showed that IRF2 might be one of major regulators to regulate KPNA2 expression in OS. To further determine if IRF2 bound to the KPNA2 promoter, a Dual-luciferase assay revealed that IRF2 was able to dramatically reduce the luciferase activity of KPNA2-WT, but not in KPNA2-MUT ( Figure 2H). Collectively, IRF2, as a functional TF, could bind to the KPNA2 in OS cells.

IRF2 de ciency may cooperate with KPNA2 to regulate cell proliferation and tumor growth of OS cells in vivo and in vitro
Since a negative correlation existed among KPNA2 and IRF2, we investigate their effects on cell proliferation, migration, invasion, and cell cycle. Firstly, in four OS cell lines, IRF2 protein expressions were lower than that in hFOB1.19 cells ( Figure 3A). Down-regulation of IRF2 or KPNA2 in U2OS cells was achieved by transfections of IRF2 or KPNA2 knockdown vectors (shKPNA2 and shIRF2), as con rmed by western blot ( Figure 3B). Regarding the malignant phenotypes of OS cells, KPNA2 knockdown inhibited the cell viability, proliferation while IRF2 knockdown had the opposite effects and partially rescued abovementioned malignant phenotypes suppressed by KPNA2 knockdown in vitro ( Figure 3C-D). In vivo, KPNA2 knockdown remarkably reduced tumor weight and tumor volumes while IRF2 knockdown promoted tumor growth, and IRF2 knockdown could weaken the KPNA2 knockdown-medicated tumor growth inhibition ( Figure 3E and 3F). The KI67 expression was reduced by KPNA2 knockdown while elevated by IRF2 knockdown ( Figure 3G). These ndings demonstrated that IRF2 silence partially attenuates the impact of KPNA2 knock-down on OS growth.

IRF2/KPNA2 might regulate migration and invasion of in osteosarcoma cells in a p65-dependent manner
Regarding the malignant phenotypes of migration and invasion, KPNA2 knockdown inhibited the abilities of migration and invasion while IRF2 knockdown had the opposite effects and partially rescued abovementioned malignant phenotypes suppressed by KPNA2 knockdown OS cells ( Figure 4A-B). Due to KPNA2 was able to activate NF-κB/p65 signaling pathway (25), we further clarify the expressions of NF-κB/p65 and p-NF-κB/p65 in OS cells after transfected or co-transfected with shKPNA2 or/and shIRF2. The protein level of p-NF-κB/p65 was apparently decreased by shKPNA2 with a slight change of NF-κB/p65 expression ( Figure 4C). However, the activation protein of p-NF-κB/p65 increased in shIRF2infected cells; more importantly, IRF2 silence could obviously attenuate the inhibitory effect of KPNA2 silence on the p-NF-κB/p65 expression ( Figure 4C). Thus, IRF2 could negatively alter KPNA2 by directly binding to the KPNA2 promoter and participated in OS progression through NF-κB/p65 signaling pathway.

Discussion
Osteosarcomas is relatively rare but devastating (26). Unfortunately, the introduction of novel adjuvant chemotherapy after aggressive surgical resection has temporarily improved overall 10-year survival, but has not signi cantly improved patient survival since the 1990s (27). Therefore, there is of great signi cance to identify novel molecules, which further helps to develop effective methods to diagnose and treat this malignant bone tumor. Here, we draw a conclusion that IRF2 binding to KPNA2 to regulate malignant biological properties of OS cells via regulating NF-κB/p65 signaling pathway. This evidence may provide new ideas for the diagnosis and treatment of osteosarcoma.
Recently, several studies have linked KPNA2 to various cancers, such as lung, breast, colon cancer. High KPNA2 was positively related to cancer invasiveness and poor prognosis, thus regarded KPNA2 as a potentially relevant therapeutic target for patients with different cancers (28). KPNA2 was involved in several cellular biological processes, including cell differentiation, development, viral infection, immune response and transcriptional regulation (29). Similarly, our study illustrated that KPNA2 was dramatically elevated in OS samples comparing to normal samples. Although KPNA2 was proved to frequently express in OS as a novel marker for the diagnosis, as well as in chondrosarcoma and Ewing sarcomas (30), the functions of KPNA2 in osteosarcoma is unclear. In the present study, data mining and bioinformatics analysis indicated that KPNA2 was overexpressed in OS patients from TARGET dataset, and the experiments veri ed high KPNA2 level in clinical OS samples and OS cell lines. Additionally, KPNA2 knockdown inhibited the proliferation, migration and invasion in two OS cell lines, and remarkably reduced tumor weight and tumor volumes in vivo. These ndings revealed that KPNA2 might play a crucial role in the biological progress of OS.
Interferon regulatory factor-2 (IRF2) exerted anti-tumor effects in several human cancers. For instance, IRF2 could suppress cell proliferation and migrate ability, and promote cell apoptosis in non-small cell lung cancer cells (31). IRF2 might play as a tumor suppressor by regulating P53 signaling in gastric cancer (32). Besides, IRF2 was proved to serve as a tumor suppressor in patients with hepatocellular carcinoma, whose inactivation led to impaired TP53 function (33). The current study highlighted that KPNA2 could negatively alter the expression of IRF2 in OS cells. Meanwhile, data mining in the GSE157322 and TARGET datasets, we discovered that IRF2 could bind to the promoter of KPNA2 and activate KPNA2 expression through bioinformatics analysis for TF prediction. This underlying mechanism was consistent with a previous report that IRF2 could bind to miR-1227 promoter, thus inhibited tumor growth (23). Besides, IRF2 was downregulated obviously, which was negatively associated with KPNA2 in OS. More importantly, IRF2 knockdown promoted malignant behaviors, which were subsequently suppressed by KPNA2 knockdown. These rescuing effect of IRF2 on KPNA2 was also re ected in tumor growth in vivo. These ndings demonstrated that IRF2 silence partially attenuated the impact of KPNA2 knock-down on osteosarcomas progressions.
Previously, researches had shown that KPNA2 played a pivotal role in melanoma development through activating NF-κB/p65 signaling pathways (25). In addition, IRF could cooperate with NF-kB p65 to promote the e cacy of T cell-related immunotherapy in neuroblastoma, which revealed a synergistic regulatory relationship between IRF and p65 (34). What's more, IRF2 could regulate NF-κB activity via the modulation of NF-κB subcellular localization (35). In our study, the deletion of KPNA2 reduced p-NF-κB/p65 expression, and futile total NF-κB/p65 expression, whereas the opposite results were found after knockdown of IRF2. KPNA2 and IRF2 had opposite regulatory effects on the activation of NF-κB/p65 pathway. Coincidentally, our results also con rmed that the inhibitory effect of KPNA2 was partially recovered by IRF2 silence. Thus, elevated KPNA2 contributes to the progression of OS by negatively regulating IRF2 via NF-κB/p65 pathway.

Availability of data and materials
All the data in the results of this study can be obtained on reasonable request from the corresponding authors.

Con ict of interest
The Authors declare that they have no con ict of interests.

Funding
No.   for detecting KI67 expression. The data was presented as the mean ± SD from three independent experiments. *p<0.05, **p<0.01 and ***p<0.001.