The role of FOSL1 in stem-like cell reprogramming processes

Cancer stem-like cells (CSCs) have self-renewal abilities responsible for cancer progression, therapy resistance, and metastatic growth. The glioblastoma stem-like cells are the most studied among CSC populations. A recent study identified four transcription factors (SOX2, SALL2, OLIG2, and POU3F2) as the minimal core sufficient to reprogram differentiated glioblastoma (GBM) cells into stem-like cells. Transcriptomic data of GBM tissues and cell lines from two different datasets were then analyzed by the SWItch Miner (SWIM), a network-based software, and FOSL1 was identified as a putative regulator of the previously identified minimal core. Herein, we selected NTERA-2 and HEK293T cells to perform an in vitro study to investigate the role of FOSL1 in the reprogramming mechanisms. We transfected the two cell lines with a constitutive FOSL1 cDNA plasmid. We demonstrated that FOSL1 directly regulates the four transcription factors binding their promoter regions, is involved in the deregulation of several stemness markers, and reduces the cells’ ability to generate aggregates increasing the extracellular matrix component FN1. Although further experiments are necessary, our data suggest that FOSL1 reprograms the stemness by regulating the core of the four transcription factors.

of the four TFs when applying SWItch Miner (SWIM) software on two datasets of gene expression data from human glioblastoma tissues and cell lines 21 . Using SWIM to search for switch genes shared by both analyzed glioblastoma datasets, the authors identified FOS-like transcription factor (FOSL1) as the most promising one.
Several findings suggest a possible role of FOSL1 in stemness. Indeed, FOSL1 was downregulated in stemlike cells and negatively correlated with the four TFs identified by Suvà and colleagues 13,21 . It is well known that CSC phenotype and epithelial-to-mesenchymal transition (EMT) are intimately interconnected 22 . Interestingly, FOSL1 was positively correlated with genes encoding proteins crucial for cell matrix adhesion (e.g. FN1 and a-actin), a key process involved in the EMT 21 . Therefore, FOSL1 expression could promote cell differentiation by repressing the four TFs and the EMT. However, few experimental data exist on the direct involvement of FOSL1 in stemness mechanisms.
In this paper we validated by in vitro experiments the role of FOSL1 in regulating the four neurodevelopmental TFs involved in the differentiation process of GSCs, as previously suggested by our computational study 21 . Moreover, we provided evidence on the effect of FOSL1 on the expression of several stemness markers and on the modulation of cell-matrix adhesion.

Results
FOSL1 negatively regulates four stem-like associated TFs in NTERA-2 and HEK293T cell lines. Given the role of the four TFs (SALL2, SOX2, OLIG2, and POU3F2) in reprogramming differentiated glioblastoma cells into stem-like cells 13 and the potential role of FOSL1 in regulating them as postulated by Fiscon and colleagues, our first goal was to investigate the molecular relationship between FOSL1 and the four TFs.
We used the expression data reported in Human Protein Atlas database 23 to select two cell lines suitable for this purpose, the NTERA-2 and HEK293T cell lines, both with low FOSL1 expression levels and discrete levels of the other four TFs. Moreover, the NTERA-2 and HEK293T are embryonic arising cell lines. In particular, NTERA-2 line arises from a malignant pluripotent embryonal carcinoma, and the HEK293T line arises from an embryonic kidney. A recent study demonstrated HEK293T stemness features 24 . The origin and characteristics of both cell lines are summarized in Fig. 1A.
As represented in Fig. 1, we verified the expression levels of all TFs in both cell lines. In order to obtain an overview of the relative expression levels of all TFs, we compared the results in NTERA-2 and HEK293T using a pool of equimolar cDNA of 8 immortalized cell lines as control sample.
The HEK293T cell line showed high levels of POU3F2, moderate levels of SALL2, and low levels of FOSL1, SOX2, and OLIG2. The NTERA-2 cell line showed high levels of SOX2, moderate levels of SALL2 and FOSL1, and low levels of OLIG2 and POU3F2.
Since we confirmed low FOSL1 expression levels in both cell lines, as reported in the Human Protein Atlas, we transfected HEK293T and NTERA-2 cell lines with a construct containing FOSL1 cDNA (pCMV-FOSL1). Using real-time polymerase chain reaction (PCR), we analyzed the expression levels of the four TFs after 48 h from transfection and compared the levels with those obtained in cell lines transfected with the control vector (empty vector).
Gene expression analysis performed before and after transfection revealed reduced expression of the four TFs in the HEK293T cell line and reduced expression of OLIG2 and POU3F2 in the NTERA-2 cell line (Fig. 2), suggesting that the switch gene FOSL1 negatively regulates the four TFs.

FOSL1 directly interacts with four TFs promoters.
In light of previous results, we used the HEK293T cell line to investigate the molecular mechanisms behind the expression level changes of the four TFs after FOSL1 overexpression. In order to determine whether FOSL1 directly or indirectly regulates expression level changes of the four TFs, we performed a chromatin immunoprecipitation (ChIP) experiment to verify whether FOSL1 binds the promoter regions of the four TFs. We selected the promoter of the four TFs considering the genomic regions upstream of the start of transcription containing the FOSL1 consensus binding sequence as previously described in 20 and reported in the Jaspar database (http:// jaspar. gener eg. net/ matrix/ MA0477. 2/). As shown in Supplementary Fig. 1, we extracted and sonicated chromatin from HEK293T cell lines transfected with the pCMV empty vector or containing FOSL1 (pCMV-FOSL1). Immunoprecipitation was performed using a specific antibody for FOSL1.
Using specific primers, we compared the quantity of the selected promoter regions, for the four TFs, in samples immunoprecipitated with FOSL1 antibody and control samples (No AB). Using real-time PCR, we found an enrichment relative to promoter regions of all four TFs in the immunoprecipitated samples if compared with controls. As shown in Fig. 3, we found an enrichment of 150-fold for SALL2 and SOX2 promoter regions, 900fold for POU3F2 promoter region and 4000-fold for OLIG2 promoter region.
Taken together, these results suggest that FOSL1 directly regulates the four TFs, binding their promoter regions.

FOSL1 effects on staminal markers.
Since we demonstrated that FOSL1 directly regulates the expression of the four TFs involved in GBM stemness, we investigated the effect of FOSL1 in the stemness process. We analyzed the expression levels of 15 pluripotent stem cell markers in the total protein extract from HEK293T cells transfected with empty vector or pCMV-FOSL1 vector.
As shown in Fig. 4, after the overexpression of FOSL1, we observed a statistically significant deregulation of 7 of the 15 stemness markers analyzed. In particular, 6 markers were down regulated (SNAIL, SOX17, VEGFR2, OTX, HCG, and TP63) and one was upregulated (NANOG).
These results highlight the role of FOSL1 in stemness processes.  21 , FOSL1 seems to directly correlate with higher levels of extracellular matrix and focal adhesion components. In order to investigate the relationship between FOSL1 and extracellular components, we analyzed the ability of cells to generate aggregates in the extracellular matrix mix from Engelbreth-Holm-Swarm cells, the Matrigel. HEK293T cell lines transfected with FOSL1 (pCMW-FOSL1) or the empty vector were photographed every two days. After one week, the cells formed aggregates. As shown in Fig. 5A, cells that overexpressed FOSL1 seemed to generate less aggregates than cells transfected with the empty vector. This suggests a lower motility of cells transfected with FOSL1, which may be due to the greater quantity of extracellular matrix components expressed in these cells when FOSL1 was overexpressed. The number of aggregates formed by each cell lines in the different conditions analyzed are reported in Fig. 5B.

Discussion
In recent decades, much attention has been paid to cancer stem-like cells, which represent the critical subset within the tumor mass responsible for perpetuating the tumor, leading to its aggressiveness and contributing to therapy resistance, recurrence, and metastasis. Several efforts have been made to discover novel therapeutic strategies that could promote the differentiation of cancer stem-like cells in order to halt cancer growth and potentially affect patient outcome.
Increasing evidence suggests that modulation of the expression of a small set of TFs is able to maintain the stem-like phenotype and prevent cell differentiation. In particular, Suvà and colleagues identified four neurodevelopmental TFs (i.e., SOX2, SALL2, OLIG2, and PUO3F2) that were selectively expressed in GSCs and whose induction was sufficient to fully reprogram differentiated cells into GSCs. Thus, the identification of an upstream pathway that regulates the four TFs could be crucial to control cancer stem-like cell differentiation in human glioblastoma and to serve as a successful target for therapeutic strategies. With this aim, we based our study on www.nature.com/scientificreports/ the results of Fiscon and colleagues 26 , who applied SWIM algorithms to two datasets of gene expression analysis from glioblastoma tissues and cell lines. SWIM was developed to computationally identify key genes (denoted as "switch genes") that are likely to be critically associated with drastic changes in cell or tissue phenotype. SWIM phenotype-specific applications are broad and include the identification of switch genes in organisms, complex diseases, and cancers [26][27][28][29] .
SWIM was able to identify important switch genes involved in the transition from a stem-like to a differentiated phenotype of glioblastoma cells. Among the switch genes common to both datasets, the authors found FOSL1, which was down-regulated in stem-like cells and highly negatively correlated with the four TFs. Moreover, the authors found a predicted FOSL1 consensus binding motif in the regulatory regions of all four TFs that showed a positive correlation with proteins crucial for cell matrix adhesion and cell motility, including actin, collagen, fibronectin, and several integrins 26 . Based on these findings, the authors identified FOSL1 as a promising candidate to orchestrate the differentiation of cancer stem-like cells by repressing expression of the four TFs and restoring the physiological equilibrium between cell adhesion and migration, thus interfering with cancer progression. FOSL1 is a well-known transcription factor. Earlier studies on FOSL1 demonstrated its involvement in embryonic development and bone formation. FOSL1 is abnormally expressed in many tumors and plays an important role in tumorigenesis and progression 30,31 . It is mainly regulated by the mitogen-activated protein kinase (MAPK) signaling pathway 30 , the most heavily involved pathway in cancer progression. Interestingly, the abnormal expression of FOSL1 in various tumors and its effects on tumor progression differ according to tumor type 30 . We then used ChIP in HEK-293T transfected cells to demonstrate that FOSL1 directly regulates OLIG2, SALL2, SOX2 and POU3F2 expression levels through the binding of their promoter regions. In order to better investigate the role of FOSL1 in the stemness processes, we analyzed the levels of 15 proteins with a commercial proteome array in FOSL1 overexpressed cells. Seven of the 15 well-known stemness markers analyzed resulted to be deregulated after the overexpression of FOSL1. All the deregulated markers are involved in the epithelialmesenchymal transition (EMT) in cancer or confer a more aggressive phenotype to the cells 32,33 . Among them, we found a reduction of SOX17, an EMT-suppressor able to transcriptionally repress FN1, a gene that codifies for a protein involved in ECM remodeling during EMT 33,34 . Interestingly, in our previous computational study, we found that FOSL1 positively correlated with genes encoding proteins crucial for cell-matrix adhesion, including FN1 and a-actin. We confirmed that the overexpression of FOSL1 induced a modulation in the cell-matrix adhesion, as demonstrated by the increase in the cell aggregates observed in a semisolid condition and by the induction of FN1 and a-actin expression.

Conclusions
Overall, our experimental results validate SWIM analysis predictions 21 , confirming the potential role of FOSL1 as a crucial factor in modulating the expression levels of the four TFs, and consequently cancer stem-like conditions. This may suggest that novel therapeutic strategies able to restore FOSL1 expression may be effective for glioblastoma treatment.   37 . SYBR Green Master Mix was used to perform quantitative analyses with the specific primers reported in Table 1 using a standard protocol according to the manufacturer's instructions (Thermo Fisher Scientific, Waltham, Massachusetts, USA). GAPDH was used as endogenous control. Final results were calculated using the 2 − ΔΔCt method and normalized to the calibrator sample. www.nature.com/scientificreports/ Chromatin immunoprecipitation (ChIP). ChIP was performed on cells transfected as described above.

RNA isolation and gene expression analysis.
The cross-link between protein and DNA was performed using formaldehyde 0.75% for 10 min at room temperature. Glycine (135 mM) was added to medium for 5 min at room temperature to quench the formaldehyde and terminate the cross-linking reaction. After 3 washes in phosphate-buffered saline, cells were harvested by the scraper of the plate. After centrifuge, cells were resuspended in ChIP lysis buffer (50 mM HEPES-KOH pH7.5, 140 mM NaCl, 1 mM EDTA pH8, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, protease inhibitors). Isolated chromatin was sonicated to obtain fragments ranging from 200 to 1000 bp. FOSL1 and cross-linked DNA were immunoprecipitated using the antibody FRA1 (D80B4) produced in rabbit (mAb #5281, Cell Signaling technologies). The antibody was conjugated to Dynabeads (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and immunoprecipitation was performed according to the manufacturer's instructions.
The de-crosslinking reaction was performed at 70 °C for 10 min and DNA was isolated with QIAmp DNA Mini Kit (Qiagen, Hilden, Germany). Targets were analyzed by PCR performed as described in 38 , and using primers as reported in Table 1.
Pluripotent stem cell array. Total proteins were isolated from HEK293T transfected with empty vector or pCMV-FOSL1, using a lysis buffer containing TrisHCl (pH 7.4, 50 mM), NaCl (150 mM), Triton (1% v/v), ethylenediaminetetraacetic acid (EDTA, 20 mM), phenylmethylsulfonyl fluoride (PMSF, 2 mM), protease and phosphatase inhibitors (Pierce, Rockford, IL, USA), leupeptin (2 ug/ml), glycerol (10% v/v), and water. Total proteins were quantified using Bradford reagent and NanoDrop 2000/2000c Spectrophotometers (Thermo Fisher Scientific). 100 μg of total proteins were then assayed with the Human Pluripotent Stem Cell Antibody Array (R&D Systems, Minneapolis, MN, USA) according to the manufacturer instructions. The acquisition of the chemiluminiscent signal of the membrane of the array was performed using ChemiDoc MP Imager (Bio-Rad). The analysis was performed using Image Lab Software (Bio-Rad) using the volume tool analysis methods.
Cell aggregation assay. HEK-293T cells were counted, transfected in a suspension of OptiMEM medium (50 μl) containing lipofectamine 3000 and P3000 (0.5 μl each) and plasmids (10 ng) and plated in 96 well plates at 10,000 cells per well of confluence. Plates were coated with a basal layer of 100% (50 μl) and a layer of 2% Matrigel Growth Factor Reduced (GFR) Basement Membrane Matrix without supplement. Cells were photographed every 2 days starting from 24 h after the transfection.
Western blot. 30 μg of total proteins, isolated from HEK293T transfected with empty vector or pCMV-FOSL1 as described before, were separated using SDS-PAGE and transferred onto polyvinyldene difluoride (PVDF) membrane. After 2 h of blocking using not-fat dry milk at 5%, proteins were detected with specific primary antibodies at dilution of 1:1000: FRA1 (D80B4) (Cell Signaling technologies), FN1 (Sigma-Aldrich), a-ACTIN (Santa-Cruz), GAPDH (Cell Signaling technologies). After the incubation of secondary antibodies, the bands were detected with chemiluminescent using Clarity Western ECL substrate (BIO-RAD) and a chargecoupled-device camera (Chemidoc, BIO-RAD).

Statistical analysis.
Results are reported as means ± standard deviation. Differences were assessed with an unpaired t-test, and p-values lower than 0.05 were considered statistically significant. All statistical analyses were performed using GraphPad Prism version 5.0 statistical software (GraphPad Software Inc., San Diego, CA, USA).

Data availability
This study did not generate new unique reagents. The samples used in this study are available under request.