Demethylation of Epiregulin Gene by Histone Demethylase FBXL11 and BCL6 Corepressor Inhibits Osteo/dentinogenic Differentiation§

Authors

  • Juan Du,

    1. Laboratory of Molecular Signaling and Stem Cells Therapy, Molecular Laboratory for Gene Therapy and Tooth Regeneration, Beijing Key Laboratory of Tooth Regeneration and Function Reconstruction, Capital Medical University School of Stomatology, Beijing, People's Republic of China
    Search for more papers by this author
  • Yushi Ma,

    1. Laboratory of Molecular Signaling and Stem Cells Therapy, Molecular Laboratory for Gene Therapy and Tooth Regeneration, Beijing Key Laboratory of Tooth Regeneration and Function Reconstruction, Capital Medical University School of Stomatology, Beijing, People's Republic of China
    Search for more papers by this author
  • Ping Ma,

    1. Laboratory of Molecular Signaling and Stem Cells Therapy, Molecular Laboratory for Gene Therapy and Tooth Regeneration, Beijing Key Laboratory of Tooth Regeneration and Function Reconstruction, Capital Medical University School of Stomatology, Beijing, People's Republic of China
    Search for more papers by this author
  • Songlin Wang,

    1. Laboratory of Molecular Signaling and Stem Cells Therapy, Molecular Laboratory for Gene Therapy and Tooth Regeneration, Beijing Key Laboratory of Tooth Regeneration and Function Reconstruction, Capital Medical University School of Stomatology, Beijing, People's Republic of China
    2. Department of Biochemistry and Molecular Biology, Capital Medical University School of Basic Medical Sciences, Beijing, People's Republic of China
    Search for more papers by this author
    • Telephone: +86-10-6706-2012; Fax: +86-10-6706-2012

  • Zhipeng Fan

    Corresponding author
    1. Laboratory of Molecular Signaling and Stem Cells Therapy, Molecular Laboratory for Gene Therapy and Tooth Regeneration, Beijing Key Laboratory of Tooth Regeneration and Function Reconstruction, Capital Medical University School of Stomatology, Beijing, People's Republic of China
    • Laboratory of Molecular Signaling and Stem Cells Therapy, Molecular Laboratory for Gene Therapy and Tooth Regeneration, Beijing Key Laboratory of Tooth Regeneration and Function Reconstruction, Capital Medical University School of Stomatology, Beijing, People's Republic of China

    Search for more papers by this author
    • Telephone: +86-10-6706-2012; Fax: +86-10-6706-2012


  • Author contributions: J.D.: data analysis and interpretation, financial support, and manuscript writing, Y.M. and P.M.: collection and/or assembly of data; S.W.: conception and design, financial support, and manuscript writing; Z.F.: conception and design, data analysis and interpretation, financial support, and final approval of manuscript.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSEXPRESS October 16, 2012.

Abstract

Mesenchymal stem cells (MSCs) are a reliable resource for tissue regeneration, but the molecular mechanism underlying directed differentiation remains unclear; this has restricted potential MSC applications. Histone methylation, controlled by histone methyltransferases and demethylases, may play a key role in MSC differentiation. Here, we investigated FBXL11, a histone demethylase, lysine (K)-specific demethylase 2A, which is evolutionarily conserved, ubiquitously expressed, and a member of the JmjC-domain-containing histone demethylase family. We tested whether FBXL11 could inhibit the osteo/dentinogenic differentiation potential in MSC cells with gain- and loss-of-function assays. We found that FBXL11 regulated osteo/dentinogenic differentiation in MSC cells. Furthermore, we found that the gene encoding the epidermal growth factor, Epiregulin (EREG), was a downstream target of FBXL11, and that EREG mediated FBXL11 regulation of MSC differentiation. Moreover, we found that the FBXL11 histone demethylase function was activated by associating with BCL6 corepressor, and this complex could repress EREG transcription by increasing histone K4/36 methylation in the EREG promoter. In conclusion, our results elucidated a new function for FBXL11 and EREG, explored the molecular mechanism underlying directed differentiation in MSC cells, and identified potential target genes for improving tissue regeneration techniques. STEM Cells2013;31:126–136

INTRODUCTION

Mesenchymal stem cells (MSCs) are multipotent cells, originally isolated from bone marrow. They can differentiate into a variety of cell types, including osteoblasts, chondrocytes, myocytes, and adipocytes. Increasing evidence indicates that MSCs are also present in non-bone marrow tissues [1–5]. Recently, a new population of MSCs has been isolated from dental and craniofacial tissues on the basis of their stem cell properties. These include stem cells from the periodontal ligament, from dental pulp, from apical papilla (SCAP), and so on [6–11]. These cells are multipotent, exhibit osteo/dentinogenic differentiation, and can self-renew. When transplanted into mice or miniature pigs, they generate bone/dentin-like mineralized tissue and are capable of repairing tooth defects [6, 12–14]. MSCs are a reliable resource for tissue regeneration, but the molecular mechanism underlying directed differentiation remains unclear; this restricts their potential applications.

Covalent histone modifications play an important role in regulating chromatin dynamics and functions [15]. One type of histone modification, methylation, occurs on both lysine and arginine residues. This modification is involved in a diverse range of biological processes, including heterochromatin formation, X-chromosome inactivation, and transcriptional regulation [16–18]. The steady-state level of covalent histone methylation is controlled by histone methyltransferases and demethylases. In a previous study, we investigated dental MSC cells derived from individuals with oculo-facio-cardio-dental (OFCD) syndrome or healthy controls. Those results demonstrated that the histone demethylase, FBXL10 (lysine (K)-specific demethylase 2B, KDM2B), lost the ability to inhibit osteo/dentinogenic differentiation in MSCs. The mechanism underlying this loss-of-function was the failure of FBXL10 to form a complex with a truncated BCL6 corepressor (BCOR) in MSC cells. This resulted in increased histone H3-Lys-4 (H3K4) trimethylation and H3-Lys-36 (H3K36) dimethylation on a target gene promoter and activation of the transcription factor, AP-2 alpha (AP-2α) [19]. However, the function of FBXL11, a paralog of FBXL10, remains unclear in MSCs.

FBXL11 (lysine (K)-specific demethylase 2A) is an uncharacterized protein, originally identified in bioinformatic searches for F-box-containing proteins [20, 21]. In addition to an F-box domain, FBXL11 also contains several interesting domains, including a JmjC domain, a CxxC zinc finger, a plant homeo-domain, and three leucine-rich repeats. The most important functional domain, the JmjC domain, has 79% homology in FBXL11 and FBXL10; JmjC domain-containing proteins are predicted to be metalloenzymes that regulate chromatin function [22–25]. Thus, presumably, the histone demethylation functions should be highly similar in FBXL11 and FBXL10.

Based on previous studies, we hypothesized that FBXL11 may have an effect on MSCs. In this study, we used MSC cells from dental root apical papilla (SCAP cells) and bone marrow (BMSCs) to investigate the functions of FBXL11 and the underlying mechanisms. Our results indicated that FBXL11 could interact with BCOR to inhibit the osteo/dentinogenic differentiation potential in MSCs by silencing the downstream target gene that encodes the epidermal growth factor (EGF), Epiregulin (EREG). This information is expected to enhance the directed differentiation of MSCs for tissue regeneration applications.

MATERIALS AND METHODS

Cell Cultures

Tooth tissues were obtained under approved guidelines set by the Beijing Stomatological Hospital, Capital Medical University with informed patient consent. Wisdom teeth were first disinfected with 75% ethanol and then washed with phosphate buffered saline (PBS). SCAP cells were isolated, cultured, and identified, as previously described [6, 9–11, 19]. Briefly, SCAP cells were gently separated from the apical papilla of the root and then digested in a solution of 3 mg/ml collagenase type I (Worthington Biochemical Corp., Lakewood, NJ, http://www.worthington-biochem.com) and 4 mg/ml dispase (Roche Diagnostics Corp., Indianapolis, IN, http://www.roche-applied-science.com) for 1 hour at 37°C. SCAP samples from different individuals were pooled, and single-cell suspensions were obtained by passing the cells through a 70 μm strainer (BD Biosciences, San Jose, CA, http://www.bdbiosciences.com). BMSCs from human bone marrow were obtained from ScienCell Research Laboratories (Carlsbad, CA, http://www.sciencellonline.com). MSC cells were grown in a humidified, 5% CO2 incubator at 37°C in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), supplemented with 15% fetal bovine serum (FBS; Invitrogen), 2 mmol/l glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen). The culture medium was changed every 3 days. Human embryonic kidney 293T cells were maintained in complete DMEM medium with 10% FBS (Invitrogen), 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen).

Plasmid Construction and Viral Infection

The plasmids were constructed with standard methods; all structures were verified by appropriate restriction digestion and/or sequencing. Human full-length FBXL11 cDNA from SCAP cells fused to a hemagglutinin (HA) tag was produced with a standard polymerase chain reaction (PCR) protocol. This sequence (HA-FBXL11) was subcloned into the pQCXIH retroviral vector with AgeI and BglII restriction sites. Similarly, the human full-length BCOR cDNA was fused with a FLAG tag (FLAG-BCOR) and subcloned into the pQCXIN retroviral vector with AgeI and BamH1 restriction sites. Short hairpin RNAs (shRNA) with the complementary sequences of the target genes were subcloned into the pLKO.1 lentiviral vector (Addgene, Cambridge, MA, http://www.addgene.org) or the pSIREN retroviral vector (Clontech Labtoratories, Mountain View, CA, http://www.clontech.com). Viral packaging was prepared as described previously [11]. For viral infections, MSCs were plated overnight, then infected with retroviruses or lentiviruses in the presence of polybrene (6 μg/ml, Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) for 6 hours. After 48 hours, infected cells were selected with different antibiotics. A scrambled shRNA (Scramsh) was purchased from Addgene. The target sequences for the shRNAs were shown in Table 1.

Table 1. The target sequences for the shRNAs
inline image

Coimmunoprecipitation Assays

The 293T cells were transfected with FuGENE® 6 (Promega, Madison, WI, http://www.promega.com) in 10 cm culture dishes with 3 μg HA-FBXL11, 3 μg FLAG-BCOR, and the same amounts of empty vectors in control dishes. SCAP cells were infected with retroviruses that expressed wild-type FLAG-BCOR or empty vector. The cells were washed twice with PBS and lysed in 500 μl of cold immunoprecipitation (IP) buffer (PBS, 10% glycerol, 0.5% Nonidet P40, and a complete protease inhibitor cocktail from Roche Diagnostics Corp.) for 10 minutes on ice. Lysates were briefly sonicated and centrifuged for 10 minutes at 4°C. The supernatants were collected and immunoprecipitated with 2 μg monoclonal anti-HA (Clone No. C29F4, Cat No. 3724, Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com), 2 μg monoclonal anti-FLAG M2 (Clone No. 9A3, Cat No. 8146, Cell Signaling Technology), or 2 μg polyclonal anti-FBXL11 (Cat No.ab31739, Abcam, Cambridge, U.K., http://www.abcam.com) antibodies. After 1 hour incubation at 4°C with gentle rotation, 40 μl of 50% protein A/G Plus-Agarose slurry (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com) was added, and the mixtures were incubated overnight at 4°C with gentle rotation. The immune complexes were collected by centrifugation and washed three times with 500 μl cold IP buffer. The immunoprecipitated proteins were detected by Western Blot with anti-HA or anti-FLAG M2 antibodies.

Western Blot Analysis

Cells were lysed in RIPA buffer (10 mM Tris-HCl, 1 mM EDTA, 1% sodium dodecyl sulfate [SDS], 1% Nonidet P40, 1:100 proteinase inhibitor cocktail, 50 mM β-glycerophosphate, 50 mM sodium fluoride). The samples were separated on a 10% SDS polyacrylamide gel and transferred to polyvinylidene difluoride membranes with a semi-dry transfer apparatus (Bio-Rad, Hercules, CA, http://www.bio-rad.com). The membranes were blotted with 5% dehydrated milk for 2 hours and then incubated with primary antibodies overnight. The immune complexes were incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG (Promega) and visualized with SuperSignal reagents (Pierce, Rockford, IL, http://www.piercenet.com). Primary antibodies against FBXL11, HA, and FLAG-M2 were described in the coimmunoprecipitation assay above. We also used a primary monoclonal antibody to detect the housekeeping protein, glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Clone No. GAPDH-71.1, Cat No. G8795, Sigma-Aldrich).

Alkaline Phosphatase and Alizarin Red Detection

MSCs were grown in mineralization-inducing media, which contained 100 μM/ml ascorbic acid, 2 mM β-glycerophosphate, and 10 nM dexamethasone. For alkaline phosphatase (ALP) staining, after induction, cells were fixed with 4% paraformaldehyde and stained with a solution of 0.25% naphthol AS-BI phosphate and 0.75% Fast red FRV in an ALP kit, according to the manufacturer's protocol (Sigma-Aldrich). ALP activity was assayed with an ALP activity kit according to the manufacturer's protocol (Sigma-Aldrich). Signals were normalized based on protein concentrations. For detecting mineralization, cells were induced for 2–3 weeks, fixed with 70% ethanol, and stained with 2% Alizarin red (Sigma-Aldrich). To quantitatively determine calcium, Alizarin Red was destained with 10% cetylpyridinium chloride in 10 mM sodium phosphate for 30 minutes at room temperature. The concentration was determined by measuring the absorbance at 562 nm on a multiplate reader and comparing to a standard calcium curve with calcium dilutions in the same solution. The final calcium level in each group was normalized to the total protein concentration detected in a duplicate plate [19].

Reverse Transcriptase-PCR (RT-PCR) and Real-time RT-PCR

Total RNA was isolated from MSCs with Trizol reagents (Invitrogen). We synthesized cDNA from 2 μg aliquots of RNA, random hexamers or oligo(dT), and reverse transcriptase, according to the manufacturer's protocol (Invitrogen). Real-time PCR reactions were performed with the QuantiTect SYBR Green PCR kit (Qiagen, Hilden, Germany, http://www.qiagen.com) and an Icycler iQ Multi-color Real-time PCR Detection System (Bio-Rad). The primers for specific genes were shown in Supporting Information Table 1.

Chromatin immunoprecipitation (ChIP) Assays

We used a ChIP assay kit (Merck Millipore, Billerica, MA, http://www.millipore.com) according to the manufacturer's protocol. Briefly, cells were incubated with a dimethyl 3,3′ dithiobispropionimidate-HCl (Pierce) solution (5 mmol) for 10 minutes at room temperature and then incubated with 1% formaldehyde for 10 minutes at 37°C. For each ChIP reaction, 2 × 106 cells were used. For DNA precipitation, 2 μg polyclonal antibodies against trimethylated H3K4 (H3K4me3; Cat No. 07-473, Merck Millipore) and dimethylated H3K36 (H3K36me2; Cat No. 07-369, Merck Millipore) were added. All resulting precipitated DNA samples were quantified with real-time PCR. Data are expressed as the percentage of input DNA. The real-time PCR primers targeted the FBXL11 binding region of the EREG promoter. These sequences were: forward, 5′-CGGGAGAGTTTCATAATTGGCCCG-3′ and reverse, 5′-CCAGGGAGGGAG GATGACTTC-3′.

Transplantation in Nude Mice

Approximately, 4.0 × 106 cells were mixed with 40 mg of hydroxyapatite/tricalcium phosphate (HAP) ceramic particles (National Engineering Research Center for Biomaterials, Sichuan University, Chengdu, People's Republic of China, http://www.aiscu.com), then transplanted subcutaneously into the dorsal surface of 10-week-old immunocompromised beige mice (nu/nu nude mice), as previously described [6, 12]. These procedures were performed in accordance with specifications of an approved animal protocol. Eight weeks after transplantation, the transplanted cells were harvested, fixed with 10% formalin, decalcified with buffered 10% EDTA (pH 8.0), and then embedded in paraffin. Sections were deparaffinized, hydrated, and stained with hematoxylin and eosin (H&E).

SCAP Nuclear Extracts

The nuclear extracts of SCAP cells were prepared with a Nuclear-Cytosol Extraction Kit (Applygen Technologies Inc., Beijing, People's Republic of China, http://www.applygen.com.cn), according to the manufacturer's instructions. Briefly, 5 × 106 SCAP cells were harvested by centrifugation. The pelleted cells were resuspended in 250 μl of cytosol extraction buffer A, incubated on ice for 10 minutes, then mixed with 15 μl cytosol extraction buffer B, and incubated on ice for 1 minute. The lysates were centrifuged, and pellets were washed with cytosol extraction buffer A, then resuspended in 50 μl of cold nuclear extraction buffer. After incubation at 4°C for 30 minutes with constant rotation, the suspension was spun at 12,000g at 4°C for 15 minutes; the nuclear extract was collected from the supernatant fraction.

Nuclear factor kappa B (NFκB) (p65) Transcription Factor DNA Binding Activity Assay

We used the NFκB (p65) Transcription Factor Assay Kit (Cayman Chemical Company, Ann Arbor, MI, https://www.caymanchem.com), according to the manufacturer's instructions. Briefly, 10 μg nuclear extracts were added into the appropriate wells and incubated overnight at 4°C without agitation. After washing five times with 1× wash buffer, 100 μl diluted NFκB (p65) antibody was added to the wells and incubated for 1 hour at room temperature without agitation. After washing five times with 1× wash buffer, 100 μl diluted goat anti-rabbit secondary antibody was added to the wells and incubated for 1 hour at room temperature without agitation. After washing five times with 1× wash buffer, 100 μl developing solution was added to the wells and incubated for 30 minutes at room temperature with gentle agitation. The reaction was stopped by adding 100 μl stop solution per well. Absorbance was measured at 450 nm.

Statistics

All statistical calculations were performed with SPSS10 statistical software. The Student's t test or one-way ANOVA were performed to determine statistical significance. A p-value ≤.05 was considered significant.

RESULTS

FBXL11 Inhibited Osteo/dentinogenic Differentiation Potential in MSC Cells

First, we tested FBXL11 expression upon osteogenic differentiation. Interestingly, when SCAP cells were induced to differentiate into osteo/dentinogenic cells, FBXL11 expression was downregulated in a time-dependent manner (Fig. 1A). A similar effect was observed in BMSCs (Supporting Information Fig. 1I). To elucidate the function of FBXL11 in MSC cells, we designed a shRNA to target FBXL11 and introduced it into SCAP cells with lentiviral infection (SCAP-FBXL11sh cells). After selection, the knockdown efficiency (60%) was verified by real-time reverse transcriptase PCR (RT-PCR) and Western Blot (Fig. 1B). Next, we examined whether FBXL11 intrinsically affected the function of MSC cells. SCAP cells were cultured in osteogenic-inducing medium, and we found that ALP activity, an early marker for osteo/dentinogenic differentiation, was induced to a greater extent in SCAP-FBXL11sh cells than in SCAP cells infected with Scramsh (SCAP-Scramsh) and uninfected SCAP cells (SCAP-Blank) (Fig. 1C). Three weeks after induction, Alizarin Red staining and quantitative calcium measurements revealed that mineralization was also significantly higher in SCAP-FBXL11sh cells than in SCAP-Scramsh and SCAP-Blank cells (Fig. 1D, 1E). Consistent with that, real-time RT-PCR results showed that the osteo/dentinogenic marker gene, bone sialoprotein (BSP), which encodes extracellular matrix proteins of bone and dentin, was induced to a greater extent in SCAP-FBXL11sh cells than SCAP-Scramsh and SCAP-Blank cells at 3 and 7 days after induction (Fig. 1F). Another osteo/dentinogenic marker gene, osteocalcin (OCN), was strongly induced at 3, 7, 10, and 21 days after induction (Fig. 1G).

Figure 1.

Silencing FBXL11 enhanced osteo/dentinogenic potential of stem cells from apical papilla (SCAP). (A):FBXL11 decreased from 1 to 3 weeks as SCAPs underwent osteo/dentinogenic differentiation. (B): SCAPs were silenced by FBXL11sh; controls: Scramsh; uninfected SCAPs (Blank). (C): Alkaline phosphatase activity. (D, E): Mineralization (Alizarin red stain, quantification, respectively). (F–I): Real-time reverse transcriptase polymerase chain reaction (RT-PCR) shows expression of (F) BSP, (G) OCN, (H) OSX, and (I) DLX2. (J): H&E stained micrographs showed mineralization after FBXL11-shRNA infected cells were transplanted into nude mice for 8 weeks. Bar = 100 μm. (K): Percent mineralized tissue. Error bars: SD. *, p ≤ .05; **, p ≤ .01 (A–I: one-way ANOVA, n = 3; K: Student's t test, n = 5). For RT-PCR, 18S RNA and GAPDH were internal controls. Abbreviations: BSP, bone sialoprotein; CT, connective tissues; D, bone/dentin-like tissue; DLX2, distal-less homeobox factor 2; HAP, hydroxyapatite tricalcium carrier; OCN, osteocalcin; OSX, osterix.

Next, we examined the expression of key transcription factors for regulating osteo/dentinogenic differentiation, including runt related transcription factor 2 (RUNX2), osterix (OSX), distal-less homeobox factors 2, 3, and 5 (DLX2, DLX3, and DLX5, respectively), FOS-like antigen 1 (Fra1), and muscle segment homeobox 2 (MSX2). We found that the mRNA levels of OSX and DLX2 were significantly higher in SCAP-FBLX11sh cells than in SCAP-Scramsh and SCAP-Blank cells (Fig. 1H, 1I), but the expression levels of the other genes were not significantly different (data not shown). In vivo transplantation experiments also demonstrated that SCAP-FBLX11sh cells generated more bone/dentin-like mineralized tissue than SCAP-Scramsh cells (Fig. 1J, 1K).

To confirm the function of FBXL11 in MSCs, we inserted the HA-FBXL11 sequence into a retroviral vector. This construct overexpressed ectopic FBLX11 when transduced into SCAP cells via retroviral infection. Ectopic FBXL11 expression was confirmed by Western Blot and real-time RT-PCR analysis (Fig. 2A). Next, transduced SCAP cells were cultured in osteogenic-inducing medium to investigate the osteo/dentinogenic differentiation potential. The results indicated that the overexpression of wild-type FBLX11 strongly inhibited ALP activity in SCAP cells (Fig. 2B). Accordingly, mineralization was markedly inhibited in SCAP cells that overexpressed HA-FBLX11 compared with cells infected with the empty vector, determined by Alizarin Red staining and quantitative calcium measurements (Fig. 2C, 2D). Real-time RT-PCR results also showed that expression of induced BSP and OCN were inhibited by overexpression of HA-FBXL11 (Fig. 2E, 2F), and the mRNA levels of OSX and DLX2 were also significantly decreased in SCAP-HA-FBLX11 cells compared to SCAP-Vector cells (Fig. 2G, 2H). Taken together, these results showed that FBXL11 expression substantially repressed the osteo/dentinogenic differentiation potential of SCAP cells (Fig. 1, 2).

Figure 2.

FBXL11 overexpression inhibited SCAPs osteo/dentinogenic differentiation. (A): Real-time reverse transcriptase polymerase chain reaction (RT-PCR) and Western blot results showed that stem cells from apical papilla overexpressed FBXL11 with HA-FBXL11 infecting. (B): Alkaline phosphatase activity. (C, D): Alizarin red stain showed mineralization. (EH): Real-time RT-PCR results showed the expression of (E) BSP, (F) OCN, (G) OSX, and (H) DLX2. Error bars: SD, (n = 3). *, p ≤ .05; **, p ≤ .01. (Student's t test). Abbreviations: BSP, bone sialoprotein; DLX2, distal-less homeobox factor 2; OCN, osteocalcin; OSX, osterix.

To determine whether the FBXL11 had similar functions in other MSC cells, we knocked-down FBXL11 in BMSC cells with lentivirus-mediated FBXL11shRNA. We found that the knockdown of FBXL11 significantly enhanced osteo/dentinogenic differentiation in BMSC cells, based on in vitro ALP activity and mineralization (Supporting Information Fig. 1A–1E). Furthermore, we found that the mRNA levels of DLX2 and OSX were significantly higher in BMSC-FBLX11sh cells than in BMSC-Scramsh cells (Supporting Information Fig. 1G, 1H), but the expression of RUNX2, DLX3, DLX5, Fra1, and MSX2 were not different (data not shown).

FBXL11 Inhibited Osteo/dentinogenic Differentiation in MSC Cells Through EREG

Next, we investigated how the FBXL11 downregulated osteo/dentinogenic differentiation in MSC cells. We hypothesized that FBXL11 may cooperate with BCOR, based on a previous study [19]. To test this hypothesis, we selected several candidate target genes of BCOR, including EREG, FBXL2, FBXL7, insulin-like growth factor binding protein 2 (IGFBP2), paired box 3 (PAX3), and S100 calcium binding protein A4 (S100A4) from microarray data [19]. Then, we silenced BCOR with a lentivirus-mediated BCOR shRNA. The real-time RT-PCR results showed increases in the expression of EREG, PAX3, and IGFBP2 after BCOR depletion in SCAP cells (Fig. 3A), but the expression of FBXL11 (Fig. 3A) and other genes were unchanged (data not shown). Taken together with previous studies [19], the results showed that EREG, PAX3, IGFBP2, and AP-2a were downstream target genes of BCOR. To determine whether FBXL11 was involved in regulating the target genes of BCOR, we knocked-down FBXL11 in SCAP cells and checked the expression of BCOR target genes with real-time RT-PCR. Interestingly, we found that the expression of EREG was increased but not the expression of AP-2a, IGFBP2, PAX3, FBXL10, or BCOR (Fig. 3B). Conversely, when FBXL11 was overexpressed, EREG was downregulated (Fig. 3C). Furthermore, when FBXL11 was downregulated by inducing osteo/dentinogenic differentiation in SCAP cells (Fig. 1A), BCOR expression was not affected (Fig. 3D), but EREG expression was upregulated (Fig. 3E). A similar result was also found in BMSC cells (Supporting Information Fig. 1F, 1J). These results indicated that EREG was a downstream target gene of FBXL11.

Figure 3.

Depletion of FBXL11 or BCOR increased EREG expression. Stem cells from apical papilla (SCAP) were infected with short hairpin RNAs (shRNA) to silence the indicated gene or with control scrambled shRNA (Scramsh). Real-time reverse transcriptase polymerase chain reaction (RT-PCR) for cells infected with (A)BCOR shRNA or (B)FBXL11 shRNA showed the expression of BCOR, EREG, PAX3, IGFBP2, FBXL11, AP-2a, and FBXL10. (C): Real-time RT-PCR showed EREG expression in SCAP cells that overexpressed HA-FBXL11. (D): Real-time RT-PCR in differentiated SCAP cells showed (D) BCOR and (E) EREG expression. 18S RNA and GAPDH were internal controls; (A, B, C) Student's t-test; (D, E) one-way ANOVA; all error bars represent SD (n = 3); *, p ≤ .05; **, p ≤ .01. Abbreviations: BCOR, BCL6 corepressor; EREG, Epiregulin; IGFBP2, insulin-like growth factor binding protein 2; PAX3, paired box 3; Scramsh, scrambled shRNA.

Next, we investigated whether EREG was involved in the FBXL11-mediated osteo/dentinogenic differentiation of MSC cells. We infected SCAP cells with retroviruses that carried two different EREG shRNAs (EREGsh1 and EREGsh2) to silence EREG and a Luciferase shRNA (Lucsh) as a control. Real-time RT-PCR results showed that EREG expression was silenced by 40% or 80% with EREGsh1 or EREGsh2, respectively, compared to EREG expression in Lucsh controls (Fig. 4A). When SCAP cells were cultured in osteogenic-inducing medium, ALP activity was reduced in EREG-depleted cells, consistent with the EREG expression level (Fig. 4B, 4C). Three weeks after induction, Alizarin Red staining and quantitative calcium also revealed that the knockdown of EREG reduced mineralization in SCAP cells; moreover, mineralization was diminished significantly more in SCAP-EREGsh2 cells than in SCAP-EREGsh1 cells (Fig. 4D, 4E). Additionally, EREG depletion caused reductions in BSP, OCN, OSX, and DLX2 mRNA expression (Fig. 4F–4I). In vivo transplantation experiments also demonstrated that SCAP cells generated less bone/dentin-like mineralized tissue after EREG depletion compared to SCAP cells infected with Lucsh (Fig. 4J, 4K).

Figure 4.

Differentiation enhanced by FBXL11 depletion was mediated by EREG. Stem cells from apical papilla (SCAP) were infected with short hairpin RNAs (shRNA) that silenced EREG (EREGsh1 and EREGsh2) or control luciferase shRNA (Lucsh). SCAP cells were analyzed after osteogenic differentiation. (A): Real-time reverse transcriptase polymerase chain reaction (RT-PCR) results showed EREG expression. (B, C): Alkaline phosphatase activity and stain. (D, E): Mineralization (Alizarin red stain, quantification, respectively). (FI): Real-time RT-PCR showed expression of (F) BSP, (G) OCN, (H) OSX, and (I) DLX2. (J): H&E stained micrographs showed mineralization after EREGsh2 infected cells were transplanted subcutaneously into the dorsal surface of 10-week-old immunocompromised beige mice for 8 weeks. Bar = 100 μm. (K): Percent mineralized tissue. All error bars represent SD. *, p ≤ .05; **, p ≤ .01 (A–I: one-way ANOVA, n = 3; K: Student's t test, n = 5). For RT-PCR, 18S RNA and GAPDH were internal controls. Abbreviations: BSP, bone sialoprotein; CT, connective tissues; D, bone/dentin-like tissue; DLX2, distal-less homeobox factor 2; EREG, Epiregulin; HAP, hydroxyapatite tricalcium carrier; Lucsh, luciferase shRNA; OCN, osteocalcin; OSX, osterix.

Next, we investigated the functions of EREG in BMSC cells. We constructed lentiviral EREGsh2, because the BMSC cells were not amenable to retrovirus infection. The EREG knockdown with lentiviral EREGsh2 in BMSC cells showed changes in ALP activity and mineralization similar to those observed in SCAP cells (Supporting Information Fig. 2A–2D). Moreover, DLX2 and OSX, but not RUNX2, expression was significantly increased in BMSCs infected with FBLX11sh compared to BMSCs infected with Scramsh (Supporting Information Fig. 2E, 2F). These results showed that, in both SCAP and BMSC cells, EREG was a key target gene of FBXL11, and EREG mediated FBXL11 regulation of osteo/dentinogenic differentiation in MSC cells.

FBXL11 Associated with BCOR and the Complex Repressed EREG Transcription by Increasing Histone K4/36 Methylation in EREG Promoter

Next, we investigated how FBXL11 regulated the expression of EREG. Because transient transfection efficiency was very low in MSC cells, we first transiently transfected HA-FBXL11 and FLAG-BCOR into 293T cells, which were readily transfected. Coimmunoprecipitation showed that the overexpression of wild-type FBXL11 and BCOR induced the formation of FBXL11-BCOR protein complexes in 293T cells (Fig. 5A). After confirming that FBXL11-BCOR protein complex was formed in 293T cells, we attempted to detect the endogenous FBXL11-BCOR protein complex in MSC cells by coimmunoprecipitation. However, the coimmunoprecipitation failed to detect the endogenous FBXL11-BCOR complex with the commercial antibodies. We suspected that the FBXL11-BCOR protein complex formed in MSC cells but at a rate too low to be detected. Thus, we attempted to increase the formation rate of FBXL11-BCOR complex by overexpressing FLAG-BCOR in SCAP cells with a retrovirus construct. This resulted in successful detection of the FBXL11-BCOR protein complex with the anti-FLAG and anti-FBXL11 antibodies (Fig. 5B). This result suggested that FBXL11 could associate with BCOR in MSC cells. We then investigated whether the FBXL11-BCOR complex could demethylate histones K36/K4 on the EREG promoter by a ChIP assay. However, despite repeated efforts, the available commercial anti-FBXL11 antibodies were ineffective in our ChIP assays. To overcome this problem, we constructed retroviruses to express HA-FBXL11 or the empty HA-vector in SCAP cells (Fig. 2A), and we performed ChIP assays with anti-HA antibody. The results indicated that, compared to the HA-vector, significantly higher levels of HA-FBXL11 were associated with the EREG promoter (Fig. 5C, 5D). Moreover, the knockdown of FBXL11 increased the H3K36me2 and H3K4me3 levels in the EREG promoter (Fig. 5E, 5F). The depletion of BCOR produced similar results, with increased K36/K4 methylation in the EREG promoter (Fig. 5G, 5H).

Figure 5.

FBXL11-BCOR complex formation increased histone K4/36 methylation in EREG promoter. FLAG and HA tags attached to wild-type BCOR and FBXL11, respectively, were ectopically expressed in 293T cells for coimmunoprecipitation. (A): Western blots showed individual signals (Input; 1% of lysate) and coimmunoprecipitated protein complexes (IP; 99% of lysate). (B): Western blot showed BCOR-FBXL11 complexes in SCAP cells that overexpressed BCOR. (C): Location of ChIP assay primers that targeted the EREG promoter at +145 to +77 bp upstream of the TSS. (D): HA-FBXL11 overexpression enhanced recruitment of FBXL11 to the EREG promoter. (EH): ChIP assays showed changes in histone methylation in the EREG promoter caused by infecting with short hairpin RNA (shRNA) to silence either (E, F) FBXL11 or (G, H) BCOR expression. All error bars represent SD (n = 3). **, p ≤ .01. (D–H: Student's t test). Abbreviations: BCOR, BCL6 corepressor; ChIP, chromatin immunoprecipitation; EREG, Epiregulin; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; H3K36me2, Histone K36 dimethylation; H3K4me3, histone K4 trimethylation; TSS, transcription start site.

FBXL11 Negatively Regulated NFκB Signaling in MSC Cells

FBXL11 has the capacity to negatively regulate NFκB signaling [23]. To elucidate whether FBXL11 have the expected effect on NFκB in MSC cells, NFκB signaling was stimulated with interleukin-1β (IL-1β) in SCAP cells that either overexpressed or lacked FBXL11. A DNA binding activity assay showed that silencing FBXL11 enhanced NFκB (p65) DNA binding activity for 0–8 hours after IL-1β treatment (Fig. 6A). In addition, real-time RT-PCR results showed that knockdown of FBXL11 stimulated the expression of NFκB target genes, IL6 and IL8, after IL-1β treatment (Fig. 6B, 6C). In contrast, overexpression of FBXL11 inhibited the NFκB (p65) DNA binding activity for 15 minutes to 2 hours after IL-1β treatment (Fig. 6D). These results showed that FBXL11 inhibited NFκB signaling in MSC cells.

Figure 6.

FBXL11 inhibited NFκB (p65) DNA binding activity in stem cells from apical papilla (SCAP). SCAP cells were cultured until they were about 90% confluent, then they were treated with 10 ng/ml IL-1β for the indicated times. (A): NFκB DNA binding activity was enhanced when FBXL11 was silenced with a short hairpin RNA (FBXL11sh) but not with a scrambled control (scramsh). (B, C): Real-time reverse transcriptase polymerase chain reaction results shows effects on IL6 and IL8 expression. GAPDH was an internal control. (D): NFκB DNA binding activity was inhibited with overexpression of FBXL11 (HA-FBXL11) in SCAP cells. All error bars represent SD (n = 3). *, p ≤ .05; **, p ≤ .01 (Student's t test). Abbreviation: IL-1β, interleukin-1β; IL6, interleukin 6; IL8, interleukin 8; NFκB, nuclear factor kappa B.

DISCUSSION

Our results indicated that the histone demethylase, FBXL11, could associate with BCOR, and this complex directly repressed the transcription of the EREG target gene by reducing promoter methylation. The result was an inhibition of the osteo/dentinogenic differentiation potential in MSC cells.

First, we found that FBXL11 was downregulated upon osteogenic differentiation in a time-dependent manner in MSC cells. This indicated that FBXL11 may be a negative regulator of osteogenic differentiation. Moreover, the gain- and loss-of-function studies showed that FBXL11 inhibited osteo/dentinogenic differentiation in MSC cells. Furthermore, we found that FBXL11 inhibited the expression of key transcription factors, OSX and DLX2, involved in osteogenic differentiation, and BSP and OCN genes, which encode extracellular matrix proteins. To investigate how FBXL11 downregulated MSCs osteo/dentinogenic differentiation, we reasoned that both FBXL10 and FBXL11 were BCL6 cofactors, and previous studies showed that BCOR could form a protein complex with FBXL10 that regulated MSCs functions [19, 25–27]. Thus, we hypothesized that FBXL11 might also cooperate with BCOR to perform its function. To test this hypothesis, we first showed that FBXL11 and BCOR targeted the same gene, EREG. EREG is a member of the EGF family, and it could activate the extracellular kinase/mitogen-activated protein kinase (ERK/MAPK) and AKT signaling pathways to stimulate proliferation of normal cells and promote smooth muscle cell dedifferentiation [28–32]. Our previous study showed that EREG was highly expressed in SCAP cells from patients with OFCD syndrome that carried a BCOR mutation. Those SCAP cells exhibited enhanced osteogenic differentiation potential [19]. In this study, FBXL11 overexpression caused downregulation of EREG. This strongly suggested that the EREG was a downstream gene, and it was negatively regulated by FBXL11 and BCOR. In MSC cells, we also showed that EREG was upregulated upon osteogenic differentiation in concert with the downregulation of FBXL11. Furthermore, silencing EREG diminished the osteo/dentinogenic differentiation potential in MSC cells in vitro and in vivo, and inhibited the key transcription factors, OSX and DLX2. These results suggested that EREG was an enhancer of osteo/dentinogenic differentiation and a negative mediator of FBXL11 regulation in MSC cells; moreover, the results suggested that EREG regulated osteo/dentinogenic differentiation potential through OSX and DLX2.

Next, we determined how FBXL11 suppressed EREG expression. First, we excluded the possibility that FBXL11 might regulate BCOR or BCOR might regulate FBXL11. Furthermore, the results from the coimmunoprecipitation and ChIP assays verified that FBXL11 could associate with BCOR and this complex could recruit the EREG promoter. A previous study showed that FBXL11 was a cofactor of BCL6 [27]; that supported the notion that BCL6, BCOR, and FBXL11 might form a functional protein complex that could regulate the EREG promoter. However, we could not find the typical BCL6 binding site in the EREG promoter with a bioinformatics analysis. This raised the question of whether BCL6 was necessary for the FBXL11-BCOR complex to suppress EREG expression. Finally, we investigated whether FBXL11 could demethylate the histones, H3K36me2 and H3K4me3, on the EREG promoter, and thus, regulate EREG expression. We discovered that depletion of either FBXL11 or BCOR increased H3K36me2 and H3K4me3 methylation on the EREG promoter and upregulated EREG expression. These results strongly suggested that FBXL11 performed histone demethylation by associating with BCOR, and this complex downregulated the transcription of EREG in MSC cells. However, further study is required to determine whether the BCL6 or other cofactors are involved in the process.

Previously, Lu et al. [23, 33] showed that FBXL11 could negatively regulate NFκB signaling in mutant human 293 cells. Our results showed that FBXL11 also negatively regulated NFκB in SCAP cells, and conversely, depletion of FBXL11 could activate the NFκB pathway, similar to results from previous reports [23, 33, 34]. Thus, we showed that FBXL11 had the expected functions in MSCs. Other studies showed that activation of the NFκB pathway could stimulate some critical osteo/dentinogenic differentiation regulators (RUNX2, BMP2, SP-1, and OSX). That stimulation resulted in enhanced mineralization of the extracellular matrix, enhanced cell differentiation, and it promoted an odontoblastic phenotype in dental pulp cells [35–38]. Therefore, we speculated that except EREG, NFκB pathway might also be involved in the osteo/dentinogenic regulation by FBXL11.

Taken together with previous studies [19], our results suggested that both FBXL11 and FBXL10 could associate with BCOR, they exerted similar histone demethylase functions on different target genes (which may be dependent on BCL6 activity), and they regulated osteo/dentinogenic differentiation in MSCs. Further study is required to determine whether these two homologs compete for BCOR association. Finally, the regulation of target genes down-stream of BCOR, like AP-2a and EREG, depended on cofactors of BCOR and on a particular binding sequence in the promoter.

CONCLUSION

Our results showed that FBXL11 inhibited osteo/dentinogenic differentiation potential in MSC cells by associating with BCOR, then increasing histone K4/36 methylation in EREG promoter to repress EREG transcription. EREG regulated osteo/dentinogenic differentiation potential through the key transcript factors, OSX and DLX2. Our work demonstrated new functions for FBXL11 and EREG, explored the molecular mechanisms underlying directed differentiation in MSC cells, and provided potential target genes for improving tissue regeneration approaches.

Acknowledgements

We thank Dr. Cun-Yu Wang, University of California, Los Angeles, School of Dentistry, for his instruction in this project, kind gifts of plasmids for FLAG-BCOR, the BCOR shRNA, and scientific review of this manuscript. This work was supported by grants from the National Natural Science Foundation of China (81070798 to Z.P.F., 81170931 to J.D.), National Basic Research Program of China (No. 2010CB944801), Funding Project to Science Facility in Institutions of Higher Learning Under the jurisdiction of Beijing Municipality (PXM2011_014226_07_000066 to Z.P.F.), the Beijing Funding Project for “Tens-Hundreds-Thousands” Outstanding Health Staff (2011, to Z.P.F.), and the Beijing Science and Technology Funding Project for Outstanding Returned Overseas Chinese Scholars (2012, to Z.P.F.). The Project Sponsored by the Scientific Research foundation for the Returned Overseas Chinese Scholars, State Education Ministry (2012-940 to Z.P.F., 2011-1568 to J.D.).

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors indicate no potential conflicts of interest.

Ancillary