Amphiregulin Regulates Differentiation of Dental Pulp Stem Cells by Activation of Mitogen-activated Protein Kinase and the Phosphatidylinositol 3-kinase Signaling Pathways

Background: Human dental pulp stem cells (hDPSCs) have received widespread attention in the elds of tissue engineering and regenerative medicine. Although amphiregulin (AREG) has been shown to play a vital function in the biological processes of various cell types, its effects on DPSCs is still unknown. The aim of this study was to explore the specic role of AREG as a biologically active factor in the regeneration of dental pulp tissue. Methods: The growth of hDPSCs, together with their proliferation and apoptosis, in response to AREG was examined by CCK-8 assay and ow cytometry. We explored the effects of AREG on osteo/odontogenic differentiation in vitro and investigated the regeneration and mineralization of hDPSCs in response to AREG in vivo. The effects of AREG gain-and loss-of-function on DPSC differentiation were investigated following transfection using overexpression plasmids and shRNA, respectively. The involvement of the mitogen-activated protein kinase (MAPK) or phosphatidylinositol 3-kinase (PI3K)/Akt pathways in the mineralization process and expression of odontoblastic marker proteins after AREG induction was investigated by using Alizarin Red S staining and western blotting. Results: AREG (0.01-0.1µg/mL) treatment of hDPSCs from 1 to 7 days minimally increased hDPSCs growth and marginally affected apoptosis compared with negative controls. AREG exposure signicantly promoted hDPSCs differentiation, shown by increased mineralized nodule formation and the expression of odontoblastic marker protein expression. In vivo micro-CT imaging and quantitative analysis showed signicantly greater formation of highly mineralized tissue in the 0.1μg/mL AREG exposure group in DPSC/NF-gelatin-scaffold composites. AREG also promoted extracellular matrix production, with collagen ber, mineralized matrix, and calcium salt deposition on the composites, as shown by H&E, Masson, and Von Kossa staining. Furthermore, AREG overexpression boosted hDPSCs differentiation while AREG silencing inhibited it. During the differentiation


Background
Mesenchymal stem cells (MSCs) are multifunctional cells found in various tissues, including bone marrow, bone, adipose tissue, Wharton's jelly, umbilical cord blood and peripheral blood, that can be used for cell therapy or organ regeneration [1][2]. Dental pulp stem cells (DPSCs) isolated from human dental pulp tissues, are speci c mesenchymal stem cells that can be induced to differentiate into odontoblasts in vitro or to form dentine-/pulp-like tissue in vivo [3][4][5][6]. DPSCs are relatively easy to obtain from waste tissue and demonstrate higher levels of proliferation, clonal potential, and mineralization than bone marrow mesenchymal stem cells (BMMSCs) [7]. Consequently, DPSCs have received widespread attention in the elds of tissue engineering and regenerative medicine. Despite this, signal transduction mechanisms involved in the oriented differentiation of DPSC are not yet fully understood.
Growth factor signaling plays pivotal regulatory roles in cell growth and differentiation. As amphiregulin (AREG) is a member of the epidermal growth factor (EGF) family, and is expressed by a variety of epithelial and mesenchymal cell types during development and in homeostasis [8][9]. Recently, AREG has been shown to play vital roles in regulating a broad range of biological processes, including cell growth, proliferation, nerve formation, cell migration and bone formation, by binding to the EGF receptor (EGFR) on the cell membrane [10]. Interestingly, AREG can regulate squamous cell differentiation and neuronal differentiation from stem/progenitor cell sources [11][12]. AREG has also been found to accumulate in multiple myeloma-derived exosomes and is involved in osteoclast differentiation through a circuitous mechanism in osteoblasts [13]. While, AREG appears to be an important growth factor in regulating the differentiation of a variety of cell types its effects on the growth and differentiation of DPSCs is still unknown.
AREG was rst identi ed in the culture supernatants of the human breast cancer cell line MCF-7 [14] and was found to bind to EGFR to activate downstream signaling pathways. Indeed, it is known to activate a variety of downstream intracellular signaling pathways, including Ras/MAPK, PI3K/AKT, mTOR, and STAT [15][16]. These signal transduction cascades regulate gene expression and initiate diverse cellular responses, including proliferation, survival, invasion, differentiation, and angiogenesis. AREG can promote migratory activity and doxorubicin resistance through activation of the MAPK pathway [17]. Furthermore, AREG-induced growth can also be regulated partially through the MAPK and PI3K-Akt/PKB pathways [18]. Currently, the role of AREG on activating downstream signaling in the regeneration of dental pulp tissue is not known. Therefore, our study aimed to explore the effects of AREG on growth, differentiation, and regeneration processes of DPSCs.

DPSC culture
Freshly extracted sound teeth were collected from patients (18-22-years-old), under the approval of the Ethics Committee of the Fourth Military Medical University (FMMU), Xi'an, China with informed consent. The hDPSCs were isolated, cultured, and identi ed as previously described [19][20]. Brie y, the pulp tissue was dissected and digested with 3 mg/mL type I collagenase and 4 mg/mL dispase (Sigma Aldrich, St Louis, MO, USA) for 45-60 min at 37℃. Single-cell suspensions were cultured in 60mm culture dishes and maintained in a-minimum essential medium (a-MEM; Invitrogen, Carlsbad, CA, USA) with 10% (v/v) fetal bovine serum (Gibco-BRL, Grand Island, NY, USA), 100 units/mL penicillin-G, and 100 mg/mL streptomycin (Invitrogen) in a humidi ed atmosphere with 5% CO 2 at 37℃. Single-cell clones of DPSCs were isolated and passaged as previously described [21]. DPSCs were grown in a 5% CO 2 incubator at 37°C and cultures at passages 3-5 were used for all studies.

Flow cytometric analysis of proliferation and apoptosis in hDPSCs
Third-passage DPSCs (3×10 6 cells/dish) were inoculated in 60 mm dishes. The experimental groups were treated with 0.01μg/mL and 0.1μg/mL AREG for 16 h, respectively. The control group received no AREG treatment. After removal from the dishes, cells were washed three times in pre-cooled phosphate-buffered saline (PBS) and were xed in 75% ice-cold ethanol. A ow cytometer (BD Biosciences, San Jose, CA, USA) was used to monitor the changes in the G0/G1, S, and G2/M phases of the cell cycle and to calculate the cell proliferation index (PI=G2/M+S). For the measurement of apoptosis, the AREG-treated DPSCs were harvested and washed as above, and 500μl of cells was diluted with 1× Annexin V Binding Buffer working solution according to the instructions of the apoptosis kit (KeyGENBioTECH, China). Then, 5μl Annexin V-APC and 5μl 7-ADD staining solution were added to the cells and the number of apoptotic cells was detected by FCM.

Alizarin Red S staining and quantitation
DPSCs (4 × 10 5 cells/well) were seeded into 6-well plates and cultured in either control medium or osteo/odontogenic induction medium with AREG for 14 days. At speci ed time, the cells were xed in 1 ml of 4% paraformaldehyde for 30 min at room temperature. After three washes with distilled water, 0.1g/ml Alizarin Red S (ARS) (Sigma-Aldrich) was used to stain the cultures for 10 min at room temperature. The unbound stain was removed by washing with deionized water until the discarded liquid appeared colorless. Five hundred microliters of water were added to each dish to keep the cells hydrated, and the cells were observed under Olympus inverted microscopy (Tokyo, Japan). For quantitation, alizarin red stain dissolved in 10% cetylpyridinium chloride (CPC) (Sigma-Aldrich) was added to the cells and the absorbance at 450 nm was read in the microplate reader.
In vivo studies All animal surgical procedures were approved by the Animal Care Committee and the Institutional Review Board (IRB) for Human Subjects Research of the Fourth Military Medical University. The high-stiffness three-dimensional (3D) nanofibrous gelatin (NF-gelatin) scaffolds were a kind gift from Prof. Tiejun Qu [22]. Initially, the NF-gelatin scaffolds were placed in 70% alcohol for half an hour and then washed three times in sterile PBS to remove residual ethanol. Before the seeding of human DPSCs (5×10 5 ), the scaffolds were soaked in α-MEM containing 10% FBS. The cell-scaffold composites were cultured in α-MEM supplemented with 10% FBS for 24 h on an orbital shaker (Orbi-shaker™, Benchmark, USA) in an incubator with 5% CO 2 at 37℃. Subsequently, the cell-scaffold composites were stimulated with 0.1μg/mL AREG in osteo/odontogenic induction medium for 7 days. Controls received no AREG exposure. The medium was changed every other day. After 7 days, the cell-scaffold composites were implanted subcutaneously on the dorsal surfaces of immune-compromised nude mice (nu/nu, 6-8 weeks old). After 4 weeks, the mice were euthanized by an anesthetic overdose and tissue growth was then harvested. Tissue samples were then immediately xed in 4% paraformaldehyde overnight. The samples (n=4 in each group) were scanned and analyzed using a micro-CT (eXplore Locus SP micro-CT; GE Healthcare, USA) as previously described [23] and the 3D micro-architectural properties of specimens were evaluated using analysis software (MicroView; GE Healthcare). After decalci cation in 17% ethylenediamine tetra-acetic acid, hematoxylin-eosin (H&E), Masson's trichrome and von Kossa staining was used for histological observation.
Overexpression and knockdown of AREG in DPSCs AREG-green uorescent protein lentivirus kits and their respective control kits were purchased from Hanbio (China). DPSCs were transfected with AREG lentivirus according to the manufacturer's instructions. In brief, third-passage DPSCs were inoculated in 12-well plates at 0.5×10 5 /mL and infected for 4 h, after which 0.5 mL fresh complete medium was added. After a further 24 h infection period, the medium was replaced with 1 mL fresh medium and after a further 72 h, the transfection rate was examined under a uorescence microscope (DMI8, Leica, German). Puromycin was used to select stably transfected cells with the puromycin concentration determined in a preliminary experiment. For simplicity, AREG-overexpressing and AREG-silenced cells were referred to as AREG (+) and AREG (-), respectively. The effects on the oriented differentiation of DPSCs were examined by ARS and Western blot analysis as described below.

Statistical analysis
All experiments were repeated separately in triplicate or quintuplicate. Data are expressed as means±SD. For statistical processing, we used SPSS software 16.0 (version 16.0; SPSS, Chicago, IL, USA). Inter-group differences were compared by the ONE-test. P<0.0.5 was considered to be statistically signi cant.

Culture and characterization of hDPSCs
Fibroblast-like clonal cells were obtained from the dental pulp tissue by limiting dilution and colony cloning (Fig. 1A a and b). Putative stem cells derived from the clonal cells (Fig. 1A c) were characterized by multiple lineage differentiation tests and flow cytometry. If cultured under inductive conditions, the cells formed mineral nodules and lipid droplets, shown by Alizarin Red and Oil Red O staining (Fig. 1B a  and b). Flow cytometry analysis showed that CD90, CD105, CD29, CD146 and STRO-1 were highly expressed in isolated cultured DPSCs (Fig.1C d-h). In contrast, the hematopoietic cell markersCD34 and CD45 (Fig. 1C b and c) were detected at minimal levels in the cultured stem cells. The morphology, colony formation, immunophenotype and ability to differentiate into multiple lineages, indicated that mesenchymal stem cells had been isolated.

Effects of AREG on growth of hDPSCs
The effects of AREG on hDPSCs numbers after1-, 3-, 5-, and 7-days incubation were assessed using the CCK-8 assay ( Fig. 2A). Data demonstrated that cell numbers in the 0.01μg/mL and 0.1μg/mL AREGtreated groups were signi cantly increased (P<0.05). In contrast, cell numbers in the 1μg/mL AREGtreated group were reduced. Cell cycle analysis showed a marginal increase in the proliferation index in the 0.01μg/mL and 0.1μg/mL AREG exposure groups (PI = G2/M+S), whereas treatment with 1μg/mL AREG decreased the proliferation index in comparison with the control group (CTRL) (Fig. 2B). In addition, FCM analysis showed that AREG did not signi cantly affect apoptosis of DPSCs compared with the CTRL group (Fig.2C).

Effects of AREG on odontogenic differentiation of hDPSCs in vitro
After two weeks of osteo/odontogenic induction, it was apparent that AREG concentrations between 0.01 and 0.1µg/mL promoted mineralized nodule formation in a dose-dependent manner, although a signi cant decrease was apparent at 1µg/mL AREG exposure ( Fig. 3A and B).
Consequently, 0.1µg/mL AREG exposure was selected as the optimal concentration for the following studies to investigate the in uence of AREG on DPSCs. When the expression of odontoblastic marker proteins was examined by western blotting, it was found that the levels of DSPP, BSP, RUNX2, and OCN were noticeably up-regulated in the AREG-treated groups compared with the controls by day 3. Levels were markedly elevated on days 7 and 14 (P<0.05) (Fig. 3C and D). Taken together, these results indicated that AREG could stimulate DPSC differentiation.

Effects of AREG on regeneration and mineralization of hDPSCs in vivo
To analyse regeneration capability in response to AREG, the cell-scaffold composites were subcutaneously implanted into nude mice. Subsequently, the specimens were scanned and images reconstructed by using micro CT analysis (Fig. 4 A and B). Micro-CT images showed the formation of more highly mineralized tissue formed in the 0.1μg/mL AREG exposure group in the composites (Fig.  4A a and b). Additionally, quantitative analysis also indicated that the bone volume fraction, trabecular thickness, and trabecular number were superior in the AREG group compared with the control group, although the trabecular separation was decreased (Fig. 4B a-d). All supported AREG enhancing odonto/osteogenic potential of the DPSCs.
The results of the H&E and Masson staining showed that the cell-scaffold composites stimulated by 0.1μg/mL AREG produced greater DPSC penetration into the pores of scaffold, as well as showing greater extracellular matrix secretion and collagen ber wrapping in the NF-gelatin scaffold, compared with the control group (Fig. 4C a, b, d, e, g, h). Von Kossa staining con rmed the presence of mineralization and further demonstrated greater deposition of mineralized matrix and calcium salts in the AREG group (Fig. 4C c, f, i). Data con rmed that AREG promoted the regeneration and mineralization capability of hDPSCs in vivo.

Effects of AREG over-expression and knockdown on hDPSCs differentiation
To further investigate AREG function during DPSC differentiation, cells were transfected with the AREG vector to induce AREG overexpression. Western blotting demonstrated increased expression of the odontoblastic markers at 3, 7, and14 days after induction compared with the control group ( Fig. 5C  and D). Furthermore, overexpression of AREG promoted mineralized nodule formation in DPSCs as detected by ARS staining and quantitation ( Fig. 5A and B). These data supported the ability of AREG to promote the differentiation of DPSCs.
To investigate whether AREG up-regulation was necessary for DPSC differentiation, AREG expression was silenced by shRNA transfection in DPSCs. Silencing was con rmed by qRT-PCR at 72 h after transfection with puromycin selection. Protein expression of DPSC odontoblast markers was decreased in response to AREG inhibition, as shown by Western blotting (Fig. 5E and F). ARS staining and quantitation showed that AREG silencing attenuated mineralized nodule formation in DPSCs (Fig. 5A and B). These results indicated that decreasing AREG expression resulted in the suppression of odontoblast differentiation.

Involvement of MAPK signaling in AREG-induced differentiation of hDPSCs
Treatment with AREG increased the protein level of p-ERK after 60 min stimulation. AREG stimulation also resulted in phosphorylated JNK in DPSCs in a time-dependent manner ( Fig. 6A and B). However, AREG showed minimal effects on phosphorylated p38 (Fig. 6A and B). Notably, incubation with the ERK, JNK, and MAPK inhibitors markedly antagonized the effect of AREG on phosphorylated ERK and JNK in hDPSCs ( Fig. 6C and D). These inhibitors also reduced the expression of the mineralization markers assayed by day 14, and mineralized nodule formation in DPSCs (Fig. 6E-H). Combined, these data indicated that the ERK, JNK, and MAPK pathways are implicated in AREG-induced differentiation of hDPSCs.
7. Involvement of PI3K/AKT signaling in AREG-induced differentiation of hDPSCs AREG phosphorylated AKT in DPSCs in a time-dependent manner ( Fig. 6A and B); this effect was inhibited by the PI3K pathway inhibitor LY294002 (Fig. 6C and D). Furthermore, Western blot data showed the levels of DSPP, BSP, RUNX2, and OCN were markedly reduced in cells treated with AREG + LY204002 (P<0.05) ( Fig. 6G and H). The PI3K pathway inhibitor LY294002 also markedly antagonized mineralized nodule formation and mineralization markers expression in DPSCs (Fig. 4E-H). These ndings suggest that the PI3K/AKT pathways are involved in the AREG-induced differentiation of hDPSCs.

Discussion
DPSC are a cellular reservoir, and have been shown to have considerable important applications in several different elds of tissue engineering. As with other somatic stem cells, a variety of studies have veri ed the potential of DPSCs in regenerative therapies, based on their multipotent differentiation capability [24]. Importantly, their cell growth and differentiation cannot be separated from the regulatory action of stimulatory growth factors and numerous studies have demonstrated the importance of the EGF family in developmental regulation processes [25][26]. EGF is known to exert its biological actions through its binding to EGFR on cell surfaces and amphiregulin (AREG) is also able to interact with the same receptor to activate downstream signaling pathways [10]. In the present study, the results indicated that AREG was able to promote odontoblastic differentiation in hDPSCs through the activation of the ERK, JNK, MAPK, and PI3K/Akt pathways, and facilitated the regeneration and mineralization of hDPSCs. This action may have potential to be harnessed for the development of novel strategies for the repair of damaged pulp and tissue regeneration.
A previous study demonstrated that AREG was able to inhibit or induce cell growth and proliferation in a range of cell types [14]. For example, AREG can induce a potent proliferative response in colon carcinoma cells [27] and it can also increase proliferation in airway epithelial and smooth muscle cells [28]. AREG also plays a vital function in skin wound healing by stimulating keratinocyte proliferation [29][30][31][32]. Furthermore, it has been reported that overexpression of AREG induced self-su cient growth and survival in lung, liver, colon, breast, and pancreatic carcinoma cells [10,[33][34][35][36][37]. In the present study, we identi ed a biphasic effect of AREG. Data demonstrated that the numbers of hDPSCs treated with 0.01µg/mL and 0.1µg/mL AREG were signi cantly increased after 3, 5, and 7 days of treatment while exposure to higher dose of 1µg/mL AREG reduced cell numbers by day 7. The analysis of cell cycle phases demonstrated that the 0.01µg/mL and 0.1µg/mL AREG treatment groups had a slightly higher proliferation index (PI = G2/M+S), in comparison with both the control group and the 1µg/mL AREG-treatment group. Notably, ow cytometry analysis showed that treatment with 0.01-1µg/ml AREG had minimal impact on the apoptosis of hDPSCs. These results are largely consistent with those of most current studies [29][30][31][32].
AREG is potentially ideally suited for use in tissue repair and regeneration applications as it is not only able to enhance proliferation but can also stimulate differentiation. These data are consistent with previous ndings as AREG has been reported to induce the differentiation of neuronal PC12 cells [38] and it has also been shown to be more effective for human mammary epithelial differentiation than other EGFR ligands [39]. Similar results have been reported for human mammary myoepithelial cells [40]. Furthermore, previous studies have demonstrated that AREG can induce cell differentiation in a variety of cell types. In our study, AREG was shown to induce mineralized nodule formation in a dose-dependent manner, although the high dose of 1 µg/mL AREG induced a marked decrease. In addition, AREG also increased DSPP, BSP, RUNX2, and OCN expression. Combined these data indicated that AREG induced odontogenic differentiation of DPSCs in vitro.
The biomimetic NF-gelatin scaffolds used here provide an excellent resource for bone tissue engineering studies due to their physical architecture and chemical composition which is similar to natural bone ECM. These constructs exhibit excellent biocompatibility, mechanical stability, and enhancement of the osteogenic differentiation [22][23]41]. To clarify the role of AREG in regeneration and mineralization in vivo, the DPSC/NF-gelatin-scaffold composites with or without AREG treatment were subcutaneously implanted in nude mice for 4 weeks. Micro CT analysis was initially used to detect the content, density, and distribution of both bone tissue and mineralized hard tissue formed in the cell-scaffold composites. The quantitative analysis using Micro CT showed that BV/TV (Bone Volume to Tissue Volume), Tb.Th. (Trabecular Thickness), and Tb.N. (Trabecular Number) were markedly increased in the AREG-stimulated group, although the trabecular separationwas decreased. These ndings indicate that AREG supplementation produced more newly mineralized tissue together with increased numbers and thicker bone trabeculae, and the trabecular structure was more compact. These data indicate that AREG can promote the formation of the mineralized tissue in vivo. Furthermore, both H&E and Masson staining showed increased DPSC penetration into the pores of scaffold, with relatively large amounts of ECM and collagen ber deposition and formation in the AREG group. Von Kossa staining indicated that biomineralization occurred in all the composites, with greater mineralized matrix and calcium salt deposition on the composites in the AREG treated group. This nding was consistent with the Micro CT quantitative analysis data. All the results demonstrated that AREG facilitated the formation of dentin-like matrix in DPSC/NF-gelatin-scaffold composites. Furthermore, the AREG over-expression experiments showed a similar promotion of odontogenic differentiation, whereas the knockdown experiments inhibited this process. These results indicated that AREG was necessary for odontoblastic differentiation of DPSCs and could facilitate the regeneration and mineralization of hDPSCs. It is yet to be determined but AREG may contribute substantially to injured pulp repair and regeneration.
AREG plays a vital role in biological processes through its interaction with EGFR and tyrosine phosphorylation of downstream proteins, activating two major signaling pathways, the MAPK and PI3K/Akt pathways [42][43]. Interestingly, a previous study has shown that AREG interacts with EGFR, activating PI3K/Akt, and subsequently the NF-κB transcription factor on the MMP-13 promoter, inducing cartilage destruction in osteoarthritis [44]. MAPK signaling has also been observed to participate in AREGinduced morphological effects in MDCK cells [45] and AREG stimulation was found to be required for the differentiation of K5+K19-hMECs through activation of ERK and MAPK but not Akt signaling [39]. Moreover, inhibition of ERK1/2 blocked AREG-induced myoepithelial differentiation [40]. In the present study, ERK, JNK, MAPK, and AKT were phosphorylated in response to AREG stimulation, an action that was markedly antagonized by speci c inhibitors of these proteins. In addition, the ERK, JNK, MAPK, and PI3K/Akt inhibitors also signi cantly reduced mineralized nodule formation and expression of protein mineralization markers. These data support the involvement of ERK, JNK, MAPK, and PI3K/AKT pathways in AREG-induced differentiation of hDPSCs.

Conclusions
In conclusion, our in vitro and in vivo data demonstrated that AREG was necessary for odontoblastic differentiation of DPSCs and facilitated the regeneration and mineralization of hDPSCs. Furthermore, the investigation of the roles of the MAPK and PI3K/AKT pathways in AREG-induced growth and osteo/odontogenic differentiation in hDPSCs indicated that activation of the ERK, JNK, MAPK and PI3K/AKT signaling pathways activation are involved in AREG-mediated differentiation of hDPSCs. Further investigation of these mechanisms may provide targets and treatment modalities for use in the repair of damaged and diseased pulp and for the regeneration of dental tissues.

Figure 2
Effects of AREG on the growth of DPSCs.