IGFBP2 enhances adipogenic differentiation potentials of mesenchymal stem cells from Wharton's jelly of the umbilical cord via JNK and Akt signaling pathways

Mesenchymal stem cell (MSC)-mediated tissue engineering represents a promising strategy to address adipose tissue defects. MSCs derived from Wharton’s jelly of the umbilical cord (WJCMSCs) may serve as an ideal source for adipose tissue engineering due to their abundance, safety profile, and accessibility. How to activate the directed differentiation potentials of WJCMSCs is the core point for their clinical applications. A thorough investigation of mechanisms involved in WJCMSC adipogenic differentiation is necessary to support their application in adipose tissue engineering and address shortcomings. Previous study showed, compared with periodontal ligament stem cells (PDLSCs), WJCMSCs had a weakened adipogenic differentiation potentials and lower expression of insulin-like growth factor binding protein 2 (IGFBP2). IGFBP2 may be involved in the adipogenesis of MSCs. Generally, IGFBP2 is involved in regulating biological activity of insulin-like growth factors, however, its functions in human MSCs are unclear. Here, we found IGFBP2 expression was upregulated upon adipogenic induction, and that IGFBP2 enhanced adipogenic differentiation of WJCMSCs and BMSCs. Moreover, IGFBP2 increased phosphorylation of c-Jun N-terminal kinase (p-JNK) and p-Akt, and activated JNK or Akt signaling significantly promoted adipogenic differentiation of MSCs. Furthermore, inhibitor-mediated blockage of either JNK or Akt signaling dramatically reduced IGFBP2-mediated adipogenic differentiation. And the JNK inhibitor, SP600125 markedly blocked IGFBP2-mediated Akt activation. Moreover, IGFBP2 was negatively regulated by BCOR, which inhibited adipogenic differentiation of WJCMSCs. Overall, our results reveal a new function of IGFBP2, providing a novel insight into the mechanism of adipogenic differentiation and identifying a potential target mediator for improving adipose tissue engineering based on WJCMSCs.


Introduction
Adipose tissue plays an essential role in the maintenance of organ contours, energy storage, metabolic balance, and immune regulation through exocrine hormones and fat cell factors. Adipose tissue defects due to tumor resection, trauma, or hereditary and congenital diseases usually lead to loss of fat tissue, poor appearance and local disordered regulatory function. A major clinical challenge is that traditional treatments such as prosthetic appliance, plastic and reconstructive surgery and fat grafting do not effectively restore adipose tissue. Adipose tissue engineering and cell-based therapies represent novel and promising approaches for regenerating adipose tissue.
MSCs have been isolated from various tissues, including bone marrow, adipose tissue, vascular tissue, dental tissue, craniofacial tissue, and umbilical cord [1][2][3][4][5][6][7]. With their convenient isolation, low immunogenicity, and ability to transdifferentiate, MSCs are considered a promising therapeutic approach for tissue regeneration [1,4,5,8,9]. WJCMSCs, which are isolated from neonatal umbilical cord tissue, are a plentiful, cost-effective, and biologically safe source of stem cells and show the significant potential for regenerative medicine [9][10][11][12]. As studied dental-derived stem cell population, PDLSCs which own the higher stemness features and preferable multi-differentiation properties, are the ideal seeding cells for tissue regeneration [8,13,14]. However, some researches suggested that compared with PDLSCs, unmodified WJCMSCs with their weaker stemness features might not be ideal seed cells for tissue regeneration [8,15]. A crucial issue for WJCMSCs-mediated adipose tissue regeneration is how to activate adipogenic differentiation and enhance regenerative ability.
As the essential member of insulin-like growth factor (IGF) axis, insulin-like growth factor binding proteins (IGFBPs) are homologs with high structural similarity but distinct functionalities. They all have similar N-terminal and C-terminal domains connected by a variable linker region [16]. IGFBPs assume a key regulatory role in many cellular processes, including proliferation, migration, differentiation, and survival. IGFBPs also play an essential role in the processes of growth, development, and tissue metabolism [17]. In the IGF axis, IGFs play influential roles in the function of IGFBPs [18,19]. Several studies indicated that IGF1 was an essential regulator of adipogenic differentiation. IGF1 was shown to upregulate phosphorylation of cAMP response element-binding protein (CREB) through the PI3K/Akt pathway, and then activated CREB increased the expression of PPARγ2, which was a crucial factor in adipogenesis through regulating specific gene expression [20][21][22]. IGFBP2 was the predominant binding protein secreted by differentiating white preadipocytes. In chickens, the IGFBP2 gene could be a candidate locus or linked to a major gene associated with abdominal fat weight and percentage of abdominal fat [23,24]. Our previous research showed that, compared with dental derived stem cells, WJCMSCs exhibited decreased adipogenic differentiation potential as well as downregulated expression of IGFBP2 [15]. These findings suggested the possible involvement of IGFBP2 in the regulation of adipogenic differentiation in MSCs.
Many events facilitate the commitment of MSC adipogenic differentiation, including the coordination of a complex network of transcription factors, co-factors, and pathway signaling intermediates. The extracellular regulated protein kinases (ERK), p38, and JNK MAPK family are a group of serine/threonine kinases that transduce extracellular signals to intracellular targets, involving a series of protein kinase cascades and long-term response that play a crucial role in regulating cell differentiation [25][26][27]. Many researches focused on the effect of the MAPK family on adipogenic differentiation. Sale et al. found that ERK1 and ERK2 were required for differentiation of 3T3-L1 fibroblasts to adipocytes [28]. And inhibited ERK pathway by specific inhibitor could restrain adipocyte differentiation ability [29]. In addition, ERK activity was essential for the expressions of the PPARγ and C/EBP [30,31]. Moreover, cells isolated from Erk -/mouse showed impaired adipogenesis capability [32]. It was previously reported that the JNK pathway also regulated adipogenesis differentiation of MSCs [33]. JNK could phosphorylate PPARγ2 by oxidized low-density lipoprotein [34]. Yet, using SP600125, a specific JNK inhibitor, could increase the expressions of CEBPα/β and PPARγ2, and stimulate adipogenesis of hASCs in a dose-dependent manner [35]. Moreover, the drug for preventing osteoporosis, alendronate, inhibited adipogenic differentiation by ERK and JNK pathway in BMSCs [36]. As for the role of p38 in adipogenic differentiation, some studies showed that using p38 inhibitors could block adipocyte differentiation [37,38]. In addition, the Akt signaling pathway was also essential for inducing PPARγ. Akt activity sustained the adipogenic differentiation of ASCs. Akt knockout mice showed impaired adipogenesis [39][40][41]. Importantly, IGFBP2 could activate multiple MAPK pathways. IGFBP2/Integrin5 interaction promoted glioma cells migration through JNK activation [42]. Exogenous IGFBP2 induced proliferation and activated the ERK pathway in NIH-OVCAR3 cells, and also promoted proliferation in rat growth plate chondrocytes via MAPK/ERK pathway [43]. In addition, the expression of IGFBP2 was positively regulated by PI3K/Akt pathway, and the Akt signal transduction was impaired in Igfbp2 -/mouse cells [44]. However, it is still unknown the effect of IGFBP2 on MAPK and Akt pathways during adipogenic differentiation of WJCMSCs. Based on the available information, we hypothesize that IGFBP2 affects the function of MSCs, but its function and mechanism remain unclear. Here, we investigate the effects and underlying mechanisms of IGFBP2 in the adipogenic differentiation of MSCs. Our results show overexpression IGFBP2 enhances adipogenic differentiation of WJCMSCs by activating JNK and Akt signaling pathway. Furthermore, we find that IGFBP2 is negatively regulated by BCOR, which represses the adipogenic differentiation potential of WJCMSCs.

Ethics statement and cell cultures
Between January and November 2012, patients were recruited from the Department of Oral and Maxillofacial Surgery of Beijing Stomatological Hospital, Capital Medical University. And human impacted third molars were collected from six healthy male patients (16-20 years old) under approved guidelines set by the Beijing Stomatological Hospital, Capital Medical University (Ethical Committee Agreement, Beijing Stomatological Hospital Ethics Review No. 2011-02), with written informed consent. In addition, we also obtained the informed consent from parent/guardian on behalf of minors (<18 years old). The authors had access to information that could identify individual participants during or after data collection.
Teeth were first disinfected with 75% ethanol and then washed with phosphate-buffered saline. PDLSCs were isolated, cultured, and identified as previously described [15,22]. Briefly, PDLSCs were separated from periodontal ligament in the middle one-third of the root. Subsequently, MSCs were digested in a solution of 3 mg/mL collagenase type I (Worthington Biochemical Corp., Lakewood, NJ, USA) and 4 mg/mL dispase (Roche Diagnostics Corp., Indianapolis, IN, USA) for 1 h at 37˚C. Single-cell suspensions were obtained by cell passage through a 70-μm strainer (Falcon, BD Labware, Franklin Lakes, NJ, USA). Human BMSCs, ASCs, and WJCMSCs were purchased from ScienCell Research Laboratories (Carlsbad, CA, USA). MSCs were grown in a humidified, 5% CO 2 incubator at 37˚C in DMEM alpha modified Eagle's medium (Invitrogen, Carlsbad, CA, USA), supplemented with 15% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, USA), 2 mmol/L glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA, USA). The culture medium was changed every 3 days. MSCs at passages 3-5 were used in subsequent experiments. Human embryonic kidney 293T cells were maintained in complete DMEM with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. For viral packaging, HEK293T cells at 80% confluency were cotransfected with plasmids and transfection reagent. For SP600125 (Cell Signaling Technology, Beverly, MA, USA), LY294002 (Cell Signaling Technology, Beverly, MA, USA), anisomycin (Cell Signaling Technology, Beverly, MA, USA), or insulin (Sigma-Aldrich, St. Louis, MO, USA) treatment, MSCs were starved for 24 h to synchronize the cells in DMEM alpha modified Eagle's medium without serum, then changed to routine culture medium and treated with appropriate agents. The studies on human MSCs were conducted between May 2012 and November 2016. All cell-based experiments were repeated three times. http://dx.doi.org/10. 17504/protocols.io.iegcbbw.

Plasmid construction and viral infection
The plasmids were constructed using standard methods; all sequences were verified by appropriate restriction digestion and/or sequencing. Human full-length IGFBP2 cDNA from ASCs fused to a M2-Flag tag was produced with a standard PCR protocol. This sequence (Flag-IGFBP2) was subcloned into the pQCXIN retroviral vector with AgeI and BglII restriction sites. Similarly, the human full-length BCOR cDNA was fused to a Flag tag (Flag-BCOR) and subcloned into the pQCXIN retroviral vector with AgeI and BamH1 restriction sites. For viral infections, MSCs were plated overnight, and then infected with retroviruses in the presence of polybrene (6 μg/ mL, Sigma-Aldrich, St. Louis, MO, USA) for 12 h. After 48 h, infected cells were selected with 600 μg/mL G418 for 10 days. http://dx.doi.org/10.17504/protocols.io.iehcbb6.

Western Blot analysis
Cells were lysed in RIPA buffer (10 mM Tris-HCl, 1 mM EDTA, 1% sodium dodecyl sulfate [SDS], 1% NP-40, 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 PVDF membranes with a semi-dry transfer apparatus (Bio-Rad, Hercules, CA, USA). The membranes were blotted with 5% dehydrated milk for 1 h and then incubated with primary antibodies overnight. The immune complexes were incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG (Promega, Madison, WI, USA) and visualized with SuperSignal reagents (Pierce, Rockford, IL, USA). Primary antibodies were purchased from following commercial sources: monoclonal anti-FLAG M2 (

Oil Red O staining
Adipogenic differentiation was induced by using the StemPro adipogenesis differentiation kit (Invitrogen, Carlsbad, CA, USA). MSCs were grown in the adipose-inducing medium for 3 weeks. For Oil Red O staining, after induction, cells were fixed with 10% formalin for at least 1 h at room temperature. Next, cells were stained with the 60% Oil Red O in isopropanol as working solution for 10 min. The proportion of Oil Red O-positive cells was determined by counting stained cells under a light microscope. The final OD value in each group was normalized with the total protein concentrations prepared from a duplicate plate. http://dx.doi.org/ 10.17504/protocols.io.iemcbc6.

Statistics
All statistical calculations were performed with SPSS20.0 statistical software (IBM, Armonk, NY). Comparisons between two groups were analysed by independent two-tailed Student's ttests, and comparisons between more than two groups were analysed by one-way ANOVA followed by a Duncan's post hoc test. Data were expressed as the mean ± standard deviation (SD) of 3 experiments per group. P values < 0.05 were considered statistically significant.
To determine whether IGFBP2 had similar functions in other MSCs, we overexpressed IGFBP2 in BMSCs via retrovirus expressing Flag-tagged wild type IGFBP2 (S1 Fig). Assessment of Oil Red O staining and real-time RT-PCR revealed that IGFBP2 significantly promoted adipogenic differentiation in BMSCs (S1 Fig). Together, these results showed that

IGFBP2 increases JNK and Akt phosphorylation in WJCMSCs
To investigate how IGFBP2 enhanced the adipogenic differentiation of WJCMSCs, we used Western Blot and quantitative analysis to examine the levels of proteins involved in MAPK signaling, including p38, ERK and JNK, and the Akt pathway. The results showed that overexpression of IGFBP2 enhanced phosphorylation of JNK and phosphorylation of Akt in WJCMSCs, while phosphorylation of ERK, and total protein levels of p38, JNK, ERK, and Akt proteins were not affected (Fig 3A and 3B). And phosphorylated p38 MAPK was not found.
To test whether activated JNK or Akt had the capability of pro-adipogenesis in MSCs, we used the JNK activator (anisomycin) or Akt activator (insulin) in WJCMSCs. WJCMSCs were treated with 100nM, 200nM or 500nM anisomycin for 24 h to activate JNK signaling, or treated with 50nM, 100nM, 200nM or 500nM insulin for 24 h to activate Akt signaling. Western Blot results showed that 100nM, 200nM or 500nM anisomycin could effectively activate JNK signaling (S2 Fig), and 50nM, 100nM,

IGFBP2-enhanced adipogenic differentiation of WJCMSCs is repressed by JNK or Akt inhibitors
First, WJCMSCs were treated with 10 μM, 20 μM, 50 μM or 100 μM specific JNK inhibitor, SP600125 for 48 h to block JNK signaling in WJCMSCs. Western Blot results ( Fig 4A) and quantitative analysis (Fig 4B) indicated that 20 μM, 50 μM or 100 μM SP600125 could effectively block JNK signaling. Then, 20 μM SP600125 was selected for further experiments. Transduced WJCMSCs were cultured in adipogenic-inducing medium with 20 μM SP600125. Three weeks after induction, Oil Red O staining revealed that 20 μM SP600125 could restrain IGFBP2-mediated enhancement of adipogenic differentiation in WJCMSCs (Fig 4C). After normalizing the data with the total protein, the results indicated that the effect of IGFBP2increased adipogenic differentiation of WJCMSCs was associated with JNK activation ( Fig  4D). To confirm this finding, we further examined the adipogenic differentiation markers PPARγ and LPL by real-time RT-PCR. The results showed that PPARγ (Fig 4E) and LPL ( Fig  4F) were significantly suppressed at 1, 2 or 3 weeks after induction in WJCMSC-Flag-IGFBP2 + SP600125 group compared with WJCMSC-Flag-IGFBP2 group.

Activated Akt signaling by IGFBP2 is repressed by the specific JNK inhibitor in WJCMSCs
To further explore the underlying mechanism, we used the specific JNK inhibitor (20 μM SP600125) or Akt inhibitor (10 μM LY294002) to block the activated JNK or Akt pathway by IGFBP2 in WJCMSCs. Western Blot results ( Fig 6A) and quantitative analysis (Fig 6B) showed that 20 μM SP600125, which could inhibit the p-JNK level, effectively abrogated the expression of phosphorylation-Akt in WJCMSC-Flag-IGFBP2 cells. However, treatment with 10 μM LY294002 had no significant effect on the expression of p-JNK in WJCMSC-Flag-IGFBP2 cells (Fig 6C and 6D).

BCOR negatively regulates IGFBP2 expression and inhibits adipogenic differentiation of WJCMSCs
Ectopic BCOR overexpression was confirmed by Western Blot analysis (Fig 7A). Real-time RT-PCR results showed that BCOR overexpression in WJCMSCs suppressed the expression of IGFBP2 (Fig 7B). Next, to investigate adipogenic differentiation, WJCMSCs were cultured in adipogenic-inducing medium. After induction for 3 weeks, Oil Red O staining (Fig 7C) and quantitative lipid deposit measurements (Fig 7D) showed there were significantly fewer lipid deposits in WJCMSC-Flag-BCOR cells than in WJCMSC-Vector cells.

Discussion
Mesenchymal stem cells derived from Wharton's jelly of the umbilical cord, which are usually discarded after birth, possess multipotent abilities between those of embryonic and adult stem   [2,45]. Studies have shown that WJCMSCs possess many attractive properties, including expanding faster than adult-derived MSCs, ample cell supply and the potential for autologous grafting if they are cryopreserved for future use [7,10,12]. Moreover, in clinical practice, WJCMSCs have been successfully used to treat autoimmune disease [11]. Our previous study PPARγ (E) and LPL (F) in WJCMSC-Flag-IGFBP2 cells following 10 μM LY294002 treatment during adipogenic induction at 1, 2, and 3 weeks. GAPDH was used as an internal control. **p < 0.01. α: anti; w: week.
https://doi.org/10.1371/journal.pone.0184182.g005 found that compared with PDLSCs, WJCMSCs exhibited decreased adipogenic differentiation potential, and unmodified WJCMSCs might not be good seed cells for tissue regeneration [15]. Therefore, a critical issue for WJCMSCs applied in tissue engineering is how to enhance the differentiation potentials and regenerative abilities. Using microarray analysis, we observed decreased expression of IGFBP2 in WJCMSCs compared with PDLSCs [15]. IGFBP2 is mainly expressed in highly proliferative fetal tissues which represent extensive cell movement and tissue remodeling [46]. Our results confirm that IGFBP2 is a potential mediator for enhancing adipogenic differentiation of WJCMSCs and BMSCs. The potentiality of IGFBP2 and the possibility of modulating specific pathways underlying biological process of WJCMSCs offer new strategies in the field of regenerative medicine.
Regulation of adipogenic differentiation by growth factors is a complex process [20,47,48]. PPARγ is a master regulator of adipogenesis, and generally most all pro-adipogenic signaling pathways associated with PPARγ [49]. JNK is one of the major sub families of MAPKs [25][26][27]. Studies revealed that JNK pathway was associated with regulating adipogenic differentiation. Previous research showed that wild-type IGFBP2-overexpressing cells showed a higher level of phosphorylation JNK [42]. In NIH-OVCAR3 cells, IGFBP2 promoted proliferation, potentiated ERK phosphorylation and activated SAPK/JNK signaling pathway [43]. Moreover, the anti-adipogenesis effect of 6-thioinosine was mediated by decreased expression of PPARγ through JNK pathway. Loss of JNK1 activity resulted in resistance to high-fat diet-induced obesity in vivo [50,51]. In addition, Akt was also essential for inducing PPARγ and adipogenic differentiation; depletion of Akt impaired adipogenesis in mice [39][40][41]. Furthermore, impaired IGF1 mitogenesis involving the Akt pathway contributed to the distinct growth phenotype of visceral preadipocytes. More importantly, many researches inferred that using the JNK specific inhibitor or siRNA led to decreased Akt phosphorylation in many cells and cell processes [52][53][54]. However, it was reported that activation of JNK decreased Akt phosphorylation in liver tissue [42,43]. Pretreatment with the JNK specific inhibitor and salvianolic acid A caused decreased p-JNK and increased p-Akt in diabetic rats with ischemia/reperfusion [55]. Our results show that IGFBP2 overexpression activates phosphorylation of JNK and Akt signaling, and activated JNK or Akt signaling enhances adipogenic differentiation of MSCs. In addition, JNK or Akt inhibitor suppresses IGFBP2-mediated enhancement of adipogenic differentiation in WJCMSCs. Separately, the results indicated that JNK and Akt signaling pathway exert an important role for IGFBP2-enhanced adipogenic differentiation. Furthermore, the specific JNK inhibitor markedly decreases the expression of phosphorylated Akt activated by IGFBP2, indicating that Akt is the downstream of the JNK in IGFBP2 mediated signaling cascade. Taken together, our results confirm that IGFBP2 enhances adipogenic differentiation of WJCMSCs via activated JNK/Akt signaling pathway. However, further study is required to investigate the regulation mechanism about JNK/Akt crosstalk in the process.
In addition to these results, we also find that BCOR negatively regulates the expression of IGFBP2; this is consistent with our previous microarray analysis, which found that IGFBP2 was highly expressed in stem cells from the apical papilla (SCAPs) from oculo-facio-cardiodental (OFCD) syndrome that had a mutation in BCOR [56]. Our results also reveal that BCOR represses adipogenic differentiation of WJCMSCs. The BCOR gene encodes a protein known as the BCL6 co-repressor, which might use an epigenetic mechanism to direct gene silencing [56][57][58]. Previous researches inferred that BCOR regulated the function of MSCs by associating with the activating enhancer binding protein 2 alpha (AP2α) promoter [56]. The 5' flanking region of IGFBP2 gene contains motifs that might be recognized by transcription factor AP2 [59]. Based on those studies, we speculate BCOR may be involved in the regulation of IGFBP2 by epigenetics or AP2. However, this is beyond the scope of the current study and will require further investigation.
In summary, our results identify a novel function of IGFBP2 in adipogenic differentiation of MSCs. Generally, BCOR negatively regulates IGFBP2, and overexpression of IGFBP2 can enhance the adipogenic differentiation of WJCMSCs through activating JNK and Akt signaling pathways. This study elucidates molecular mechanisms underlying adipogenic differentiation of WJCMSCs, and suggests that IGFBP2 may be a potential target to promote the adipose tissue regeneration.