Comparative proteomic profiling of human dental pulp stem cells and periodontal ligament stem cells under in vitro osteogenic induction

Archives of Oral Biology 89 (2018) 9–19

Contents lists available at ScienceDirect

Archives of Oral Biology journal homepage: www.elsevier.com/locate/archoralbio

Comparative proteomic profiling of human dental pulp stem cells and periodontal ligament stem cells under in vitro osteogenic induction He Wanga,b, Dandan Maa,b, Xiaoyi Zhanga,b, Shuaimei Xuc, Tingting Ninga,b, Buling Wua,b,

T ⁎

a

Department of Stomatology, Nanfang Hospital, Southern Medical University, No. 1838 North Guangzhou Avenue, Guangzhou, 510515, China College of Stomatology, Southern Medical University, No. 1838 North Guangzhou Avenue, Guangzhou, 510515, China c Department of Endodontics and Operative Dentistry, Stomatological Hospital, Southern Medical University, No. 366 South Jiangnan Avenue, Guangzhou, 510280, China b

A R T I C L E I N F O

A B S T R A C T

Keywords: Comparative proteomics Dental pulp stem cells Isobaric tag for relative and absolute quantitation Osteogenesis/odontogenesis Osteogenic induction Periodontal ligament stem cells

Objective: This study aimed to compare the proteomic profiling of human dental pulp stem cells (DPSCs) and periodontal ligament stem cells (PDLSCs) under in vitro osteogenic induction, which imitates the microenvironment during osteo-/odontogenesis of DPSCs and PDLSCs. Design: The proteomic profiles of osteoinduced DPSCs and PDLSCs from a single donor were compared using the isobaric tag for relative and absolute quantitation (iTRAQ) technique and subsequent bioinformatics analysis. Results: A total of 159 differentially expressed proteins in PDLSCs and DPSCs were identified, 82 of which had a higher expression level in PDLSCs, while 77 were more highly expressed in DPSCs. Among these enriched proteins, certain members from the collagen, heat shock protein and protein S100 families may distinguish osteoinduced PDLSCs and DPSCs. Gene ontology (GO) classification revealed that a large number of the enriched terms distinguishing PDLSCs and DPSCs are involved in catalytic activity, protein binding, regulation of protein metabolic processes and response to stimulus. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis indicated several involved pathways, including the fatty acid biosynthesis pathway, pantothenate and CoA biosynthesis pathway, arachidonic acid metabolism pathway and PPAR signaling pathway. Further verification showed that the mineralization and migration capacities of PDLSCs were greater than those of DPSCs, in which heat shock protein beta-1, Protein S100-A10 and S100-A11 may play a part. Conclusions: Less than 5% of the differentially expressed proteins make up the comparative proteomic profile between osteoinduced PDLSCs and DPSCs. This study helps to characterize the differences between osteoinduced PDLSCs and DPSCs in vitro.

1. Introduction Human dental pulp stem cells (DPSCs) and periodontal ligament stem cells (PDLSCs) are two of the most important adult mesenchymal dental-derived stem cells that can be isolated from adult dental pulp and periodontal ligament tissues, respectively. Although these cell types are both derived from neural crest and share a similar immunophenotype in vitro, DPSCs and PDLSCs are considered to be distinct from each other due to their disparate origins and functions. DPSCs lie in dental pulp tissue, which is surrounded by dentin and plays a vital role in tooth nourishment, inhibiting bacterial invasion and reacting to mechanical and chemical stimuli. As undifferentiated mesenchymal cells, DPSCs can differentiate into odontoblasts and form reparative dentin under an exogenous stimulus (Gronthos, Mankani, Brahim, Robey, & Shi, 2000). Periodontal ligament tissue possesses

high regenerative capacity and a rapid turnover rate. Fibroblasts, comprising most of the PDLSCs, produce extracellular matrix and other substances, which are key to maintaining periodontal homeostasis (Hinz, 2013) and periodontal ligament width over the lifetime (Lim et al., 2014). Similar to dental pulp, the periodontal ligament also contains undifferentiated mesenchymal cells, PDLSCs (Ivanovski, Gronthos, Shi, & Bartold, 2006; Seo et al., 2004), which function as regenerative cell resources under unfavorable conditions. One of the most widely adopted methods to study the osteogenic or odontogenic differentiation of dental-derived cells is in vitro osteogenic induction using osteogenic induction medium (Cui et al., 2014; Qu et al., 2016). As a well-recognized method, in vitro osteogenic induction, to a certain degree, imitates the microenvironment needed for osteogenesis/odontogenesis of mesenchymal cells. Under such conditions, although DPSCs and PDLSCs can form osteoid tissue, which is a

⁎ Corresponding author at: Department of Stomatology, Nanfang Hospital, Guangzhou; College of Stomatology, Southern Medical University, No. 1838 North Guangzhou Avenue, Guangzhou, 510515, China. E-mail address: [email protected] (B. Wu).

https://doi.org/10.1016/j.archoralbio.2018.01.015 Received 21 September 2017; Received in revised form 16 January 2018; Accepted 23 January 2018 0003-9969/ © 2018 Elsevier Ltd. All rights reserved.

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

was used to obtain single cell–derived colonies (Seo et al., 2004). Periodontal ligament and dental pulp cells of passage 1 at logarithmic phase were selected. The multiple proportion dilution method was performed to modulate cell density to 10 to 15 cells/mL. Cells were seeded into 96-well plates with a volume of 100 μL/well, and the medium was changed every 2 to 3 days. After reaching approximately 80% confluence, single cell–derived colonies were subcultured using 0.25% (w/v) trypsin-0.02% ethylenediamine tetraacetic acid digestion. All primary cells were used at 3 to 4 passages, and the same passage of PDLSCs or DPSCs were used for each experiment.

trait of osteogenesis/odontogenesis, DPSCs differentiate into odontoblast-like cells (Gronthos et al., 2000) while certain PDLSCs differentiate into cementoblast-like cells, adipocytes, and collagen-forming cells (Seo et al., 2004). This finding suggests that they have distinct reactions towards in vitro osteogenic induction, and gaining insight into these reactions may contribute to further exploration and utilization of these two types of dental mesenchymal cells for clinical application. Clinically, the loss of dental pulp tissue and periodontal ligament tissue due to pulpitis (Peng et al., 2017) and periodontitis (Slots, 2017), respectively, usually leads to poor prognosis of the affected teeth. The core idea of tissue regeneration of dental pulp and periodontal ligament lies in inducing DPSCs and PDLSCs to differentiate into odontoblasts and cementoblasts, thereby restoring the damaged tissue and ensuring its functionality. However, DPSCs and PDLSCs are difficult to isolate from patients with pulpitis or periodontitis patients, and cannot be replaced by each other or by other cell sources due to their unique differentiation characters (Miran, Mitsiadis, & Pagella, 2016). Therefore, we are less likely to find suitable autologous cell sources for tissue regeneration, let alone have the opportunity to isolate and expand DPSCs and PDLSCs for autotransplantation in patients who have lost their pulp or periodontal ligament tissue. Understanding out how DPSCs and PDLSCs are different and seeking a mutual replacement method for these two types of cells would benefit their clinical application. Even though there has been intensive in vitro osteogenic induction researches on DPSCs (Wei et al., 2008) and PDLSCs, either collectively or separately, the difference in the underlying mechanisms between these two types of cells remains poorly understood. Understanding the exact differences in proteomic profiling and relevant pathways would help to elucidates their osteogenic mechanisms. However, to the best of our knowledge, only one comparative proteomic study regarding human DPSCs and PDLSCs has been reported (Eleuterio et al., 2013), and that study was conducted under a standard culture condition using a two-dimensional electrophoresis approach. Thus, this study aimed to compare the differentially expressed proteins in DPSCs and PDLSCs, along with the underlying mechanisms, under in vitro osteogenic induction using an iTRAQ proteomic approach and bioinformatics analysis.

2.2. Protein preparation Samples of DPSCs and PDLSCs were cultured in one of the most widely adopted recipes for osteogenic induction for 14 days until protein preparation: complete medium supplemented with 10 mM/L βglycerol phosphate, 50 mg/mL ascorbic acid, and 10−7 M dexamethasone. For protein digestion, cells were centrifuged at 1,200g and washed in ice-cold phosphate buffered saline three times. Subsequently, frozen cell pellets were resuspended in 8 M urea, 6 M Gua, 6 M urea/ 2 M Gua, 1% RG, 2% sodium deoxycholate, or 0.5% RG/0.5% sodium deoxycholate (all in 100 mM NH4HCO3) in the presence of 5 mM Tris (2-carboxyethyl) phosphine in triplicate experiments. Following sonication, all urea-containing samples were incubated for 1 h at 37 °C, whereas all other samples were incubated at 60 °C for 30 min. After the extraction and reducing steps, all samples were incubated with 10 mM iodoacetamide at 25 °C for 30 min. Following bicinchoninic acid measurement, the protein samples were quenched using 20 mM N-acetylcysteine, and 50 μg of total protein was used for protein digestion. After diluting the chaotropic salt concentration to 6 M and diluting sodium deoxycholate to 1% using 100 mM NH4HCO3, 0.5 μg LysC was added to the protein samples. LysC was allowed to cleave for 4 h at 37 °C, followed by overnight digestion at 37 °C using 1 μg of trypsin. For the second digestion step, the LysC-digested samples were further diluted to a chaotropic salt concentration of 1.6 M. Before liquid chromatography–mass spectrometry analysis, sodium deoxycholate and RG were precipitated using 1% trifluoroacetic acid, and all samples were desalted using C18 microspin columns (Harvard Apparatus, USA) according to the manufacturer’s instructions. For digestion, we followed the previous published method of Wisniewski et al. (Wisniewski, Zougman, Nagaraj, & Mann, 2009). For the filter aided sample preparation experiments, cells were reconstituted in 5% sodium deoxycholate and 4% sodium dodecyl sulfate, respectively, sonicated and incubated for 15 min at 95 °C in the presence of 5 mM Tris (2-carboxyethyl) phosphine. Following an additional sonication step, the samples were enabled to cool and further incubated with 10 mM iodoacetamide at 25 °C for 30 min. A total of 50 μg of cleared protein isolate was later transferred into spin filters (Microcon YM-30, Millipore, USA). Sodium dodecyl sulfate – containing samples were mixed with 8 M urea and centrifuged for 15 min at 14,000g. Two additional urea wash/centrifugation cycles were performed, and tandem digest was performed as described earlier. Upon protein digestion, the filter was rinsed twice with 0.1% trifluoroacetic acid, and flow throughs were collected in the same tube. For sodium deoxycholate – filter aided sample preparation, the protein isolate was washed once with 1% sodium deoxycholate prior to tandem digest. Postdigest sodium deoxycholate was precipitated using 1% trifluoroacetic acid.

2. Materials and methods 2.1. Isolation and culture of human DPSCs and PDLSCs Human dental pulp and periodontal ligament tissue from healthy and intact tooth samples were isolated and cultured as described previously (Ma et al., 2012; Qu et al., 2016). Briefly, four premolars were collected from a single male donor (12 years of age), who was undergoing tooth extraction due to orthodontic treatment in the Oral and Maxillofacial Surgery Department, Nanfang Hospital, Guangzhou, China. Informed consent was obtained from the patient, and this subject was approved by the Ethics Committee of Nanfang Hospital. Tooth samples were rinsed three times with phosphate buffered saline containing 100 units/mL penicillin and 100 mg/mL streptomycin. Dental periodontal ligament tissue was separated from the surface of the tooth and dental pulp tissue was isolated from the pulp chamber. Each tissue was minced into pieces and then later digested in 3 mg/mL collagenase type I (SigmaAldrich, USA) and 4 mg/mL dispase (Sigma-Aldrich, USA) at 37 °C for 1 h as described previously (Cui et al., 2014). Single-cell suspensions were obtained by passing the cells through a 70 mm strainer (Carrigtwohill Co, Ireland). Single-cell suspensions were seeded into six-well plates at a density of 104 cells/well with Dulbecco’s Modified Eagle’s Medium supplemented with 1) 15% fetal calf serum (Gibco-BRL, Grand Island, NY); 2) 100 mmol/L ascorbic acid 2-phosphate; 3) 2 mmol/L glutamine; 3) 100 U/mL penicillin; and 4) 100 mg/mL streptomycin, at 37 °C in 5% carbon dioxide. The medium was changed every 2 days, and a limited dilution technique

2.3. iTRAQ labeling of peptides The supernatant containing precisely 100 μg protein of each sample was digested with Trypsin Gold (Promega, USA) at 37 °C for 16 h. After trypsin digestion, peptides were dried by vacuum centrifugation. Desalted peptides were labeled with iTRAQ reagents (sigma, iTRAQ@ Reagent-8PLEX Multiplex Kit, 4381663) according to the manufacturer's instructions (AB Sciex, CA). For 100 μg peptide 1units of labeling 10

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

reagent were used. Peptides were dissolved in 20 μL of 0.5 M triethylammonium bicarbonate pH 8.5 solution and labeling reagent was added in 70 μL of isopropanol. After 1 h incubation the reaction was stopped with 50 mM Tris/HCl pH 7.5. The labeled peptides were incubated for 2 h at room temperature. Differentially labeled peptides were mixed and subsequently desalted on 100 mg strong cation exchange (strata-x-c, Phenomenex; 8B-S029-EBJ) columns.

organism. GO has 3 ontologies which can describe molecular function, cellular component, and biological process, respectively. KEGG is a collection of manually drawn pathway maps which represents our current knowledge on the molecular interaction and reaction networks. 2.8. Alizarin red staining and alkaline phosphatase (ALP) activity staining DPSCs and PDLSCs were seeded in 500 μL of complete culture medium in 24-well plates and cultured to 70% confluence. Differentiation was induced by culturing cells in complete medium supplemented with 10 mM β-glycerol phosphate, 50 μg/mL ascorbic acid, and 10−7 mol/L dexamethasone for 3 weeks for alizarin red staining and 10 days for alkaline phosphatase activity staining. The induced cells were fixed in 4% paraformaldehyde for 20 min at room temperature and then stained with 2% alizarin red S or using a BCIP/ NBT Alkaline Phosphatase Color Development Kit (Beyotime, China) according to the manufacturer’s instructions, and the assay was repeated 3 times.

2.4. High performance liquid chromatography (HPLC) fractionation A 700 μg iTRAQ-labeled peptide mix was fractinated using a Durashell reverse phase column (5 μm, 150 Å, 250 mm × 4.6 mm i.d., Agela) on a rigol L3000 high performance liquid chromatography operating at 1 mL/min. The column oven was set as 40C. Mobile phases A (2% acetonitrile, 20 mM NH4FA, adjusted pH to 10.0 using NH3·H20) and B (98% acetonitrile, 20 mM NH4FA, adjusted pH to 10.0 using NH3ÀH2O) were used to develop a gradient elution. The solvent gradient was set as follows: 5–8% B, 5 min; 18–32% B, 30 min; 32–95% B, 10 min; 95% B, 10 min; 95–5% B, 1 min. The tryptic peptides were monitored at UV 214 nm. Eluent was collected every minute and then merge to 6 fractions. The samples were dried under vacuum and reconstituted in 20 μL of 0.1% (v/v) FA, 3% (v/v) acetonitrile in water for subsequent analyses.

2.9. Transwell migration assay The migration capacity of DPSCs and PDLSCs were evaluated using Transwell chambers (8 μm pore size, BD Falcon, USA). Following overnight serum starvation, cells were harvested and re-suspended in Dulbecco’s Modified Eagle’s Medium without fetal calf serum, then added to the upper transwell chamber in a volume of 200 μL. The migration ability of DPSCs and PDLSCs were stimulated by adding 600 μL Dulbecco’s Modified Eagle’s Medium supplemented with 20% fetal calf serum in the lower chamber. Following 24 h of culture, the cells that transversed through the membrane were fixed in methanol and stained using a Wright-Giemsa staining kit (Baso, China). Cells were counted in ten random fields visualized under a light microscope and expressed as the average number of cells per field.

2.5. Nano liquid chromatography-mass spectrum/mass spectrum (LC–MS/ MS) analysis Fractions from the first dimension reverse phase liquid chromatography were dissolved with loading buffer and were separated by a C18 column (75 μm inner-diameter, 360 μm outer-diameter × 10 cm, 1.9 μm C18, Dr. Maisch GmbH). Mobile phase A consisted of 0.1% formic acid in water solution, and mobile phase B consisted of 0.1% formic acid in acetonitrile solution. A series of adjusted 90 min gradients according to the hydrophobicity of fractions eluted in one dimension liquid chromatography with a flow rate of 300 nL/min was applied. Q exactive high performance mass spectrometer was operated in positive polarity mode with a capillary temperature of 320C. Full mass spectrum scan resolution was set to 60000 with an automatic gain control target value of 3 × 106 for a scan range of 350–1500 m/z. A data-dependent top 20 method was operated during which an higherenergy collisional dissociation spectra was obtained at 15000 mass spectrum 2 resolution with an automatic gain control target of 1 × 105 and maximum injection time of 45 ms, 1.6 m/z isolation window, and normalized collision energy of 30, dynamically excluded of 60 s.

2.10. Western blot analysis Representative data of proteomic results were validated by western blot analysis. Three pairs of DPSCs and PDLSCs were obtained from orthodontic extraction of healthy and intact premolars, from 3 donors aged 14, 19 and 21 years. Cells were cultured in osteogenic induction medium for 14 days and then cells were collected and lysed (Cell lysis buffer kit, KeyGEN, China). Cell extracts were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membrane. The membranes were blocked with 5% skim milk or 5% bovine serum albumin for 1 h and then incubated overnight at 4 °C with their respective primary antibodies, i.e. anti-COL1A2 (Collagen alpha-2(I) chain), anti-CAP-G (Macrophagecapping protein), anti-HSPB1 (Heat shock protein beta-1), antiS100A10 (Protein S100-A10), anti-POSTN (Periostin), anti-COL8A1 (Collagen alpha-1(VIII) chain), anti-HSPA5 (78 kDa glucose-regulated protein), anti-ASL (Argininosuccinate lyase), anti-S100A9 (Protein S100-A9) (1:1000, ProteinTech, USA), anti-PTGIS (Prostacyclin synthase) (1:250; Abcam, UK), or anti-GAPDH (1:1000, GNI, Japan), followed by incubation with horseradish peroxidase conjugated antirabbit/mouse IgG secondary antibody (1:5000, GNI, Japan) for 1 h at room temperature. Immunoreactive proteins were visualized using WESTAR ETAC ULTRA (Cyanagen, Italy). Each assay was repeated 3 times.

2.6. Protein identification The mass spectrum raw data files (Wiff files) from Triple-Time of Flight Mass Spectrometer 5600 were first searched by ProteinPilot version 4.2 using the Paragon search engine against the SwissProt protein database. To reduce the probability of false peptide identification, we counted only peptides at the 95% confidence interval by a ProteinPilot probability analysis greater than “identity” as identified, and each confident protein identifications was supported by at least one unique peptide. We only used ratios with p-values ≤ 0.05, and only fold changes of > 1.5 or < 0.67 were considered to be significant. 2.7. Function method description We used two databases to predict gene functions. They were GO (Gene Ontology, http://www.geneontology.org (Ashburner et al., 2000)) and KEGG (Kyoto Encyclopedia of Genes and Genomes (Kanehisa et al., 2006; Kanehisa, Goto, Kawashima, Okuno, & Hattori, 2004)). As an international standardization of gene function classification system, GO provides a set of dynamic updating controlled vocabulary to describe genes and gene products attributes in the

2.11. Statistical analysis The data of osteogenic capacity and migration assays were analyzed and expressed as the means and standard deviations. Statistical significance was evaluated by the independent samples t-test using SPSS 13.0 software. Statistical significance was set at P < 0.05. 11

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

(caption on next page)

12

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

Fig. 1. PDLSCs and DPSCs comparative proteome. (A) Differentially expressed proteins were analyzed by hierarchical clustering. Heat map indicates that expression patterns varied among different groups. Red indicates a high relative expression level, whereas blue indicates a low relative expression level. (B) Differentially expressed proteins related by fold change (FC) and P value were shown by volcano plot. P < 0.05 and FC < 0.67 or ˃2 were considered statistically significant. Red dots indicate significant up-regulated proteins, while green dots indicate significant down-regulated proteins. (C) GO molecular function (MF) classification of differentially expressed proteins of PDLSCs and DPSCs. (D) GO biological process (BP) classification of differentially expressed proteins of PDLSCs and DPSCs. (E) Bubble chart for the KEGG pathway significant enrichment analysis of differentially expressed proteins of PDLSCs versus DPSCs. The color of dots represents the P value of hypergeometric test. The larger number of −log10 (P value) stands for the greater the reliability of the test and the higher statistical significance. The scale of the dots represents the number of differentially expressed proteins in the term. x = number of identified differentially expressed proteins in this term, y = number of identified total proteins in this term.

3. Results

increase in the number of cells that crossed the membrane (25.13 ± 4.14 vs. 32.8 ± 4.02, **P ˂ 0.0005; Fig. 2C, F and G), when compared to DPSCs.

3.1. iTRAQ analysis and identification of differentially expressed proteins By the liquid chromatography-mass spectrum/mass spectrum analysis, 4461 proteins were identified from 12,872 peptides, and a common set of 3305 proteins was detected in all six samples (n = 3 for each group). In total, 159 differentially expressed proteins from PDLSCs and DPSCs were characterized according to specific criteria. Heat map analysis demonstrated a cluster of 82 proteins with higher expression levels in PDLSCs compared to that in DPSCs, while a cluster of 77 proteins had lower expression levels in PDLSCs compared to that in DPSCs (Fig. 1A). Identified differentially expressed proteins related by fold change (FC) and P value are shown by volcano plot (Fig. 1B). All the up-regulated and down-regulated proteins (PDLSCs vs. DPSCs) are presented in Table 1 and Table 2, respectively.

3.5. Validation of the proteomic results To validate the proteomic results, five higher expression and five lower expression representative proteins in PDLSCs were selected. Analysis showed that the expression levels of COL1A2, PTGIS, CAP-G, HSPB1 and S100A10 (Fig. 2H) were higher in PDLSCs than in DPSCs, while POSTN, COL8A1, HSPA5, ASL and S100A9 (Fig. 2I) were lower in PDLSCs, which is consistent with the proteomic results. 4. Discussion To clarify the molecular mechanisms of dental stem cell differentiation, a two-dimensional electrophoresis mass spectrograph-based approach was adopted for many proteomic studies (Ma et al., 2014; Patil et al., 2014). As a global analysis tool, however, this approach has many drawbacks, e.g., it is low-throughput, time-consuming and laborintensive. With comparatively higher sensitivity and reduced labor, iTRAQ proteomic is a preferable high-throughput quantitative technique in the area of dental research (Wu, Wang, Baek, & Shen, 2006). As a result, a common set of over 3000 proteins were detected in both PDLSCs and DPSCs, while only less than 5% of them were differentially expressed. This suggests PDLSCs and DPSCs share a wide range of their proteome, which may be due to their similar origins and close localizations. However, notably, these differentially expressed proteins may reflect the ultimate distinct differentiation potential and diverse functions under the respective microenvironments of these two cells. With subsequent GO and KEGG enrichment analysis, we identified possible roles and involved pathways that distinguish osteoinduced PDLSCs from DPSCs. Measuring the mineralization capacity by alizarin red staining and alkaline phosphatase activity is widely utilized in stem cells osteogenic differentiation research. Our results showed that PDLSCs possess a stronger mineralization capacity than DPSCs, which is a strong indicator of osteo-/odontogenic differentiation. Additionally, the present comparative proteomic analysis of the differentially expressed proteins between PDLSCs and DPSCs provided clues to the essence of this phenomenon. Ten differentially expressed proteins were selected and validated by western blot, which was consistent with the proteome data. For example, Prostacyclin synthase (PTGIS) and Argininosuccinate lyase (ASL) are two of the most significant differentially expressed proteins between PDLSCs and DPSCs. These enzymes are important to many metabolism and biosynthetic processes as revealed by GO and KEGG analysis (see Supplemental Tables). Among many of the differentially expressed proteins, several protein families drew our attention. Compared with the DPSC group, some of the family members were up-regulated in the PDLSC group, while several were down-regulated, such as the collagen, heat shock protein and protein S100 families. For the collagen family, COL1A1 (Collagen alpha-1(I) chain) and COL1A2 were up-regulated, while COL4A2 (Collagen alpha-2(IV) chain) and COL8A1 were down-regulated. As the main component of connective tissue, collagen is the most abundant

3.2. GeneOntology (GO) enrichment analysis Among 3488 identified proteins that fit in the GO protein spectrum, 132 identified differentially expressed proteins between PDLSCs and DPSCs were examined using GO enrichment analysis. As a result, a total of 80 statistically significant GO terms were enriched. According to the classification of GO terms, 39 enriched GO terms belonged to molecular function, 35 terms belonged to biological process, and 6 terms belonged to cellular component. Most of the enriched terms (82.5%, n = 66) lay in GO level 3 (15%, n = 12), 4 (15%, n = 12), 5 (25%, n = 20), and 6 (27.5%, n = 22). All of the enriched GO terms, along with the up-/ down-regulated proteins in the enriched terms are shown in Supplemental Table 1. The integrated molecular function (MF) and biological process (BP) classification are shown in Fig. 1C and D, respectively. 3.3. Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis Among 2656 identified proteins that fit in the KEGG protein spectrum, 98 identified differentially expressed proteins between PDLSCs and DPSCs were examined using KEGG enrichment analysis. A total of 14 KEGG pathway terms were enriched: 1) Amoebiasis, 2) Arachidonic acid metabolism, 3) PPAR signaling pathway, 4) Protein digestion and absorption, 5) Meiosis – yeast, 6) Cell cycle – yeast, 7) Fatty acid biosynthesis, 8) Adipocytokine signaling pathway, 9) Pentose and glucuronate interconversions, 10) ECM-receptor interaction, 11) Alanine, aspartate and glutamate metabolism, 12) Ascorbate and aldarate metabolism, 13) Arginine biosynthesis, and 14) Pantothenate and CoA biosynthesis (see Fig. 1E). All of the enriched KEGG terms, along with the up-/down-regulated proteins in the enriched terms are shown in Supplemental Table 2. 3.4. Capacity of differentiation and migration of DPSCs and PDLSCs Compared to DPSCs, PDLSCs formed more calcium deposition (Fig. 2A and D) and had a more prominent Alkaline phosphatase activity (Fig. 2B and E), suggesting a greater mineralization capacity of PDLSCs, which is one of the benchmarks of mesenchymal stem cell osteogenic differentiation. Additionally, PDLSCs showed about a 30% 13

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

Table 1 (continued)

Table 1 Up-regulated proteins of PDLSCs versus DPSCs. Protein#

Gene

Description

FC

P-value

Q16647 P26447 P52943 P14384 O14907 P04792 Q8IVF2 P40121 Q03169

PTGIS S100A4 CRIP2 CPM TAX1BP3 HSPB1 AHNAK2 CAPG TNFAIP2

6.55 3.69 2.89 2.53 2.43 2.43 2.43 2.36 2.16

0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Q13190 P60903 P36269 Q9H8H3 P08571 Q13740 P01034 O94832 P63313 O95298

STX5 S100A10 GGT5 METTL7A CD14 ALCAM CST3 MYO1D TMSB10 NDUFC2

2.11 2.09 2.08 2.07 2.03 2.03 2.00 1.98 1.98 1.98

0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.009

P60981 Q01469 P02452 P31949 P58546 Q9NZU5 Q5JPH6

DSTN FABP5 COL1A1 S100A11 MTPN LMCD1 EARS2

1.95 1.95 1.95 1.93 1.93 1.91 1.91

0.000 0.000 0.000 0.002 0.001 0.000 0.000

P12277 P48357 O43301 Q96D15 Q13643 P17661 Q9UNL2

CKB LEPR HSPA12A RCN3 FHL3 DES SSR3

1.89 1.89 1.88 1.86 1.86 1.85 1.84

0.000 0.000 0.000 0.000 0.001 0.002 0.000

P02511 O60701 P29373 Q9Y3D6 P51888 Q8WUJ3

CRYAB UGDH CRABP2 FIS1 PRELP CEMIP

1.84 1.83 1.83 1.82 1.79 1.78

0.000 0.000 0.000 0.000 0.000 0.000

Q01995 P00966 Q9NRM1 P04080 O15075 O75915 O95864 Q96CX2

TAGLN ASS1 ENAM CSTB DCLK1 ARL6IP5 FADS2 KCTD12

1.78 1.77 1.76 1.75 1.73 1.71 1.68 1.68

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

P04083 Q9NTJ3

ANXA1 SMC4

1.66 1.64

0.000 0.000

Q8TED1 Q86 ´ 02

GPX8 CDR2L

1.64 1.64

0.000 0.002

Q96EK6

GNPNAT1

1.63

0.000

O00192

ARVCF

1.63

0.000

Q9P299 Q9H425 Q92520 Q3MHD2 Q8NBS9

COPZ2 C1orf198 FAM3C LSM12 TXNDC5

1.63 1.63 1.62 1.61 1.61

0.000 0.000 0.000 0.000 0.000

P54687

BCAT1

1.61

0.000

P84157 P46087

MXRA7 NOP2

Prostacyclin synthase Protein S100-A4 Cysteine-rich protein 2 Carboxypeptidase M Tax1-binding protein 3 Heat shock protein beta-1 Protein AHNAK2 Macrophage-capping protein Tumor necrosis factor alpha-induced protein 2 Syntaxin-5 Protein S100-A10 Gamma-glutamyltransferase 5 Methyltransferase-like protein 7A Monocyte differentiation antigen CD14 CD166 antigen Cystatin-C Unconventional myosin-Id Thymosin beta-10 NADH dehydrogenase [ubiquinone] 1 subunit C2 Destrin Fatty acid-binding protein, epidermal Collagen alpha-1(I) chain Protein S100-A11 Myotrophin LIM and cysteine-rich domains protein 1 Probable glutamate–tRNA ligase, mitochondrial Creatine kinase B-type Leptin receptor Heat shock 70 kDa protein 12A Reticulocalbin-3 Four and a half LIM domains protein 3 Desmin Translocon-associated protein subunit gamma Alpha-crystallin B chain UDP-glucose 6-dehydrogenase Cellular retinoic acid-binding protein 2 Mitochondrial fission 1 protein Prolargin Cell migration-inducing and hyaluronan-binding protein Transgelin Argininosuccinate synthase Enamelin Cystatin-B Serine/threonine-protein kinase DCLK1 PRA1 family protein 3 Fatty acid desaturase 2 BTB/POZ domain-containing protein KCTD12 Annexin A1 Structural maintenance of chromosomes protein 4 Probable glutathione peroxidase 8 Cerebellar degeneration-related protein 2-like Glucosamine 6-phosphate Nacetyltransferase Armadillo repeat protein deleted in velo-cardio-facial syndrome Coatomer subunit zeta-2 Uncharacterized protein C1orf198 Protein FAM3C Protein LSM12 homolog Thioredoxin domain-containing protein 5 Branched-chain-amino-acid aminotransferase, cytosolic Matrix-remodeling-associated protein 7 Probable 28S rRNA (cytosine(4447)C(5))-methyltransferase

1.61 1.60

0.000 0.000

Protein#

Gene

Description

FC

P-value

P09936

UCHL1

1.59

0.000

Q6UVK1 Q12882

CSPG4 DPYD

1.59 1.59

0.000 0.000

P33992 Q16658 Q86WV6 P46821 Q7Z3B1 Q8N8S7 P04179

MCM5 FSCN1 TMEM173 MAP1B NEGR1 ENAH SOD2

1.57 1.57 1.57 1.56 1.55 1.55 1.55

0.000 0.000 0.000 0.000 0.000 0.000 0.000

P08123 Q8IXL7 Q06278 P28845

COL1A2 MSRB3 AOX1 HSD11B1

1.55 1.55 1.54 1.54

0.001 0.000 0.000 0.000

Q13162 P33991 P15144 Q8TDX7 Q9BWS9 P17931 Q9BVC6

PRDX4 MCM4 ANPEP NEK7 CHID1 LGALS3 TMEM109

Ubiquitin carboxyl-terminal hydrolase isozyme L1 Chondroitin sulfate proteoglycan 4 Dihydropyrimidine dehydrogenase [NADP(+)] DNA replication licensing factor MCM5 Fascin Stimulator of interferon genes protein Microtubule-associated protein 1B Neuronal growth regulator 1 Protein enabled homolog Superoxide dismutase [Mn], mitochondrial Collagen alpha-2(I) chain Methionine-R-sulfoxide reductase B3 Aldehyde oxidase Corticosteroid 11-beta-dehydrogenase isozyme 1 Peroxiredoxin-4 DNA replication licensing factor MCM4 Aminopeptidase N Serine/threonine-protein kinase Nek7 Chitinase domain-containing protein 1 Galectin-3 Transmembrane protein 109

1.54 1.53 1.53 1.52 1.52 1.51 1.50

0.000 0.001 0.000 0.000 0.000 0.000 0.000

Notes: #Protein codes from UniProt database. (www.uniprot.org), FC = Fold change.

protein in mammals (Di Lullo, Sweeney, Korkko, Ala-Kokko, & San, 2002) and abundant in cartilage, bones, blood vessels, and dentin. In the human body, over 90% of the collagen is type I, which is encoded by the COL1A1and COL1A2 genes (Retief, Parker, & Retief, 1985). Their products, two fibrillar pro-alpha1(I) chain and one pro-alpha2(I) chain, make a molecule of type I procollagen. The periodontal ligament is known to be exposed to mechanical forces in response to mastication, speech and orthodontic tooth movement (Beertsen, McCulloch, & Sodek, 1997; Zhang, Du, Zhou, & Yu, 2014). Type I collagen expression is controlled by mechanical forces (Kook, Jang, & Lee, 2011). Thus, higher expression levels of COL1A1 and COL1A2 in PDLSCs, to some degree, may be explained by exposure to mechanical tension in the periodontal ligament. COL1A1 was reported to contribute to in vivo odontogenesis of dental pulp (Braut. Kollar, & Mina, 2003) and in vitro mineralization of dental pulp cells (Balic, Rodgers, & Mina, 2009). COL8A1 (Muragaki, Mattei, Yamaguchi, Olsen, & Ninomiya, 1991) and COL4A2 (Griffin, Emanuel, Hansen, Cavenee, & Myers, 1987) are part of type IV collagen and type VIII collagen, respectively. As non-fibrillar collagens, type IV collagen forms the basal lamina and type VIII collagen is a short chain collagen, both of which are related to the basement membrane. It is not surprising that PDLSCs and DPSCs both express abundant collagens because they are similar to fibroblasts, which are the most common cells that create collagen. This cell-specific deposition of different types of collagen may contribute to tissue-specific regeneration and damage repair of dental pulp and the periodontal ligament. For the heat shock protein family (HSPs), HSPB1 and HSPA12A (Heat shock 70 kDa protein 12A) were up-regulated in PDLSCs while HSPA2 (Heat shock-related 70 kDa protein 2) and HSPA5 were downregulated. HSPs are a family of proteins which are produced by cells responding to the stress condition exposure. They are expressed during stresses such as during heat shock, wound healing or tissue remodeling (Laplante et al., 1998). Many members of this family perform chaperone function either by helping to refold proteins that were damaged by cell stress, or by stabilizing new proteins to ensure correct folding (De Maio, 1999). Previous research (Eleuterio et al., 2013) also found that HSPs were abundantly expressed in both PDLSCs and DPSCs, and the authors suggested that HSPs could be crucial for stem cell biology during cellular homeostasis and development. HSPB1 is a chaperone of 14

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

Table 2 Down-regulated proteins of PDLSCs versus DPSCs. Protein#

Gene

Description

FC

O94875 P04424 Q15582 Q9BYD1 P23219 P29317 O60488 P11166 P07305 P49961 Q16719 Q9Y4F1 P13473 Q13043 Q9H4 × 1 P08572 Q15293 Q9H1A4 P15121 P47895 P51911 P50583 P54652 P21980 Q9UKG9 O15427 Q06033 P18887 Q15063 Q15526 Q9UDR5 P0C0L4 P35749 P07942 Q9HBG6 Q16537 Q9HAU0 Q13620 P19823 Q8N2K0 P01023 Q14657 Q02127 Q16527 O14773 Q6PCE3 O15439 Q9BV38 Q9Y613 P33121 O76041 Q15031 P07093 Q9H9J2 P60174 P27658 Q02952 P23381 Q5TFE4 Q96B97 Q6UN15 P42229 Q9Y2V2 Q9H936 Q96QG7 P35237 Q9NYI0 Q7Z460 Q9HAT2 P11021 Q8WVV9 P05091 Q15758 Q9P0M6

SORBS2 ASL TGFBI MRPL13 PTGS1 EPHA2 ACSL4 SLC2A1 H1F0 ENTPD1 KYNU FARP1 LAMP2 STK4 RGCC COL4A2 RCN1 ANAPC1 AKR1B1 ALDH1A3 CNN1 NUDT2 HSPA2 TGM2 CROT SLC16A3 ITIH3 XRCC1 POSTN SURF1 AASS C4A MYH11 LAMB1 IFT122 PPP2R5E PLEKHA5 CUL4B ITIH2 ABHD12 A2M LAGE3 DHODH CSRP2 TPP1 PGM2L1 ABCC4 WDR18 FHOD1 ACSL1 NEBL LARS2 SERPINE2 MRPL44 TPI1 COL8A1 AKAP12 WARS NT5DC1 SH3KBP1 FIP1L1 STAT5A CARHSP1 SLC25A22 MTMR9 SERPINB6 PSD3 CLASP1 SIAE HSPA5 HNRNPLL ALDH2 SLC1A5 H2AFY2

Sorbin and SH3 domain-containing protein 2 Argininosuccinate lyase Transforming growth factor-beta-induced protein ig-h3 39S ribosomal protein L13, mitochondrial Prostaglandin G/H synthase 1 Ephrin type-A receptor 2 Long-chain-fatty-acid–CoA ligase 4 Solute carrier family 2, facilitated glucose transporter member 1 Histone H1.0 Ectonucleoside triphosphate diphosphohydrolase 1 Kynureninase FERM, RhoGEF and pleckstrin domain-containing protein 1 Lysosome-associated membrane glycoprotein 2 Serine/threonine-protein kinase 4 Regulator of cell cycle RGCC Collagen alpha-2(IV) chain Reticulocalbin-1 Anaphase-promoting complex subunit 1 Aldose reductase Aldehyde dehydrogenase family 1 member A3 Calponin-1 Bis(5′-nucleosyl)-tetraphosphatase [asymmetrical] Heat shock-related 70 kDa protein 2 Protein-glutamine gamma-glutamyltransferase 2 Peroxisomal carnitine O-octanoyltransferase Monocarboxylate transporter 4 Inter-alpha-trypsin inhibitor heavy chain H3 DNA repair protein XRCC1 Periostin Surfeit locus protein 1 Alpha-aminoadipic semialdehyde synthase, mitochondrial Complement C4-A Myosin-11 Laminin subunit beta-1 Intraflagellar transport protein 122 homolog Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit epsilon isoform Pleckstrin homology domain-containing family A member 5 Cullin-4B Inter-alpha-trypsin inhibitor heavy chain H2 Monoacylglycerol lipase ABHD12 Alpha-2-macroglobulin EKC/KEOPS complex subunit LAGE3 Dihydroorotate dehydrogenase (quinone), mitochondrial Cysteine and glycine-rich protein 2 Tripeptidyl-peptidase 1 Glucose 1,6-bisphosphate synthase Multidrug resistance-associated protein 4 WD repeat-containing protein 18 FH1/FH2 domain-containing protein 1 Long-chain-fatty-acid–CoA ligase 1 Nebulette Probable leucine–tRNA ligase, mitochondrial Glia-derived nexin 39S ribosomal protein L44, mitochondrial Triosephosphate isomerase Collagen alpha-1(VIII) chain A-kinase anchor protein 12 Tryptophan–tRNA ligase, cytoplasmic 5′-nucleotidase domain-containing protein 1 SH3 domain-containing kinase-binding protein 1 Pre-mRNA 3′-end-processing factor FIP1 Signal transducer and activator of transcription 5A Calcium-regulated heat-stable protein 1 Mitochondrial glutamate carrier 1 Myotubularin-related protein 9 Serpin B6 PH and SEC7 domain-containing protein 3 CLIP-associating protein 1 Sialate O-acetylesterase 78 kDa glucose-regulated protein Heterogeneous nuclear ribonucleoprotein L-like Aldehyde dehydrogenase, mitochondrial Neutral amino acid transporter B(0) Core histone macro-H2A.2

0.30 0.32 0.35 0.36 0.36 0.37 0.38 0.38 0.39 0.42 0.42 0.43 0.46 0.46 0.47 0.49 0.49 0.49 0.49 0.49 0.53 0.53 0.53 0.55 0.55 0.56 0.56 0.57 0.58 0.58 0.58 0.59 0.59 0.59 0.59 0.60 0.60 0.61 0.61 0.61 0.61 0.62 0.62 0.62 0.62 0.62 0.63 0.63 0.63 0.63 0.63 0.63 0.64 0.64 0.64 0.64 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.65 0.66 0.66 0.66 0.66 0.66 0.66 0.66

15

P-value 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.003 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 (continued on next page)

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

Table 2 (continued) Protein#

Gene

Description

FC

P-value

Q8N335 P30520 P06702

GPD1L ADSS S100A9

Glycerol-3-phosphate dehydrogenase 1-like protein Adenylosuccinate synthetase isozyme 2 Protein S100-A9

0.66 0.66 0.67

0.000 0.000 0.000

Notes: #Protein codes from UniProt database. (www.uniprot.org), FC = Fold change.

mineralization capacity of PDLSCs versus DPSCs. Unlike HSPB1, the possible roles of other identified proteins of the HSPs seem more difficult to explain with limited information. HSPA12A (Han, Truong, Park, & Breslow, 2003) and HSPA2 are two members of the Hsp70 family. The Hsp70 system interacts with extended peptide segments of proteins and also with partially folded proteins to prevent remodel folding pathways, aggregation, and regulate activity (Mashaghi et al., 2016) when not interacting with a substrate peptide. Similarly, HSPA5 (Hendershot, Valentine, Lee, Morris, & Shapiro, 1994) takes part in protein folding and holding, endoplasmic reticulum translocation and endoplasmic reticulum-associated degradation. The reason that HSPA12A, HSPA2, and HSPA5 were differently expressed in PDLSCs and DPSCs needs to be further studied. For S100 protein family, S100A4 (Protein S100-A4), S100A10 and

the small heat shock protein group. The common functions of small heat shock protein include chaperone activity, inhibition of apoptosis, regulation of cell development, cell differentiation, and signal transduction. HSPB1 interacts with actin and intermediate filaments by preventing the formation of non-covalent filament/filament interactions of the intermediate filaments and protecting actin filaments from fragmentation. It also preserves the focal contacts fixed at the cell membrane (Vargas-Roig et al., 1997). Among other chaperones, HSPB1 is involved in the process of cell differentiation (Arrigo, 2005). Taken together with the fact that cytoskeletal binding proteins distinguish cultured dental follicle cells and periodontal ligament cells (Li et al., 2016), we inferred that HSPB1may distinguish cultured osteoinduced PDLSCs and DPSCs by interacting with actin and intermediate filaments, and a higher level of HSPB1 may contribute to the higher

Fig. 2. Comparison of differentiation and migration capacities between PDLSCs and DPSCs. (A, B, D, E) Alizarin red staining and ALP activity staining of PDLSCs and DPSCs. (C, F) The migration capacity of cells was evaluated by transwell assay. The number of PDLSCs group that migrated through the membrane was significantly increased than the cells in the DPSCs group. (**P = 0.000). (G) One representative statistical result of transwell assay. (H) Representative western blot results of higher expression level proteins in osteoinduced PDLSCs than DPSCs. COL1A2 = Collagen alpha-2(I) chain, PTGIS = Prostacyclin synthase, CAPG = Macrophage-capping protein, HSPB1 = Heat shock protein beta-1, S100A10 = Protein S100-A10. (I) Representative western blot results of lower expression level proteins in osteoinduced PDLSCs than DPSCs. POSTN = Periostin, COL8A1 = Collagen alpha-1(VIII) chain, HSPA5 = 78 kDa glucose-regulated protein, ASL = Argininosuccinate lyase, S100A9 = Protein S100-A9, GAPDH served as internal reference.

16

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

of the enriched terms belonged to metabolic processes and response to stimulus. On one hand, as a significant change event of living stem cells, differentiation inevitably causes changes in many basic and vital metabolic processes. The substances involved in these enriched terms ranged from simple and small molecular substances, such as glucose and urea, to complex, large molecular substances, such as protein and DNA, suggesting that PDLSCs and DPSCs may undergo distinct metabolic changes during the differentiation process. On the other hand, the abundant enriched items related to response to stimulus hinted that the differentiation induction environment could also act as a stress condition, causing a series of relevant reactions. Among the above enriched items, many of them were related to the cytoskeleton. There were more up-regulated actin- and cytoskeletonrelated proteins in the PDLSC group than in the DPSC group, suggesting a more active cytoskeleton transformation of osteoinduced PDLSCs. Another iTRAQ proteomics study showed that periodontal ligament cells displayed enhanced actin cytoskeletal dynamics relative to dental follicle cells (Li et al., 2016). As mentioned above, the periodontal ligament is known to be exposed to mechanical forces and the mechanical forces may lead to changes in cell shape and cytoskeletal structure. During the differentiation process, the cytoskeleton of mesenchymal stem cells undergoes distinct changes (Rodriguez, Gonzalez, Rios, & Cambiazo, 2004). Again, the enhanced actin cytoskeletal dynamics of PDLSCs may be partially explained by exposure to mechanical tension in the periodontal ligament, and this may contribute to the higher migration capacity of PDLSCs. CAPG, also known as actin regulatory protein CAP-G (Dabiri, Young, Rosenbloom, & Southwick, 1992), contributes to the control of actin-based motility in non-muscle cells. In the present study, CAPG had a higher expression level in PDLSCs than in DPSCs while a contrary result was found in a previous study (Eleuterio et al., 2013). This may be due to either different culture conditions, i.e., the first was under an osteoinduced environment while the second was a non-osteoinduced environment, or to the heterogeneities of PDLSCs and DPSCs used in these two studies. Furthermore, there are many other questions waiting to be resolved, such as why Thymosin beta-10 (TMSB10) (Shiotsuka et al., 2013), which inhibits actin polymerization, and Destrin (DSTN) (Hatanaka et al., 1996), which takes part in actin depolymerizing, all had a relatively higher expression level in PDLSCs. Taken together, the exact functions and involved mechanisms of the cytoskeleton in the differentiation process of PDLSCs and DPSCs needs to be further studied, which would shed new light on our comprehensive understanding of stem cell differentiation. A previous study (Eleuterio et al., 2013) had performed a two-dimensional electrophoresis proteomic study of human PDLSCs and DPSCs under a standard culture condition. Compared to the current study, it is interesting to note similarities and differences between the two. In the previous study, 113 proteins were identified, including Heat shock protein family, ubiquitin C-terminal hydrolase (USP); differentiation markers versus myocyte and hepatic phenotype such as Actin related protein 3 (ACTR3), Vinculin; cytoskeleton proteins such as class III intermediate filament, Laminin and Myosin. Most of these were identified in our study as well, although only few of them were found differentially expressed between PDLSCs and DPSCs. This suggests that PDLSCs and DPSCs share a wide range of their proteome under osteogenic and non-osteogenic conditions. However, the identified proteins involved in the cell cycle, stress response, homing and neurogenesis in that study, including Putative ATP-dependent Clp protease proteolytic subunit (CLPP), NAD(P)H dehydrogenase (quinone) 1 (NQO1), Succinyl-CoA:3-ketoacid coenzyme A transferase 1 (SCOT1), N(G),N(G)-dimethylarginine dimethylaminohydrolase 1 (DDAH1) and an additional new isoform of tubulin (TBB5) were not identified in our study. Further, the proteins exclusively expressed in PDLSCs found in their study, i.e., Galectin 1 (LEG1), Adaptin ear-binding coat-associated protein 2 (NECP2), Aldo-keto reductase family 1 member C1(AK1C1), Septin11 (SEP11), and Annexin 10

S100A11 (Protein S100-A11) were more highly expressed in PDLSCs, while S100A9 was more highly expressed in DPSCs. The S100 protein family is a group of low-molecular-weight proteins that are characterized by two calcium-binding sites with a helix-loop-helix conformation (Marenholz, Heizmann, & Fritz, 2004). Considered to be damage-associated molecular pattern molecules, S100 proteins localize in the cytoplasm and/or nucleus of various cells and take part in regulating many cellular processes, such as differentiation and cell cycle progression. S100A10 facilitates Ca2+ uptake, nociception, and cell polarization. Complexed with the annexin II, S100A10 binds channel proteins and receptors and guides them to the cell surface, causing increased membrane localization and subsequent magnified functional expression of these proteins (Rescher & Gerke, 2008). S100A11 may function in motility, tubulin polymerization, and apoptosis (Kanamori et al., 2004). Thus, it is reasonable to suggest that S100A10 and S100A11 may be involved in the differentiation process of PDLSCs and DPSCs. As a significant trait of in vitro osteoinduced dental stem cells, calcium deposition needs Ca2+ uptake. Additionally, cell motility, such as migration (Xu, Cui, Ma, Sun, & Wu, 2015), is essential for tissue repair and regeneration as dental stem cells. These could partially contribute to the higher mineralization capacity of PDLSCs which was consistent with a similar study (Eleuterio et al., 2013), and the higher migration capacity of PDLSCs, which were both verified in the present study. Unlike S100A10 and S100A11, however, S100A9 may function in a more complex way in PDLSCs and DPSCs. S100A9 and S100A8 are two members of the S100 family of calcium-modulated proteins. S100A9 complexes with S100A8 and thus regulates myeloid cell function by binding to Toll-like receptor 4 (Vogl et al., 2007) and the receptor for advanced glycation end products (Boyd, Kan, Roberts, Wang, & Walley, 2008). Toll-like receptor 4 is well known for recognizing lipopolysaccharide and acts as a trigger for the NF-kB intracellular signaling pathway and inflammatory cytokine production (Vaure & Liu, 2014), both in immune cells and DPSCs (He et al., 2014). Advanced glycation end products are implicated in calcification of dental pulp cells (Nakajima, Inagaki, Hiroshima, Kido, & Nagata, 2013) and dental pulp inflammation (Tancharoen et al., 2014). Therefore, S100A9 may play a part in the inflammatory reaction and calcification of PDLSCs and DPSCs. Interestingly, we observed that periostin (POSTN) was comparatively lower in PDLSCs than in DPSCs. As a matricellular protein, Periostin is strongly expressed in collagen-rich connective tissues, such as the periodontal ligament. Periostin mediates and augments collagen fibrillogenesis (Norris et al., 2007), cell migration, adhesion (Horiuchi et al., 1999), the response to mechanical stress, and wound healing (Jackson-Boeters, Wen, & Hamilton, 2009; Padial-Molina et al., 2013; Padial-Molina, Volk, & Rios, 2014). It has been reported to be important in regulating periodontal disease pathogenesis and repair (Rios et al., 2005). Periostin may play different roles in the differentiation processes of PDLSCs and DPSCs. Periostin was reported to act as a negative regulator of mineralization in the dental pulp tissue via Notch signals. Furthermore, the expression of Periostin in DPSCs can be regulated by TGF-β1 and biomechanical stimulation (Wiesen et al., 2015). By coincidence, Transforming growth factor-beta-induced protein ig-h3 (TGFBI), which is induced by TGF-β, was also detected in this study and had a higher expression level in the DPSC group. Thus, the higher expression level of Periostin in DPSCs may be induced by TGF-β, a possibility which needs to be further clarified. To categorize the possible involved functions of identified differentially expressed genes in osteoinduced PDLSCs and DPSCs, GO enrichment analysis and annotation were performed. For the classification of molecular function, 75% of the enriched terms belonged to catalytic activity, protein binding and negative regulation of protein metabolic processes, indicating active protein interactions, as well as different biosynthetic and regulatory processes of PDLSCs and DPSCs towards differentiation. For the classification of biological processes, two thirds 17

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

observed that the mineralization and migration capacities of PDLSCs were greater than those of DPSCs, which is partially consistent with the proteomic results. Though further verification is needed, the present study provides detailed comparative proteomic data on osteoinduced PDLSCs and DPSCs which would fuel subsequent researches in this area.

(ANX10), were also not identified in our study. These differences may be mainly due to osteo-/odontogenic differentiation of PDLSCs and DPSCs. Dental stem cells are known as possessing multilineage potential (Patil et al., 2014), and the proteomic changes found in Eleuterio et al. (Eleuterio et al., 2013) and the present study may reflects commitment to osteogenic differentiation of these two cells. KEGG pathway analysis revealed several involved pathways distinguishing the osteogenic differentiation process of PDLSCs and DPSCs. Notably, two identified differentially expressed proteins, i.e., longchain-fatty-acid-CoA ligase 4 and long-chain-fatty-acid-CoA ligase 1, related to the fatty acid biosynthesis pathway, all had a higher expression level in DPSCs when compared with that in PDLSCs. While branched-chain-amino-acid aminotransferase and dihydropyrimidine dehydrogenase [NADP(+)], which are the only two identified differentially expressed proteins related to the pantothenate and CoA biosynthesis pathway, all had a higher expression level in PDLSCs. This indicated that the fatty acid biosynthesis pathway may be more prominent in DPSCs while the pantothenate and CoA biosynthesis pathway may be more significant in PDLSCs. Among the other 12 enriched KEGG pathways, the arachidonic acid metabolism pathway and PPAR signaling pathway greatly drew our attention. In mammalian cells, lipid-rich organelles, including lipid bodies and lipid droplets, are sites for biosynthesis of arachidonic acidderived inflammatory mediators, i.e., eicosanoids (Toledo et al., 2016). Thus, the arachidonic acid metabolism pathway may act as a potential inflammation modulator of the osteogenic differentiation process of PDLSCs and DPSCs. PPARs are in control of modulating the expression of genes involved in lipid metabolism, adipogenesis, energy homeostasis, cell proliferation, apoptosis, inflammation, immune tolerance, and inducing/anticancer effects (Fanale, Amodeo, & Caruso, 2017). Still, the possible role of PPARs in PDLSCs and DPSCs needs to be further investigated since there are many possible roles the PPAR signaling pathway may play in the differentiation process. As a one-sample-pooling proteomics study, the PDLSCs and DPSCs in the present study were obtained from a single donor; thus, the possible impact of biological variation was reduced. Considering the heterogeneity of different donors and their isolated stem cells, this onesample-pooling strategy helped to minimize the variation in proteomic studies. However, due to limited samples, this strategy could also lead to false negative and false positive proteins, though some selected proteins were validated. Therefore, in order to reach a higher repeatability, enlarging the pooling samples from different donors seems more effective.

Conflicts of interest None. Acknowledgements This work was supported by the National Natural Science Foundation of China (Grant No. 81670986 and 81400495, D.D.M); Guangdong Natural Science Foundation (2015A03030101, D.D.M). We thank Yilin Hao, Yihong Ge, Ci Song and Jianjia Li for their contribution in cell culture in this work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.archoralbio.2018.01.015. References Arrigo, A. P. (2005). In search of the molecular mechanism by which small stress proteins counteract apoptosis during cellular differentiation. Journal of Cellular Biochemistry, 94, 241–246. Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. M., et al. (2000). Gene ontology: Tool for the unification of biology: The Gene Ontology Consortium. Nature Genetics, 25, 25–29. Balic, A., Rodgers, B., & Mina, M. (2009). Mineralization and expression of Col1a13.6GFP transgene in primary dental pulp culture. Cells Tissues Organs, 189, 163–168. Beertsen, W., McCulloch, C. A., & Sodek, J. (1997). The periodontal ligament: A unique, multifunctional connective tissue. Periodontology 2000, 13, 20–40. Boyd, J. H., Kan, B., Roberts, H., Wang, Y., & Walley, K. R. (2008). S100A8 and S100A9 mediate endotoxin-induced cardiomyocyte dysfunction via the receptor for advanced glycation end products. Circulation Research, 102, 1239–1246. Braut, A., Kollar, E. J., & Mina, M. (2003). Analysis of the odontogenic and osteogenic potentials of dental pulp in vivo using a Col1a1-2.3-GFP transgene. International Journal of Developmental Biology, 47, 281–292. Cui, L., Xu, S., Ma, D., Gao, J., Liu, Y., Yue, J., et al. (2014). The role of integrin-alpha5 in the proliferation and odontogenic differentiation of human dental pulp stem cells. Journal of Endodontics, 40, 235–240. Dabiri, G. A., Young, C. L., Rosenbloom, J., & Southwick, F. S. (1992). Molecular cloning of human macrophage capping protein cDNA: A unique member of the gelsolin/villin family expressed primarily in macrophages. Journal of Biological Chemistry, 267, 16545–16552. De Maio, A. (1999). Heat shock proteins: Facts, thoughts and dreams. Shock, 11, 1–12. Di Lullo, G. A., Sweeney, S. M., Korkko, J., Ala-Kokko, L., & San, A. J. (2002). Mapping the ligand-binding sites and disease-associated mutations on the most abundant protein in the human, type I collagen. Journal of Biological Chemistry, 277, 4223–4231. Eleuterio, E., Trubiani, O., Sulpizio, M., Di Giuseppe, F., Pierdomenico, L., Marchisio, M., et al. (2013). Proteome of human stem cells from periodontal ligament and dental pulp. PLoS One, 8, e71101. Fanale, D., Amodeo, V., & Caruso, S. (2017). The interplay between metabolism PPAR signaling pathway, and cancer. PPAR Research, 2017, 1830626. Griffin, C. A., Emanuel, B. S., Hansen, J. R., Cavenee, W. K., & Myers, J. C. (1987). Human collagen genes encoding basement membrane alpha 1 (IV) and alpha 2 (IV) chains map to the distal long arm of chromosome 13. Proceedings of the National Academy of Sciences of the United States of America, 84, 512–516. Gronthos, S., Mankani, M., Brahim, J., Robey, P. G., & Shi, S. (2000). Postnatal human dental pulp stem cells (DPSCs) in vitro and in vivo. Proceedings of the National Academy of Sciences of the United States of America, 97, 13625–13630. Han, Z., Truong, Q. A., Park, S., & Breslow, J. L. (2003). Two Hsp70 family members expressed in atherosclerotic lesions. Proceedings of the National Academy of Sciences of the United States of America, 100, 1256–1261. Hatanaka, H., Ogura, K., Moriyama, K., Ichikawa, S., Yahara, I., & Inagaki, F. (1996). Tertiary structure of destrin and structural similarity between two actin-regulating protein families. Cell, 85, 1047–1055. He, W., Wang, Z., Zhou, Z., Zhang, Y., Zhu, Q., Wei, K., et al. (2014). Lipopolysaccharide enhances Wnt5a expression through toll-like receptor 4, myeloid differentiating factor 88, phosphatidylinositol 3-OH kinase/AKT and nuclear factor kappa B pathways in human dental pulp stem cells. Journal of Endodontics, 40, 69–75. Hendershot, L. M., Valentine, V. A., Lee, A. S., Morris, S. W., & Shapiro, D. N. (1994). Localization of the gene encoding human BiP/GRP78, the endoplasmic reticulum cognate of the HSP70 family, to chromosome 9q34. Genomics, 20, 281–284.

5. Conclusions We have completed a quantitative iTRAQ proteomic comparison between PDLSCs and DPSCs from a single donor. First, we have identified a total of 159 differentially expressed proteins in PDLSCs and DPSCs. Among these enriched proteins, several members of the collagen, heat shock protein and protein S100 families may distinguish osteoinduced PDLSCs from DPSCs, though their exact roles need to be further clarified. Second, GO classification terms that distinguish osteoinduced PDLSCs from DPSCs were identified. For molecular function, 75% of the enriched terms belonged to catalytic activity, protein binding and negative regulation of protein metabolic processes, indicating different biosynthetic and regulatory processes of PDLSCs and DPSCs towards differentiation. For biological processes, two thirds of the enriched GO terms belonged to metabolic processes and response to stimulus, suggesting that PDLSCs and DPSCs may undergo distinct metabolic changes during the differentiation process and that the differentiation induction environment could act as a stress condition. Third, KEGG analysis revealed several pathways distinguishing osteoinduced PDLSCs from DPSCs, including the fatty acid biosynthesis pathway, pantothenate and CoA biosynthesis pathway, arachidonic acid metabolism pathway and PPAR signaling pathway. Last, we 18

Archives of Oral Biology 89 (2018) 9–19

H. Wang et al.

factor – alpha and Porphyromonas gingivalis lipopolysaccharides. Journal of Periodontal Research, 49, 405–414. Patil, R., Kumar, B. M., Lee, W. J., Jeon, R. H., Jang, S. J., Lee, Y. M., et al. (2014). Multilineage potential and proteomic profiling of human dental stem cells derived from a single donor. Experimental Cell Research, 320, 92–107. Peng, C., Zhao, Y., Wang, W., Yang, Y., Qin, M., & Ge, L. (2017). Histologic findings of a human immature revascularized/regenerated tooth with symptomatic irreversible pulpitis. Journal of Endodontics, 43, 905–909. Qu, Q., Fang, F., Wu, B., Hu, Y., Chen, M., Deng, Z., et al. (2016). Potential role of long non-coding RNA in osteogenic differentiation of human periodontal ligament stem cells. Journal of Periodontology, 87, e127–e137. Rescher, U., & Gerke, V. (2008). S100A10/p11: Family friends and functions. Pflugers Archiv, 455, 575–582. Retief, E., Parker, M. I., & Retief, A. E. (1985). Regional chromosome mapping of human collagen genes alpha 2(I) and alpha 1(I) (COLIA2 and COLIA1). Human Genetics, 69, 304–308. Rios, H., Koushik, S. V., Wang, H., Wang, J., Zhou, H. M., Lindsley, A., et al. (2005). Periostin null mice exhibit dwarfism, incisor enamel defects: And an early-onset periodontal disease-like phenotype. Molecular and Cellular Biology, 25, 11131–11144. Rodriguez, J. P., Gonzalez, M., Rios, S., & Cambiazo, V. (2004). Cytoskeletal organization of human mesenchymal stem cells (MSC) changes during their osteogenic differentiation. Journal of Cellular Biochemistry, 93, 721–731. Seo, B. M., Miura, M., Gronthos, S., Bartold, P. M., Batouli, S., Brahim, J., et al. (2004). Investigation of multipotent postnatal stem cells from human periodontal ligament. Lancet, 364, 149–155. Shiotsuka, M., Wada, H., Kiyoshima, T., Nagata, K., Fujiwara, H., Kihara, M., et al. (2013). The expression and function of thymosin beta 10 in tooth germ development. International Journal of Developmental Biology, 57, 873–883. Slots, J. (2017). Periodontitis: Facts, fallacies and the future. Periodontology 2000, 75, 7–23. Tancharoen, S., Tengrungsun, T., Suddhasthira, T., Kikuchi, K., Vechvongvan, N., Tokuda, M., et al. (2014). Overexpression of receptor for advanced glycation end products and high-mobility group Box 1 in human dental pulp inflammation. Mediators of Inflammation, 2014, 754069. Toledo, D. A., Roque, N. R., Teixeira, L., Milan-Garces, E. A., Carneiro, A. B., Almeida, M. R., et al. (2016). Lipid body organelles within the parasite trypanosoma cruzi: A role for intracellular arachidonic acid metabolism. PLoS One, 11, e160433. Vargas-Roig, L. M., Fanelli, M. A., Lopez, L. A., Gago, F. E., Tello, O., Aznar, J. C., et al. (1997). Heat shock proteins and cell proliferation in human breast cancer biopsy samples. Cancer Detection and Prevention, 21, 441–451. Vaure, C., & Liu, Y. (2014). A comparative review of toll-like receptor 4 expression and functionality in different animal species. Frontiers in Immunology, 5, 316. Vogl, T., Tenbrock, K., Ludwig, S., Leukert, N., Ehrhardt, C., van Zoelen, M. A., et al. (2007). Mrp8 and Mrp14 are endogenous activators of Toll-like receptor 4, promoting lethal: Endotoxin-induced shock. Nature Medicine, 13, 1042–1049. Wei, X., Wu, L., Ling, J., Liu, L., Liu, S., Liu, W., et al. (2008). Differentially expressed protein profile of human dental pulp cells in the early process of odontoblast-like differentiation in vitro. Journal of Endodontics, 34, 1077–1084. Wiesen, R. M., Padial-Molina, M., Volk, S. L., McDonald, N., Chiego, D. J., Botero, T., et al. (2015). The expression of periostin in dental pulp cells. Archives of Oral Biology, 60, 760–767. Wisniewski, J. R., Zougman, A., Nagaraj, N., & Mann, M. (2009). Universal sample preparation method for proteome analysis. Nature Methods, 6, 359–362. Wu, W. W., Wang, G., Baek, S. J., & Shen, R. F. (2006). Comparative study of three proteomic quantitative methods, DIGE, cICAT, and iTRAQ: using 2D gel – or LCMALDI TOF/TOF. Journal of Proteome Research, 5, 651–658. Xu, S., Cui, L., Ma, D., Sun, W., & Wu, B. (2015). Effect of ITGA5 down-regulation on the migration capacity of human dental pulp stem cells. International Journal of Clinical and Experimental Pathology, 8, 14425–14432. Zhang, Y. R., Du, W., Zhou, X. D., & Yu, H. Y. (2014). Review of research on the mechanical properties of the human tooth. International Journal of Oral Science, 6, 61–69.

Hinz, B. (2013). Matrix mechanics and regulation of the fibroblast phenotype. Periodontology 2000, 63, 14–28. Horiuchi, K., Amizuka, N., Takeshita, S., Takamatsu, H., Katsuura, M., Ozawa, H., et al. (1999). Identification and characterization of a novel protein, periostin: With restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. Journal of Bone and Mineral Research, 14, 1239–1249. Ivanovski, S., Gronthos, S., Shi, S., & Bartold, P. M. (2006). Stem cells in the periodontal ligament. Oral Diseases, 12, 358–363. Jackson-Boeters, L., Wen, W., & Hamilton, D. W. (2009). Periostin localizes to cells in normal skin: But is associated with the extracellular matrix during wound repair. Journal of Cell Communication and Signaling, 3, 125–133. Kanamori, T., Takakura, K., Mandai, M., Kariya, M., Fukuhara, K., Sakaguchi, M., et al. (2004). Increased expression of calcium-binding protein S100 in human uterine smooth muscle tumours. Molecular Human Reproduction, 10, 735–742. Kanehisa, M., Goto, S., Kawashima, S., Okuno, Y., & Hattori, M. (2004). The KEGG resource for deciphering the genome. Nucleic Acids Research, 32, D277–D280. Kanehisa, M., Goto, S., Hattori, M., Aoki-Kinoshita, K. F., Itoh, M., Kawashima, S., et al. (2006). From genomics to chemical genomics: New developments in KEGG. Nucleic Acids Research, 34, D354–D357. Kook, S. H., Jang, Y. S., & Lee, J. C. (2011). Involvement of JNK-AP-1 and ERK-NF-kappaB signaling in tension-stimulated expression of type I collagen and MMP-1 in human periodontal ligament fibroblasts. Journal of Applied Physiology (1985), 111, 1575–1583. Laplante, A. F., Moulin, V., Auger, F. A., Landry, J., Li, H., Morrow, G., et al. (1998). Expression of heat shock proteins in mouse skin during wound healing. Journal of Histochemistry & Cytochemistry, 46, 1291–1301. Li, J., Li, H., Tian, Y., Yang, Y., Chen, G., Guo, W., et al. (2016). Cytoskeletal binding proteins distinguish cultured dental follicle cells and periodontal ligament cells. Experimental Cell Research, 345, 6–16. Lim, W. H., Liu, B., Cheng, D., Williams, B. O., Mah, S. J., & Helms, J. A. (2014). Wnt signaling regulates homeostasis of the periodontal ligament. Journal of Periodontal Research, 49, 751–759. Ma, D., Gao, J., Yue, J., Yan, W., Fang, F., & Wu, B. (2012). Changes in proliferation and osteogenic differentiation of stem cells from deep caries in vitro. Journal of Endodontics, 38, 796–802. Ma, D., Cui, L., Gao, J., Yan, W., Liu, Y., Xu, S., et al. (2014). Proteomic analysis of mesenchymal stem cells from normal and deep carious dental pulp. PLoS One, 9, e97026. Marenholz, I., Heizmann, C. W., & Fritz, G. (2004). S100 proteins in mouse and man: From evolution to function and pathology (including an update of the nomenclature). Biochemical and Biophysical Research Communications, 322, 1111–1122. Mashaghi, A., Bezrukavnikov, S., Minde, D. P., Wentink, A. S., Kityk, R., ZachmannBrand, B., et al. (2016). Alternative modes of client binding enable functional plasticity of Hsp70. Nature, 539, 448–451. Miran, S., Mitsiadis, T. A., & Pagella, P. (2016). Innovative dental stem cell-based research approaches: The future of dentistry. Stem Cells International, 2016, 7231038. Muragaki, Y., Mattei, M. G., Yamaguchi, N., Olsen, B. R., & Ninomiya, Y. (1991). The complete primary structure of the human alpha 1 (VIII) chain and assignment of its gene (COL8A1) to chromosome 3. European Journal of Biochemistry, 197, 615–622. Nakajima, Y., Inagaki, Y., Hiroshima, Y., Kido, J., & Nagata, T. (2013). Advanced glycation end-products enhance calcification in cultured rat dental pulp cells. Journal of Endodontics, 39, 873–878. Norris, R. A., Damon, B., Mironov, V., Kasyanov, V., Ramamurthi, A., Moreno-Rodriguez, R., et al. (2007). Periostin regulates collagen fibrillogenesis and the biomechanical properties of connective tissues. Journal of Cellular Biochemistry, 101, 695–711. Padial-Molina, M., Volk, S. L., Rodriguez, J. C., Marchesan, J. T., Galindo-Moreno, P., & Rios, H. F. (2013). Tumor necrosis factor-alpha and Porphyromonas gingivalis lipopolysaccharides decrease periostin in human periodontal ligament fibroblasts. Journal of Periodontology, 84, 694–703. Padial-Molina, M., Volk, S. L., & Rios, H. F. (2014). Periostin increases migration and proliferation of human periodontal ligament fibroblasts challenged by tumor necrosis

19