JOURNAL OF TISSUE ENGINEERING AND REGENERATIVE MEDICINE RESEARCH J Tissue Eng Regen Med (2015) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/term.2077

ARTICLE

Tissue-specific composite cell aggregates drive periodontium tissue regeneration by reconstructing a regenerative microenvironment Bin Zhu1,2†, Wenjia Liu2,3†, Hao Zhang2†, Xicong Zhao1,2, Yan Duan1,2, Dehua Li1* and Yan Jin2,3* 1 State Key Laboratory of Military Stomatology, Department of Implantation, School of Stomatology, Fourth Military Medical University, Xi’an, Shaanxi, People’s Republic of China 2 State Key Laboratory of Military Stomatology, Centre for Tissue Engineering, School of Stomatology, Fourth Military Medical University, Xi’an, Shaanxi, People’s Republic of China 3 Research and Development Centre for Tissue Engineering, Fourth Military Medical University, Xi’an, Shaanxi, People’s Republic of China

Abstract Periodontitis is the most common cause of periodontium destruction. Regeneration of damaged tissue is the expected treatment goal. However, the regeneration of a functional periodontal ligament (PDL) insertion remains a difficulty, due to complicated factors. Recently, periodontal ligament stem cells (PDLSCs) and bone marrow-derived mesenchymal stem cells (BMMSCs) have been shown to participate in PDL regeneration, both pathologically and physiologically. Besides, interactions affect the biofunctions of different derived cells during the regenerative process. Therefore, the purpose of this study was to discuss the different derived composite cell aggregate (CA) systems of PDLSCs and BMMSCs (iliac-derived or jaw-derived) for periodontium regeneration under regenerative microenvironment reconstruction. Our results showed although all three mono-MSC CAs were compacted and the cells arranged regularly in them, jaw-derived BMMSC (JBMMSC) CAs secreted more extracellular matrix than the others. Furthermore, PDLSC/JBMMSC compound CAs highly expressed ALP, Col-I, fibronectin, integrin-β1 and periostin, suggesting that their biofunction is more appropriate for periodontal structure regeneration. Inspiringly, PDLSC/JBMMSC compound CAs regenerated more functional PDL-like tissue insertions in both nude mice ectopic and minipig orthotopic transplantation. The results indicated that the different derived CAs of PDLSCs/JBMMSCs provided an appropriate regenerative microenvironment facilitating a more stable and regular regeneration of functional periodontium tissue. This method may provide a possible strategy to solve periodontium defects in periodontitis and powerful experimental evidence for clinical applications in the future. Copyright © 2015 John Wiley & Sons, Ltd. Received 12 January 2015; Revised 17 May 2015; Accepted 16 June 2015

Keywords periodontium regeneration; jaw-derived mesenchymal stem cells; iliac-derived mesenchymal stem cells; periodontal ligament-derived stem cells; cell aggregate; regenerative microenvironment

1. Introduction Periodontitis is a highly prevalent chronic inflammatory disease that leads to loss of supporting tissues, including *Correspondence to: Y. Jin, Research and Development Centre for Tissue Engineering, Fourth Military Medical University, Xi’an 710032, China. E-mail: [email protected]; D. Li, Department of Implantation, School of Stomatology, Fourth Military Medical University,Xi’an 710032, China. E-mail: [email protected] † These authors contributed equally to this study. Copyright © 2015 John Wiley & Sons, Ltd.

periodontal ligament and alveolar bone, which then causes loss of teeth (Liu et al., 2011). Importantly, the periodontium has a limited capacity for regeneration after it is damaged. Therefore, once inflammation is established, only therapeutic intervention may have the potential to induce periodontium regeneration (Bartold et al., 2000). Procedures commonly used in the clinic have included guided tissue regeneration, bone graft placement, root surface conditioning and growth factor application. However, these methods have limitations in attaining complete and predictable regeneration,

B. Zhu et al.

especially in serious periodontium defects (Pejcic et al., 2013). ’Periodontium tissue’ means the periodontal ligament (PDL) linking the cementum and alveolar bone. Therefore, the regeneration of PDL fibers inserting into the surface of the cementum and alveolar bone is a key point. Thanks to the study of stem cell biology and regenerative medicine, tissue-engineering and cell-based approaches in periodontal regeneration have become possible (Kawaguchi et al., 2004; Lin et al., 2008). Some research has shown that only cells residing in the periodontal ligament are capable of forming new inserting collagen fibres on exposed root surfaces, which indicates that functional PDL fibres have been regenerated successfully (Karring et al., 1980; Nyman et al., 1982; Gottlow et al., 1984). Fortunately, periodontal ligament-derived mesenchymal stem cells (PDLSCs), isolated as multipotent stem cells, provided a good choice for stem cell-based periodontal regeneration (Seo et al., 2004; Shi et al., 2005). Research has shown that PDLSCs possess the capacity to regenerate PDL, cementum and alveolar bone (Ding et al, 2010; Tsumanuma et al., 2011). However, compared with PDLSCs, BMMSCs have been prioritized in bone tissue regeneration referred in the periodontium (Huang et al., 2009). Therefore, some other cells for regenerating whole periodontium may be necessary. A previous study demonstrated that bone marrow mesenchymal stem cells (BMMSCs) could not only promote the proliferation and differentiation of PDLSCs (Xie and Liu, 2012) but also be engrafted into periodontal tissue and differentiate into periodontium-specific cells to participate in periodontium regeneration (Yang et al., 2010; Zhou et al., 2011). However, how to regenerate the functional periodontium is still unclear. It has been shown that the regeneration of complex tissues depends on the capacities of the seed cells and the regulation of environment factors (Paschos et al., 2014). The regenerative microenvironment affects tissue-regeneration issues including cell migration, proliferation, differentiation and regulation by cytokines. Cytokines such as BMP-2, VEGF and FGF in vitro have been certified beneficial to complex tissue regeneration (Szpalski et al., 2013; Chen et al., 2013). However, the regenerative local area lacks persistent and active beneficial stimulus in vivo and the co-culture strategy provides this possibility. With co-culture, different derived cells respectively differentiated to the expected tissues, and cytokines secreted by different derived cells promoted the regeneration (Xie and Liu, 2012; Razavi et al., 2013). Meanwhile the interactions between different derived cells, mimicking tissue developmental crosstalk in vitro, induce synergetic biological behaviour, including migration, proliferation, differentiation and regulation. The co-culture strategy of MSCs and other cells brings great benefits in blood vessels, peripheral nerve and cartilage regeneration (Andrew et al., 2011; Meretoja et al., 2012; Mehnert et al., 2014). However, as mentioned above, functional regenerated periodontium has to include inserting collagen fibres on the exposed material surface. To realize this goal, more appropriate regenerative microenvironment will be Copyright © 2015 John Wiley & Sons, Ltd.

needed. Cell aggregate (CA) technology means that dissociated cells at high density are cultured in vitro, secreting abundant extracellular matrix (ECM), and tend to group themselves with cells of their own type. This is similar to generating a cell sheet, but it is easier to form multilayered cells and contain more cell sources without exogenous material. It was reported that in temperature-responsive culture dishes, PDL cell sheets were made to regenerate a cementum–PDL complex (Flores et al., 2008; Washio et al., 2010). However, the adhesion of multilayer cell sheets by fibrin gel may cause residual degradation affecting PDL regeneration, and the repeatability and stability of PDL regeneration by mono-PDL cell sheet in nude mice ectopic transplantation remain to be discuss. Besides, CAs contain plenty of self-produced ECM, which is the basis for bioregeneration. Importantly, composite CAs provide a continuous and active regenerative stimulus for complex tissue regeneration. Recently, the transplantation of aggregates of synovial MSCs regenerated meniscus more effectively in a rat massive meniscal defect (Katagiri et al., 2013). In this study, we designed a composite cell aggregate (CA) system including two layers of PDLSC aggregate and one layer of BMMSC (jaw- or iliac-derived) aggregate outside for periodontium regeneration. Additionally, surface dentine was removed from dentine block to expose a cavernous dentinal tubule, called the treated dentine matrix (TDM), and bovine spongy bone was calcinated to remove organic matter and obtain a multiporous inorganic structure called ceramic bovine bone (CBB); TDM and CBB were respectively used as the transplanted material. The results showed that although all three mono-MSC CAs were compacted and the cells arranged regularly in them, the PDLSC/JBMMSC compound CAs regenerated more bone-like tissue and inserting PDL-like fibres on the TDM and CBB surfaces in both nude mice ectopic and minipig orthotopic transplantation. The objective of this study was to develop and test the effect of a composite CAs system on periodontium regeneration. It should provide a feasibility method for future clinical applications.

2. Materials and methods 2.1. Sample collection and cell culture Sample collection conformed to the protocol approved by the ethical authorities at the Fourth Military Medical University. Healthy human tooth samples were collected from 10 donors, aged 15–24 years, undergoing routine premolar or third molar extractions for orthodontic reasons. Iliac bone marrow samples were collected from 10 donors aged 18–30 years, undergoing the alveolar bone cleft repair by auto-ilium transplantation. Human jaw bone marrow samples were obtained from patients who had undergone orthognathic surgery (10 donors aged 18–28 years). Four 2 year-old minipigs were used for culturing pig-related cells. J Tissue Eng Regen Med (2015) DOI: 10.1002/term

Regenerative microenvironment for periodontium tissue

The isolation of PDLSCs, JBMMSCs and IBMMSCs has been described previously (Seo et al., 2004; Xie and Liu, 2012; Cicconetti et al., 2007). In brief, PDL tissues were gently separated from the surface of the mid-third of the root and subsequently digested with 3 mg/ml collagenase type I (Sigma, St. Louis, MO, USA) for 30 min at 37°C. Single-cell suspensions were obtained by passing the samples through a 70 μm strainer (Falcon, BD Labware, Franklin Lakes, NJ, USA). The cells were seeded into six-well plates and cultured with α-minimum essential medium (α-MEM; Gibco BRL, Gaithersburg, MD, USA) supplemented with 15% fetal bovine serum (FBS; Gibco) and incubated in 5% CO2 at 37°C. Bone marrow aspirate (1 ml) was seeded into a 10 cm culture dish and 20 ml α-MEM supplemented with 15% FBS (Gibco) was added, then incubated in 5% CO2 at 37°C. Jaw bone debris was rinsed with α-MEM containing 100 U/ml penicillin and 100 mg/ml streptomycin (Invitrogen, Fredick, USA) for 3 times. The debris were cut into almost 1 × 1 mm2 pieces and passed through a 70 mm strainer to obtain single-cell suspensions. Both the cell suspensions and jaw bone pieces were seeded into 10 cm dish and cultured as iliac BMMSCs. All samples from the different donors (both humans and pigs) were mixed to obtain a pool of cells for further culture. All the experiments in this study were replicated at least three times.

2.2. Cell evaluation 2.2.1. Colony-forming efficiency and flow cytometry To assess the ability to produce colony-forming units (CFUs), single-cell suspensions (1 × 103 cells) were seeded in 10 cm diameter culture dishes. Day 14 cultures were fixed in 4% paraformaldehyde and stained with 0.1% toluidine blue. Then, fluorescence-activated cell sorting (FACS) analysis was used to analyse STRO-1, CD29, CD90, CD105, CD31, CD34 and CD45 (R&D Systems) (Seo et al., 2004; Xie and Liu, 2012; Cicconetti et al., 2007). Multiple colony-derived hMSCs at passages 2–4 were used in our experiments.

2.2.2. Cell proliferation assay To test cell proliferation, we used the 3-(4,5-dimethylthiazol2yl)-2,5-diphenyltetrazolium bromide (MTT) assay. hPDLSCs, hJBMMSCs and hIBMMSCs at passage 2 were plated at a density of 5 × 103 cells/well in 96-well plates and cultured in α-MEM (10% FBS). The MTT assay was carried out for 7 days, according to the manufacturer´s protocol (Sigma). Absorbance was determined at 490 nm using a microplate reader (Bio-TEK Instruments, Winooski, VT, USA).

2.2.3. Adipogenic differentiation To induce adipogenic differentiation, at 100% confluence the human cells were cultured in α-MEM supplemented with 10% FBS, 2 μm insulin, 0.5 M isobutylmethylxanthine Copyright © 2015 John Wiley & Sons, Ltd.

and 10 nM dexamethasone for 21 days. Then the cells were washed twice with PBS and fixed in 4% paraformaldehyde for 60 min. Oil red O (Sigma) staining was performed, as previously described. The staining was dissolved by isopropyl alcohol in each well and the absorbance was determined at 520 nm, using a microplate reader (Bio-TEK). Expressions of the specific adipogenic marker peroxisome proliferator-activated receptor-γ (PPARγ) and lipoprotein lipase (LPL) were determined by real-time PCR.

2.2.4. Osteogenic differentiation To induce osteoblast differentiation, after reaching 80% confluence human cells were cultured in α-MEM supplemented with 10% FBS, 100 nM dexamethasone, 50 μg/ml ascorbic acid and 5 mM β-glycerophosphate for 28 days. The medium was changed every 3 days. Calcium accumulation was detected by 2% alizarin red staining and dissolved by 1 ml sodium dodecyl sulphate solution; the light absorption of sodium dodecyl sulphate solution with alizarin red was read at 570 nm with a microplate reader (Bio-TEK). The cells were washed twice in phosphate-buffered saline (PBS) after fixation in 4% paraformaldehyde for 20 min. ALP staining and activity were determined using the BCIP/NBT Alkaline Phosphatase Colour Development Kit (Beyotime Co., Shanghai, China) and the Alkaline Phosphatase (AKP/ALP) Detection Kit (Zhong Sheng Co., Beijing, China). Expressions of the osteoblastic genes Runt-related transcription factor 2 (Runx2), alkaline phosphatase (ALP), osteocalcin (OCN), type I collegen (Col-I) and bone sialoprotein (BSP) were determined by real-time PCR.

2.2.5. Induction of hPDLSCs by hJBMMSCs/ hIBMMSCs hPDLSCs were seeded in a six-well plate and hJBMMSCs/ hIBMMSCs were seeded separately in Transwells at a density of 1 × 105 cells/well. After 1 day in normal culture medium, the cells were cultured in osteogenic medium and harvested after 3 and 14 days of induction. Real-time PCR was performed to test the expression of Runx2, ALP, OCN, Col-I and BSP.

2.2.6. Real-time polymerase chain reaction (RT– PCR) assay Human cells were seeded separately at a density of 1 × 105 cells/culture bottle. After 1 day in normal culture medium, the cells were cultured in osteogenic medium and harvested after 0, 3, 7 and 14 days of induction. Meanwhile, after 1 day in normal culture medium, the cells were cultured in adipogenic medium and harvested after 7 days of induction. Total cellular RNA was isolated with Trizol reagent (Invitrogen, Carlsbad, CA, USA). Then, isolated RNA was used as a template for cDNA synthesis, prepared using a Superscript II first-strand cDNA synthesis kit (Invitrogen Life Technologies, Carlsbad, CA, USA). Real-time PCR was performed according to the manufacturer’s protocol. The primers used are listed in Table 1. J Tissue Eng Regen Med (2015) DOI: 10.1002/term

B. Zhu et al. Table 1. Primer sequences Gene RUNX2 OCN COL-I ALP BSP LPL PPARγ β-ACTIN

Forward primer

Reverse primer

5′-CCCGTGGCCTTCAAGGT-3′ 5′-CCCAGGCGCTACCTGTATCAA-3′ 5′-CCAGAAGAACTGGTACATCAGCAA-3′ 5′-TAAGGACATCGCCTACCAGCTC-3′ 5′-GATTTCCAGTTCAGGGCAGTAG-3′ 5′-AGGACCCCTGAAGACAC-3′ 5′-CAAGACAACCTGCTACAAGC-3′ 5′-TGGCACCCAGCACAATGAA-3′

5′-CGTTACCCGCCATGACAGTA-3′ 5′-GGTCAGCCAACTCGTCACAGTC-3′ 5′-CGCCATACTCGAACTGGAATC-3′ 5′-TCTTCCAGGTGTCAACGAGGT-3′ 5′-CCCAGTGTTGTAGCAGAAAGTG-3′ 5′-GGCACCCAACTCTCATA-3′ 5′-TCCTTGTAGATCTCCTGCAG-3′ 5′-CTAAGTCATAGTCCGCCTAGAGCA-3′

2.3. Scaffold materials

2.4. Cell aggregates

2.3.1. Human TDM preparation

2.4.1. Construction of compound cell aggregates containing three layers

In this study, human treated dentine matrix (TDM) was fabricated as described previously (Guo et al., 2012). Briefly, premolar teeth removed for clinical reasons at the School of Stomatology, Fourth Military Medical University, were collected. Periodontal tissue was completely scraped away, using a curette, along with removal of the outer cementum and part of the dentine. Dental pulp tissues were removed using a barbed broach and pre-dentine was mechanically removed using a carborundum bit. For the fabrication of human TDM, dentine matrix was formed to a length of 5.0 mm and a thickness of up to 2.0 mm and mechanically cleaned using an ultrasonic cleaner. Human dentine matrices were then treated with 17% ethylene diamine tetra-acetic acid (EDTA; Sigma) for 5 min, 10% EDTA for 5 min, 5% EDTA for 10 min. Human TDMs were maintained in sterile PBS with 100 U/ml penicillin (Hyclone, USA) and 100 mg/ml streptomycin (Hyclone) for 72 h, then washed in sterile deionized water for 10 min in an ultrasonic cleaner and finally stored in α-MEM at 4°C.

Multiple colony-derived PDLSCs, JBMMSCs and IBMMSCs from both humans and minipigs at passage 3 were seeded at approximately 1 × 105/ml into a 12-cell plate. The PDLSCs were cultured in normal α-MEM containing 10% FBS for 3 days. After reaching 90% confluence, the medium was changed to aggregate induction medium containing 10% FBS and ascorbate (50 μg/ml). Finally, PDLSC aggregates were formed and easily detached from the culture plates with a cell scraper. The construction of IBMMSC and JBMMSC CAs were similar to the PDLSC CA. The compound CAs contained two layers of PDLSCs and one layer of JBMMSCs or IBMMSCs outside. The TDM or CBB was coated by one layer of PDLSC CA to form the complex. Then this complex was again coated by another PDLSC CA, and this complex was next coated by an outer layer of JBMMSC or IBMMSC CA (Figure 1).

2.4.2. H&E and immunohistochemical staining for the cell aggregates

2.3.2. Minipig TDM preparation Minipig TDM (pTDM) was fabricated as described previously (Ji et al., 2014). Four canine teeth of each minipig were extracted and the dentine matrix was formed to a length of 1.5 cm and a diameter of 5.0 mm, with an empty dental pulp cavity. The average thickness of pTDM was about 2–3 mm. Finally, the pTDMs were stored in α-MEM at 4°C.

2.3.3. Ceramic bovine bone preparation Ceramic bovine bone (CBB; Research and Development Centre for Tissue Engineering, Fourth Military Medical University, Xi’an, China) was produced from fresh bovine rib bones, which were subsequently cut into blocks, washed in normal saline and soaked in H2O2 to remove proteins. Next, the blocks were formed to a length of 5.0 mm and a thickness of up to 2.0 mm. These blocks were washed with running water, heated at 900°C for 1 h and sterilized in a high-temperature and -pressure environment. Then they were washed in sterile deionized water for 10 min in an ultrasonic cleaner and finally stored in α-MEM at 4°C (Xie and Liu, 2012). Copyright © 2015 John Wiley & Sons, Ltd.

The CAs were fixed in 4% phosphate-buffered paraformaldehyde for 24 h, embedded in paraffin, longitudinally sectioned and stained with haematoxylin and eosin (H&E). Other sections were incubated with primary antibodies as follows: anti-ALP (1:200; Abcam, UK); anti-Col-I (1:200; Santa Cruz, USA); anti-fibronectin (1:200; Abcam); antiperiostin (1:200; Abcam); and anti-integrin-β1 (1:200; Abcam). PBS was used for the negative controls instead of the primary antibodies. Biotinylated secondary antibodies (1:1000) were purchased from Dako (USA). The stained sections were observed using a light microscope (Nikon, Japan).

2.4.3. Observation of cell aggregates around the TDM–CBB surface via SEM The entire complex of TDM–CBB with compound CAs was cultured in normal medium for 3 days to ensure its integrity, then fixed in 4% paraformaldehyde. The samples were anodized in an electrolyte containing 0.5 wt% hydrofluoric acid and 1 M phosphoric acid for 1 h. After that, the whole complex was observed by scanning electron microscopy (SEM; Hitachi, S-4800. Japan). J Tissue Eng Regen Med (2015) DOI: 10.1002/term

Regenerative microenvironment for periodontium tissue

Figures 1. Construction of compound CA-bioscaffolds. (A) The bioscaffold TDM was coated by one layer of PDLSC CA to form the complex; then this complex was again coated by another PDLSC CA; next, this complex was covered by one layer of IBMMSC/JBMMSC CA on the outside. (B) Schematic diagram for the CBB/TDM compound CAs system

2.4.4. Differentiation of compound CAs Entire complexes of TDM–CBB with compound CAs were cultured in normal medium for 3 days to ensure their integrity, then cultured in osteogenic medium and harvested after 3 or 14 days of induction. Expressions of Runx2, ALP, OCN, Col-I and BSP were determined by real-time PCR.

2.5. In vivo transplantation 2.5.1. Transplantation surgery In vivo transplantation comprised two parts. The whole complex was implanted into six 8-week nude mice (BALB/c-nu; FMMU Medical Laboratory Animal Centre, Xi’an, China). All the animal procedures complied with the guidelines provided by the Animal Care Committee of the Fourth Military Medical University.

2.5.1.1. Nude mice ectopic transplantation. Constructs of hPDLSC/hJBMMSC plus TDM were transplanted into subcutaneous pockets on the left side of 10 nude mice, and constructs of hPDLSC/hIBMMSC plus TDM were implanted into the other side. Another 10 nude mice were provided for the CBB with compound CAs. Eight weeks after the transplantation, all 20 nude Copyright © 2015 John Wiley & Sons, Ltd.

mice were euthanized and the 40 complexes were removed for analysis.

2.5.1.2. Minipig orthotopic transplantation. Under local anaesthesia, four canine teeth were extracted from each of five 2 year-old minipigs, 3 months before the present experiments. All five pigs were intramuscularly injected with 0.1 mg/kg medetomidine (Sigma) and 15 mg/kg ketamine (Sigma) for anaesthetic premedication and then subjected to an intravenous injection of 2.5 mg/kg propofol (Sigma). An endotracheal tube was inserted and anaesthesia was maintained with sevoflurane (Sigma). pTDM with pJBMMSC/pPDLSC CAs were randomly implanted into the left side of the maxillary and mandibular jaw bones; pTDM with pIBMMSC/pPDLSC CAs were randomly implanted into the right side of the same bones. Following alveolar bone exposure, the implant site, where the canine tooth had been extracted three months previously, was prepared with a round burr (5.0 mm in diameter, to a total depth of 1.5 cm) and with a custom-made drill (5.2 mm in diameter, to an incisal depth of 5.0 mm). pTDMs (5.0 mm × 1.5 cm) coated by three layers of autologous CAs were implanted into fabricated moulded wells (5.0 mm diameter to a root depth of 1.0 cm, and 5.2 mm in diameter to a top depth of 5.0 mm). Implant beds were prepared to avoid friction during implantation, and the implants were quite stable. Finally, the gingival incision was tightly closed. All 20 complexes of each minipig were placed for 12 weeks and then removed for analysis. J Tissue Eng Regen Med (2015) DOI: 10.1002/term

B. Zhu et al.

2.5.2. Morphological and histomorphometric evaluation of regenerated PDL tissues The complexes were fixed in a solution of 10% formaldehyde at room temperature for 7 days. Then they were demineralized with 10% EDTA, pH 6.9, in 30 days, then dehydrated in a ascending series of ethanols (70–100%), infiltrated and finally embedded in paraffin. Paraffin sections were stained with H&E and Masson’s trichrome. The immunohistochemical primary antibody was antiCol-I (1:200; Santa Cruz, USA). Biotinylated secondary antibody (1:1000) was purchased from Dako (USA). A microscope (Leica Microsystems AG, Wetzlar, Germany) was used for histological evaluation, which was based on morphological observation of three sections/complex.

2.6. Statistical analysis All statistical analyses were performed using ANOVA, followed by Fisher’s protected least significant difference (PLSD) test and Student’s t-test using SPSS v. 15.0 software (SPSS, USA). All values are expressed as mean ± SD; p < 0.05 was considered to be statistically significant.

3. Results 3.1. Comparison of biology characteristics of the three cells Human PDLSCs, JBMMSCs and IBMMSCs are heterogeneous populations of stem cells and express the mesenchymal cell lineage marker STRO-1 and also adhesion molecules CD29 and CD105 and extracellular matrix protein CD90, but haematopoietic lineage CD31, CD34 and CD45 are negatively expressed (see supporting information, Figure S1). The three cell types were capable of forming CFUs generated from single cells; percentage CFUs of hPDLSCs (27.8 ± 2.4%) was greater than that of hJBMMSCs (24.5 ± 2.3%) and hIBMMSCs (19.6 ± 1.7%). The MTT assay results were similar to those of the CFU assay, suggesting that the viability of hPDLSCs was greater than that of bone marrow-derived MSCs, especially IBMMSCs (see supporting information, Figure S2). To investigate the differentiation potential of the three cell types, they were cultured in osteogenic differentiation medium. ALP staining was performed at day 14 and mineralized nodules were stained with alizarin red at day 28 (Figure 2A). The expression levels of osteoblast-related genes, including Runx2, ALP, OCN and Col-I, were analysed after osteogenic induction for 14 days (Figure 2D). Both staining and gene expression analysis showed the same results, that the osteogenic differentiation capacity of hJBMMSCs was the strongest among the three cell types (Figure 2B–D). However, the adipogenic differentiation of hJBMMSCs was weaker than that of the other two cell Copyright © 2015 John Wiley & Sons, Ltd.

types; oil red O staining and the expression of the specific adipogenic marker peroxisome PPARγ and LPL supported this result (Figure 2E–G).

3.2. Biological properties of different derived cell aggregates 3.2.1. Identification of biological characteristics of the three mono-CAs H&E staining showed that all three mono-CAs were dense and contained plenty of cells that could ensure collagen secretion (Figure 3A). SEM analysis showed that all the cells were spread in the CAs and made tight contact with each other. Among the three CAs, hJBMMSC CA contained the most compacted cell arrangement and collagen secretion (Figure 3B). Immunohistochemical analyses showed that all three CAs positively expressed Col-I and ALP (early markers for osteogenic differentiation), periostin (a special marker of PDL) and integrin β1 and fibronectin (markers related to the biofunction of CA) (Figure 4). Among the three CAs, hJBMMSCs expressed ALP and Col-I more highly than the other CAs. In addition, the expression of periostin in hJBMMSC CA was weaker than in hPDLSC CA but significantly stronger than hIBMMSC CA (Figure 4).

3.2.2. Morphology of compound CAs grown on scaffolds in vitro SEM analysis of the scaffold CBB showed a porous structure (Figure 5A). In contrast, the surface of hTDM showed that dentinal tubules were sufficiently exposed after being treated with EDTA (Figure 5B). Both compound hPDLSC/ hJBMMSC and hPDLSC/hIBMMSC CAs were separately seeded on hTDM and CBB for 7 days. SEM pictures showed that two compound CAs could adhere to the scaffolds well, proliferate adequately and extend excessively on the surface of hTDM and CBB (Figure 5C–F). There is no obviously difference between these two compound CAs.

3.2.3. hPDLSC/hJBMMSC CAs possessed a stronger collagen secretion and osteogenic differentiation capacity in vitro After 3 days of osteogenic induction, the expression levels of Runx2, Col-I, ALP and BSP in the hPDLSC/hJBMMSC CAs were greater than those in the hPDLSC/hIBMMSC CAs (Figure 6A, C). After 14 days, Col-I expression in hPDLSC/hJBMMSC CAs with hTDM exceeded more than two-fold compared to hPDLSC/hIBMMSC CAs. ALP expression in hPDLSC/hJBMMSC CAs with CBB was almost twofold compared to hPDLSC/hIBMMSC CAs. Furthermore, the OCN level also increased in hPDLSC/hJBMMSC with hTDM (Figure 6B, D). These data suggest that the J Tissue Eng Regen Med (2015) DOI: 10.1002/term

Regenerative microenvironment for periodontium tissue

Figure 2. Verification of related stem cell characteristics in hPDLSCs, hJBMMSCs and hIBMMSCs. (A) Osteogenic differentiation of the three cell types, assessed by alizarin red and ALP staining; scale bars = 25 and 50 μm. (B, C) ALP activity and quantity of mineralized nodes were analysed by absorptiometry. (D, G) Expression of osteoblast-related genes and adipogenic-specific genes were investigated by real-time PCR. (E) Adipogenic differentiation of the three cell types, assessed by oil red O staining; scale bar = 100 μm. (F) Quantity of lipid droplets, tested by absorptiometry; data shown as mean ± SD; *p < 0.05; n = 3

hPDLSC/ hJBMMSC compound CAs not only secrete more collagen but also have a stronger osteogenic differentiation potential.

3.3. Regenerative PDL-like tissue around scaffold surfaces in vivo 3.3.1. Ectopic transplantation in nude mice

3.2.4. hJBMMSCs/hIBMMSCs increased the collagen secretion and osteogenic differentiation of hPDLSCS by paracrine function After 3 and 14 days of osteogenic induction, PDLSCs induced by JBMMSCs highly expressed Runx2, ALP, and Col-I, > two-fold those induced by IBMMSCs (Figure 6E, F). In particular, in PDLSCs induced by JBMMSCs, ALP expression after 3 days and Runx2 expression after 14 days of induction reached almost three-fold of those of PDLSCs induced by IBMMSCs. Copyright © 2015 John Wiley & Sons, Ltd.

Eight weeks after transplantating whole complexes into the subcutaneous spaces of nude mice, we harvested 28 of 33 regenerated tissue complex specimens (Table 2A) and examined their morphology after staining with H&E, Masson’s trichrome and anti-Col-1 immunohistochemical stain. In all transplantation samples of the two types of CAs and biomaterials, PDL-like tissue regeneration was found. In the hPDLSC/hJBMMSC compound CAs, typically arranged collagen fibre bundles (PDL-like fibres) were formed, inserted vertically into the surfaces of hTDM J Tissue Eng Regen Med (2015) DOI: 10.1002/term

B. Zhu et al.

Figure 3. Morphological characteristics of three mono-hMSC CAs. (A) H&E staining showed that all three mono-CAs were dense and contained plenty of cells that could ensure collagen secretion. (B) The arrangement of cells in all CAs was observed by SEM; scale bar = 100 μm

and CBB. The newly inserted collagen fibres and bonelike matrix were observed while bone-like matrix formed on the surface of the hTDM (Figure 7A) and CBB (Figure 8A). These regenerated structures closely resembled the physiological structure of the real periodontium. In contrast, the hPDLSC/hIBMMSC compound CAs formed fewer inserted fibres and more parallel fibres around the hTDM (Figure 7B) and CBB (Figure 8B). Besides, in nude mice ectopic transplantation, the success rate of PDLSCs/JBMMSCs CAs was higher than that of PDLSCs/JBMMSCs CAs on both biomaterials (p < 0.05) (Table 2A).

3.3.2. Orthotopic transplantation in minipigs To further investigate the periodontal regeneration capacity of the compound CAs, we isolated PDLSCs, JBMMSCs and IBMMSCs from five minipigs and cultured CAs separately (see supporting information, Figure S3B, C). We implanted pig TDM coated with three layers of CAs into the alveolar bone of minipigs (see supporting information, Figure S3A). Four experimental complexes grew well and solidly. After 12 weeks, 17 of 20 samples were harvested. Among eight samples of pIBMMSCs/pPDLSCs CAs with pTDM, six samples showed PDL-like tissue regeneration; seven of 10 samples of pJBMMSCs/ pPDLSCs CAs with pTDM also showed PDL-like tissue regeneration (Table 2B). H&E, Masson’s trichrome and anti-Col-1 immunohistochemical staining showed that the autologous pigPDLSC/pig-JBMMSC compound CAs regenerated a dozen arranged inserted PDL-like fibres on the TDM and alveolar bone side surfaces. Moreover, plenty of bone-like matrix was formed on the alveolar bone surface. These PDL-like fibres could integrate TDM and alveolar bone together closely (Figure 9A). However, there were only some parallel collagen fibres and little bone tissue formed between TDM and alveolar bone in the autologous pPDLSC/pIBMMSC compound CAs Copyright © 2015 John Wiley & Sons, Ltd.

(Figure 9B). Taking these results together, we inferred that the PDLSC/JBMMSC compound CAs not only regenerated PDL-like inserted fibres but also formed more bone tissue. Therefore, the PDLSC/JBMMSC compound CAs appear to be more appropriate for periodontal regeneration than the PDLSC/IBMMSC compound CAs.

4. Discussion In this study, we aimed to regenerate functional periodontium via a different derived composite CAs technology and to inverstigate the regenerative microenvironmental effects. Our data showed that collagen fibres with good contact and well organized PDL-like fibres were successfully regenerated, vertically inserted into the surface of TDM and CBB in the PDLSC/JBMMSC composite CAs. Our data suggested that this method may provide a possibility for regenerating periodontium in the periodontitis defects. The periodontium includes periodontal ligament, alveolar bone, cementum and gingiva. Among these, the functional PDL, which is a dense, fibrous and highly vascular soft connective tissue linking two mineralized tissues together, is the hardest to regenerate. Therefore, a single-cell method could not achieve this goal. Advances in co-culture strategy, which activates various signals of different derived cells to inhibit cell apoptosis, promote cell differentiation and encourage complex tissue regeneration (Meretoja et al., 2012), have laid the foundation for periodontal regeneration. Nyman et al. (1982) have demonstrated that the cells of the periodontal ligament possess the ability to regenerate a connective tissue attachment on denuded root surfaces, with the formation of a new cementum layer with inserted fibres. PDLSCs were shown to have a strong ability to regenerate periodontal tissue, both in vitro and in vivo. Considering bone formation, BMMSCs are also necessary for bone-like J Tissue Eng Regen Med (2015) DOI: 10.1002/term

Regenerative microenvironment for periodontium tissue

Figure 4. Immunohistochemical analyses of three mono-hMSC CAs. All three hMSC CAs positively expressed ALP, Col-I, fibronectin, periostin and integrin-β1: ALP and Col-I indicated osteogenic differentiation capacity; integrin-β1 and fibronectin were markers of biofunctions; periostin was a special marker of PDL; scale bar = 100 μm

tissue regeneration. Therefore, IBMMSCs, JBMMSCs and PDLSCs were chosen for constructing compound CAs in our study. Our data showed that each MSC CA had at least five layers of cells immersed in ECM, with abundant MSCs arranged in an orderly and regular way. Besides being the framework supporting cell sheet structure, ECM proteins have important biofunctions during tissue regeneration (Kim et al., 2011). It is believed that, compared to cell sheet, CA technology could contain more cells easily, and the extensibility is perfectly suitable to ensure plenty of ECM secretion for tissue repair and regeneration. Immunohistochemical results showed that three monoderived MSC CAs expressed ECM-related proteins, such as Col-I, integrinβ1, fibronectin and periostin. Abundant Copyright © 2015 John Wiley & Sons, Ltd.

Col-I expression, supporting the entire CA structure, suggested that the MSC CAs had favourable mechanical characteristics for maintaining integrity and stability (Takahashi et al., 2011). Integrin-β1 and fibronectin are ubiquitously expressed major cell surface receptors/ ligands for the ECM, and activate the intracellular signalling pathways for controlling cell morphology, proliferation, survival and differentiation (Hynes, 2002). Moreover, the integrin-β1–fibronectin interaction is the dynamic response of the complexes to a pulling force (Streuli and Akhtar, 2009). Thus, the high expression of integrinβ1/fibronectin ensures that our CAs had sufficient mechanical strength to wrap up the treated dentine matrix tightly without being thoroughly torn. Periostin plays an important role as a cell adhesion molecule for J Tissue Eng Regen Med (2015) DOI: 10.1002/term

B. Zhu et al.

Figure 5. Microscopic appearance of the scaffold materials and adhesion of the compound CAs to the scaffolds. (A, B) SEM images of CBB showed a porous structure, while TDM showed a micropore facial structure. (C–F) The attachment of compound CAs to the surface of CBB or TDM was analysed by SEM

preosteoblasts that are involved in osteoblast recruitment, attachment and spreading (Li et al., 2003). Additionally, it is a matricellular protein that is preferentially expressed in the fibroblastic cells of PDL and osteoblasts on the alveolar bone surface (Kashima et al., 2009). Therefore, deficiency of periostin usually leads to rapid alveolar bone resorption and the loss of periodontium homeostasis (Rios et al., 2008). Compared to the IBMMSC CA, the strong periostin expression of the PDLSC and JBMMSC CAs indicated that they have greater potential to form alveolar bone and periodontal ligament than IBMMSC CA. All these factors indicate that hJBMMSCs possess the original tissue’s specification and that hJBMMSCs may be more appropriate for periodontal structure regeneration than hIBMMSCs. We co-cultured PDLSCs with JBMMSCs or IBMMSCs indirectly by Transwell in osteogenic medium for 3 and 14 days. PDLSCs induced by JBMMSCs highly expressed almost three-fold more Runx2, ALP and Col-I than those induced by IBMMSCs. Therefore, we believe that paracrine factors may be important in the interaction between PDLSCs and JBMMSCs/IBMMSCs. In this study, we found that compound CAs using PDLSCs/JBMMSCs expressed high levels of OCN, Runx2, BSP and Col-I compared to those using PDLSCs/ IBMMSCs. The results were similar to those of two cell types compared previously (Caterson et al., 2002; Matsubara et al., 2005). Jawbone and periodontal tissue are adjacent to one another and interact with each other frequently in a similar microenvironment (Aubin et al., 1995; Ona and Wakabayashi, 2006). BMMSCs were influenced by interactions between the Copyright © 2015 John Wiley & Sons, Ltd.

jaw and periodontium to facilitate periodontal regeneration based on similar biological properties. However, BMMSC CA in vitro lacked induction by the periodontal microenvironment. Combined with the immunohistochemical results, we believe that PDLSCs/JBMMCSs are superior to PDLSCs/ IBMMSCs for periodontal regeneration. We designed and constructed CBB–CA and TDM–CA complexes to mimic periodontal regeneration of the tooth root side and alveolar bone side heterotopic transplantation in nude mice. The results of the hPDLSC/hJBMMSC compound CAs showed a mineral matrix deposit on the surface of the scaffolds, and PDL-like fibres were arranged in an orderly manner on the surface of TDM and CBB, with the ends inserting into them. Previous researches on periodontium regeneration only used a single type of cell, including PDLSCs, dental follicle cells (DFCs) and IBMMSCs, therefore, new cementum formation and new connective tissue attachment were observed on some areas of the scaffold surfaces in ectopic experiments (Ding et al., 2010; Yang et al., 2010; Guo et al., 2012). Unfortunately the resources of DFCs and IBMMSCs are limited. In a recent study, TDM with platelet-rich fibrin were planted into canine fresh tooth extraction sockets, and cementum–PDL-like tissue were observed (Ji et al., 2014). However, the explanation for this tissue generation without seed cells seems unclear. On the basis of these studies, we transplanted pTDM coated with three-layered CAs into the alveolar bones of minipigs. H&E and Masson’s trichrome staining showed that the autologous pPDLSC/pJBMMSC compound CAs regenerated bone-like tissue, with PDL-like fibres inserting onto surfaces of both J Tissue Eng Regen Med (2015) DOI: 10.1002/term

Regenerative microenvironment for periodontium tissue

Figure 6. Comparison of the osteogenic differentiation capacity of hPDLSCs directly co-cultured with hJBMMSCs/hIBMMSCs and hPDLSCs indirectly co-cultured with hJBMMSCs/hIBMMSCs. After compound CAs with (A, B) TDM and (C, D) CBB had been cultured, and hPDLSCs were indirectly co-cultured with (E) hJBMMSCs and (F) hIBMMSCs in Transwells by osteogenic induction for 3 and 14 days, the expressions of osteoblast-related genes were investigated by real-time PCR; expression levels were normalized to β-actin; data are shown as mean ± SD; *p < 0.05; n = 3

Table 2. A) Ectopic transplantation success rate of compound CAs with biomaterials in nude mice. B) Orthotopic transplantation success rate of compound CAs with biomaterials in minipigs

Group Implanted Died Lost Harvested Regenerated Success rate (%)

hJBMMSCs/ hPDLSCs with TDM

hIBMMSCs/ hPDLSCs with TDM

hJBMMSCs/ hPDLSCs with CBB

hIBMMSCs/ hPDLSCs with CBB

hJBMMSCs/ hPDLSCs with TDM

hIBMMSCs/ hPDLSCs with TDM

10 1 0 9 8 88.9

10 1 2 7 5 71.4

10 1 1 8 7 87.5

10 1 0 9 7 77.8

10 0 1 9 7 77.8

10 0 2 8 6 75.0

TDM and alveolar bone. These PDL-like fibres could integrate TDM and alveolar bone closely together. In the process of specific complex tissue regeneration, including hard and soft tissues, MSCs that are derived from the aimed regeneration tissues have the advantage to differentiate and engraft to the original tissues (Dimarino et al., 2013; Wang et al., 2011). These MSCs may Copyright © 2015 John Wiley & Sons, Ltd.

participate in tissue-specific regeneration and provide a microenvironment for tissue regeneration (Keating, 2012). Therefore, although it is believed that iliac-derived BMMSs have high ’stemness’ compared to jaw-derived BMMSCs, PDLSC/JBMMSC CAs regenerated bone–PDLlike complex tissue but not bone-like or collagen fibre tissues, respectively, under the regenerative J Tissue Eng Regen Med (2015) DOI: 10.1002/term

B. Zhu et al.

Figure 7. Regeneration of PDL–bone-like tissue around CBB in nude mice. (A) More bone-like matrix and inserting PDL-like fibres around CBB were observed in the hPDLSC/hJBMMSC compound CAs group by H&E and Masson’s trichrome staining. (B) Bone-like matrix and parallel fibres around CBB were observed in the hPDLSC/hIBMMSC compound CAs group by H&E and Masson’s trichrome staining; BLM, bone-like matrix; CBB, ceramic bovine bone; CF, collagen fibre; PLF, PDL-like fibres; scale bars = 50 and 100 μm

Figure 8. Regeneration of PDL/bone-like tissue around hTDM in nude mice. (A) More bone-like matrix and inserting PDL-like fibres around hTDM were observed in the hPDLSC/hJBMMSC compound CAs group by H&E and Masson’s trichrome staining. (B) Less bonelike matrix and parallel fibres around CBB were observed in the hPDLSC/hIBMMSC compound CAs group by H&E and Masson’s trichrome staining; TDM, treated dentine matrix; BLM, bone-like matrix; PLF, PDL-like fibres; CF, collagen fibre; scale bars = 50 and 100 μm

microenvironment reconstruction of PDLSCs and JBMMSCs. This new strategy may become a possible solution for periodontitis in oral clinic treatments. We successfully regenerated PDL-like fibres on treated dentine matrix and ceramic bovine bone Copyright © 2015 John Wiley & Sons, Ltd.

surface in both ectopic and orthotopic transplantation in vivo, and none of the bioscaffold fell off. Several problems remain unsolved, such as whether a TDM with regenerated cementum–PDL-bone-like tissue is strong enough to support occlusion forces, and whether J Tissue Eng Regen Med (2015) DOI: 10.1002/term

Regenerative microenvironment for periodontium tissue

Figure 9. Regeneration of PDL–bone-like tissue between pTDM and alveolar bone in minipig orthotopic transplantation. (A) More bone-like matrix and inserting PDL-like fibres on the TDM and jawbone surfaces were observed in the pPDLSC/pJBMMSC compound CAs group by H&E and Masson’s trichrome staining. (B) Less bone-like matrix and parallel fibres on the TDM and jawbone surfaces were observed in the pPDLSC/pIBMMSC compound CAs group by H&E and Masson’s trichrome staining; TDM, treated dentine matrix; BLM, bone-like matrix; PLF, PDL-like fibres; CF, collagen fibre; JB, jawbone; scale bars = 50 and 100 μm

this system is effective in an inflammatory microenvironment. These questions are our key points for research in future work. In conclusion, differently derived CAs of PDLSCs/ JBMMSCs can be used to reconstruct an appropriate regenerative microenvironment facilitating a more stable and regular regeneration of functional periodontium tissue. This method may provide a possible strategy for solving periodontium defects in periodontitis, and powerful experimental evidence for clinical applications in the future.

Conflict of Interest The authors declare no conflicts of interest.

Acknowledgements This work was supported by the National Basic Research Programme (973 Programme; Grant No. 2011CB964700) and the National Natural Science Foundation of China (Grant Nos 31030033, 31200972 and 81171001).

References Andrew K, Ekaputra GD, et al. 2011; The three-dimensional vascularization of growth factor-releasing hybrid scaffold of poly(3-caprolactone)/collagen fibers and hyaluronic acid hydrogel. Biomaterials 32: 8108–8117. Aubin JE, Liu F, Malaval L, et al. 1995; Gupta AK. Osteoblast and chondroblast differentiation. Bone 17: S77–83. Bartold PM, McCulloch CA, Narayanan AS, et al. 2000; Tissue engineering: a new paradigm for periodontal regeneration based on molecular and cell biology. Periodontology 24: 253–269. Caterson EJ, Nesti LJ, Danielson KG, et al. 2002; Human marrow-derived mesenchymal progenitor cells: isolation, culture expansion, and analysis of differentiation. Mol Biotechnol 20: 245–256. Chen J, Crawford R, Chen C, et al. 2013; The key regulatory roles of the PI3K/Akt signaling pathway in the functionalities of mesenchymal stem cells and applications in tissue regeneration. Tissue Eng B Rev 19: 516–528.

Copyright © 2015 John Wiley & Sons, Ltd.

Cicconetti A, Sacchetti B, Bartoli A, et al. 2007; Human maxillary tuberosity and jaw periosteum as sources of osteoprogenitor cells for tissue engineering. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 104: 618.e1–12. Dimarino AM, Caplan AI, Bonfield TL 2013; Mesenchymal stem cells in tissue repair. Front Immunol 4: 201. Ding G, Liu Y, Wang W, et al. 2010; Allogeneic periodontal ligament stem cell therapy for periodontitis in swine. Stem Cells 28: 1829–1838. Flores MG, Hasegawa M, Yamato M, et al. 2008; Cementum–periodontal ligament complex regeneration using the cell sheet technique. J Periodont Res 43: 364–371. Gottlow J, Nyman S, Karring T, et al. 1984; New attachment formation as the result of controlled tissue regeneration. J Clin Periodontol 11: 494–503. Huang GTJ, Gronthos S, Shi S 2009; Mesenchymal stem cells derived from dental tissues vs. those from other sources: their

biology and role in regenerative medicine. J Periodont Res 88: 792–806. Guo W, Gong K, Shi H, et al. 2012; Dental follicle cells and treated dentin matrix scaffold for tissue engineering the tooth root. Biomaterials 33: 1291–1302. Hynes RO 2002; Integrins: bidirectional, allosteric signaling machines. Cell 110: 673–687. Ji B, Sheng L, Chen G, et al. 2014; The combination use of platelet-rich fibrin and treated dentin matrix for tooth root regeneration by cell homing. Tissue Eng A 21: 26–34. Karring T, Nyman S, Lindhe J 1980; Healing following implantation of periodontitis affected roots into bone tissue. J Clin Periodontol 7: 96–105. Kashima TG, Nishiyama T, Shimazu K, et al. 2009; Periostin, a novel marker of intramembranous ossification, is expressed in fibrous dysplasia and in c-Fosoverexpressing bone lesions. Hum Pathol 40: 226–237. Katagiri H, Muneta T, Tsuji K, et al. 2013; Transplantation of aggregates of synovial

J Tissue Eng Regen Med (2015) DOI: 10.1002/term

B. Zhu et al. mesenchymal stem cells regenerates meniscus more effectively in a rat massive meniscal defect. Biochem Biophys Res Commun 435: 603–609. Kawaguchi H, Hirachi A, Hasegawa N 2004; Enhancement of periodontal tissue regeneration by transplant of bone marrow mesenchymal stem cells. J Periodont 75: 1281–1287. Keating A 2012; Mesenchymal stromal cells: new directions. Cell Stem Cell 10: 709–716. Kim SH, Turnbull J, Guimond S 2011; Extracellular matrix and cell signalling: the dynamic cooperation of integrin, proteoglycan and growth factor receptor. J Endocrinol 209: 139–151. Li F, Redick SD, Erickson HP, et al. 2003; Force measurements of the α5β1 integrin–fibronectin interaction. Biophys J 84: 1252–1262. Lin NH, Gronthos S, Bartold PM 2008; Stem cells and periodontal regeneration. Aust Dent J 53: 108–112. Liu Y, Liu W, Hu C, et al. 2011; MiR-17 modulates osteogenic differentiation through a coherent feed-forward loop in mesenchymal stem cells isolated from periodontal ligaments of patients with periodontitis. Stem Cells 29: 1804–1816. Matsubara T, Suardita K, Ishii M, et al. 2005; Alveolar bone marrow as a cell source for regenerative medicine: differences between alveolar and iliac bone marrow stromal cells. J Bone Miner Res 20: 399–409. Mehnert JM, Kisch T, Brandenburger M 2014; Co-culture systems of human sweat gland derived stem cells and peripheral nerve cells: an in vitro approach for peripheral nerve regeneration. Cell Physiol Biochem 34: 1027–1037.

Nyman S, Lindhe J, Karring T, et al. 1982; New attachment formation following surgical treatment of human periodontal disease. J Clin Periodontol 9: 290–296. Ona M, Wakabayashi N 2006; Influence of alveolar support on stress in periodontal structures. J Dent Res 85: 1087–1091. Paschos NK, Brown WE, Eswaramoorthy R, et al. 2014; Advances in tissue engineering through stem cell-based co-culture. J Tissue Eng Regen Med 9: 488–503. Pejcic A, Kojovic D, Mirkovic D, et al. 2013; Minic I. Stem cells for periodontal regeneration. Balkan J Med Genet 16: 7–12. Razavi S, Razavi MR, Kheirollahi-Kouhestani M, et al. 2013; Co-culture with neurotrophic factor secreting cells induced from adipose-derived stem cells: promotes neurogenic differentiation. Biochem Biophys Res Commun 440: 381–387. Rios HF, Ma D, Xie Y, et al. 2008; Periostin is essential for the integrity and function of the periodontal ligament during occlusal loading in mice. J Periodontol 79: 1480–1490. Seo BM, Miura M, Gronthos S, et al. 2004; Investigation of multipotent postnatal stem cells from human periodontal ligament. Lancet 364: 149–155. Shi S, Bartold PM, Miura M, et al. 2005; The efficacy of mesenchymal stem cells to regenerate and repair dental structures. Orthodont Craniofac Res 8: 191–199. Streuli CH, Akhtar N 2009; Signal cooperation between integrins and other receptor systems. Biochem J 418: 491–506. Szpalski C, Sagebin F, Barbaro M, et al. 2013; The influence of environmental factors on

bone tissue engineering. J Biomed Mater Res B Appl Biomater 101: 663–75. Takahashi H, Nakayama M, Shimizu T, et al. 2011; Anisotropic cell sheets for constructing three-dimensional tissue with wellorganized cell orientation. Biomaterials 32: 8830–8838. Tsumanuma Y, Iwata T, Washio K, et al. 2011; Comparison of different tissuederived stem cell sheets for periodontal regeneration in a canine one-wall defect model. Biomaterials 32: 5819–5825. Meretoja VV, Dahlin RL, Kasper FK, et al. 2012; Enhanced chondrogenesis in co-cultures with articular chondrocytes and mesenchymal stem cells. Biomaterials 33: 6362–6369. Wang L, Shen H, Zheng W, et al. 2011; Characterization of stem cells from alveolar periodontal ligament. Tissue Eng A 17: 1015–1026. Washio K, Iwata T, Mizutani M, et al. 2010; Assessment of cell sheets derived from human periodontal ligament cells: a preclinical study. Cell Tissue Res 341: 397–404. Xie H, Liu H 2012; A novel mixed-type stem cell pellet for cementum/periodontal ligament-like complex. J Periodontol 83: 805–815. Yang Y, Rossi FM, Putnins EE 2010; Periodontal regeneration using engineered bone marrow mesenchymal stromal cells. Biomaterials 31: 8574–8582. Zhou J, Shi S, Shi Y, et al. 2011; Role of bone marrow-derived progenitor cells in the maintenance and regeneration of dental mesenchymal tissues. J Cell Physiol 226: 2081–2090.

Supporting information on the internet The following supporting information may be found in the online version of this article: Figure S1. Identification of stem cell-related cell surface markers on hPDLSCs, hIBMMSCs and hJBMMSCs. (A–C) Detection of cell surface markers characteristic of mesenchymal stem cells on the three cell types. Analyses were performed via flow cytometry, detecting PE, FITC or APC conjugated monoclonal antibodies for human CD29, CD90, CD105, CD31, CD34, CD45, Stro-1 or isotype-matched control IgGs; CONT, control Figure S2. Comparison of proliferation capacity among hPDLSCs, hIBMMSCs and hJBMMSCs. (A) The original cell culture and passage 3 MSCs. (B) Analysis of the single-colony cluster formation capacity via a single-colony cluster stained with 0.1% toluidine blue in all the three MSCs. (C) For quantification of the single-colony clusters, five random fields were captured of each section. Data are expressed as the total number of single-colony clusters/group (×200); scale bar = 50 μm. (D) Indentification of similar proliferation rates of the three cell types by means of MTT assay Figure S3. ALP staining of hPDLSCs indirectly co-cultured with hJBMMSCS/hIBMMSCs. After hPDLSCs were indirectly co-cultured with hJBMMSCs and hIBMMSCs in Transwells in osteogenic induction for (A) 3 and (B) 14 days, ALP activity of the mineralized nodes was analysed by absorptiometry; data shown as mean ± SD; *p < 0.05; n = 3 Figure S4. Oral surgery and MSC CAs of minipig. (A) a hole was drilled on the alveolar bone of a minipig, MSC CA–TDM was prepared and the complex was implanted. (B) Original cell culture of minipig MSCs. (C) MSC CAs were obtained

Copyright © 2015 John Wiley & Sons, Ltd.

J Tissue Eng Regen Med (2015) DOI: 10.1002/term