Original Article

TISSUE ENGINEERING: Part A Volume 00, Number 00, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/ten.tea.2014.0221

Engineering the Periodontal Ligament in Hyaluronan–Gelatin–Type I Collagen Constructs: Upregulation of Apoptosis and Alterations in Gene Expression by Cyclic Compressive Strain Aarthi Saminathan, PhD,1 Gopu Sriram, PhD,1 Jayasaleen Kumar Vinoth, PhD,1,2 Tong Cao, DDS, PhD,1 and Murray C. Meikle, DDSc, PhD1,3

To engineer constructs of the periodontal ligament (PDL), human PDL cells were incorporated into a matrix of hyaluronan, gelatin, and type I collagen (COLI) in sample holders (13 · 1 mm) of six-well Biopress culture plates. The loading dynamics of the PDL were mimicked by applying a cyclic compressive strain of 33.4 kPa (340.6 gm/cm2) to the constructs for 1.0 s every 60 s, for 6, 12, and 24 h in a Flexercell FX-4000C Strain Unit. Compression significantly increased the number of nonviable cells and increased the expression of several apoptosis-related genes, including initiator and executioner caspases. Of the 15 extracellular matrix genes screened, most were upregulated at some point after 6–12 h deformation, but all were downregulated at 24 h, except for MMPs1–3 and CTGF. In culture supernatants, matrix metalloproteinase-1 (MMP-1) and tissue inhibitor of metalloproteinases-1 (TIMP-1) protein levels were upregulated at 24 h; receptor activator of nuclear kappa factor B (RANKL), osteoprotegerin (OPG) and fibroblast growth factor-2 (FGF-2) were unchanged; and connective tissue growth factor (CTGF) not detected. The low modulus of elasticity of the constructs was a disadvantage—future mechanobiology studies and tissue engineering applications will require constructs with much higher stiffness. Since the major structural protein of the PDL is COLI, a more rational approach would be to permeabilize preformed COLI scaffolds with PDL-populated matrices.

dimensional extracellular network of collagens, proteoglycans, and noncollagenous proteins, and to address this deficiency, hydrogels have been widely used as scaffolds to create a variety of three-dimensional tissue constructs.10,11 Not only do hydrogels have material properties that mimic more closely the environment experienced by cells in vivo, but they also confer beneficial effects on gene expression, cell adhesion, and phenotype.12–14 Nevertheless, there are some formidable problems to overcome in successfully engineering three-dimensional tissue constructs, including not only the replication of tissue complexity, but also the mechanical and viscoelastic characteristics of the native tissue.15,16 To create a tissue more closely resembling the structure of the ligament, we have seeded human PDL cells into a commercially-available hydrogel composed of hyaluronan (HA) and gelatin (GLN) marketed as Extracel,17,18 in formats designed for use with the Flexercell system for mechanically deforming cells. Our aim has been to engineer constructs with optimal biophysical properties for studies of

Introduction

T

he periodontal ligament (PDL) is a thin fibrous connective tissue, embedded between the cementum covering the roots of the teeth and supporting alveolar bone that evolved in mammals to provide for the eruption and attachment of teeth to the bones of the jaws.1 Together with its supporting bone, the PDL develops from the dental follicle, a connective tissue composed of ectomesenchymal cells surrounding the developing roots of the teeth derived from the cephalic neural crest.2 In addition to its attachment role, the PDL also serves as a barrier, preventing the fusion or ankylosis of teeth to the bone—the consequence of which is progressive resorption of the cementum and dentine of the root, and its replacement by bone as the adjoining osseous tissue undergoes normal remodeling activity.3 Since the PDL functions in a mechanically active environment, two-dimensional culture systems have been widely used to investigate the mechanobiology of PDL cells.4–9 However, cells are normally embedded in a complex three1 2 3

Faculty of Dentistry, National University of Singapore, Singapore, Singapore. National Dental Centre, Singapore, Singapore. Faculty of Dentistry, University of Otago, Dunedin, New Zealand.

1

2

cell–cell and cell–matrix interactions in tooth support, and in the longer-term, the regeneration of a functional PDL in vivo—an essential requirement for the success of alveolar bone augmentation procedures. While we have shown that thin films (80–100 mm) of PDL/HA-GLN hydrogel are suitable for studies of tensile mechanical strain,19 thicker constructs are required for investigating the effects of compression. To address this prerequisite and increase the complexity of the scaffold, we have incorporated type I collagen (COLI) into the HA-GLN matrix to provide extra Arg-GlyAsp (RGD) binding sites on the triple helices of intact collagen. In this study, we report on the effect of cyclic compressive strain on the viability and biological activity of PDL cells cultured in three-dimensional HA-GLN-COLI matrices, and make suggestions regarding the composition of the next generation of artificial PDL constructs. Materials and Methods Preparation of PDL cell/hydrogel constructs

Primary human PDL cells were purchased from a commercial source (Lot number: 5145; ScienCell Research Laboratories, Carlsbad, CA). The cells were isolated from tissue samples from several donors, cryopreserved at passage 1, and on delivery on dry ice immediately transferred to liquid nitrogen. The cells were subsequently thawed and expanded in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Gibco, Invitrogen, Singapore, Singapore) and antibiotic–antimycotic reagents as described below. Expanded cells were harvested, aliquoted, and stored in liquid N2 until required and passage three cells used in all experiments. To form the constructs, rat tail COLI (Sigma-Aldrich, Singapore, Singapore) at a concentration of 1.3 mg/mL was mixed with Extracel (Glycosan Biosystems, Salt Lake City, UT) a mixture of Glycosil (thiol-modified

FIG. 1. Biopress culture plates for applying compression to cell and tissue samples. Circular foam discs attached to a rubber piston confine the constructs to the central region of each well. The constructs were cultured in 5 mL medium and compression applied from below by compressed air, the constructs being restrained by a stationary platen as shown in the top two and middle left wells. Color images available online at www.liebertpub .com/tea

SAMINATHAN ET AL.

HA) and Gelin-S (thiol-modified GLN). The three components of Extracel that come as lyophilized solids and DG Water (degassed, deionized water) were used to dissolve the Glycosil, Gelin-S, and the cross-linking agent Extralink (PEGDA) in individual vials. COLI, Glycosil, and Gelin-S were mixed in the ratio of 1:1:1 and the PDL cells added. To form the hydrogel, Extralink was added to the Glycosil–Gelin-S–COLI mix at a ratio of 1:6 and after blending in a pipette, 300 mL of the gel mixture (containing 5.0 · 106/mL cells) pipetted into the sample holders (foam rings; internal diameter 13 mm) of six-well Biopress plates (Flexcell International Corporation, Hillsborough, NC), and incubated at 37C; gelation occurs after about 20 min (Fig. 1). The constructs were then incubated for 2 weeks before loading in 5 mL DMEM (Gibco, Invitrogen) supplemented with 10% fetal calf serum (Gibco), antibiotic–antimycotic reagent (10,000 units penicillin, 10,000 mg streptomycin, and 25 mg/ mL amphotericin B; Invitrogen), 100 mM l-glutamine (Invitrogen) and gentamicin (10 mg/mL; Gibco) at 37C in a humidified atmosphere of 5% CO2/95% air. Application of compressive mechanical strain to PDL constructs

The circular foam discs attached to a rubber piston confine the constructs to the central region of each well and compression is applied from below by positive air pressure, the gel disc being compressed between the piston and a platen fitting over each well (Fig. 1). The minimum force recommended by the Flexcell International Corporation for operating the Flexercell FX-4000C unit is 0.5 lbs; in view of the likelihood that about half the applied strain will be transmitted to the cells,20 a force of 0.453 kg was programmed into the computer. Based on a formula supplied by the Flexcell International Corporation this translates into a compressive strain of 33.4 kPa (340.6 gm/cm2). This was applied for 1.0 s

ENGINEERING THE PERIODONTAL LIGAMENT IN 3D CONSTRUCTS

(1 Hertz) every 60 s, for 6, 12, and 24 h, to Biopress plates clamped in a standard Bioflex baseplate, linked to a Flexercell FX-4000C Compression Plus Strain Unit.

3

2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA); the RNA integrity number (RIN) values were 9.90. Gene expression using real-time PCR

Analysis of cell viability

The effect of compression on cell viability was determined by the fluorescein diacetate (FDA)–propidium iodide (PI) method.21 At the end of each time point the foam ring was removed and the gel carefully removed with a spatula and transferred to a 60 mm Petri dish, and stained with 0.1 mL (2 mg) FDA and 0.3 mL (6 mg) PI from stock solutions of both dyes (Sigma-Aldrich) in the dark. Constructs were then viewed with a Carl Zeiss, LSM 510 META confocal imaging system—viable cells fluoresce bright green, nonviable cells bright red. Three triangulated fields (1200 · 1200 mm) were chosen for analysis, and the number of viable and nonviable cells counted using the Image-Pro Plus (version 6.1.0.346) software (Media Cybernetics, Bethesda, MD). Caspase 3/7 assay

Caspases are a family of cysteine proteinases that specifically cleave protein substrates within the cell to trigger apoptosis. Two executioner caspases (3 and 7) were measured by the Caspase-Glo 3/7 Assay (Promega Corporation, Madison, WI), which provides a luminescent Caspase 3/7 substrate containing the tetrapod sequence (Asp-Glu-ValAsp) in a reagent optimized for caspase activity, luciferase activity, and cell lysis. At the end of each time point the constructs were removed and treated with the Caspase Glo 3/7 reagent for 1 h in the dark at room temperature according to the manufacturer’s instructions. Each sample was aliquoted into triplicate tubes and luminescence, expressed as relative light units, measured with a Sirius Single Tube Luminometer (Berthold Detection Systems GmbH, Pforzheim, Germany). Cell recovery from Extracel

At the end of the experimental period the culture medium was aspirated from each well and the hydrogel surface washed with phosphate buffered saline. Based on a protocol developed and recommended by Glycosan Biosystems, 1.5 mL trypsin-EDTA (Invitrogen) was added and incubated at 37C for 3 h. After a further wash, 1.5 mL of 10X collagenase/hyaluronidase in DMEM (StemCell Technologies, Singapore, Singapore) was added and incubated overnight at 37C. At the end of the second incubation, 3 mL DMEM supplemented with 10% FBS was added to the hydrolyzed gel and the cell suspension centrifuged at 1200 rpm for 5 min. The resulting pellet was resuspended in 1.5 mL DMEM and centrifuged again at 10,000 rpm for 2 min. RNA extraction

Total mRNA extraction was carried out using an RNeasy Plus Mini kit (Qiagen, Singapore, Singapore) according to the manufacturer’s instructions. This yielded RNA with an A260/280 value > 1.95 and an A230/280 value > 1.2 in a NanoDrop 2000 Spectrophotometer (Thermo Scientific, Wilmington, DE), indicative of a pure sample. Integrity of the RNA was assessed by gel electrophoresis on an Agilent

A known amount (500 ng) of total RNA was reverse transcribed using the iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA) and a My Cycler thermal cycler (Bio-Rad). Real-time polymerase chain reaction (PCR) was performed in triplicate using Fast SYBR Green Master Mix (Applied Biosystems, Singapore, Singapore) with a StepOnePlus Real-Time PCR System (Applied Biosystems). The primer sequences (First Base, Singapore, Singapore) and the annealing temperatures used to screen the samples for extracellular matrix and apoptosis gene expression are listed in Tables 1 and 2. The amplification was carried out through the first step at 95C for 10 min, followed by 40 cycles with 15 s at 95C, 10 s at the particular annealing temperature, and 20 s at 72C. The fluorescence signal was acquired at 72C. The relative expression of these genes was normalized to the expression of two calibrator genes: b-actin (ACTB) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using the standard curve method. Critical threshold (Ct) values were calculated using the StepOnePlus Version2.1 software (Applied Biosystems); a gene was excluded if it had a threshold value exceeding 33–35. Relative quantitation (RQ) was used to measure differences in expression levels of specific target genes and represented as fold-change (Applied Biosystems). RQ values >1.00 signify an increase in gene expression by compressed cells over controls. Culture supernatant proteomics

Conditioned media were harvested and enzyme-linked immunosorbent assays (ELISAs) used to quantitate levels of: (1) a selection of proteolytic enzymes, matrix metalloproteinase-1 (MMP-1; collagenase-1), MMP-2 (gelatinase-A), MMP-3 (stromelysin-1), and tissue inhibitor of metalloproteinases-1 (TIMP-1) obtained from (R&D Systems China, Shanghai, China); (2) The cell–cell signaling molecules receptor activator of nuclear kappa factor B (RANKL) and osteoprotegerin (OPG) from Cusabio Biotech (Wuhan, China); (3) The growth factors connective tissue growth factor (CTGF) from Aviscera Bioscience (Santa Clara, CA) and fibroblast growth factor-2 (FGF-2) from R&D Systems China. Each protein was assayed according to the manufacturer’s instructions. Statistical analysis

Significant differences between compressed and control cultures were determined by Student’s t-test (two-tailed) using the GraphPad Prism (GraphPad Software, Inc., San Diego, CA) with the level of significance set at p < 0.05. To compare differences in mean levels of gene expression across the 6–24 h time scale, a one-way analysis of variance (ANOVA) was used. Results

Our first attempts at engineering PDL constructs for compression studies involved incorporating PDL cells into Extracel in a circular format (13 · 1.0 mm), designed to fit the Biopress culture plates. FDA-PI staining and confocal

4

SAMINATHAN ET AL.

Table 1. Extracellular Matrix Genes and Primer Sequences Used for Real-Time RT-PCR HGNC gene symbol

Description

ACTB

Actin, beta

GAPDH P4HB

Glyceraldehyde-3-phosphate dehydrogenase Prolyl-4-hydroxylase, beta subunit

RUNX2

Runt-related transcription factor 2

SOX9

SRY (sex determining region Y)-box 9

PPAR-c MYOD

Peroxisome proliferator-activated receptor-gamma Myogenic differentiation antigen1

COL1A1

Collagen type I, alpha-1

COL2A1

Collagen type II, alpha-1

COL3A1

Collagen type III, alpha-1

MMP-1

TIMP-1

Matrix metalloproteinase-1; collagenase-1 Matrix metalloproteinase-2; gelatinase A (72 kDa) Matrix metalloproteinase-3; stromelysin-1 Tissue inhibitor of metalloproteinases-1

TGFB1

Transforming growth factor, beta-1

BGLAP SP7

Gamma carboxyglutamic acid protein; osteocalcin Transcription factor Sp7; osterix

BMP2

Bone morphogenetic protein 2

TNFSF11

Tumor necrosis factor ligand superfamily, member 11; RANKL TNF receptor superfamily member 11B; osteoprotegerin

MMP-2 MMP-3

TNFRSF11B

Primer sequences F: CCAAGGCCAACCGCGAGAAGATGAC R: AGGGTACATGGTGGTGCCGCCAGAC F: ACCACAGTCCATGCCATCAC R: TCCACCACCCTGTTGCTGTA F: GTCTTTGTGGAGTTCTATGCCC R: GTCATCGTCTTCCTCCATGTCT F: TGAGAGCCGCTTCTCCAACC R: GCGGAAGCATTCTGGAAGGA F: GAACGCACATCAAGACGGAG R: TCTCGTTGATTTCGCTGCTC F: ATTGACCCAGAAAGCGATTC R: CAAAGGAGTGGGAGTGGTCT F: CGGCGGAACTGCTACGAAG R: GCGACTCAGAAGGCACGTC F: GAACGCGTGTCATCCCTTGT R: GAACGAGGTAGTCTTTCAGCAACA F: TTCAGCTATGGAGATGACAATC R: AGAGTCCTAGAGTGACTGAG F: AACACGCAAGGCTGTGAGACT R: GCCAACGTCCACACCAAATT F: GGGAGATCATCGGGACAACTC R: GGGCCTGGTTGAAAAGCAT F: TGATCTTGACCAGAATACCATCGA R: GGCTTGCGAGGGAAGAAGTT F: TGGCATTCAGTCCCTCTATGG R: AGGACAAAGCAGGATCACAGTT F: CTGTTGTTGCTGTGGCTGATA R: CCGTCCACAAGCAATGAGT F: GCAACAATTCCTGGCGATACCTC R: AGTTCTTCTCCGTGGAGCTGAAG F: ATGAGAGCCCTCACACTCCTC R: GCCGTAGAAGCGCCGATAGGC F: TGGCGTCCTCTCTGCTTGA R: TCAGTGAGGGAAGGGTGGGT F: CAGAGACCCACCCCCAGCA R: CTGTTTGTGTTTGGCTTGAC F: TCCCATCTGGTTCCCATAAA R: GGTGCTTCCTCCTTTCATCA F: CGTCAAGCAGGAGTGCAATC R: CCAGCTTGCACCACTCCAA

Annealing temperature 58 60 62 58 58 62 60 60 58 60 60 60 60 60 60 60 60 58 60 60

RT-PCR, reverse transcription-polymerase chain reaction.

microscopy showed that on encapsulation, the cells adopted a rounded morphology and retained high viability over a 4week time course (Fig. 2). Nevertheless, it was clear from side views that the body of the matrix was relatively sparsely populated. Only at the hydrogel surface and the gel– substrate interface did the cells appear to achieve confluence and assume a fibroblastic phenotype, suggesting the HA-GLN scaffold was insufficiently stiff to facilitate cell attachment (Fig. 3A). We, therefore, added type I rat tail collagen to provide additional RGD binding sites on the triple-helices of intact collagen, and this enabled the cells to populate the full thickness of the gel (Fig. 3B). However, by the third and fourth weeks of culture, in common with all collagen-containing hydrogels, the constructs showed a tendency to contract. Two-week HA-GLN-COLI constructs were, therefore, used in all loading experiments which minimized this complication, although we did find that owing to their low modulus of elasticity, by the end of each

time course, the compression regimen had reduced the thickness of the constructs to around 300–400 mm. Cyclic compressive strain resulted in a significant increase in the number of nonviable cells after 12 and 24 h loading; however, it was only possible to count the dead cells—the dense fibrillary growth of the viable cells made their quantification impossible (Figs. 4,5). The increase in the number of cells undergoing programmed cell death was confirmed by the Caspase 3/7 assay, which was significantly upregulated across all three time points (Fig. 6). Real-time reverse transcription-polymerase chain reaction (RT-PCR) was then used to screen a panel of 18 extracellular matrix-related genes expressed by PDL and mesenchymal stem cells (MSCs). We found that while the osteoblast-specific transcription factor RUNX2 and its chondrogenic counterpart SOX9 were initially expressed, and in the case of SOX9, significantly upregulated at 12 h. However, they were not detectable in 24 h cultures, and the adipocytes-

ENGINEERING THE PERIODONTAL LIGAMENT IN 3D CONSTRUCTS

5

Table 2. Apoptosis Genes and Primer Sequences Used for Real-Time RT-PCR HGNC gene symbol

Description

ACTB

Actin, beta

GAPDH

IL1A

Glyceraldehyde-3-phosphate dehydrogenase Tumor necrosis factor, alpha cachectin Tumor necrosis factor receptor superfamily, member 1A Tumor necrosis factor receptor superfamily, member 1B Tumor necrosis factor receptor (ligand) superfamily, member 10 Interleukin 1-alpha

IL1B

Interleukin 1-beta

IL6

Interleukin 6 (interferon, beta 2)

CASP1

Caspase 1, apoptosis-related cysteine proteinase Caspase 2, protein phosphatase 1, regulatory subunit 57 Caspase 3, apoptosis-related cysteine proteinase Caspase 6, apoptosis-related cysteine proteinase Caspase 7, apoptosis-related cysteine proteinase Caspase 8, apoptosis-related cysteine proteinase Caspase 9, protein phosphatase 1, regulatory subunit 56 Caspase 10, apoptosis-related cysteine proteinase

TNF TNFRSF1A TNFRSF1B TNFSF10

CASP2 CASP3 CASP6 CASP7 CASP8 CASP9 CASP10

Primer sequence F: CCAAGGCCAACCGCGAGAAGATGAC R: AGGGTACATGGTGGTGCCGCCAGAC F: ACCACAGTCCATGCCATCAC R: TCCACCACCCTGTTGCTGTA F: GAGCACTGAAAGCATGATCC R: CGAGAAGATGATCTGACTGCC F: TGCCTACCCCAGATTGAGAA R: ATTTCCCACAAACAATGGAGTAG F: GCGAGTTACAGCTAAAGCAT R: TTATGCAACATGACCTTGGA F: GAGCTGAAGCAGATGCAGGAC R: TGACGGAGTTGCCACTTGACT F: CCCAACACCTGGACCTCGGC R: CGTAGGCACGGCTCCTCAGC F: AAGCTGAGGAAGATGCTG R: ATCTACACTCTCCAGCTG F: TGCGTCCGTAGTTTCCTTCT R: GCCTCAGACATCTCCAGTCC F: AATACTGTCAAATTCTTCATTGCAGATAAT R: AAGTCGGCAGAGATTTATCCAATAA F: GCACTGAGGGAGACCAAGCA R: CACAGCTCAACGGTGGGAGTA F: AGAACTGGACTGTGGCATTGAG R: GCTTGTCGGCATACTGTTTCAG F: GGCAGTTCCCTGGAGTTCAC R: GACCTTCCTGTTCACCAGCG F: AGTGACAGGTATGGGCGTTCG R: GCATCTATCCCCCCTAAAGTGG F: CTCCCCAAACTTGCTTTATG R: AAGACCCCAGAGCATTGTTA F: CGAACTAACAGGCAAGCAGC R: ACCTCACCAAATCCTCCAGAAC F: AATCTGACATGCCTGGAG R: ACTCGGCTTCCTTGTCTAC

and myoblast-specific transcription factors PPARG and MYOD were not detected at any time point. P4HB a gene abundantly expressed by cells synthesizing collagen and a useful marker for the fibroblast phenotype was expressed at 6 and 12 h at RQ values either side of 1.00, but had declined by

Annealing temperature 58 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60

24 h (Table 3). BGLAP, the gene encoding the bone matrix protein osteocalcin, SP7 (Osterix) the osteoblast transcription factor acting downstream of RUNX2, and the cartilagespecific collagen gene COL2A1, all failed to be identified (Table 3). Of the 15 extracellular matrix genes screened, 12

FIG. 2. Cell viability in Extracel. Human periodontal ligament (PDL) cells (5.0 · 106/mL) were incorporated into 300 mL of Extracel in 13 mm plastic rings and three discs cultured in 5 cm Petri dishes. At the end of each time point the culture supernatants were discarded and the cells stained with 0.1 mL (2 mg) fluorescein diacetate (FDA) and 0.3 mL (6 mg) propidium iodide (PI) from stock solutions of both dyes (Sigma-Aldrich) in the dark and viewed with a Carl Zeiss, LSM 510 META confocal imaging system; viable cells (green); nonviable cells (red). Three fields were selected from three wells and the number counted using the Image-Pro Plus (version 6.1.0.346) software. Color images available online at www .liebertpub.com/tea

6

SAMINATHAN ET AL.

FIG. 3. Profile images of human PDL cells in three-dimensional constructs, captured with a Carl Zeiss, LSM 510 META confocal imaging system. (A) Human PDL cells in hyaluronan (HA)–GLN constructs; the body of the matrix is sparsely populated with most cells aggregating at the gel surface and the gel– substrate interface. (B) With the addition of type I rat tail collagen providing additional Arg-Gly-Asp binding sites, the cells are evenly distributed throughout the full thickness of the gel matrix. Viable cells fluoresce bright green and nonviable cells bright red. Color images available online at www.liebertpub .com/tea were detected at Ct values < 35. At 6 h compressive strain significantly increased the expression of COL1A1, MMP-1, and MMP-3 by RQ values > 2.00, and in the case of CTGF > 3.00 at both 6 and 12 h; significant differences in RQ values for P4HB, COL1A1, COL3A1, MMP-1, and BMP2 at various points across the time scale were also detected. Although most genes were upregulated at some point with RQ values > 1.00 after 6 h and/or 12 h deformation, all were downregulated in the 24 h cultures, except for the three major extracellular matrix-degrading enzymes MMPs1–3 and the matricellular protein CTGF (Table 3).

FIG. 4. The effect of compression on cell viability. At the end of each time point the foam ring was removed and the gel carefully removed with a spatula and transferred to a 60 mm Petri dish, and stained with FDA and PI in the dark. Constructs were then viewed with a Carl Zeiss, LSM 510 META confocal imaging system. Viable cells fluoresce bright green and nonviable cells bright red. Each field measures 1200 · 1200 mm. (A) Control; (B) 6 h cyclic compression; (C) 12 h cyclic compression; (D) 24 h cyclic compression. Color images available online at www .liebertpub.com/tea

To further investigate the molecular basis for the increase in nonviable cells, we next measured the expression of 15 genes involved in apoptosis (Table 2). Insufficient RNA remained for more than two runs, which precluded the use of a statistical analysis; these findings although informative are, therefore, provisional. TNFRSF1A, TNFSF10 (TRAIL), and IL1A were all upregulated after 24 h, the latter two by RQ values > 2.00. TNF, IL1B, and the caspase inhibitor gene IL6 in contrast, were downregulated at the same time point; the other member of the tumor necrosis factor (TNF) receptor superfamily TNFRSF1B was not detected. Each of

ENGINEERING THE PERIODONTAL LIGAMENT IN 3D CONSTRUCTS

7

Table 3. Effect of Cyclic Compressive Strain on Gene Expression by PDL Cells in HA-GLN-COLI Constructs 6H

FIG. 5. The effect of cyclic compression on the number of dead cells. A compressive force of 33.4 kPa (340.6 gm/cm2) was applied for 1.0 s (1 Hertz) every 60 s in a rectangular wave form to PDL/HA-gelatin (GLN)-type I collagen (COLI) constructs for 6, 12, and 24 h in six-well Biopress plates linked to a Flexercell FX-4000C Compression Plus Strain Unit. At the end of each time point the constructs were removed, the cells stained with FDA and PI, and viewed with a Carl Zeiss LSM 510 META confocal imaging system. Three fields were selected from three wells and the number of nonviable cells counted using the Image-Pro Plus software. The data are cross-sectional and expressed as mean – SEM. *p < 0.05. the eight caspases (cysteine-dependent aspartate-directed proteinases) involved in the apoptosis signaling cascade showed a progressive increase in expression (Table 4). These included the interleukin-1-converting enzyme CASP1, and the initiator caspases CASP2, CASP8, CASP9, and CASP10,

P4HB RUNX2 SOX9 PPARG MYOD COL1A1 COL2A1 COL3A1 MMP-1 MMP-2 MMP-3 TIMP-1 TGFB1 BGLAP SP7 BMP2 RANKL OPG CTGF FGF-2

a

1.05 – 0.14 0.43 – 0.07 0.76 – 0.19 ND ND 2.89 – 0.54c ND 1.61 – 0.23c 2.20 – 0.28c 1.51 – 0.20 2.41 – 0.98 0.71 – 0.16 0.50 – 0.14 ND ND 1.41 – 0.08c 1.39 – 0.39 0.42 – 0.16 3.15 – 0.56 1.56 – 0.23

12 H

24 H

0.96 – 0.11 0.96 – 0.09a 1.60 – 0.18b ND ND 1.08 – 0.15 ND 1.20 – 0.09b 0.45 – 0.12 0.81 – 0.16 1.08 – 0.19 0.78 – 0.13 1.41 – 0.66 ND ND 1.17 – 0.07c 0.82 – 0.23 1.13 – 0.51 3.09 – 0.97 1.77 – 0.63

0.54 – 0.05 ND ND ND ND 0.93 – 0.29 ND 0.31 – 0.04 1.12 – 0.23b 1.23 – 0.21 1.30 – 0.42 0.59 – 0.06 0.47 – 0.06 ND ND 0.34 – 0.08 ND 0.38 – 0.11 1.92 – 0.56 0.63 – 0.21

Human PDL cells in three-dimensional constructs were subjected to a compressive strain of 32.4 kPa (340.6 gm/cm2) for 1.0 s (1 Hertz) every 60 s for 6–24 h, and gene expression quantified by realtime RT-PCR. The data are cross-sectional, expressed as relative quantity and represents the mean – SEM of four separate determinations; RQ values >1.00 signify an increase in gene expression by compressed cells over controls. A one-way ANOVA was used to compare mean levels of gene expression across the time course. a p < 0.05. b p < 0.01. c p < 0.001. ANOVA, analysis of variance; COLI, type I collagen; GLN, gelatin; HA, hyaluronan; ND, not detected, PDL, periodontal ligament; RQ, relative quantitation.

Table 4. Effect of Cyclic Compressive Strain on the Expression of a Panel Apoptosis-Related Genes

TNF TNFRSF1A TNFRSF1B TNFSF10 IL1A IL1B IL6 CASP1 CASP2 CASP3 CASP6 CASP7 CASP8 CASP9 CASP10 FIG. 6. The effect of cyclic compression on the activity of two executioner caspases (caspase 3 and caspase 7). A compressive strain of 32.4 kPa (340.6 gm/cm2) was applied for 1.0 s (1 Hertz) every 60 s to PDL/HA-GLN-COLI constructs for 6, 12, and 24 h in six-well Biopress plates. Data are cross-sectional and expressed in relative light units as means – SEM for six wells. ***p < 0.001.

6H

12 H

24 H

0.28 0.26 ND 0.81 0.64 0.39 1.17 0.53 0.68 0.35 0.23 0.40 0.67 0.49 0.24

1.56 0.38 ND 0.70 0.52 1.17 0.60 1.41 0.96 0.73 1.08 0.49 0.55 0.20 0.68

0.24 1.24 ND 2.45 2.00 0.74 0.71 1.76 1.48 0.89 1.17 1.03 1.59 1.11 2.18

Human PDL cells in three-dimensional constructs were subjected to a compressive strain of 32.4 kPa (340.6 gm/cm2) for 1.0 s (1 Hertz) every 60 s for 6–24 h, and gene expression quantified by realtime RT-PCR. The data are cross-sectional, expressed as relative quantity and the mean of two separate determinations; RQ values > 1.00 signify an increase in gene expression by compressed cells over controls.

8

SAMINATHAN ET AL.

FIG. 7. Matrix metalloproteinase (MMP) and tissue inhibitor of metalloproteinases-1 (TIMP-1) synthesis by PDL/HA-GLNCOLI constructs subjected to a compressive strain of 32.4 kPa applied for 1.0 s (1 Hertz) every 60 s for 6, 12, and 24 h in six-well Biopress plates. At each endpoint the culture media from six wells were pooled in pairs and assayed in triplicate. The antibodies to MMP-1 recognize pro-MMP-1 only, not active enzyme or MMP-1 complexed to TIMPs. The TIMP-1 immunoassay only recognizes natural TIMP-1. The data are cross-sectional and expressed as means – SEM. *p < 0.05. which cleave the inactive proforms of the executioner caspases CASP3, CASP6, and CASP7; these in turn cleave other intracellular protein substrates to trigger the apoptotic process. Interestingly, although CASP3 and CASP7 expression were increased across the time scale, CASP3 was the only caspase gene not to reach an RQ value > 1.00 at 24 h and CASP7 only just (Table 4). Protein levels of all three MMPs plus TIMP-1 were identified by ELISAs in culture supernatants from both experimental and control cultures, although the concentrations of MMP-2 and MMP-3 were low in comparison with MMP1 and TIMP-1. Both of these proteins were significantly upregulated by cyclic compressive strain (Fig. 7), although after 24 h, the concentration of free TIMP-1 (760 ng/mL) in compressed cultures was more than twice that of the latent pro-form of MMP-1 (360 ng/mL). RANKL, OPG, and FGF2 were identified in all cultures, but remained unchanged by compressive strain, although the concentration of OPG (ng/ mL) was an order of magnitude greater than either RANKL or FGF-2 (pg/mL). CTGF protein, in contrast, despite RQ values > 2–3.00 could not be detected in culture supernatants from either compressed or control constructs (Table 5).

Discussion

Two-dimensional in vitro culture systems have provided valuable information regarding cell behavior at its most reductionist level. However, they lack the physical and other environmental signals provided by three-dimensional microenvironments that are important determinants of phenotype, proliferation, gene expression, and matrix turnover—in other words, information derived from two-dimensional models is unlikely to accurately reflect the complex physiology of cells in vivo.22 The present study, therefore, represents a significant advance on previous reports of the effect of compressive strain on the mechanobiology of PDL cells, which have been limited by culturing PDL cells, either in the form of two-dimensional monolayers6,23,24 or thin COLI films,25,26 and by applying static loads to mechanically deform the cells—essentially variations on a weighted surface pressing down on the cells. In such model systems, as well as those subjected to tensile strain, the cells will interact with substrates that are much stiffer than normally experienced in vivo, which, given the key role of integrin receptors in mediating outside-in signaling at focal

ENGINEERING THE PERIODONTAL LIGAMENT IN 3D CONSTRUCTS

Table 5. Effect of Cyclic Compressive Strain on Cytokines and Growth Factors in Culture Supernatants Protein

6H

RANKL (pg/mL) Control 73.69 – 1.20 Experimental 62.50 – 4.18 OPG (ng/mL) Control 82.28 – 5.39 Experimental 91.94 – 8.36 FGF-2 (pg/mL) Control 14.40 – 0.73 Experimental 14.92 – 0.83 CTGF Con/Exp ND

12 H

24 H

61.66 – 2.41 59.88 – 2.27

62.97 – 2.89 64.64 – 4.30

82.98 – 4.49 82.42 – 7.92 94.38 – 1.26 106.21 – 6.25 14.01 – 0.71 15.28 – 0.08

13.29 – 0.20 16.07 – 1.01

ND

ND

Human PDL cells in three-dimensional constructs were subjected to cyclic compressive strain for 6–24 h in six-well Biopress plates. At each time point the culture media from six wells were pooled in pairs and assayed in triplicate. The data are cross-sectional and expressed as means – SEM. No significant differences in individual protein levels were detected over the 6–24 h time course, or between experimental and control cultures. CTGF, connective tissue growth factor; FGF-2, fibroblast growth factor-2; ND, not detected; OPG, osteoprotegerin; RANKL, receptor activator of nuclear kappa factor B.

adhesions,27–29 will significantly influence their response to mechanical loading. We have previously shown that PDL cells can be incorporated into HA–GLN hydrogel films (80–100 mm) with high viability; the cells remained rounded over the first few days and gradually acquired a fibroblastic phenotype.19 However, when incorporated into gels 1 mm in thickness, we found that the addition of type I rat tail collagen was required to enable the cells populate the full thickness of the gel. To mimic the mechanical environment of the PDL provided by mastication and other forms of occlusal loading, we then applied a cyclic compressive strain to the constructs and investigated its effect on apoptosis and the expression of a selection of genes involved in matrix turnover. Apoptosis, a form of programmed cell death in which cells are induced to activate their own death or suicide, plays an important role in many pathophysiological processes, including embryonic development, tissue remodeling, and homeostasis.30,31 The increase in the number of dead cells and Caspase 3/7 activity, plus preliminary data on the expression of a number of apoptosis-related genes, including four initiator (CASP2, CASP8, CASP9, and CASP10) and three executioner (CASP3, CASP6, and CASP7) caspases,32 strongly suggests that apoptotic signaling pathways had been upregulated by compressive strain. This contrasts with the effect of cyclic tensile strain on cultured PDL cells in twodimensional culture, which we have previously shown to have a positive, albeit transient effect on apoptosis,9 unless more vigorous strain regimens are applied.33–37 In other words, the induction of apoptosis in mechanically deformed cells appears to be related as one might expect, to the magnitude and frequency of the applied strain, whether it is largely compressive or tensile, and in the case of the PDL, will vary depending on its anatomical relationship with the roots of the various teeth and their supporting bone. The

9

further investigation of the relationship between mechanical strain and apoptosis-related genes in cells in threedimensional microenvironments is clearly warranted. While PDL cells are predominantly fibroblastic, they constitute a heterogeneous cell population that includes MSCs with the potential to differentiate into osteoblasts, chondroblasts, adipocytes, and myoblasts.38–40 While the osteoblastspecific transcription factor RUNX2 and its chondrogenic counterpart SOX9 were initially expressed, neither were detectable in 24 h cultures, nor were PPARG or MYOD at any time point. The two downstream mediators of osteogenesis, SP7 (osterix) and BGLAP (osteocalcin), and the cartilagespecific collagen gene COL2A1 were not expressed either. This suggests that the differentiation pathway of the stem cell pool remained directed toward fibrosis, a conclusion supported by the elevated expression of CTGF, the gene found to be most responsive to the applied force. The upregulation of SOX9 (6 h) and BMP2 (6–12 h) is of particular clinical interest, since it provides a potential mechanism for the appearance of localized cell-free areas at compression sites in the PDL during orthodontic tooth movement; an adaptation referred to as hyalinization because of the histological similarity to hyaline cartilage.41 The role of SOX9 and its regulatory target BMP2 in determining chondrogenic lineage development is well established,42,43 and in the skeleton, cartilage is typically found at sites of compressive mechanical loading. The present findings, therefore, suggest that hyalinization might arise from the directed differentiation of localized MSCs down the chondrogenic pathway, followed by their eventual removal by programmed cell death. CTGF, originally identified as a polypeptide growth factor and downstream mediator of transforming growth factorbeta (TGF-b),44,45 has been reclassified as a matricellular protein situated in the pericellular matrix, and a member of the CCN family that serves to regulate cell–matrix interactions and cell function.46,47 Previous investigations have shown that in monolayer culture, CTGF is a mechanoresponsive gene,9,48–50 which suggests that CTGF functions in the PDL to maintain the fibroblast phenotype. Given that CTGF was the most responsive gene, our inability to identify CTGF protein in culture media in the present investigation came as a surprise –, particularly as we have previously identified CTGF in culture supernatants from HA–GEL films populated with PDL cells over a 1–3 week time course with the same ELISA.19 Amongst its many properties, the multiple functional domains of CTGF enable it to bind to integrin receptors on the cell surface and heparin sulfate proteoglycans and fibronectin in the ECM,46 which suggests that 6–24 h was insufficient time for a matrix-bound protein to leach out of the gel in sufficient amounts to reach the detection limit of the assay. The major enzymes involved in extracellular matrix turnover are the MMPs (matrixins), a family of proteolytic enzymes that play key roles in connective tissue resorption during growth, morphogenesis, and pathophysiological remodeling.51 The actions of MMPs are closely regulated by TIMPs through the formation of high affinity, essentially irreversible complexes with the activated forms of the enzymes – TIMP-1 is the major form, but four have been described to date.52 In addition to increased expression of MMP-1, MMP-2, and MMP-3 following compression,

10

MMP-1, MMP-2, MMP-3, and TIMP-1 protein were all released into the culture media; MMP-1 and TIMP-1 levels were significantly higher in 24 h cultures despite TIMP-1 expression remaining unchanged. Nevertheless, because the ELISAs only recognize latent proMMPs and free TIMP-1, we are not in a position to say what proportion of the enzymes are in the active, latent, or complexed forms. The finding that PDL cells express RANKL and OPG, members of the RANKL/RANK/OPG triad that constitutes a ligand–receptor system directly regulating the final steps of the bone resorptive cascade,53 has resulted in their being widely investigated for possible roles in tooth support. RANKL, which exists in both membrane-bound and soluble forms, stimulates the differentiation and function of osteoclasts, an effect mediated by RANK, a member of the TNF receptor family expressed primarily on cells of the monocyte/macrophage lineage, including osteoclasts and their precursor cells.54 OPG is a secreted protein that inhibits bone resorption by acting as a decoy receptor, binding to and neutralizing both cell-bound and soluble RANKL.55 Interestingly, in each of the studies in which static compression has been applied to PDL cells in two-dimensional culture, RANKL expression was upregulated,6,23,56,57 which contrasts with the present findings, highlighting differences between culturing cells on rigid two-dimensional substrates, and enclosed in a three-dimensional microenvironment. The challenge in engineering three-dimensional constructs is to replicate not only tissue complexity, but also the mechanical and viscoelastic characteristics (resistance to elastic deformation or stiffness) of the native tissue, and it is clear that further improvements in the composition of the present constructs are required. Although close to the minimum level recommended for use with the Flexercell Compression Plus system, a force of 33.4 kPa was at the limit appropriate for deforming the constructs in their present format, highlighting one of the disadvantages of hydrogels—their poor mechanical properties, resulting in tissue constructs with significantly poorer mechanical strength than the real tissue.15 Analysis of the rheological properties of cross-linked HA-GLN hydrogels has shown that the elastic moduli range from 11 Pa to 3.5 kPa depending on the concentration of HA; increasing the ratio of gelatin reduced gel stiffness by diluting the concentration of the HA component. The shear elastic modulus of Extracel is around 70 Pa,58 and although it has not been measured, the elastic modulus of the present constructs is likely to be similar to the 90 Pa reported recently for a three-dimensional collagen gel populated with PDL cells.59 While this may be suitable for imaging cell– cell and cell–matrix interactions in studies of tooth support and periodontitis by immunological methods and confocal microscopy, for tissue engineering applications a much higher stiffness is required. The difficulty of trying to reconstruct the biophysical properties of the human PDL, a thin (0.1–0.4 mm), complex fiber-reinforced tissue sandwiched between the bone and cementum that responds to mechanical loading in a viscoelastic and nonlinear manner,60 is that the elastic modulus is essentially unknown. A recent systematic review of 23 studies, for example, in which finite element analysis had been used, found that Young’s modulus ranged from 10 kPa to 1750 MPa, a difference approaching the extraordinary figure of six orders of magnitude.61 In comparison, the

SAMINATHAN ET AL.

elastic moduli of soft mammalian tissues range from near 100 Pa for soft organs such as the brain, to tens of thousands in muscle and around 300 MPa for Achilles tendon.62 This suggests that recapitulating the biophysical properties of the next generation of artificial PDL constructs requires a much more robust extracellular matrix. Since COLI is the major structural protein of the PDL, rather than add collagen and other structural macromolecules to hydrogels to increase their complexity, a more rational approach for the next generation of PDL constructs would be to permeabilize preformed fibrillar COLI scaffolds,63 with PDL cell-populated matrices, and in thinner formats to replicate the physical dimensions of the ligament in vivo. Acknowledgments

This research was supported by the Academic Research Fund of the National University of Singapore (R222-000029-112, R221-000-043-112, and R221-000-063-750). The authors thank Mr. Chan Swee Hen, Faculty of Dentistry Research Laboratory, DSO Building and Ms. Lee Shu Ying of the Confocal Microscopy Unit, Yong Loo Lin School of Medicine, for their invaluable help and advice during the course of this research. Disclosure Statement

The authors wish to state that no personal or financial conflicts of interest arose during the conduct of this investigation. References

1. Meikle, M.C. Craniofacial Development, Growth and Evolution. Bressingham, Norfolk, England: Bateson Publishing, 2002. pp. 311–342. 2. Ten Cate, A.R. The development of the periodontium–a largely ectomesenchymally derived unit. Periodontology 2000 13, 8, 1997. 3. Beertsen, W., McCullough, C.G.A., and Sodek, J. The periodontal ligament: a unique, multifunctional connective tissue. Periodontology 2000 13, 20, 1997. 4. Shimizu, N., Yamaguchi, M., Goseki, T., Ozawa, Y., Saito, K., Takiguchi, H., et al. Cyclic-tension force stimulates interleukin-1b production by human periodontal ligament cells. J Periodontol Res 29, 328, 1994. 5. Long, P., Hu, J., Piesco, N., Buckley, M., and Aggarwal, S. Low magnitude of tensile strain inhibits IL-1b-dependent induction of pro-inflammatory cytokines and induces synthesis of IL-10 in human periodontal ligament cells in vitro. J Dent Res 80, 1416, 2001. 6. Kanzaki, H., Chiba, M., Shimizu, Y., and Mitani, H. Periodontal ligament cells under mechanical stress induces osteoclastogenesis by receptor activator of nuclear factor kB ligand up-regulation via prostaglandin E2 synthesis. J Bone Miner Res 17, 210, 2002. 7. Wescott, D.C., Pinkerton, M.N., Gaffey, B.J., Beggs, K.T., Milne, T.J., and Meikle, M.C. Osteogenic gene expression by human periodontal ligament cells under cyclic tension. J Dent Res 86, 1212, 2007. 8. Pinkerton, M.N., Wescott, D.C., Gaffey, B.J., Beggs, K.T., Milne, T.J., and Meikle, M.C. Cultured human periodontal ligament cells constitutively express multiple osteotropic cytokines and growth factors, several of which are re-

ENGINEERING THE PERIODONTAL LIGAMENT IN 3D CONSTRUCTS

9.

10. 11. 12.

13.

14. 15.

16. 17.

18.

19.

20. 21.

22. 23.

24.

sponsive to mechanical deformation. J Periodontol Res 43, 343, 2008. Saminathan, A., Vinoth, K.J., Wescott, D.C., Pinkerton, M.N., Milne, T.J., Cao, T., et al. The effect of cyclic mechanical strain on the expression of adhesion-related genes by periodontal ligament cells in two-dimensional culture. J Periodontol Res 47, 212, 2012. Lee, K.Y., and Mooney, D.J. Hydrogels for tissue engineering. Chem Rev 101, 1869, 2001. Schacht, E.H. Polymer chemistry and hydrogel systems. J Phys Conf Ser 3, 22, 2004. Roskelley, C.D., Desprez, P.Y., and Bissell, M.J. Extracellular matrix-dependent tissue specific gene expression in mammary epithelial cells requires both physical and biochemical signal transduction. Proc Natl Acad Sci U S A 91, 12378, 1994. Wang, F., Weaver, V.M., Petersen, O.W., Larabell, C.A., Dedhar, S., Briand, P., et al. Reciprocal interactions between beta 1-integrin and epidermal growth factor receptor in three-dimensional basement membrane breast cultures: a different perspective in epithelial biology. Proc Natl Acad Sci U S A 95, 14821, 1998. Cukierman, E., Pankov, R., Stevens, D.R., and Yamada, K.M. Taking cell-matrix adhesions to the third dimension. Science 294, 1708, 2001. Aherne, M., Yang, Y., El Haj, A.J., Then, K.Y., and Liu, K.-K. Characterizing the viscoelastic properties of thin hydrogel-based constructs for tissue engineering applications. J R Soc Interface 2, 455, 2005. Serban, M.A., and Prestwich, G.D. Modular extracellular matricies: solutions for the puzzle. Methods 45, 93, 2008. Pike, D.B., Cai, S., Pomraning, K.R., Firpo, M.A., Fisher, R.J., Shu, X.Z., et al. Heparin-regulated release of growth factors in vitro and angiogenic response in vivo to implanted hyaluronan hydrogels containing VEGF and bFGF. Biomaterials 27, 5242, 2006. Serban, M.A., Scott, A., and Prestwich, G.D. Use of hyaluronan-derived hydrogels for three-dimensional cell culture and tumor xenografts. Curr Protoc Cell Biol Chapter 10, Unit 10.14, 2008. Saminathan, A., Vinoth, K.J., Low, H.H., Cao, T., and Meikle, M.C. Engineering three-dimensional constructs of the periodontal ligament in hyaluronan-gelatin hydrogel films and a mechanically-active environment. J Periodontol Res 48, 790, 2013. Wall, M.E., Weinhold, P.S., Siu, T., Brown, T.D., and Banes, A.J. Comparison of cellular strain with applied substrate strain in vitro. J Biomech 40, 173, 2007. Jones, K.H., and Senft, J.A. An improved method to determine cell viability by simultaneous staining with fluorescein diacetate–propidium iodide. J Histochem Cytochem 33, 77, 1985. Baker, B.M., and Chen, C.S. Deconstructing the third dimension–how 3D culture microenvironments alter cellular cues. J Cell Sci 125, 3015, 2012. Nakajima, R., Yamaguchi, M., Kojima, T., Takano, M., and Kasai, K. Effects of compression force on fibroblast growth factor-2 and receptor activator of nuclear factor kappa B ligand production by periodontal ligament cells in vitro. J Periodontol Res 43, 168, 2008. Kim, S.J., Park, K.H., Park, Y.G., Lee, S.W., and Kang, Y.G. Compressive stress induced the up-regulation of MCSF, RANKL, TNF-a expression and the down-regulation

25.

26.

27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

38. 39.

40. 41.

42.

43.

11

of OPG expression in PDL cells via the integrin-FAK pathway. Arch Oral Biol 58, 707, 2013. Lee, Y.H., Nahm, D.S., Jung, Y.K., Choi, J.Y., Kim, S.G., Cho, M., et al. Differential expression of periodontal ligament cells after loading of static compressive force. J Periodontol 78, 446, 2007. de Araujo, R.M.S., Oba, Y., and Moriyama, K. Identification of genes related to mechanical stress in human periodontal ligament cells using microarray analysis. J Periodontol Res 42, 15, 2007. Sastry, S.K., and Burridge, K. Focal adhesions: a nexus for intracellular signaling and cytoskeletal dynamics. Exp Cell Res 261, 25, 2000. Wang, N., Butler, J.P., and Ingber, D.E. Mechanotransduction across the cell surface and through the cytoskeleton. Science 260, 1124, 1993. Clarke, E.A., and Brugge, J.S. Integrins and signal transduction pathways: the road taken. Science 268, 233, 1995. Wyllie, A.H., Kerr, J.F.R., and Currie, A.R. Cell death: the significance of apoptosis. Int Rev Cytol 68, 251, 1980. Elmore, S. Apoptosis: a review of programmed cell death. Toxicol Pathol 35, 495, 2007. Chowdhury, I., Tharakan, B., and Bhat, G.K. Caspases–an update. Comp Biochem Physiol B Biochem Mol Biol 151, 10, 2008. Zhong, W., Xu, C., Zhang, F., Jiang, X., Zhang, X., and Ye, D. Cyclic stretching force-induced early apoptosis in human periodontal ligament cells. Oral Dis 14, 270, 2008. Hao, Y., Xu, C., Sun, S., and Zhang, F. Cyclic stretching force induces apoptosis in human periodontal ligament cells via Caspase-9. Arch Oral Biol 54, 864, 2009. Xu, C., Hao, Y., Wei, B., Ma, J., Li, J., Huang, Q., et al. Apoptotic gene expression by human periodontal ligament cells following cyclic stretch. J Periodontol Res 46, 742, 2011. Boccafoschi, F., Bosetti, M., Gatti, S., and Cannas, M. Dynamic fibroblast cultures: response to mechanical stretching. Cell Adh Migr 1, 124, 2007. Boccafoschi, F., Sabbatini, M., Bosetti, M., and Cannas, M. Overstressed mechanical stretching activates survival and apoptotic signals in fibroblasts. Cells Tissues Organs 192, 167, 2010. Lekic, P., Rojas, J., Birek, C., Tenenbaum, H., and McCulloch, C.A. Phenotypic comparison of periodontal ligament cells in vivo and in vitro. J Periodontol Res 36, 71, 2001. Seo, B.-M., Miura, M., Gronthos, S., Bartold, P.M., Batouli, S., Brahim, J., et al. Investigation of multipotential postnatal stem cells from human periodontal ligament. Lancet 364, 149, 2004. Nagatomo, K., Komaki, M., Sekiya, I., Sakaguchi, Y., Noguchi, K., Oda, S., et al. Stem cell properties of human periodontal ligament cells. J Periodontol Res 41, 303, 2006. Bister, D., and Meikle, M.C. Re-examination of ‘Einige Beitra¨ge zur Theorie der Zahnregulierung’ (Some contributions to the theory of the regulation of teeth) published in 1904–1905 by Carl Sandstedt. Eur J Orthod 35, 160, 2013. Akiyama, H., Chaboissier, M.-C., Martin, J.F., Schedl, A., and de Crombrugghe, B. The transcription factor Sox9 has essential roles in successive steps of the chondrocyte differentiation pathway and is required for expression of Sox5 and Sox6. Genes Dev 16, 2813, 2002. Schmitt, B., Ringe, J., Ha¨upl, T., Notter, M., Manz, R., Burmester, G.R., et al. BMP2 initiates chondrogenic lineage development of adult human mesenchymal stem cells in high-density culture. Differentiation 71, 567, 2003.

12

44. Bradham, D.M., Igarashi, A., Potter, R.L., and Grotendorst, G.R. Connective tissue growth factor: a cysteine-rich mitogen secreted by human vascular endothelial cells is related to the SRC induced immediate gene product CEF-10. J Cell Biol 114, 1285, 1991. 45. Igarashi, A., Okochi, H., Bradham, D.M., and Grotendorst, G.R. Regulation of connective tissue growth factor gene expression in human skin fibroblasts and during wound repair. Mol Biol Cell 4, 637, 1993. 46. Chen, C.-C., and Lau, L.F. Functions and mechanisms of action of CCN matricellular proteins. Int J Biochem Cell Biol 41, 771, 2009. 47. Bornstein, P., and Sage, E.H. Matricellular proteins: extracellular modulators of cell function. Curr Opin Cell Biol 14, 608, 2002. 48. Miyake, Y., Furumatsu, T., Kubota, S., Kawata, K., Ozaki, T., and Takigawa, M. Mechanical stretch increases CCN2/ CTGF expression in anterior cruciate ligament-derived cells. Biochem Biophys Res Commun 409, 247, 2011. 49. Guo, F., Carter, D.E., and Leask, A. Mechanical tension increases CCN2/CTGF expression and proliferation in gingival fibroblasts via a TGFb-dependent mechanism. PLoS One; 6, e19756, 2011. 50. Kanazawa, Y., Nomura, J., Yoshimoto, S., et al. Cyclical cell stretching of skin-derived fibroblasts downregulates connective tissue growth factor (CTGF) production. Connect Tissue Res 50, 323, 2009. 51. Murphy, G., and Nagase, H. Progress in metalloproteinase research. Mol Aspects Med 29, 290, 2008. 52. Murphy, G. Tissue inhibitors of metalloproteinases. Genome Biol 12, 233, 2011. 53. Filvaroff, E., and Derynck, R. Bone remodelling: a signalling system for osteoclast regulation. Curr Biol 8, R679, 1998. 54. Nakagawa, N., Kinosaki, M., Yamaguchi, K., Shima, N., Yasuda, H., Yano, K., et al. RANK is the essential signaling receptor for osteoclast differentiation factor in osteoclastogenesis. Biochem Biophys Res Commun 253, 395, 1998. 55. Simonet, W.S., Lacey, D.L., Dunstan, C.R., Kelley, M., Chang, M.S., Luthy, R., et al. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell 89, 309, 1997. 56. Tsuji, K., Uno, K., Zhang, G.X., and Tamura, M. Periodontal ligament cells under intermittent tensile stress regu-

SAMINATHAN ET AL.

57.

58.

59.

60.

61.

62. 63.

lates mRNA expression of osteoprotegerin and tissue inhibitor of metalloproteinase-1 and - 2. J Bone Miner Res 22, 94, 2004. Yamamoto, T., Kita, M., Kimura, I., Oseko, F., Terauchi, R., Takahashi, K., et al. Mechanical stress induces expression of cytokines in human periodontal ligament cells. Oral Dis 12, 171, 2006. Vanderhooft, J.L., Alcoutlabi, M., Magda, J.J., and Prestwich, G.D. Rheological properties of cross-linked hyaluronangelatin hydrogels for tissue engineering. Macromol Biosci 9, 20, 2009. Kim, S.G., Kim, S.-G., Viechnicki, B., Kim, S., and Nah, H.-D. Engineering of a periodontal ligament construct: cell and fibre alignment induced by shear stress. J Clin Periodontol 38, 1130, 2011. Jo´nsdo´ttir, S.H., Giesen, E.B., and Maltha, J.C. Biomechanical behaviour of the periodontal ligament of the beagle dog during the first 5 hours of orthodontic force application. Eur J Orthodont 28, 547, 2006. Fill, T., Carey, J.P., Toogood, R.W., and Major, P.W. Experimentally determined mechanical properties of, and models for, the periodontal ligament: critical review of current literature. J Dent Biomech 2011, 312980, 2011. Levental, I., Georges, P.C., and Janmey, P.A. Soft biological materials and their impact on cell function. Soft Matter 2, 1, 2006. Glowacki, J., and Mizuno, S. Collagen scaffolds for tissue engineering. Biopolymers 89, 338, 2008.

Address correspondence to: Murray C. Meikle, DDSc, PhD Faculty of Dentistry University of Otago PO Box 647 Dunedin 9054 New Zealand E-mail: [email protected] Received: April 20, 2014 Accepted: August 25, 2014 Online Publication Date: October 2, 2014