lable at ScienceDirect

Biomaterials 32 (2011) 4525e4538

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Human treated dentin matrix as a natural scaffold for complete human dentintissue regeneration

Rui Li a,b,1, Weihua Guo a,1, Bo Yang a,b, Lijuan Guo a,b, Lei Sheng a,b, Gang Chen a,b, Ye Li a, Qing Zou a,Dan Xie a, Xiaoxue An a, Yali Chen a, Weidong Tian a,b,*

a State Key Laboratory of Oral Diseases, Sichuan University, Chengdu 610041, PR ChinabDepartment of Oral and Maxillofacial Surgery, West China College of Stomatology, Sichuan University, No.14, 3rd Section, Renmin South Road, Chengdu 610041, PR China

a r t i c l e i n f o

Article history:Received 11 February 2011Accepted 4 March 2011Available online 31 March 2011

Keywords:Dentin matrixStem cellsMicroenvironmentNatural scaffoldToothRegeneration

* Corresponding author. Department of Oral andChina College of Stomatology, Sichuan University, No.1Road, Chengdu 610041, PR China. Tel./fax: þ86 28 85

E-mail address: drtwd@sina.com (W. Tian).1 These authors contributed equally to this work.

0142-9612/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.biomaterials.2011.03.008

a b s t r a c t

An essential aspect of tooth tissue engineering is the identification of suitable scaffolding materials tosupport cell growth and tissue regeneration. Treated dentin matrix (TDM) from a rat has recently beenshown to be a suitable scaffold for rat dentin regeneration. However, due to species-specific differences,it remains unclear whether a similar fabrication method can be extended to human TDM and humandentin regeneration. Therefore, this present study explored the biological response to a human TDM(hTDM) created using a modified dentin treatment method. Various biological characteristics, includingcell proliferation, cell migration, cell viability, and cytotoxity were investigated. To assess the inductivecapacity of hTDM, dental follicle cells (DFCs) were combined with hTDM and were implanted in vivo for 8weeks in a mouse model. The resulting grafts were studied histologically. The results showed hTDMreleased dentinogenic factors, indicating that hTDM could play a sustained role in odontogenesis. DFCattachment, growth, viability, and cytotoxicity on the surface of hTDM showed a notable improvementover those on calcium phosphate controls. Most importantly, in vivo hTDM induced and supportedregeneration of complete dentin tissues, which expressed dentin markers DSP and DMP-1. As cells in andaround the regenerated dentin were positive for human mitochondria, implanted DFCs and hTDM wereresponsible for the regenerated dentin tissues. In conclusion, hTDM is indicated as an ideal biomaterialfor human dentin regeneration.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The loss of teeth from carries and accidents affects manyhumans at some point in their lives. While it is not usually a lifethreatening condition, tooth loss greatly impacts quality of life.Conventionally, lost teeth are replaced with dentures or syntheticdental implants. However, the developing field of tissue engi-neering has recently indicated that replacement teeth comprised ofnatural materials can be created in vitro [1e4]. Tissue engineeringinvolves the harvest of stem cells from the patient (in this casedental follicle stem cells, DFCs), growing and inducing these cellstowards a relevant lineage (i.e. odontoblasts) on a suitable scaf-folding material, and reintroducing the cell-scaffold construct tothe patient.

Maxillofacial Surgery, West4, 3rd Section, Renmin South50 3499.

All rights reserved.

The identification and development of suitable scaffoldingmaterials capable of regenerating tooth structures, such as dentin,are essential aspects of tooth tissue engineering [5,6]. A suitablescaffolding material is non-cytotoxic, bioactive, and capable ofproviding a three-dimensional microenvironment supporting cellgrowth, differentiation, and cellular organization into a tissuestructure. In addition, the scaffolding material should have soundmechanical properties and be physiologically compatible whenreintroduced to the patient. A number of synthetic and naturalpolymers, as well as calcium phosphate-based materials, have beentrialed for dentin tissue engineering. However, few were able toregenerate complete dentin tissue [7e9]. It may be the case thatwhile these materials support cell growth and mineralization, theyare not capable of inducing differentiation towards an odontogenicspecialization.

Dentin is a calcified tissue which makes up most of the tooth,thereby making it a key component for tissue engineering of toothstructure. Dentin, which is less mineralized and more elastic thanenamel, is comprised of approximately 70% hydroxyapatite, anextracellular matrix component, and the odontoblasts that once

R. Li et al. / Biomaterials 32 (2011) 4525e45384526

made the dentin [10,11]. The soluble proteins of human dentin arebioactive proteins considered to be necessary for dentinogenesis[12]. Scaffolds mimicking the structure of natural dentin have beenshown to be useful in dentin regeneration in vivo [13]. It was alsofound that treated dentin matrix (TDM) derived from rats couldinduce dental precursor cells to differentiate into dentin in a ratmodel [14]. While our previous study suggests rat TDM to bea suitable scaffold and inductive microenvironment for dentinregeneration [14], it is not clear whether the production processand material would be extendable to humans due to speciesdifferences.

Therefore, in this study we investigated the capacity of humanteeth, which were extracted out of clinical necessity and thrownaway as medical waste, to act as a suitable scaffolding material forhuman dentin tissue regeneration. In the present study, we fabri-cated human TDM by modifying our previous method used withrats and studied the odontogenic characteristics of hTDM. Ulti-mately, this work is an assessment of whether hTDM can providea suitable scaffold and inductive microenvironment for regenerat-ing complete dentin tissue in humans.

2. Materials and methods

2.1. Cell culture, evaluation, and proof of multipotentiality

2.1.1. Cell isolationIsolation of human DFCs has been described previously [15]. Human DFCs were

obtained from third molars, which were extracted for clinical reasons. Isolated DFCswere cultured in Dulbecco’s modified Eagle’s medium (DMEM, Hyclone, USA) sup-plemented with 10% fetal bovine serum (FBS, Hyclone, USA) in a humidified atmo-sphere at 37oC and 5% CO2. Cell culture medium was refreshed every two days.

2.1.2. Immunofluorescent cell stainingOne day prior to staining, DFCs were released and seeded onto 0.8 cm � 0.8 cm

coverslips in a 24-well plate for further culture. DFCs were fixed with 4% polyoxy-methylene for 30 min. Subsequent steps were performed according to the manu-facturer’s recommendations. Antibodies used in immunofluorescent stainingincluded: stro-1, vimentin, and CK-14. All antibodies were from Abcam, USA. Allsamples were examined under a fluorescence microscope (Leica Optical, Germany).

2.1.3. Osteogenic differentiationA total of 1 � 105 DFCs were seeded into each well of a six-well plate. At 80%

confluence, DFCs were cultured in osteogenic medium containing 10% FBS, 5 mM

L-glycerophosphate (Sigma, USA), 100 nM dexamethasone (Sigma, USA), and 50 mg/ml ascorbic acid (Sigma, USA) for 15 days. The control group was cultured in DMEMwith 10% FBS. The medium was changed every two days. After 15 days, cells werewashed twice in PBS after being fixed in 4% paraformaldehyde for 10 min and thenincubated in 0.1% alizarin red solution (Sigma, USA) in TriseHCl (pH 8.3) at 37 �C for30 min. After being washed twice in PBS, cells were routinely observed and pho-tographed under a light microscope (Nikon, Japan).

2.1.4. Adipogenic differentiationPrimary adipogenic conditions were prepared as described previously [16].

Adipogenic medium consisted of DMEM supplemented with 10% FBS, 2 mM gluta-mine (Sigma, USA), 100 U/ml penicillinestreptomycin (Hyclone, USA), 100 mM

ascorbic acid (Sigma, USA), 0.5 mM methylisobutylxantine (Sigma, USA), 0.5 mM

hydrocortisone (Sigma, USA) and 60 mM indomethacin (Sigma, USA). The cells weretreated as mentioned above. After four weeks, the cells growing under adipogenicconditions were washed twice with PBS and fixed in 70% ethanol for 15 min. Oil redO (Sigma, USA) staining was performed as previously reported [16].

2.1.5. Neurogenic differentiationAt 80% confluence, DFCs were cultured in neurogenic medium containing 2%

Dimethyl Sulphoxide (DMSO), 200 mM butylated hydroxyanisole (Sigma, USA),25 mM KCl (Kelong, China), 2 mM valporic acid (Sigma, USA), 10 mM forskolin (Sigma,USA), 1 mM hydroxycortisone (Sigma, USA), 5 mg/mL insulin (Gibco, USA), and 2 mM

L-glutamine (Sigma, USA). After 4 h, cells were analyzed by immunocytofluorescencefor expression of the neural cell marker, bIII-tubulin (Abcam, USA). Images werefixed and analyzed under a fluorescence microscope (Leica Optical, Germany).

2.2. Fabrication of human treated dentin matrix (hTDM)

Twenty mandibular and twenty maxillary premolars were harvested from tenpatients requiring their removal for clinical reasons at the West China Stomatology

Hospital of Sichuan University. Periodontal tissues were completely scraped awaywith a curette. By grinding along the tooth profile, outer cementum and part of thedentinwas removed. Dental pulp tissues and pre-dentinwere mostly removed usingmechanical means. The resulting human dentin matrix was soaked in deionizedwater for 5 h and mechanically cleaned for 20 min every hour using an ultrasoniccleaner. The deionized water was changed every hour. Human dentin matrices werethen soaked in 17% Ethylene Diamine Tetra-acetic Acid (EDTA, Sigma, USA) for 5min,washed in deionized water for 10 min in an ultrasonic cleaner, and exposed to 10%EDTA for 5 min. They were then washed in deionized water for 10 min in anultrasonic cleaner, exposed to 5% EDTA for 10 min, and washed in deionized waterfor 10 min in an ultrasonic cleaner. Human treated dentin matrices were maintainedin sterile phosphate buffered saline (PBS; Hyclone, USA) with 100 units/ml penicillin(Hyclone, USA) and 100 mg/ml streptomycin (Hyclone, USA) for 72 h, were washed insterile deionized water for 10 min in an ultrasonic cleaner, and were finally stored inDMEM (Hyclone, USA) at 4 �C. In comparison to our previous study in rats, the abovemethod has been modified with respect to the concentration of EDTA and thetreatment time [14]. The procedure for untreated human dentin matrix (hUDM) wasthe same as for the hTDM, except that there was no exposure to EDTA.

In this experiment, we used biphasic hydroxyapatite (HA)/tricalcium phosphate(TCP) scaffolds as a control group. HA/TCP are considered a suitable scaffold fortissue engineering, with demonstrated good biocompatibility. HA/TCP was providedby the National Engineering Research Center for Biomaterials of Sichuan University.

2.3. Evaluation of human treated dentin matrix

2.3.1. Morphological observation of human treated dentin matrixTDMs were observed by scanning electron microscope (SEM) (Inspect F, FEI,

Netherlands). Briefly, TDMs were washed in PBS three times, fixed with 2.5%glutaraldehyde at 0oC, dehydrated and dried in a critical-point dryer, and finallyobserved under SEM. Histological evaluation was also employed, which involveda conventional hematoxylin-eosin staining (H&E staining).

2.3.2. Investigation of factors secreted by human treated dentin matrixAn enzyme linked immunosorbent assay (ELISA) was used to investigate the

suitability and effectiveness of hTDM fabrication. A liquid extract of the dentinmatrix was collected according to the protocol of the International StandardizationOrganization (ISO 10993). Liquid extracts were made of hTDM, hUDM, and theHA/TCP controls. Briefly, scaffold samples were pulverized using a cold mortar andpestle after freezing in liquid nitrogen. DMEM was added to the pulverized scaffoldwith a ratio of 20 g scaffold powder per 100 mL DMEM. Slurries were incubated for1e6 days at 37oC and the samples were filtered using a 0.22 mm filter. The resultingconcentration of different proteins in the liquid extracts of hTDM and hUDM weremeasured using ELISA kits for COL-1, TGF-b1, DSP, DMP-1, biglycan, and decorin. Inaddition, following the above procedure, we collected the liquid extract of wholehTDM and hUDM scaffolds without converting them into powder.

Protein concentrations were analyzed from liquid extracts collected every dayfrom day 1 to day 6. ELISA kits for COL-1, TGF-b1, DSP, DMP-1, biglycan and decorinwere obtained from RD (RD, USA). All ELISA procedures were followed according tothe manufacturer’s recommendations. Absorbance readings at 450 nm wereobtained using a spectrophotometer (Thermo VARIOSKAN FLASH, Thermo, USA).Concentrations were determined by comparing with a standard curve preparedfrom a range of concentrations of COL-1, TGF-b1, DSP, DMP-1, biglycan and decorin(RD, USA). DMEM was used as negative control. The experiments were repeated atleast three times.

Immunohistochemistry was used to detect the presence of proteins related tothe regeneration of dentin on the hTDM. Dentin tissues were demineralized by 10%EDTA (Sigma, USA) and paraffin embedded. The primary antibodies included (1)anti-TGF beta 1 (TGF-b1) at a 1:100 dilution, (2) anti-decorin (decorin) at a 1:100dilution, (3) anti-DMP-1 at a 1:100 dilution, (4) anti-COL-1 at a 1:100 dilution, (5)anti-biglycan at a 1:100 dilution, (6) anti-DSP at a l:100 dilution. The antibodies forTGF-b1, COL-1 and biglycan were purchased from Abcam Biotechnology (Abcam,UK) and the others were obtained from Santa Cruz Biotechnology (Santa Cruz, USA).PBS was used as a negative control instead of the primary antibody. Biotinylatedsecondary antibodies (1:1000) were purchased from Dako (Dako, USA).

2.4. Effect of hTDM on biological characteristics of DFCs in vitro

In order to determine the biocompatibility of hTDM, and to further analyzebiological characteristics of DFCs induced by hTDM, we analyzed the morphology,proliferation, viability, migration, and cytotoxity of DFCs affected by hTDM. All of theexperiments were repeated at least three times.

2.4.1. Morphology of DFCs on hTDMDFCs were seeded onto hTDM at a concentration of 1 �104 and were incubated

in a 24-well plate for one to nine days. DMEMwith 10% FBS was changed once everythree days. Cell morphology and growth on the hTDM were studied using SEM(Inspect F, FEI, Netherlands). We also examined the cell morphology and growtharound hTDM under a light microscope (Nikon, Japan). All procedures were thesame for the control group.

R. Li et al. / Biomaterials 32 (2011) 4525e4538 4527

2.4.2. Cell viabilityA cell count kit-8 (CCK-8, Dojindo, Japan) was used to quantitatively evaluate

DFC viability. After DFCs were grown on hTDM for 24, 48, 72, 96, or 120 h, theoriginal culture medium was replaced by 500 ml DMEM with 10% FBS containing50 ml CCK-8. After incubating at 37 �C for 3 h, 100 ml of the above solution was takenfrom each sample and added to one well of a 96-well plate. Six parallel replicateswere prepared. The absorbance at 450 nm was determined using a spectropho-tometer (Thermo VARIOSKAN FLASH, Thermo, USA). HA/TCP controls underwentthe same treatment. The test was repeated at least three times.

2.4.3. Cell proliferationFlow cytometric analysis was performed as described previously [14] to analyze

differences in cell-cycle when grown on different biomaterials. Briefly, cells werecultured with hTDM for two days and were then released from the material. Cellswere washed twice with 0.01 M PBS in order to make a single-cell suspension. Colddehydrated alcohol (2 ml) was quickly mixed with the cell suspension to fix cells at4 �C for 24 h. Finally, the cells were washed twice again with PBS, stained with100 mg/ml propidium iodide (PI, Sigma, USA) at 4 �C for 30 min, and subjected tocell-cycle analysis by using flow cytometry (Cytomic FC500, BECKMAN, USA). Theprocess was the same for the control group. This experiment was repeated at leastthree times.

2.4.4. Cell migrationDFCs were seeded into 6-well dishes (5 � 104 cells per well). Using a pipette tip,

a scrapewas produced in themonolayer on each plate. The adherent monolayer wasthen washed two times in PBS to remove non-adherent cells. Liquid extracts ofhTDM or HA/TCP controls were then added to the wells. After 0 and 24 h, eachscraped monolayer surface area was inspected using light microscopy (Nikon,Japan). This experiment was repeated three times.

2.4.5. In vitro cytotoxicity of human treated dentin matrixBefore transplantation, we analyzed the cytotoxicity of hTDM and HA/TCP

scaffolds. MTTwas used to assess cytotoxicity. DFCs at the 3rd passagewere culturedin 24 well plates with 1�104 cells per well. The liquid extract of hTDMwith 10% FBSwas added as the test group. DMEMwas used as a blank group, and HA/TCPwas usedas the control group. After incubation for 1, 2, 3, 4, or 5 days, 5 g/L MTT solution wasadded into each well for another 4 h of incubation. Mediumwas removed and 150 mldimethyl sulphoxide (DMSO)was added to eachwell. The absorbance at 490 nmwasdetermined using a spectrophotometer (Thermo VARIOSKAN FLASH, Thermo, USA).This test was repeated at least three times.

We divided the cytotoxicity level into five stages by analyzing relative growthrate (RGR). The relative growth rate was calculated according to the ISO 10993protocol as %RGR¼(absorbance of text group/absorbance of blank group)� 100% Thecalculated RGRs are summarized in Table 1.

2.5. Dentin regeneration in vivo using hTDM as an inductive scaffold

In order to observe the biocompatibility and inductive dentin regeneration ofhTDM in vivo, we implanted test and control groups into the dorsum of immuno-deficient mice. All animal experiments were conducted in accordance with thecommittee guidelines of Sichuan University for animal experiments, which also,meets the NIH guidelines for the care and use of laboratory animals. Immunodefi-cient mice were obtained from the Laboratory Animal Research Centre of SichuanUniversity and were maintained on a daily ration of Purina rodent chow in housingquarters with cycled light (12 h on/off), regulated temperature, and sterile water.

Nine immunodeficient mice were divided into three groups including one testgroup (TDM combined with DFCs) and two control groups (HA/TCP combined withDFCs and single TDM). DFCs were seeded onto scaffolds at a density of 1 �104 cells/scaffold and incubated at 37oC for three days. Implantation of scaffolds was per-formed under deep anesthesia. Eight weeks later, all samples were obtained fromthe immunodeficient mice under deep anesthesia. Implants were fixed with 4%paraformaldehyde overnight at 4 �C, demineralized with 10% EDTA (pH 8.0), andembedded in paraffin. Paraffin sections were stained with hematoxylin and eosinstain (H&E), Mason stain, and immunohistochemical stains.

Table 1Cytotoxicity levels scored according to the ISOstandard.

RGR (%) Cytotoxicity

�100 075e99 150e74 225e49 31e24 40 5

Immunohistochemical antibodies included DMP-1, DSP (Santa Cruz, USA) andhuman mitochondria (Millipore, USA). All antibodies were use at a dilution of1:100, according to the manufacturers’ protocol. DMP-1 and DSP were used toidentify whether the newly-formed tissues were dentin tissue. Anti-humanmitochondria antibodies reacted only with cells from a human source, and wereused to identify whether the human seeding cells participated in the newly-formed tissues.

2.6. Statistical analysis

All data were expressed as the mean � SD. Statistical significance was analyzedusing SPSS 11.5 software (SPSS, USA). A value of p < 0.05 was considered statisticallysignificant

3. Results

3.1. Cell culture

Human DFCs are mesenchymal cells with typical spindle shapemorphology and were the dominant cell population at the 3rd invitro passage (Fig. 1A). DFCs were positive for vimentin and stro-1,which are proteins characteristic of mesenchymal stem cells, butwere negative for CK-14, which is a marker of epithelial cells(Fig. 1B, C, D).

After culture under adipogenic conditions for 20 days, humanDFCs formed lipid droplets when stained with oil red (Fig. 1E). Withculture in osteogenic medium for 15 days, human DFCs formedmineralized nodules by alizarin red staining (Fig. 1F). Interestingly,the mineralized nodules were polymorphic in shape, which is inagreement with human DFCs. DFCs underwent morphologicalchanges (Fig. 1G) and expressed the neurogenic makers bIII-tubulin(Fig. 1H) under neurogenic induction. Therefore, the multi-potentiality of human DFCs was demonstrated.

3.2. Fabrication and analysis of human treated dentin matrix

Human TDM was fabricated using the tooth profile (Fig. 2A andB) H&E staining showed that periodontal and dental pulp tissue hadbeen removed completely (Fig. 2C and D). Examination under SEMdemonstrated that the dentinal tubules were sufficiently exposedand fiber bundles of intertubular and peritubular dentin wereloosened (Fig. 2E and F). There were obvious differences betweenhTDM and hUDM (Fig. 3A and B).

Most notably, the method used in a previous study with ratdentin [14] did not work on human dentin matrix (supplement).Therefore, in the present study the demineralization method wasaltered by changing the EDTA concentration and incubation time.The treatment time of 17% EDTA was extended to 5 min, and thetreatment time of 5% EDTA was extended to 10 min. An additionaltreatment step of 10% EDTA for 5minwas added between these twosteps. The step-by-step process of demineralization was shown tobe effective and necessary.

Using immunohistochemistry, we detected proteins related todentinogenesis. The results showed that hTDM was positive forCOL-1, DSP, TGF-b1, DMP-1, biglycan and decorin, but negative forPBS (Fig. 4). The ELISA kit results also indicated that the concen-tration of proteins and factors contained in the hTDM liquidextract showed no statistically significant difference with thosereleased by hUDM (Fig. 5). Fig. 5 shows the average concentrationof these proteins, with COL-1 and TGF-b1 having the highestlevels. As expected, these proteins were not detected in the liquidextract of the HA/TCP or DMEM controls (data not shown). Usingthe whole hTDM and hUDM structure to create liquid extractsdetected higher protein concentrations from hTDM than hUDM(Fig. 3).

Fig. 1. Human DFCs assumed spindle shapes under light microscope examination (A), and were positive for stro-1 (B) and vimentin (C), but negative for epithelial marker CK-14 (D).After being cultured under adipogenic conditions for 20 days, lipid droplets were observed (E). Alizarin red staining showed mineralized nodules for cells grown under osteogenicconditions (F). Under neurogenic conditions, DFCs underwent morphologic changes (G) and expressed the neurogenic maker bIII-tubulin (H).

R. Li et al. / Biomaterials 32 (2011) 4525e45384528

Fig. 2. Details of the treatment process to generate human dentin from extracted teeth and morphology of treated dentin. First the crown, pulp and cementum are removed (A andB). H&E staining (C and D) and examination by SEM (E and F) illustrates that hTDM has dentinal tubules exposed sufficiently and loosened fiber bundles of intertubular andperitubular dentin.

R. Li et al. / Biomaterials 32 (2011) 4525e4538 4529

3.3. Effect of hTDM on biological characteristics of DFCs in vitro

3.3.1. Influence of hTDM on cell morphologyAfter DFCs were grown on hTDM for 1, 3, 5, 7, or 9 days, cell

morphology was examined by SEM. On day 1 on the hTDM, cellnumber was low; however, the attached cells had a polygonalshape which is characteristic of viable attachment. On day 3, cellproliferation on the hTDM was obvious and dentinal tubules werehardly seen due to coverage by cells. On day 5, 7, and 9, no dentinaltubules could be seen due to large quantities of cells and theextracellular matrix synthesized by these cells. Interestingly, thecells on the surface of hTDM showed a pattern of proliferation in

a regular direction (Fig. 6). In the HA/TCP control group, cellproliferation and differentiation did not appear to be as high as inthe experimental groups (Fig. 6) until day 5 and later. Under lightmicroscope, it was apparent that DFCs were growing normally,which demonstrated good biocompatibility of the hTDM. Overall,both the hTDM and HA/TCP scaffolds showed good biocompati-bility. The cells on hTDM showed cellular polarity in a regulardirection, which suggested differentiation potential (Fig. 7).

3.3.2. Cell viabilityThe time-dependent cell viability data on different scaffolds is

represented as a histogram in Fig. 8. From day 1 to day 5, DFCs on

R. Li et al. / Biomaterials 32 (2011) 4525e45384530

Fig. 4. Immunohistochemical examination of TDM indicates that it was positive for COL-1, TGF-b1, DMP-1, DSP, decorin, and biglycan, but negative for the control of PBS substitutedfor the primary antibody.

Fig. 3. SEM examination of hTDM (A) and untreated dentin matrix (B). The dentinal tubules were exposed sufficiently and fiber bundles of intertubular and peritubular dentinbecame loose in TDM (A). These morphological features were not found in untreated dentin matrix (B). Using the whole hTDM and hUDM structure to create liquid extracts detectedhigher protein concentrations from hTDM than hUDM.

R. Li et al. / Biomaterials 32 (2011) 4525e4538 4531

Fig. 5. Comparison of TDM and UDM, showing average concentration of proteins and factors contained in TDM was the same as that of UDM, indicating the modified methodpreserved dentinogenesis-related proteins and factors.

R. Li et al. / Biomaterials 32 (2011) 4525e45384532

hTDM showed a statistically significant higher viability then thoseon HA/TCP. This indicates a better cell viability of DFCs on hTDMand that the hTDM possessed excellent biocompatibility.

3.3.3. Cell proliferationTo evaluate the effect of hTDM on the proliferation of DFCs,

changes to the cell-cycle of experimental and control groups wereevaluated by FCM analysis. The results showed that test cells(hTDM) presented a significantly higher percentage of cells inG2 þ M þ S phases (54.80 � 0.14%) than the HA/TCP control group(48.50 � 0.14%) with p < 0.05 (Fig. 9). The result indicated thathTDM was able to promote proliferation of the DFCs.

3.3.4. Cell migrationTo assess cell motility, a scratch assay was performed on the DFC

monolayer. Interestingly, 24 h after injury of the cell layer, a clearmotility difference was observed (Fig. 10). DFCs, treated by hTDMliquid extract, migrated more quickly than those treated by normalmedium or HA/TCP liquid extract. Furthermore, the cells of thehTDM-treated group moved in a uniform direction, which was notfound in the control groups. There was no obvious difference inbehavior for treatment with normal medium or the liquid extract ofHA/TCP (data not shown).

3.3.5. In vitro cytotoxicity of hTDMAccording to the ISO standard, both hTDM and HA/TCP showed

good biocompatibility and could be used as in vivo scaffolds. Thecytotoxicity of hTDM was very low at a level from 0 to 1, while thecytotoxicity of HA/TCP was higher at a level from 1 to 2. Interest-ingly, the cytotoxicity level of hTDM at day 1 was 0, which indicatesthat this material may be beneficial to cell proliferation. Table 2summarizes the cytotoxicity level and RGR of hTDM and HA/TCPmaterials.

3.4. Dentin regeneration in vivo by using TDM as scaffold andinductive microenvironment

All experimental animals recovered quickly, showing no signs ofdiscomfort after implant surgery. Histological examination showedthat hTDM could integrate with host tissues without any inflam-matory response, further demonstrating the good biocompatibility

of TDM. Complete dentin tissues were regenerated on hTDM, withdentinal tubule structure visible in H&E and Mason staining (Fig. 11A, B, D, and E). In the regenerated dentin tissues, distinctive dentinstructures were clearly observed, such as dentinal tubules, pre-dentin, spherically shaped mineralized nodules, and polarizingodontoblast-like cells visibly secreting matrix. Importantly, cellnuclei positive for hematoxylin could be found in the regeneratedtissues, which could not be seen in the raw hTDM (Fig. 11C). In theHA/TCP control groups, these typical features of dentin tissues werenot seen. Only bone-like tissues were formed around HA/TCPcontrols (Fig. 11G). Likewise, no dentin tissue formed by usinga TDM without DFC (Fig. 11F).

Expression of DSP and DMP-1, markers of dentin, were detectedin the tissue formed in hTDM (Fig.11H and I). Therefore, this furtherindicates that the regenerated tissues were dentin. In order toinvestigate whether the implanted human DFCs were responsiblefor the regenerated tissues, anti-humanmitochondria, which couldonly react with a human source, were employed. The regeneratedtissue and its surrounding cells were positive for human mito-chondria, indicating that implanted TDM and DFCs participated inhuman dentin regeneration in the mouse model (Fig. 11J).

4. Discussion

Many materials have been adopted as scaffolds for dentinregeneration [3,6,8,17]; however, few have succeeded in obtainingcomplete dentin tissue. Studies have indicated that acellular dentinmatrix may be a suitable scaffold for tooth tissue engineering dueto its non-immunogenicity, suitable mechanical properties, andrichness in potentially dentinogenetic factors [14,18,19]. Theorganic matrix of dentin has been reported to contain approxi-mately 30 volume percent of collagen, noncollagenous proteins(NCPs), and growth factors [20e22], and many of these proteinsand factors have been shown to be important in dentin develop-ment, mineralization, and regeneration [12,21]. However, in orderto realize the odontogenic potential of dentin matrix, its inductiveproteins and factors need to be liberated in order to induce seededcells towards an ondontogenic specialization.

In endodontics, EDTA is a widely known chelating agent capableof loosening dentin structure and effectively removing the smearlayer of organic and inorganic debris created in root canal therapy

Fig. 6. SEM examination of cell morphology on different scaffolds. From day 1 to day 9, cells on hTDM showed higher cell proliferation than those on HA/TCP. This indicated thathTDM had good biocompatibility and supported cell growth. Interestingly, cells on TDM showed a pattern of proliferation in an orderly, regular direction visible from day 3 to day 9.

R. Li et al. / Biomaterials 32 (2011) 4525e4538 4533

Fig. 7. Appearance of DFCs growing around different scaffolds. From day 1 to day 9, the cells surrounding TDM generally grew more quickly than those around HA/TCP, althoughafter 5 days, the number of cells around the two groups were not significantly different. Furthermore, cells around TDM were oriented and polarized in a common direction.

R. Li et al. / Biomaterials 32 (2011) 4525e45384534

[23,24]. The combined use of EDTA and ultrasonic cleaning hasbeen shown to be particularly effective for smear layer and debrisremoval [25]. The effect of different EDTA concentrations has beenpreviously explored in smear layer removal and dentin tubuleexposure [26e29]. Treatment time is also known to affect the

resulting dentin structure [29]. The dentin treatment protocol mustbe highly optimized as excessive demineralization can destroy thestructure of dentin and also affect the containment and function-ality of odontogenic factors, while insufficient demineralizationgenerates a less odontogenic scaffold. Therefore, our general

Fig. 8. Cell viability of DFCs cultured on hTDM or HA/TCP from day 1 to day 5. Cell viability on hTDM showed a obvious improvement over that on HA/TCP. * means p < 0.05.

R. Li et al. / Biomaterials 32 (2011) 4525e4538 4535

treatment strategy for dentin matrix involved combined treatmentwith EDTA and ultrasonic cleaning optimized in EDTA concentra-tion and incubation time, which was intended to liberate poten-tially odontogenic factors and loosen dentin structure.

Our previous study [14] showed that the TDM of rats was aneffective scaffold in dentin regeneration by using a two-stepprotocol with 17% and 5% EDTA. However, in a preliminary study,we found the methodology designed for rat dentin was not capableof sufficiently exposing dentin tubules, or loosening intertubularand peritubular dentin structure, in human dentin matrix. There-fore, because of these species-specific differences between humanand rat dentin matrix, different concentrations of EDTA anddifferent treatment times were utilized in this study. An additionalstep of adding 10% EDTA to human dentin matrix treatmentprotocol was taken, and incubation times were increased slightlyover those used to treat rat tissue. Moreover, a step-by-stepdemineralization was found to be very important in order tomaintain functional proteins and factors. Our results indicate thatthe modified protocol was effectively optimized, with key factors

Fig. 9. Cell cycle test showing that approximately 45% of cells treated by hTDM arrested in GTDM enhanced proliferation of DFCs.

preserved and capable of being released into liquid extracts frompulverized and whole hTDM structures.

While it is possible that proteins and other elements of thedentin matrix are damaged during treatment, the overall concen-tration of key factors was found to be statistically equivalent tothose of untreated human dentin matrix, thereby indicating thatthe treatment did not significantly affect key factors. Most impor-tantly, when measuring factor release from a whole scaffold, it wasfound that a whole hTDM released more key factors than the wholehUDM. This is likely due to the exposure of dentin tubules andloosening of the intertubular and peritubular dentin and indicatesan optimized treatment protocol.

In this study, ELISA kits were used to analyze the concentrationof various proteins. In a wide array of studies, ELISA assays havebeen used quantitatively and qualitatively to detect factors orproteins contained in serum or liquid extracts with high sensitivity[30,31]. In the present study, hTDM was found to express COL-1,TGF-b1, decorin, biglycan, DMP-1, and DSP. These dentinogenesis-related factors not only play important roles in inducing DFC

0 þ G1, compared with 52% which were not treated by TDM. The result indicated that

Fig. 10. Cell motility of DFCs treated by TDM liquid extract (experimental group) in comparison with DFCs treated by HA/TCP liquid extract. Arrows indicate the range of cellmigration in 24 h.

R. Li et al. / Biomaterials 32 (2011) 4525e45384536

proliferation and differentiation into odontoblasts, but also forma network to form dentin tissues and control mineralization duringdentin regeneration and development [32e37].

HA/TCP has been used as a scaffold in dentin and bone tissueengineering [38e40]. However, it was not possible to obtaincomplete dentin tissues with HA/TCP scaffolds. In our study, hTDMand HA/TCP scaffolds were seeded with DFCs and cultured in vitro.Examination of scaffolds by SEM and light microscopy indicatedgreater early DFC cell proliferation on TDM than on HA/TCP. The cellmobility test demonstrated that DFCs have a greater mobility inresponse to hTDM liquid extract than to HA/TCP liquid extract.Cytotoxicity tests indicated hTDM had better biocompatibility thanHA/TCP. This evidence indicates that in vitro, hTDM scaffolds aremore suitable and bioactive for DFCs than HA/TCP scaffolds.

To further investigate the dentinogenetic functions of hTDM inan in vivo environment, hTDM scaffolds seeded with DFCs wereimplanted into mice at a site with no mineralization capacity.Implantation immediately after seeding DFCs to the hTDM scaffoldsdid not generate the most satisfactory results. Rather, a 3 day invitro culture of hTDM scaffolds and seeded DFC cells generated thebest outcome for in vivo implantations. As the risk of contaminationfor in vitro culture likely increases with time, a minimal 3 day invitro culture prior to implantation was deemed most suitable.

Histological examination of implanted hTDM/DCF constructsshowed newly regenerated tissues which were positive for twoidentified markers for dentin and odontoblasts: DMP-1 and DSP

Table 2Cytotoxicity levels and RGR of hTDM and HA/TCP materials. Group A indicates TDM; gro

Group Day 1 Day 2 Day 3

RGR (%) cytotoxicity level RGR (%) cytotoxicity level RGR (%)

A 103.75 0 83.3 1 87.5B 90.4 1 76.5 1 79.9

[41,42]. The regenerated complete dentin included obvious pre-dentin, dentin, and odontoblast layers. Cell nuclei positive forhematoxylin were only observed in the regenerated dentin tissues,an observationwhich was not found in our previous study [14]. Thisevidence suggests that the cells supported by the hTDM scaffoldsecrete extracellular matrix and are responsible for the regenerateddentin. In addition, the regenerated tissues and surrounding cells ofthe hTDM were positive for human mitochondria, indicating thatthe implanted seeding cells were responsible for the regeneratedtissues.

In contrast with hTDM, only mineralized tissues or bone-liketissues were formed around the implanted HA/TCP-DFC constructs,demonstrating that factors inherent in hTDM are required forcompleted dentin tissue regeneration, and that hTDM is a moresuitable scaffold for dentin tissue regeneration than HA/TCP. Inaddition, our results show that the seeding cells (in our case DFCs)are main co-contributors in the tissue engineering of dentin, as nodentin tissues could form using hTDM without cells. Previousresearch has implied dental papilla cells and dental pulp cells arethe best seeding cells for dentin regeneration, but few studies havesuccessfully regenerated complete dentin [39,43,44]. In this andour previous study, it was demonstrated that while DFCs areregarded as precursor cells of periodontal tissues [45,46], DFCs canregenerate complete dentinwhen combined with TDM in vivo. Thissuggests DFCs may be the most suitable seeding cells for completedentin tissue regeneration.

up B indicates HA/TCP.

Day 4 Day 5

cytotoxicity level RGR (%) cytotoxicity level RGR (%) cytotoxicity level

1 77.9 1 92 11 74.5 2 67.2 2

Fig. 11. Histological examination of TDM combined with DFCs implanted in vivo. Panels A, B, D and E show formation of distinctive dentin structures. Most importantly, the cellnuclei in these samples stained positive for hematoxylin (Fig. C). No dentin tissue was formed using TDM without seeding cells (F). On HA/TCP, only mineralized matrix-like tissues(MM) were formed (G). Regenerated tissues on TDM were positive for DSP and DMP-1, indicating that they were dentin tissues (H and I). Furthermore, regenerated tissue and itssurrounding cells were positive for human mitochondria, indicating that implanted TDM and DFCs positively participated in dentin regeneration (J). The regenerated tissues werenegative for the control of PBS substituted for the primary antibody (K). (B) Enlargement of dashed boxed area in (A). (C) Enlargement of solid boxed area in (A). (E) Enlargement ofsolid boxed area in (D).

R. Li et al. / Biomaterials 32 (2011) 4525e4538 4537

5. Conclusions

Human treated dentin matrix, in combination with humandental follicle cells, is a suitable composition for complete dentintissue regeneration using tissue engineering methodologies. Aprevious methodology designed for treating rat dentin was opti-mized for human dentin. Treated human dentin scaffolds showedsuperior biocompatibility and bioactivity in comparison to calciumphosphate controls. In addition, hTDM scaffolds were capable of

inducing complete human dentin tissue regeneration in an in vivoimplantation. However, some problems remain to be solved. Moreinformation is needed to understand how the hTDM microenvi-ronment induces tooth regeneration in order to utilize themodifiedtreatment method for clinical tooth reconstruction. In addition,assessment of whether hTDM has the capacity to induce non-odontogenic cells, such as adipose-derived stem cells or bonemarrow stromal cells, into an odontogenic specialty are of interest.These issues will be addressed in future investigations. The concept

R. Li et al. / Biomaterials 32 (2011) 4525e45384538

of utilizing depleted human teeth in tissue engineering mayprovide a new and more viable source for dentin tissue regenera-tion in the future and may even support efforts to reconstruct thetooth root.

Acknowledgements

This study was supported by grants from the Nature ScienceFoundation of China (project nos. 2010CB944800 and 30973348)and theKey Technology R&D Program of Sichuan Province (projectnos. 10ZC1618).

Appendix. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.biomaterials.2011.03.008.

References

[1] Young CS, Abukawa H, Asrican R, Ravens M, Troulis MJ, Kaban LB, et al. Tissue-engineered hybrid tooth and bone. Tissue Eng 2005;11:1599e610.

[2] Huang GTJ, Sonoyama W, Liu Y, Liu H, Wang SL, Shi ST. The hidden treasure inapical papilla: the potential role in pulp/dentin regeneration and biorootengineering. J Endodont 2008;34:645e51.

[3] SonoyamaW, LiuY, FangD, YamazaT, SeoBM, ZhangC, et al.Mesenchymal stemcell-mediated functional tooth regeneration in swine. Plos One 2006;1:e79.

[4] Abukawa H, Zhang W, Young CS, Asrican R, Vacanti JP, Kaban LB, et al.Reconstructing mandibular defects using autologous tissue-engineered toothand bone constructs. J Oral Maxil Surg 2009;67:335e47.

[5] Nakashima M, Akamine A. The application of tissue engineering to regener-ation of pulp and dentin in endodontics. J Endod 2005;31:711e8.

[6] Scheller EL, Krebsbach PH, Kohn DH. Tissue engineering: state of the art in oralrehabilitation. J Oral Rehabil 2009;36:368e89.

[7] Duailibi MT, Duailibi SE, Young CS, Bartlett JD, Vacanti JP, Yelick PC. Bio-engineered teeth from cultured rat tooth bud cells. J Dent Res 2004;83:523e8.

[8] Duailibi SE, Duailibi MT, Zhang W, Asrican R, Vacanti JP, Yelick PC. Bio-engineered dental tissues grown in the rat jaw. J Dent Res 2008;87:745e50.

[9] Kuo TF, Huang AT, Chang HH, Lin FH, Chen ST, Chen RS, et al. Regeneration ofdentin-pulp complex with cementum and periodontal ligament formationusing dental bud cells in gelatin-chondroitin-hyaluronan tri-copolymer scaf-fold in swine. J Biomed Mater Res A 2008;86A:1062e8.

[10] Thesleff I, Vaahtokari A. The role of growth factors in determination anddifferentiation of the odontoblastic cell lineage. Proc Finn Dent Soc 1992;88(Suppl. 1):357e68.

[11] Stern MM, Myers RL, Hammam N, Stern KA, Eberli D, Kritchevsky SB, et al. Theinfluence of extracellular matrix derived from skeletal muscle tissue on theproliferation and differentiation of myogenic progenitor cells ex vivo.Biomaterials 2009;30:2393e9.

[12] Chun SY, Lee HJ, Choi YA, Kim KM, Baek SH, Park HS, et al. Analysis of thesoluble human tooth proteome and its ability to induce dentin/tooth regen-eration. Tissue Eng Part A 2011;17:181e91.

[13] Lluch AV, Fernandez AC, Ferrer GG, Pradas MM. Bioactive scaffolds mimickingnatural dentin structure. J BiomedMater Res B Appl Biomater 2009;90:182e94.

[14] Guo WH, He Y, Zhang XJ, Lu W, Wang CM, Yu H, et al. The use of dentin matrixscaffold and dental follicle cells for dentin regeneration. Biomaterials 2009;30:6708e23.

[15] Morsczeck C, Gotz W, Schierholz J, Zeilhofer F, Kuhn U, Mohl C, et al. Isolationof precursor cells (PCs) from human dental follicle of wisdom teeth. MatrixBiol 2005;24:155e65.

[16] Jo YY, Lee HJ, Kook SY, Choung HW, Park JY, Chung JH, et al. Isolation andcharacterization of postnatal stem cells from human dental tissues. Tissue Eng2007;13:767e73.

[17] Honda MJ, Tsuchiya S, Sumita Y, Sagara H, Ueda M. The sequential seeding ofepithelial and mesenchymal cells for tissue-engineered tooth regeneration.Biomaterials 2007;28:680e9.

[18] Zaman KU, Sugaya T, Hongo O, Kato H. A study of attached and orientedhuman periodontal ligament cells to periodontally diseased cementum anddentin after demineralizing with neutral and low pH etching solution.J Periodontol 2000;71:1094e9.

[19] Fukazawa E, Nishimura K. Superficial cemental curettage: its efficacy inpromoting improved cellular attachment on human root surfaces previouslydamaged by periodontitis. J Periodontol 1994;65:168e76.

[20] Marshall Jr GW, Marshall SJ, Kinney JH, Balooch M. The dentin substrate:structure and properties related to bonding. J Dent 1997;25:441e58.

[21] Park ES, Cho HS, Kwon TG, Jang SN, Lee SH, An CH, et al. Proteomics analysis ofhuman dentin reveals distinct protein expression profiles. J Proteome Res2009;8:1338e46.

[22] Liu J, Jin T, Chang S, Ritchie HH, Smith AJ, Clarkson BH. Matrix and TGF-beta-related gene expression during human dental pulp stem cell (DPSC) miner-alization. In Vitro Cell Dev Biol Anim 2007;43:120e8.

[23] Lottanti S, Gautschi H, Sener B, Zehnder M. Effects of ethylenediaminetetra-acetic, etidronic and peracetic acid irrigation on human root dentine and thesmear layer. Int Endod J 2009;42:335e43.

[24] Hulsmann M, Heckendorff M, Lennon A. Chelating agents in root canaltreatment: mode of action and indications for their use. Int Endod J 2003;36:810e30.

[25] Kuah HG, Lui JN, Tseng PSK, Chen NN. The effect of EDTA with and withoutultrasonics on removal of the smear layer. J Endodont 2009;35:393e6.

[26] Nakashima K, Terata R. Effect of pH modified EDTA solution to the propertiesof dentin. J Endod 2005;31:47e9.

[27] Perez F, Rouqueyrol-Pourcel N. Effect of a low-concentration EDTA solution onroot canal walls: a scanning electron microscopic study. Oral Surg Oral MedOral Pathol Oral Radiol Endod 2005;99:383e7.

[28] Di Lenarda R, Cadenaro M, Sbaizero O. Effectiveness of 1 mol L-1 citric acidand 15% EDTA irrigation on smear layer removal. Int Endod J 2000;33:46e52.

[29] Sen BH, Erturk O, Piskin B. The effect of different concentrations of EDTA oninstrumented root canal walls. Oral Surg Oral Med Oral Pathol Oral RadiolEndod 2009;108:622e7.

[30] Vigano L, Benfenati E, van Cauwenberge A, Eidem JK, Erratico C, Goksoyr A,et al. Estrogenicity profile and estrogenic compounds determined in riversediments by chemical analysis, ELISA and yeast assays. Chemosphere 2008;73:1078e89.

[31] Aung FM, Cano P, Fernandez-Vina M, Lichtiger B. Results of HLA antibodytesting using ELISA vs the fluorescent bead method and retrospective reviewof data for recipients of packed RBCs and platelets from male HLA-immunizeddonors. Am J Clin Pathol 2011;135:90e5.

[32] Piattelli A, Rubini C, Fioroni M, Tripodi D, Strocchi R. Transforming growthfactor-beta 1 (TGF-beta 1) expression in normal healthy pulps and in thosewith irreversible pulpitis. Int Endod J 2004;37:114e9.

[33] Reed CC, Iozzo RV. The role of decorin in collagen fibrillogenesis and skinhomeostasis. Glycoconj J 2002;19:249e55.

[34] Yamaguchi Y, Mann DM, Ruoslahti E. Negative regulation of transforminggrowth factor-beta by the proteoglycan decorin. Nature 1990;346:281e4.

[35] Schonherr E, Broszat M, Brandan E, Bruckner P, Kresse H. Decorin core proteinfragment Leu155-Val260 interacts with TGF-beta but does not compete fordecorin binding to type I collagen. Arch Biochem Biophys 1998;355:241e8.

[36] Ferdous Z, Wei VM, Iozzo R, Hook M, Grande-Allen KJ. Decorin-transforminggrowth factor-interaction regulates matrix organization and mechanicalcharacteristics of three-dimensional collagen matrices. J Biol Chem 2007;282:35887e98.

[37] Goldberg M, Septier D, Rapoport O, Iozzo RV, Young MF, Ameye LG. Targeteddisruption of two small leucine-rich proteoglycans, biglycan and decorin,excerpts divergent effects on enamel and dentin formation. Calcif Tissue Int2005;77:297e310.

[38] Ando Y, Honda MJ, Ohshima H, Tonomura A, Ohara T, Itaya T, et al. Theinduction of dentin bridge-like structures by constructs of subcultured dentalpulp-derived cells and porous HA/TCP in porcine teeth. Nagoya J Med Sci2009;71:51e62.

[39] Gronthos S, Mankani M, Brahim J, Robey PG, Shi S. Postnatal human dentalpulp stem cells (DPSCs) in vitro and in vivo. Proc Natl Acad Sci U S A 2000;97:13625e30.

[40] Arinzeh TL, Tran T, McAlary J, Daculsi G. A comparative study of biphasiccalcium phosphate ceramics for human mesenchymal stem-cell-induced boneformation. Biomaterials 2005;26:3631e8.

[41] Papagerakis P, Berdal A, Mesbah M, Peuchmaur M, Malaval L, Nydegger J, et al.Investigation of osteocalcin, osteonectin, and dentin sialophosphoprotein indeveloping human teeth. Bone 2002;30:377e85.

[42] Toyosawa S, Okabayashi K, Komori T, Ijuhin N. mRNA expression and proteinlocalization of dentin matrix protein 1 during dental root formation. Bone2004;34:124e33.

[43] Huang GT. Pulp and dentin tissue engineering and regeneration: currentprogress. Regen Med 2009;4:697e707.

[44] Peng L, Ye L, Zhou XD. Mesenchymal stem cells and tooth engineering. Int JOral Sci 2009;1:6e12.

[45] Yao S, Pan F, Prpic V, Wise GE. Differentiation of stem cells in the dentalfollicle. J Dent Res 2008;87:767e71.

[46] Yalvac ME, Ramazanoglu M, Gumru OZ, Sahin F, Palotas A, Rizvanov AA.Comparison and optimisation of transfection of human dental follicle cells,a novel source of stem cells, with different chemical methods and electro-poration. Neurochem Res 2009;34:1272e7.