Dental filling materials come into long-term contact with living tissues. Therefore – according to national and international medical device legislation – they must pass a certification process before they are allowed to be marketed (1). Part of this certification is a clinical risk assessment including a biocompatibility evaluation. Basic biological information in this context is – amongst others – derived from cytotoxicity tests and, therefore, cytotoxicity test methods are included into the relevant international standards which are commonly used in the course of clinical risk assessment (1-3).
Such cytotoxicity tests can be classified in (1) those with direct contact between the target cells and the material/eluate from a material and (2) those with a barrier between them (2, 3). For cytotoxicity testing of dental filling materials, bovine dentin disks with varying thickness have been described as a suitable barrier simulating the in vivo situation in the mouth (4-8). Such a test method (dentin barrier test) has been included into the relevant standard for the preclinical evaluation of the biocompatibility of dental materials ISO-7405 (2).
With this test method, three dimensional cell cultures were recommended for recording cell reactions as a surrogate for the pulp tissue. For this purpose, immortalized bovine pulp-derived fibroblasts were cultured on a polyamide mesh (2, 6). In order to further approach the clinical situation, immortalized human pulp-derived (9) cells could be used with the additional advantage that for these cells a multitude of diagnostic cell markers are available – in contrast to bovine cells. Immortalized human pulp-derived pulp-cells have been generated by introducing the SV 40 large T antigen into primary pulp-derived cells, by which the tumor suppressor protein p53 was inactivated (9). The expression of a number of cell markers showed close similarity of the immortalized cells with those of primary pulp-derived fibroblast without immortalization (9). However, initial experiments showed that human immortalized pulp-derived cells could not be cultured on the polyamide meshes, as was successfully done with the immortalized bovine cells (3, 4, 6). Therefore, the aim of this study was to evaluate the suitability of three different hydrogels commonly used in tissue engineering (fibrin, collagen and peptides) (10) for establishing a three-dimensional culture with immortalized human pulp-derived cells to be used in the dentin barrier test. Furthermore, the cytotoxicity of a new resin-based composite material, for which data have not yet been reported, was evaluated using these new cell cultures in the dentin barrier test and compared to relevant controls and to a reference material. The underlying hypotheses were that (1) all tested hydrogels are suitable for the intended purpose and (2) the cytotoxicity of the new resin based material is not significantly different from a reference (=commonly used) resin based filling materials.
Materials and Methods
Pulp derived fibroblasts from human wisdom teeth transfected with the SV40 large T antigen  were used as a continuously proliferating (“immortalized”) cell-line. Cells from a frozen stock (up to passage 43) were thawed and used for the experiments. Cells were cultured in MEMα (GIBCO) with 20% FCS (GIBCO). The experiments for hydrogel evaluation were performed in 48 well plates, the material evaluation in the split perfusion chamber as recommended for the dentin barrier test by ISO 7405  (see below). All incubations were performed at 37°C and 5% CO2.
Fibrinogen from bovine plasma (Sigma Aldrich) was solubilized in TBS buffer (10, 15 and 20 mg/ml) and thrombin (Sigma Aldrich) in 20 mM CaCl2 (15, 100 and 150 U/ml). Equal amounts of fibrinogen and the cell suspension (cell densities ranging from 2x105 to 6.25x102 and without cells) were mixed (60 µl) and used with and without a polyamide mesh. Subsequently, the thrombin-solution (twice the volume of fibrinogen) was mixed with fibrinogen and gelation occurred immediately. The gel was then overlaid with culture medium.
For the peptide hydrogel (Puramatrix, BD Biosciences) 10 mg/ml was liquefied in an ultrasonic bath and then mixed in equal amounts with 20% sucrose. Cells were added and each gel (50 µl) contained 105 cells, again with and without the polyamide mesh. 100 µl of culture medium was carefully pipetted on the peptide which led to a gelation within one hour.
For the collagen hydrogels, type I collagen from rat tails (BD Biosciences) was mixed with 1 N NaOH and 10 x PBS to a final concentration of 3.5 mg/ml at physiological pH. Cell suspensions in culture medium with different cell concentrations resulting in final cell numbers ranging from 105 to 1.25x104 cells per gel were added to a final volume of 10 µl with or without a polyamide mesh. Gelation occurred within 15 minutes and then the gel was overlaid with culture medium. Incubation time was up to 28 days with culture medium replacement three times a week with three replicates for each incubation time. In order to define the optimal cell seeding density in the collagen hydrogel, cells numbers ranging from 1.25x104 to 105 were seeded onto the meshes and incubated for 24 hours. Additionally, tissue samples after 14 and 28 days of incubation were fixated in 2.5% formalin for 24 h and embedded in Epon. The semi-thin sections were stained with toluidine-blue.
Polyamide meshes (Reichelt Chemietechnik, Heidelberg, Germany) were used for both the evaluation of the hydrogels and in the dentin barrier test. Specifications are given in ISO-7405 (2): mesh width: 150 µm, thickness of the filaments of 95 µm and a total thickness of the mesh: 155 µm. The meshes were cut into round samples with 0.5 cm2 surface on each side. Before use, the meshes were cleaned for 30 minutes in 0.1 M acetic acid and then rinsed in distilled water 5 times.
Mechanical stability of the gels was regarded sufficient if the gel could be transported without destruction from the 48-well plates to the split perfusion chamber of the dentin-barrier-test. Contractions of the gels or degradation were evaluated visually under an inverted microscope after different periods of incubation for any changes of form.
Cell vitality of the cultures was evaluated using the MTT-assay and the WST-1 cell proliferation assay. The MTT-test is recommended in ISO-7405 for the dentin barriers test (2). Briefly after the removal of the supernatant cell culture medium, 100 µl of MTT solution (Sigma Aldrich) was added to the cultures and incubated for 2 hours. Cells were then solubilized with 200 µl DMSO and 100 µl/gel was transferred to a 96-well plate. Color development was determined photometrically at 540 nm. Accordingly, WEST-1 reagent (Roche) was mixed with 5 ml of the corresponding electrolyte solution and 10 µl in 90 µl culture medium were pipetted onto the gels, after the original culture medium had been removed. After 2 hours incubation, the cultures were shaken for 5 minutes (100/min) and then 100 µl of the supernatant was transferred to a 96-well plate. Photometric measurements were performed at 450 nm. Optimal initial cell seeding density in collagen hydrogels on the polyamide meshes was evaluated microscopically and the criterion for optimal cell density was the smallest cell number but the lack of cell voids (confluent cell layer) between the filaments of the meshes.
Dentin barrier test
The test was performed according to ISO 7405 (2). For a short time, the test cells were seeded in the collagen hydrogel on the polyamide mesh as described above and then incubated for two weeks with culture medium replacement three times per week. Dentin disks from lower anterior bovine front teeth were prepared as described elsewhere  and 200 µm thick slices closest to the pulp were used. The dentin slices and the two weeks three-dimensional cultures were mounted into the split-diffusion chamber (Minucells&Minutissues, Bad Abbach, Germany), as recommended by ISO 7495 . The chamber was perfused with culture medium for 24 hours at 0.3 ml/h. Then the test materials were applied on the other side of the dentin disk according to the instructions of the manufacturers and the system was incubated for another 24 hours with the same perfusion. The cultures were carefully removed from the split perfusion chamber and the cell reaction was recorded using the WST-1 assay as described above.
The tested materials, manufacturers and Lot-numbers are listed in Table 1. The two resin-based composites – one commonly used reference material (Tetric EvoCeram) and the test material (N’Durance) were applied with and N’Durance also without a commonly used two-bottle self-etch adhesive (Clearfil SE Bond). Negative controls (non-toxic control) and positive controls (Meshes without cells) as well as a toxic reference material (standard toxic material) were used according to ISO 7405 (2). The resin based composites and the dentinal adhesives were light-cured for 30 seconds with a light curing unit (Optilux 400, Demetron, Danbury, USA) with an irradiance of 699 mW/cm2 .
For the dentin barrier test three independent experiments were performed with 4 replicates for each material (including controls). For each experiment, photometrical readings of the materials were related to the negative control (= 100%) and the meshes without cells (= 0%) and expressed as % relative growth. Data are presented as medians with 25/75 percentiles. Statistical analysis was performed using the non-parametrical Kruskal-Wallis-test (U-test). The level for significance was set to p< 0.05.
For the fibrin and the peptide hydrogels with and without the polyamide mesh the cultures could not be transferred from the culture plate to the split perfusion chamber without the destruction of the gels, thus the mechanically stability was considered insufficient. Furthermore, high cells numbers (2x105 to 5x104) caused fibrin gel degradation after 5 days of incubation. Also, the fibrin and peptide gel could not be sufficiently dissolved in DMSO (MTT-test) or generated data were not reproducible (WST-1). Therefore, the evaluation of cell reaction using the MTT or the WST-1 assay was problematic in these two types of hydrogels.
In contrast, the collagen hydrogel proved to have sufficient mechanical stability. Depending on the cell number, collagen gels contracted. Light microscopy observations, however, showed that only with a cell number of 2x105 after 5 days of incubation, a visible contraction occurred. Lower cell numbers did not cause microscopically detectable contraction after 14 days of incubation. The MTT assay could not be performed, because the gel could not be sufficiently dissolved. However, the WST-1 assay could be performed showing a linear relation of the photometric reading on the cell numbers ranging from 12 500 to 100 000 cells per gel (data not shown) with a correlation coefficient of r2= 0.9.
Optimal cell seeding density for the collagen gels was determined directly under the inverted microscope after 24 h incubation time. With an initial cell density of 5x104 an even distribution of the cells (Fig. 1) with no empty spaced between the meshes could be observed. Growth kinetics under these conditions (three replicates for each time point) show an continuous cell growth until an incubation time of 10 days, then a plateau phase until day 21 was observed and then a decline in cell numbers (Fig. 2). The histological sections (28 days of incubation) show a three-dimensional multi-layer culture between the mesh filaments (Fig. 3).
Dentin barrier test
Results of the dentin barrier test are shown in Fig. 4 and Table 2. No statistically significant difference could be observed between the results for the two resin-based composites with or without the dentinal adhesive and the non-toxic control material. On the other side, statistically significant differences were observed between the test materials or the non-toxic control and the toxic standard reference material.
The first aim of this study was to test three different hydrogel based scaffolds – commonly used in tissue engineering (10) – for their suitability to grow human immortalized pulp-derived fibroblast in three dimensional cultures, which are then used in the dentin barrier test as described in to ISO 7405 (2). Fibrin and peptide based hydrogels as used here were discarded primarily because of their insufficient mechanical properties. Despite the general suitability of these scaffolds for tissues engineering (10), mechanical properties play a special role here due to the fact that the cell scaffold constructs have to be transferred from the culture plate for growth to the split perfusion chamber for the dentin barrier test. Mechanical properties of such constructs are not only dependent upon the hydrogel type, but also on the incubation times required and upon the cell types (11). Poor mechanical properties have also been described for fibrin (12), and the additional support by the polyamide mesh did not sufficiently improve the stability. A second reason for not using fibrin and peptide based hydrogels was the fact that cell reactions in these hydrogels could not be evaluated using commonly used vitality tests such as the MTT or the WST-1 assay as described above. Modifications of the vitality test, e.g. by changing solvents (13) were not further followed because of the poor mechanical properties of the two hydrogels.
On the other hand, type I collagen as used in this study proved to be suitable for the given purpose. The general suitability of collagen (especially type I) as scaffold for tissue engineering has long been known (see survey in (10)). One reason is that collagen offers chemical and structural information of the natural extracellular matrix (ECM) (14). Collagen type I has also been described to be superior to other scaffolds (gelatin or chitosan) for cultivating and differentiating dental stem cells (15).
Collagen is biodegraded by enzymes (10), but this process apparently does not impair the suitability of collagen I for the purposes described here. Growth kinetics (Fig. 2) showed a constant increase of cell vitality for the first ten days with a plateau up to 21 days. Only after that period, a decline of cell vitality may indicate degradation of the scaffold. For the dentin barrier test the total time for cells growth and test is normally less than three weeks. Another point of interest is that constructs of collagen hydrogels containing (fibroblastic) cells are known to contract (16). This is dependent on the collagen concentration and the cell number (17). In this study, with a cell number of less than 2x105 up to 14 days of incubation no contraction could be observed microscopically. The cell proliferation assay conducted with an incubation period of up to 28 days showed a decrease of culture vitality only after the third week. Apparently, any possible contraction of the collagen cell construct did not influence the suitability of the construct until the third week. Histology confirmed the presence of a multilayer three dimensional culture even after a 28 days incubation period.
The WST-1 assay was technically easy to perform with cells in the collagen hydrogels and showed a linear relation of the photometric reading with the seeded cell density for cell numbers relevant in the dentin barrier test. Therefore, the WST-1 assay was considered suitable for the evaluation of the cell reaction. Both, the MTT and the WST-1 assay are based on the use of tetrazolium salts that are metabolically reduced to highly colored end products called formazans (18). In contrast to MTT, the reduction of Wst-1 (so-called second generation tetrazolium salts) into soluble formazan is apparently extracellular or plasma membrane associated (18). Therefore, the cell disruption and solubilization of insoluble MTT-derived formazan are not necessary, which apparently makes the Wst-1 assay more suitable for these hydrogel cell constructs than the MTT assay.
Due to the reasons mentioned above, the hypothesis that all three tested hydrogels under the given conditions are suitable as cell carriers has to be rejected. Collagen proved to be a suitable cell scaffold used together with a polyamide mesh for cultures to be used in the dentin barrier test. Advantages of cells from human origins are that they are closer to the clinical situation and with those cells a multitude of analytical tools developed for cell biology, which are mainly developed to be used for cells from rodents or humans, can be applied.
Data for the non-toxic control material, and for the standard toxic control are in accordance with data from the dentin barrier test using immortalized bovine pulp-derived cells in three dimensional cultures (19). The tested dentinal self-etch adhesive (Clearfil SE Bond) proved to be slightly toxic in the dentin barrier test with 500 µm dentin disks using bovine derived pulp cells (8). However, the dentinal bonding was tested alone without a composite, which might influence the reaction. Kim et al. (20) also tested Clearfil SE Bond alone in a dentin barrier test and showed approximately 50% of reduction of cell viability for 500 µm dentin disks. However, other cells were used by Kim et al. (L-929 mouse fibroblasts) and the control did not involve the placement of a (non-toxic) material showing the influence of the technical manipulations of the cultures during the experiment and again no composite resin was placed on top of it. In this study, data for the toxic reference standard material (Tetric EvoCeram) used with a dentinal adhesive showed no toxic cellular reaction. This is in agreement with non-toxic reactions observed with TetricEvCeram in the same test using immortalized bovine pulp-derived cells (21), although another self-etch dentinal adhesive (Excite) was used in those experiments. Altogether, these data indicate that the use of immortalized human derived pulp cells render similar results as those derived from bovine cells in the same test device.
Data for the new restorative material (N’Durance) showed no cytotoxic reaction both with and without the use of a self-etch dentinal adhesive. No other reports in the literature on the cell reaction towards this material have been available, so far. The test material contains instead of bis-GMA a new resin based on dimer acid-based dimethacrylates (22), which were proven to have a significantly higher final double-bond conversion than common dimethacrylate monomers (22, 23). Furthermore, ethoxylated bisphenol A dimethacrylate (bis-EMA) and urethane dimethacrylate (UDMA) are part of the resin phase, which altogether amounts to 41% of the material (manufacturer information). In comparison with other resin based composites, N’Durance being a strongly hydrophobic material showed a low solubility in distilled water (24). These data are in line with the non-toxic reaction found in this study.
The second hypothesis of this study that the test and the reference materials elicit the same cell reaction in the dentin barrier test cannot be rejected. The extrapolation of data from cell culture experiments to the patient situation is generally very problematic (1). However, for the dentin barrier test using three dimensional cultures of pulp-derived cells, a good correlation of results from this test and clinical data has been shown for a limited number of materials with bovine cells (4, 6). Therefore, the results for the test material indicate that in the clinical situation no pulp damage is expected if the pulp is covered with an intact dentin layer.