2.1

Cements

Three experimental calcium-silicate hydraulic cements were designed and prepared. A calcium-silicate powder (CEM I white Aalborg, Aalborg, Denmark) subjected to thermal and mechanical treatments , was added with 5 wt% calcium chloride (wTC) or 5 wt% calcium chloride and 17 wt% bismuth oxide (wTC-Bi), or 5 wt% calcium chloride and 5 wt% alpha-tricalcium phosphate (wTC-TCP). The cements were finally grinded in an agate ball mill. MTA (White ProRoot MTA, Dentsply, Tulsa, USA) was used as control.

The cements were prepared by mixing for 15 s the powders with DPBS (Dulbecco’s phosphate buffered saline, Lonza, Lonza Walkersville Inc., Walkersville, MD, USA, cat. no. BE17-512) in a mortar with pestle, using a 1 g/300 μL powder/liquid ratio, and grinded to produce a homogeneous paste free from agglomerates and air bubbles. DPBS is a physiological-like buffered Ca- and Mg-free solution with the following composition (mM): K ⁺ (4.18), Na ⁺ (152.92), Cl ⁻ (139.48) PO ₄ ³⁻ (9.56) (sum of H ₂ PO ₄ ⁻ 1.5 and HPO ₄ ²⁻ 8.06).

Each material was layered on a 13 mm diameter Thermanox™ Plastic coverslip (Nalgene, Nunc International, NY, USA) with a 133 ± 0.1 mm ² exposed surface area, to obtain standard disks. Mechanical vibrations were used to make the surface flat and regular . MTA was mixed with its liquid (water) and prepared according to manufacturer directions.

2.2

Experimental design

After preparation, the disks were immediately immersed in 5 mL of DPBS (20 mL of medium for 1 g of cement paste) in hermetically sealed containers and maintained at 37 °C for 5 h (fresh samples = t 0) or for 14 days (14-day aged samples = t 14) or 28 days (28-day aged samples = t 28) to obtain samples of different age.

The storage medium was renewed every week. At the established time-points (5 h, 24 h, 14 and 28 days), the medium was carefully removed and its pH was evaluated. The outer surface and the inner structure, following sample fracturing, of the cements were analyzed using SEM/EDX, micro-Raman and ATR-FTIR, and assayed using human marrow stromal cells.

2.3

Setting time of cements

The setting time of calcium-silicate cements was determined using Gilmore needles (ASTM International, West Conshohocken, PA, USA) in accordance with ASTM standard C266-07 with the following modifications. Ten grams of cement were mixed instead of 650 g, and the physiological temperature of 37 °C was used instead of 25 °C.

The experimental cements were mixed as previously described and applied on a glass slab with a stainless-steel spatula. The mixtures were placed into a mould measuring 10 mm in diameter and 2 mm in thickness in accordance with ADA specification 2008 . The mould was completely filled and the excess material was removed to obtain a flat surface. The samples were then stored in a sealed curing chamber at 37 °C and 95 ± 5% relative humidity (RH).

The initial and final setting times were assessed by penetration/indentation test. The moulds were removed from the curing chamber and immediately tested for setting time to prevent the dehydration of cement surface. Each sample was used only for one penetration/indentation test and then discarded. Approximately hundred samples were used to assess the setting time of each experimental material.

The Gilmore initial time of setting end point was the elapsed time (minutes) between the mixing of cement with water and the first penetration measurement that does not mark the specimen surface with a complete circular impression. Verification measurements must be obtained within 90 s of the first initial set measurement. The Gilmore initial setting time was established when any indentation was left by a Gilmore needle weighing 113.4 g with a tip diameter of 2.12 mm.

After the initial setting time was measured, the specimens were tested every 10 min with a Gilmore needle weighing 453.6 g with a tip diameter of 1.06 mm. As the final setting time approached (i.e. no indentation), specimens were tested every minute to determine the exact Gilmore final setting time.

2.4

Radiopacity of cements

The radiopacity of cements was determined following the ISO 9917-2007 for water-based cements. Briefly, samples of 10 ± 0.1 mm diameter and 1.0 ± 0.1 mm height were prepared. Dental X-ray film of speed group D was used. Aluminum step wedge (60 mm long × 10 mm wide, having a thickness range from 1 to 6 mm in equally spaced steps of 1 mm) was used as reference. The cements were positioned on a X-ray film with the radiographic unit positioned at 3 cm and exposure time 0.13 s. The film was scanned, the equivalent thickness of aluminium was estimated from the film and examined using an image analysis software (Image J).

2.5

ESEM–EDX and SEM–EDX analyses

Samples were examined by means of an environmental scanning electron microscope equipped with an energy dispersive X-ray instrument (ESEM–EDX XL30, Phillips, Eindhoven, The Netherlands), using an accelerating voltage of 20–25 kV. Cement discs were placed directly onto the ESEM stub and examined without any preparation and manipulation (i.e. the samples were neither coated nor dehydrated for the analysis). All samples were initially analyzed under low vacuum conditions (9.9 Torr) at 100% RH and 4 °C. After the initial examination each sample was inspected at 40% RH, 4 °C and 4.9–2.9 Torr.

In addition, a SEM equipped with EDX (SEM–EDX 515, Philips, Eindhoven, The Netherlands) was used for the morphological analysis of outer and internal fractured cement surfaces and to determine the elemental composition of the samples. At 25 kV acceleration, the X-ray electron beam penetration of ESEM–EDX (inside a material with a density of about 3 g/cm ³ ) resulted 2.98 μm and consequently the volume excited and involved in the emission of characteristic X-rays from the constituting elements must be considered 10 μm ³ .

The samples were carefully rinsed with deionized water, dehydrated/desiccated at 37 °C, fractured to expose both the outer and the inner surfaces, mounted on a stub, and then coated with a very thin carbon film.

2.6

Scanning white-light interferometry

Quantitative three-dimensional topographical analysis was performed by the calculation of three different roughness parameters, namely S _a , S _z and S _sk . S _a denotes the average roughness of the surface. Despite its usefulness in the comparison of different surfaces, it can be deceptive, as completely different surfaces may present the same S _a mean depending on specific characteristics of their relief. S _z values define the average of the 5 largest peak to valley heights in the area of interest. The closer this value is to S _a , the more homogeneous are the irregularities on the surface. The relatively high S _z values define an irregularly rough surface, with considerable variations between peaks and valleys. S _sk parameter tells about the symmetry of the surface in relation to a theoretical mean plane that cuts the relief. Therefore, negative S _sk values indicate a predominance of valleys and positive ones reveal a prevalence of peaks.

Sample discs were prepared by incorporating 1 g of cement to 0.3 mL of DPBS, immediately immersed in DPBS and stored at 37 °C for 5 h, 14 or 28 days. After each storage period, the samples were removed from the solution and analyzed using a scanning white-light interferometer (Wyko NT 3300; Veeco Metrology Inc., Tucson, AZ, USA). The samples were positioned on circular acetate plates to acquire the shape of discs with a diameter of 1.5 cm. Nine samples per cement were prepared, totalizing 27 specimens, and three different samples for cement after different ageing were analyzed. During all the measurements, the samples were prevented from dehydration, which guaranteed that they were analyzed while still humid. The results were obtained by employing a vertical scanning interferometry technique, a scan length of 60 μm, and a working distance of 3.4 mm. An association of two lenses with magnifications of 50× and 1× was set up, which allowed each measurement to be performed within a field-of-view of 91.4 μm × 120.0 μm. On each disc, seven equidistant locations were measured, starting from its center and moving towards its periphery. Once each experimental group was comprised of 3 samples, the average surface roughness per cement and per storage timing was determined from 21 spots on the disc surfaces (7 measurements/sample; 3 samples/group). Considering the non-homogeneity of the variances and the independency of the samples, pairwise Kruskal–Wallis analysis and multiple comparisons test were used to verify statistical differences among the experimental groups at a significance level of 5%.

2.7

pH measurements

Cement discs were prepared and immersed in 5 mL of DPBS (pH 7.4) in a polypropylene sealed container and stored at 37 °C. The DPBS was renewed every week. The pH of the soaking solutions was collected after 24 h, 14 and 28 days and measured using a pH meter (Denver Instrument Basic pH meter, equipped with a Hamilton liq-glass electrode and ±0.01 resolution) previously calibrated with acidic–neutral–alkaline standard solutions. The pH was registered when the value was stable. The electrode was inserted into the soaking solutions at room temperature (24 °C) and each measurement was repeated three times. The mean pH was then plotted against recording time.

2.8

ATR/FTIR analyses

IR spectra were recorded on a Nicolet 5700 FTIR, equipped with a Smart Orbit diamond attenuated total reflectance (ATR) accessory and a DTGS detector; the spectral resolution was 4 cm ⁻¹ and 128 the number of scans for each spectrum. The ATR area had a 2 mm diameter. The IR radiation penetration was about 2 μm.

To minimize problems deriving from possible lack of homogeneity of the sample, five spectra were recorded on each specimen both on the upper surface and on the inner fractured side. IR spectra were recorded also on the unhydrated cement powders as well as on alpha-TCP used for preparing the WTC-TCP cement. The reported IR spectra were the average of the spectra recorded on five different points.

2.9

Micro-Raman analyses

Micro-Raman spectra were obtained using a Jasco NRS-2000C instrument connected to a microscope with 20× magnification (in these conditions the laser spot size was about 5 μm). All the spectra were recorded in back-scattering conditions with 5 cm ⁻¹ spectral resolution using the 488 nm line (Innova Coherent 70) with a power of about 50 mW. The detector was a 160 K frozen CCD from Princeton Instruments Inc.

To minimize problems deriving from the possible sample inhomogeneity, five spectra at least were recorded on each specimen area (i.e. upper surface, inner fractured side). Raman spectra were recorded on unhydrated cement powders (five spectra on five different points of each sample) as well as alpha-TCP used for preparing the wTC-TCP cement. The reported Raman spectra were the average of the spectra recorded on five different points.

2.10

Cell culture

Before cell seeding, the material disks were placed in 24-well culture plate and treated for 2 h with 1% antibiotic/antimycotic solution (10 000 U penicillin, 10 μg streptomycin, 25 μg amphotericin B per mL), washed twice with phosphate buffered saline (PBS) and finally pre-wetted with culture medium containing 10% fetal bovine serum for 24 h at 37 °C, to mimic the in vivo situation at implantation when proteins from blood rapidly coat the material surface .

Human marrow stromal cells (Cambrex, Charles City, IA, USA) were cultured with D-MEM supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS), 1% penicillin–streptomycin, 50 μg/mL ascorbic acid, 10 ⁻⁸ M dexamethasone and 2 mM l -glutamine, in 5% CO ₂ humidified atmosphere at 37 °C.

On day 0, 1 × 10 ⁴ cells/cm ² were seeded on the cement disks of different age in 24-well microplates, and cultured in 1 mL of complete medium up to 28 days.

2.11

Cell viability

Alamar blue (Biosource International) dye was used for cell viability. At 24 h, 7, 14, 21, and 28 days, the Alamar reagent was added 1:10 (v/v) to the culture wells for 4 h. Then the fluorescence was read at 490 nm excitation—540 nm emission wavelength using a Cytofluor 2350 fluorimeter (Millipore Corporation, Bedford, MA, USA). The results were expressed as relative fluorescence units (RFU). Cells plated on polystyrene culture wells (TCPS) provided the controls. Statistical analysis was performed with StatView™ 5.0.1 software for Windows (SAS Institute Inc., Cary, NC, USA). Results in the graphs are reported as mean ± standard error of 6 replicates, and the differences were analyzed using Wilcoxon test with a significance level of p < 0.05.

Conversely, cell death was measured by the lactate dehydrogenase assay (LDH, Roche Diagnostics, Italy), which is a marker of cell membrane damage. The assay was performed following the manufacturer instructions and the mean and standard deviation of six replicates was expressed as optical density (OD).

2.12

Cell functions

Alkaline phosphatase released in the medium at 14 days of culture was assayed utilizing the conversion of a colorless p-nitrophenyl phosphate [5.3_10_3 m] to a colored p-nitrophenol according to the manufacturer’s protocol (Sigma, St. Louis, MO). The color change was measured spectrophotometrically at 405 nm (Spectra III, TECAN, Austria), and the alkaline phosphatase levels were normalized to the total protein content of cells (mM per protein microgram).

Collagen released in the medium was assayed at 24 h, 7, 14, and 28 days of culture using the CICP Metra EIA kit (Medical System S.p.A., Italy) following the manufacturer instructions, and the mean and standard deviation of four replicates was expressed as ng/mL. The collagen production was normalized to the total protein content of cells (nanogram per protein microgram).

2.13

Cell morphology by SEM/EDX

The cell colonization of the cements of different age was assessed after 28 days of culture using a conventional scanning electron microscope (SEM – Jeol 5200, JEOL, Japan). The samples were washed in 0.2 M phosphate buffer pH 7.3 and fixed in Karnovsky’s fixative (1% paraformaldehyde, 1.5% glutaraldehyde and 0.1 M sodium cacodylate buffer, pH 7.2–7.4).