Normal impact third molars were collected from young adults (n = 4) (21 - 25 years of age) with their informed consent at Finnish Student Health Service, Tampere, Finland. Three of the pulps were from male and one (pulp 1) from a female patient. The donors did not smoke nor had diabetes or asthma. Two pulps (from maxilla and mandible) from one patient were pooled. The pulp tissue was separated from the crown and the root and placed on 2 ml of DMEM/F12 (Gibco Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS, Invitrogen), 1% antibiotic-antimycotic (Gentaur, Belgium), and 1% GlutaMax (Invitrogen). Tissue was digested with 3 mg/ml collagenase type I and 4 mg/ml dispase for 1 hour at 37˚C, centrifuged, and resuspended with medium described above. Cell suspensions were filtered with 100 μm cell strainer and cells were seeded in 6-well plates (Nunc, Thermo Fisher Scientific, Rochester, NY, USA). Subconfluent cultures were passaged using trypsin in phosphate buffered saline (PBS). Cells were transferred to T-25 cell culture flasks when wells were confluent. Cells were cultured 3 - 6 weeks after isolation to increase the number of cells for the experiments.

Neural differentiation was conducted with protocol published previously [16]. Briefly, dental pulp stem cells (DPSCs) at passage 2 were seeded at 20,000 cells/cm² in 10 µg/ml mouse laminin (Sigma-Aldrich, St Louis, MO, USA)-coated 24- and 48-well plates (Nunc) and 20,000 cells/well in 0.1% polyethyleneimine (PEI) and 20 µg/ml mouse laminin-coated 6-well MEA-plates (MultiChannel Systems, Reutlingen, Germany). The used differentiation protocol consisted of three stages. Epigenetic reprogramming was induced for 48 h with DMEM/F12 supplemented with 10 µM 5-azacytidine (Sigma-Aldrich), 2.5% FBS (Invitrogen), and 10 ng/ml bFGF (R&D Systems, Minneapolis, MN, USA) starting 24 h after cell seeding. Next, neural induction was conducted with DMEM/F12 supplemented with 250 µM IBMX, 50 µM forskolin, 200 nM TPA, 1% ITS (all from Sigma-Aldrich), 1 mM dbcAMP, 10 ng/ml bFGF, 10 ng/ml NGF, and 30 ng/ml NT-3 (all from R&D Systems) for 3 days. Cells were washed with PBS before neural maturation with Neurobasal medium (Gibco Invitrogen) supplemented with 1 mM dbcAMP (Sigma-Aldrich), 1% N2, 1% B27 without vitamin A (both from Gibco Invitrogen), and 30 ng/ml NT-3 for 7 days. All solutions were freshly prepared prior to use. Control cells were maintained in control medium (DMEM/F12 supplemented with 2.5% FBS and 1% penicillin/streptomycin) which was changed as differentiation mediums.

RNA samples from DPSCs was collected at three time points, at time 0 (before cell plating), after 3 days of neural induction, and after 7 days of neural maturation. Similar samples were also collected from control cells. Total RNA was isolated using NucleoSpin^® RNA XS kit (Macherey-Nagel, Germany) according to manufacturer’s instructions. Next, 50 ng of total RNA per sample was used for cDNA synthesis using random primers (High Capacity cDNA Reverse Transcription Kit, Applied Biosystems) in a reaction volume of 20 μl. Next, quantitative real-time PCR was performed with Taqman^® gene expression assays (Applied Biosystems) with 3 µl cDNA according to manufacturer’s instructions in a reaction volume of 15 μl for Musashi for neural progenitor cells (Hs01045984_m1), light neurofilament for neuronal cells (NF68, Hs00196245_m1), glial fibrillary acid protein for astrocytes (GFAP, Hs00909236_m1), brain lipid-binding protein for radial glial cells (BLBP, Hs00361426_ m1), and Olig2 for oligodendrocytes (Hs00377820_m1). GAPDH was used as the internal control (4352934E). Quantitative real-time PCR was performed using the following conditions: 50˚C for 2 min, 95˚C for 10 min, and 40 cycles of 95˚C for 15 s and 60˚C for 1 min with ABI 7300. All samples were analyzed as technical triplicates (variation required less than 0.5 CT) and no-template control was used. The data was analyzed with a 7300 System SDS Software (Applied Biosystems). Relative quantification was calculated using 2^–ΔΔCT-method [23]. Gained data was normalized to the expression of GAPDH. The data is presented as mean fold change compared to start point.

Samples for immunocytochemistry were collected after neural induction and neural maturation stages. Differentiating DPSCs were fixed with 4% PFA in PBS for 20 min in room temperature (RT). Blocking was conducted with 10% normal donkey serum (NDS), 0.1% TritonX-100, and 1% bovine serum albumin (BSA) in PBS for 45 min at RT. Cells were washed once with 1% NDS, 0.1% TritonX-100, and 1% BSA in PBS and incubated with primary antibodies diluted to the same solution at 4˚C overnight. Antibodies used were mouse anti-nestin (1:100, Chemicon, Temecula, CA, USA) for neural progenitor cells, mouse anti-β-tubulin₃ (1:1200, Sigma-Aldrich) and rabbit anti-microtubule associated protein 2 (MAP-2, 1:600, Chemicon) for neuronal cells, sheep anti-GFAP (1:600, R&D Systems) for astrocytes, and mouse anti-GalC (1:200, Chemicon) for oligodendrocytes. All solutions for GalC stainings were made without TritonX-100. After washing three times with 1% BSA in PBS cells were incubated for 1 hour with Alexa Fluor-488 and/or -568 conjugated with anti-mouse, antirabbit, and anti-sheep antibodies (1:400, Molecular Probes/ Invitrogen, Carlsbad, CA, USA). After washing with PBS and phosphate buffer the cells were mounted with Vectashield containing DAPI (4’,6-diamidino-2-phenylindole) (Vector Laboratories Inc., Burlingame, CA, USA). Stained cells were imaged with a phase contrast microscope with fluorescence optics (Olympus IX51, Olympus, Finland) and Olympus DP30BW camera. Images were edited with Adobe Photoshop CS2.

HDPSCs were also cultured on planar microelectrode array (MEA) 6-well plates (6 × 9 electrode layout, electrode diameter 30 µm, inter-electrode distance 200 µm, MultiChannel Systems, Reutlingen, Germany) during neural differentiation (n = 8 wells/differentiating cells/pulp and n = 4 wells/control cells/pulp) as previously described for hESC-derived neurons [8]. Measurements were performed 1 - 2 times/week. Measurements were performed with MEA 1060-Inv-BC-amplifier with integrated TPC Temperature controller adjusted to +37˚C and data was recorded with MC_Rack software (all from Multichannel systems). Prior to measurements, MEA plates were sealed with PDMS discs in laminar hood to keep the cultures sterile for repeated measurements. After executing the differentiation protocol, the cells were maintained on MEA plates for additional 1 - 2 weeks to prolong the measurement period to 4 weeks.

As a positive control we used hESC-derived neuronal cells differentiated with previously published protocol [24]. Briefly, the cells were fixed and stained as described above after 8 weeks of differentiation as neurospheres and then 3 days on mouse laminin-coated 24- wells. 8 weeks old neurospheres were also plated on PEI and mouse laminin-coated MEA-plates and measured 1 - 2 times/week for 4 weeks.

The culture time before actual neural differentiation experiment varied from 3 to 6 weeks depending on the cell number after isolation. Morphology of dental pulp stem cells in general varied from spindle-shaped fibroblasts to flat cells as previously shown [25].

After seeding the cells mostly resembled fibroblasts. There were no differences in the morphology after the epigenetic reprogramming. The control cells had, however, proliferated notably more efficiently. After initiating the neural induction stage, the morphology of the cells of three hDPSC lines over went a change; the cells became more round-shaped and some processes were developed. During the following three days the morphology changed further. Many processes could be detected in the cell populations as well as branched cells with round soma resembling astrocytes, neuronal cells, or oligodendrocytes. Differences between hDPSC lines could be observed. Figure 1(a) represents the induction stages of all 4 hDPSC lines. In induction stage, according to morphology, pulp 2 derived populations resembled mostly neuronal cells. Pulp 4 was clearly different from the others as no clear neuronal morphologies were detected. The neuronal cell morphologies were detected only during neural induction stage. In the neural maturation stage the cell morphology changed again more into fibroblast-like as the number of processes decreased, as represented in Figure 1(b) with pulps 1 and 2.

The neural progenitor cell marker Musashi was detectable already in non-differentiated cells and the expression increased with all 4 hDPSC lines during differentiation experiment (Figure 1(c)). With three hDPSC lines (pulps 2, 3, and 4) the increment in expression of Musashi was at least 10-fold. In line with hDPSC line morphologies during differentiation experiment, the expression of NF68 was not detected in any of the DPSC lines expect with pulp 2. In this hDPSC line NF68 was expressed already before the differentiation protocol was started and a 3-fold increase in expression could be detected at the maturation stage (Figure 1(d)). The expression of GFAP, Olig2, or BLBP was not detected in any of the hDPSC lines at any timepoints.

Variation between hDPSC lines was also detected in protein expression. Neural precursor marker nestin could be detected in cultures after maturation stage with all the studied pulps. Also, control cultures were positive for nestin but the morphology of the cells was different (Figure 2(a)). Neuronal marker β-tubulin₃ was detected only with pulp 2 after maturation stage but the morphology did not resemble that of neurons (Figure 2(a)). MAP-2 staining was negative in all the cultures. Oligodendrocyte marker GalC stained cells after neural induction and maturation stages in pulp 2 as well as in control cultures (Figure 2(a)). GFAP-positive astrocytes were not detected among any of the differentiated DPSC lines nor control cultures. As a positive control for immunostaining we used hESC-derived neuronal cells which stained positive for both β-tubulin₃ and MAP-2 (Figure 2(c)).

Neuronal network signaling was not detected during 4 weeks follow-up on MEA among hDPSC-derived control or differentiated cultures (Figure 2(b)). Neuronal cells derived from hESCs as a positive control formed functional neuronal networks after 2 weeks of culturing on MEA dishes which could be detected as spikes (Figure 2(c)).