Human Dental Pulp Stem Cells (hDPSC) are multipotent stem cells harvested from soft living pulp tissue inside adult teeth. hDPSC were provided (courtesy of Dr. Petros Papagerakis, from The University of Saskatchewan, Canada). The cells were grown in αMEM medium + GlutaMAX-I with 1 g/L of D-glucose and sodium pyruvate (Gibco, Thermo Fisher Scientific, Illkirch-Graffenstaden, France), supplemented with 15% fetal calf serum (FCS) and 100 U/mL penicillin and 100 mg/mL of streptomycin (this medium will be further referred to as GM) at 37°C in a humid environment with 5% CO₂. These cells were further differentiated into odontoblast-like cells in the organotypic constructs. Human ameloblast-like cells were initially characterized as ameloblastoma cells (AM-1) and were originally obtained from a 20-year-old female, immortalized with HPV-16 vector, and clone selection was performed with 1 mg/mL G418. Cells were provided by Dr. Sylvie Babajko and Prof. Ariane Berdal (UMR-S 1333 Santé Orale, Universite Paris Cité, and, Physiopathological basis of skeletal dysplasia, Université Paris Cité, Inserm UMR-S 1163-IHU Imagine, Paris, France) and were grown in Keratinocyte SFM medium (Thermo Fisher Scientific, Illkirch-Graffenstaden, France) complemented with 1 mg/mL G418 (Roche) for clone selection and 50 μg/mL of Gentamicin at 37°C in a humidified atmosphere with 5% CO₂.

Immunofluorescence on 2D cell cultures was performed at room temperature after plating hDPSC and AM1 in a 24 well-plate (5 × 10⁴ cells/well) and fixing with PFA 4%. Following a 15 min PBS-Tween 0.2% treatment, cells were incubated with Bovine serum albumin (BSA) 0.5% and Triton X-100 0.1% solution in PBS (PBS-BT) for 15 min. Antibodies (primary: COL1A1: mouse anti-human, Santa Cruz, sc-293182, dilution 1/100; DSPP: rabbit anti-human, Bioss, bs-10316R, dilution 1/200; FAM83H: rabbit anti-human, Invitrogen, PA5–55094, dilution 1/100; AMELX: mouse anti-human, Santa Cruz, sc-365284, dilution 1/100; secondary: Alexa Fluor 488, donkey anti-mouse IgG ThermoFisher, A-21202, dilution 1/200; Alexa Fluor 488, donkey anti-rabbit IgG, ThermoFisher, A-21206, dilution 1/250) were incubated for 1 h, followed by incubation in phalloidin 1/1000 in PBS-1% BSA solution for 15 min. Finally, one drop 1 µg/mL of 4′,6-diamidino-2-phenylindole (DAPI) solution was added to each slide. The slides were observed using Leica Dmi 8 + Yokogawa CSU W1 - ILAS2 confocal microscope.

The cells were detached from culture dishes using trypsin 0.25% EDTA 0.02% and washed with GM. The coating was performed as previously described (3). Briefly, the cells were resuspended in Tris-HCl buffer (Tris-HCl 50 mM pH 7.4) and subsequently incubated for 1 min with 0.04 mg/mL FN, G (Sigma-Aldrich) solutions in Tris-HCl buffer or with Tris-HCl buffer alone (rinsing step). To remove the solutions, the cells were centrifuged at 200 g for 1 min, and the supernatant was removed. After (FN/G)₄FN nanofilms were formed, the cells were resuspended in GM at a desired concentration.

To verify the deposition of (FN/G)₄FN nanofilms onto the cells, 5 × 10⁴ cells were deposited into a 24-well insert with a semipermeable membrane (Corning 3470, 0.4 μm pore size) and placed into 24-well plates. One milliliter of GM was added in a 24-well plate outside the inserts. The cells were incubated for 4 h at 37 °C, and then another 1 mL of GM was added to the 24-well plate to connect the media between the inside and the outside of the insert. After 24 h the medium was removed and the cells were fixed with 4% PFA, then stained for FN (primary antibody: 1/200 rabbit anti-human, Sigma-Aldrich, F3648, secondary antibody: 1/250 Alexa Fluor 488, donkey anti-rabbit IgG, ThermoFisher, A-21206), actin, and nuclei as previously described (4). To assess the viability of the coated cells, 10⁵, 2.5 × 10⁵ cells were deposited into the inserts coated with 100 μl of 0.04 mg/mL FN at 37°C for 30 min, cultured for 4 days in GM, and submitted to Live/Dead staining according to the manufacturer's instructions (Figure 1A).

3D construct formation was performed using ultra-low attachment plates (PrimeSurface® 96V-shaped bottom, S-bio, Japan). First, odontoblast-like cells were cultured at a concentration of 5 × 10³ cells/well for 7 days in GM to allow cell attachment, then for 15 days in differentiation medium (DM). The differentiation medium was composed of the GM supplemented with 0.1 M dexamethasone, 5 mM glycerophosphate, 50 mg/mL ascorbic acid, and 10 ng/mL TGF-β1, all from Sigma-Aldrich (Merck, Darmstadt, Land de Hesse, Germany). At day 22, ameloblast-like cells were added into the wells containing differentiated odontoblast constructs at a concentration of 3 × 10³ cells per well and were incubated for 7 more days, ending at day 29 (Figures 1B,C). Next, the resulting structures were fixed using 4% PFA, then incorporated into a block of Epredia™ Cryomatrix™ and placed at −80°C to freeze. The block was then cut with a cryostat, and 25 µm sections were deposited on glass slides for fluorescent staining (Figure 1D).

For fluorescent staining of organoids sections, they were first treated with Triton X-100 0.1% in PBS for 15 min, then saturated with 0.1% BSA for 1 h, then incubated with primary and secondary antibodies (primary: COL1A1: mouse anti-human, Santa Cruz, sc-293182, dilution 1/100; DSPP: rabbit anti-human, Bioss, bs-10316R, dilution 1/200; FAM83H: rabbit anti-human, Invitrogen, PA5-55094, dilution 1/100; AMELX: mouse anti-human, Santa Cruz, sc-365284, dilution 1/100; secondary: Alexa Fluor 488, donkey anti-mouse IgG ThermoFisher, A-21202, dilution 1/200; Alexa Fluor 488, donkey anti-rabbit IgG, ThermoFisher, A-21206, dilution 1/250; Alexa Fluor 568, donkey anti-rabbit IgG, ThermoFisher, A-10042, dilution 1/200; Alexa Fluor 568, donkey anti-mouse IgG, ThermoFisher, A-10037, dilution 1/250) for 1 h and 30 min, respectively, and rinsed 3 times with PBS after each step. ProLong™ Diamond antifade reagent (Molecular Probes) was used for mounting. Mounted slides were observed with the Leica Dmi 8 + Yokogawa CSU W1—ILAS2 confocal microscope.

Fibronectin labeling and cell viability experiment were performed 2 times, two samples per experimental condition. Cell viability values correspond to the average number of dead cells per image quantified from 10 images, and the error bars to standard deviation. Organotypic 3D construct culture and immunofluorescence results correspond to 3 independent experiments, 10 constructs per experimental condition. Odontoblast core diameter was quantified from 5 images, an average of 10 diameters measured using ImageJ software (v1.44p, NIH, Bethesda, US) was calculated. Data are reported as means ± standard deviation. Student's t-test was performed to compare coated and uncoated samples. Statistical significance was set at p < 0.05.

The cell-accumulation method, which consists of cell coating with FN/G nanomatrix, was developed by the group of Prof. Akashi and applied to different cell types, including human dermal fibroblasts (10), skeletal muscle progenitors (11) and hepatocytes (12). Here, we used it for the first time to coat hDPSC cells. The results showed the presence of FN network in monolayer samples of hDPSC coated with (FN/G)₄FN (Figure 2A). Some FN can also be detected in uncoated samples, but in much lower quantities. This corresponds to the FN secreted by the cells.

Next, we tested cell viability in two types of constructs of different thickness (10⁵ or 2.5 × 10⁵ cells seeded into the inserts, as indicated in the Figure 1A) and found that (FN/G)₄FN coating improved viability in thicker constructs (Figure 2B), leading to a significant decrease in the number of dead cells compared to uncoated samples (Figure 2C). These results demonstrate the importance of the cell environment for cell survival in 3D constructs and indicate that FN/G coating is appropriate to maintain the viability of hDPSC in 3D structures.

The basic level of expression of proteins associated with amelogenesis and odontogenesis was first assessed in 2D cultures. Type I collagen (COL1) forms most of the organic material (∼85%) in dentin, and dentin sialophosphoprotein (DSPP) is the most abundant non-collagenous protein in dentin (13). Amelogenin (AMELX) is the main enamel matrix protein, and Family with Sequence Similarity 83 Member H (FAM83H) has a role in the structural development and mineralization of enamel (14, 15). The results showed expression of these proteins in 2D cultures of AM-1 (ameloblast-like cells) and hDPSC (Figure 3), indicating that the cells were appropriate to use for formation of bilayered organoids mimicking odontogenesis. More detailed characterization of the cell types can be found in our previous work (9).

Organotypic 3D bilayered construct formation from hDPSC and AM-1 cells was monitored day by day; selected time points are presented in Figure 4A. At Day 1 in GM, some differences are visible between the samples, but they disappear by Day 4. At Day 13, in DM, the structures appear dense, which is probably due to the mineralization of the hDPSC during differentiation. Finally, after AM-1 addition, at Day 4 AM-1, bilayered structures can be observed, with a denser odontoblast-based core and semi-transparent AM-1 layer (Figure 4A).

Fluorescent staining of odontoblast and ameloblast markers, COL1A1/DSPP and FAM83H/AMELX, respectively, allowed to visualize odontoblast core and the surrounding ameloblast layer in 3D construct cross-sections (Figures 4B,D). The odontoblast core was significantly larger in FN/G-coated samples (Figure 4C), suggesting that FN/G coating was favorable for odontogenic differentiation of hDPSC, making the core more stable over the 1-month culture. These results are in accordance with the observations using phase contrast: while some degradation of organoids could be observed in uncoated samples, FN/G-coated samples remained unaffected (data not shown).

The epithelial ameloblast-associated protein FAM83H showed robust staining in the ameloblast-like cell layer (Figure 4B) in both non-coated and FN/G-coated samples. While no difference of protein expression was visually observed, the variation has been detected at mRNA level (Supplementary Table S1; Supplementary Figure S1), with a 3-fold increase in FAM83H expression in FN/G-coated samples compared to non-coated samples, indicating that a well-differentiated and stable odontoblast core is favorable for FAM83H expression.

AMELX protein expression was observed in both non-coated and FN/G-coated samples. However, while it was randomly distributed in AM-1 cell layer in non-coated samples, in FN/G-coated samples it was mostly localized in contact with the odontoblast core (Figure 4D, arrows), which reflects what happens in vivo during odontogenesis: during enamel maturation, ameloblasts attach to the dentin surface (13). Precise molecular mechanisms of epithelio-mesenchymal crosstalk during tooth formation are difficult to study in vitro using 2D cell culture, therefore the development of appropriate 3D odontogenesis models can help to better understand these processes. In the study by Alghadeer et al. AMELX and ENAM were visualized at the interface between 3D co-cultured human iPSC-derived epithelial organoids and hDPSC (8). Here, we propose a different 3D system with a stable odontoblast-like core, a surrounding ameloblast-like layer and AMELX expression at the interface. As AMELX localized expression is very challenging to precisely quantify, we present only qualitative imaging results in the present work. In the future, 3D construct permeabilization will be applied to image odontoblast-ameloblast interactions in 3D and provide quantitative results. In addition, different batches of hDPSC could be tested to evaluate the potential variability between hDPSC from different donors. Another limitation of our study concerns using immortalized AM-1 cells, which could be replaced by human-induced pluripotent stem cells (hiPSCs) differentiated into human dental epithelial cells using established protocol (7).