Fabrication of a Vertically Stacked Dentin–Pulp Complex Organ-on-a-Chip Device Using a Human Dentin Disc

Summary

To describe the fabrication and assembly of a vertically stacked dentin–pulp complex organ-on-a-chip device incorporating a human dentin disc and photolithography-fabricated microchannels, followed by dental pulp stem cell seeding and culture.

Abstract

Conventional preclinical models used in dental research, including animal models and static cell cultures, present important limitations in reproducing the physiological conditions of the dentin–pulp interface and predicting human biological responses to dental biomaterials. Organ-on-a-chip (OoC) technology provides an alternative approach by enabling the recreation of tissue-specific microenvironments within controlled microfluidic systems. This manuscript describes a step-by-step protocol for the fabrication and assembly of a dentin–pulp complex OoC device incorporating a human dentin disc within a three-layer polydimethylsiloxane (PDMS) microfluidic platform. The protocol includes dentin disc preparation, fabrication of microstructured molds using photolithography, PDMS casting and curing, plasma-assisted layer bonding, device assembly, leak testing, and dental pulp stem cell seeding under static culture conditions. Critical fabrication parameters influencing device performance are discussed, including control of PDMS layer thickness, plasma surface treatment, and integration of the dentin disc within the device to minimize leakage. Cell adhesion and viability within the device were evaluated using scanning electron microscopy and confocal live/dead staining. Representative results demonstrate successful integration of the dentin disc within the microfluidic assembly, maintenance of leak-free conditions, and adhesion of viable dental pulp stem cells to the dentin surface. This protocol provides a reproducible approach for the fabrication of a dentin–pulp complex OoC platform that may support future studies investigating cellular responses to dental biomaterials and microenvironmental stimuli.

Introduction

Animal experimentation has long been instrumental in advancing the understanding of pathological and physiological processes and has played a central role in the early stages of drug development¹. Over the course of the 20th century, human life expectancy increased substantially, and while it is not possible to precisely quantify the extent to which this improvement is attributable solely to animal research, it is widely acknowledged that many contemporary medical treatments have been developed or validated using animal models². Notable examples include the development of vaccines against poliomyelitis, measles, and hepatitis B, the evaluation of the safety and efficacy of penicillin, and the production of insulin. In the field of dental and biomaterials research, these models have been critical for exploring the caries process, biology and regeneration of the dentin-pulp complex, and physiopathology of the periodontal tissues^3,4.

Despite the extensive use of these methodologies, the current drug development pipeline continues to exhibit high failure rates, with many drug candidates failing during clinical trials despite demonstrating efficacy and safety in preclinical laboratory settings and animal models⁵. This discrepancy has been associated with interspecies differences that may limit the translational predictability of animal models owing to anatomical, genetic, and physiological differences between species⁶. In addition, the development of new therapeutic approaches remains time-consuming and costly⁷. Furthermore, conventional in vitro models predominantly employ simplistic two-dimensional (2D) monolayer cultures on non-physiological plastic substrates, failing to recapitulate the biochemical and physical cues, cell–cell interactions, and extracellular matrix (ECM) signaling present in the human oral cavity^8,9. Although 2D models offer certain experimental advantages, they are inherently unable to recapitulate the complex interactions that occur among multiple components simultaneously, such as biofilms, biomaterials, dentin, and pulp cells¹⁰. Consequently, there is increasing interest in the development of alternative experimental platforms capable of more accurately reproducing human biological responses. In 2022, the FDA Modernization Act 2.0 removed the mandatory requirement for animal testing in the drug approval process and recognized the potential use of alternative platforms, including organ-on-a-chip (OoC) technologies, in preclinical research¹¹.

An OoC is a microfluidic device designed to replicate the structure, function, and microenvironment of a specific organ or tissue component¹². These devices are populated with living cells and tissues cultured within engineered microchannels that recapitulate key aspects of the physiological microenvironment to which cells are naturally exposed. Beyond structural mimicry, microchannels enable spatiotemporal control over the cellular environment and support the maintenance of tissue-specific functions, thereby providing experimental conditions that may more closely resemble in vivo physiology than conventional culture systems¹². Of particular relevance, microchannels in OoC devices facilitate the controlled flow of culture medium, which exerts shear stress upon resident cells¹³. These mechanical stimuli influence cellular responses including adhesion, proliferation, migration, gene expression, and cell–cell interactions^14,15. In addition, continuous medium flow contributes to maintenance of the microenvironment through nutrient supply and removal of metabolic waste products¹⁶. Furthermore, compared with conventional 2D cultures, some OoC platforms have been reported to support improved control of oxidative stress-related readouts, including intracellular reactive oxygen species and cell viability, in three-dimensional (3D) on-chip models¹⁷. OoC systems have therefore emerged as promising platforms for studying tissue physiology, biomaterial interactions, and disease processes under controlled experimental conditions¹⁸. In dental research, OoC technology has been progressively applied to model increasingly complex oral tissue interfaces¹⁹. Early tooth-on-a-chip platforms emulated the biomaterial–dentin–pulp interface²⁰, while more advanced iterations have since incorporated vasculature and trigeminal innervation to more faithfully replicate the structural and functional complexity of the pulp-dentin complex²¹. Beyond the pulp, 3D models of immature root canals have enabled the assessment of endodontic irrigants and angiogenic sprouting on stem cells from the apical papilla^22,23. At the periodontal level, gum-on-a-chip platforms have been employed to investigate host-microbe interactions and to evaluate the immunomodulatory potential of probiotics as therapeutic agents²⁴, whereas periodontal ligament-on-chip devices have demonstrated how interstitial fluid flow modulates vascular network formation and the osteogenic differentiation of periodontal ligament stem cell spheroids^25,26. Complementing these models, specialized oral mucosa-on-a-chip and dental implant-on-a-chip systems have further broadened the scope of dental microfluidic research by characterizing host-material interactions and the cytotoxicity of dental monomers under physiologically relevant conditions^27,28.

While OoC systems have been increasingly applied to dental, oral, and craniofacial research, continued efforts toward standardization and reproducibility of fabrication workflows remain important. Several tooth-on-a-chip models have been reported in the literature^20,29,30,31, demonstrating different device configurations and experimental applications. However, variations in device architecture, fabrication methods, incorporation of dentin, and reporting of experimental parameters may limit reproducibility and direct comparison between studies. In particular, differences in microchannel geometry, dentin integration strategies, and fabrication approaches may influence the biological and mechanical microenvironment within the device. Some reported models incorporate dentin to reproduce aspects of the dentin–pulp interface^20,29,30, whereas others use alternative configurations³¹. In addition, fabrication methods such as photolithography, laser cutting, micromilling, and 3D printing each present specific advantages and limitations depending on the intended device design and application. Furthermore, a critical determinant of biological utility is the architectural orientation of the device, specifically whether the channels and dentin slice are arranged in a horizontal or vertical configuration^19,32. In horizontally stacked configurations, compartments arranged side by side facilitate lateral compartmentalization and real-time visualization of cells, making them particularly well suited for modeling tissues with inherently lateral spatial organization, such as the periodontal complex^24,32. However, this design is inherently limited in its capacity to recreate perpendicular biological interfaces, such as the biomaterial-dentin-pulp complex, as the lateral arrangement of compartments does not faithfully replicate the in vivo orientation of the dentin barrier between the cavity and the pulp^20,32. Vertically stacked configurations address these limitations by superimposing two compartments along the y-axis, separated by a native dentin slice, enabling the simultaneous quantification of molecular transport across intact dentin and real-time monitoring of pulp cell responses measured in the outlet medium³⁰. However, a recognized limitation of this orientation concerns cross-sectional imaging of the dentin-cell interface, as acquiring cross-sectional images of the dentin-cell interface within the same focal plane remains technically more challenging than in horizontally stacked configurations, where cells and extracellular matrices are arranged laterally. Although multiple dentin-pulp complex on-a-chip models have been reported in the literature, the majority adopt a horizontally stacked configuration in which compartments are arranged side by side within the same plane^10,20,21,29, while relatively few have incorporated a native dentin slice within a vertically stacked configuration³⁰. Nevertheless, variations in fabrication approaches and reporting parameters across these models continue to limit reproducibility and cross-study comparisons³³. The present manuscript aims to improve methodological reproducibility by providing a detailed fabrication protocol for a vertically stacked tooth-on-a-chip device, thereby facilitating broader adoption of the platform within the dental research community.

This manuscript describes a step-by-step fabrication and assembly protocol for a vertically stacked dentin–pulp complex OoC device incorporating a human dentin disc within a three-layer polydimethylsiloxane (PDMS) microfluidic platform fabricated using a photolithography-produced mold. The protocol also describes dental pulp stem cell seeding and culture within the device under static conditions.

Protocol

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Universidad de La Frontera for studies involving human samples.

1. Acquisition of dentin disk

NOTE: Prepare the device design before beginning the protocol (Figure 1A–1C).

1. Collection and storage of teeth
   1. Collect healthy human molars or premolars after obtaining informed patient consent.
   2. Maintain all specimens in 0.5% chloramine solution at 4°C to prevent dentin dehydration and protein denaturation.
   3. Replace the chloramine solution once per week during storage.
2. Preparation of dentin discs
   1. Section each tooth horizontally at the middle third of the crown using a high-speed handpiece operating at 350,000 rpm and equipped with a cylindrical diamond bur (1 mm diameter; 100–150 µm grit size).
   2. Perform sectioning under continuous water irrigation at approximately 50 mL/min.
   3. Position the section at the middle third of the crown to avoid the pulp chamber and the enamel present at the occlusal grooves and fissures, thereby obtaining a disc composed predominantly of dentin tissue.
      CAUTION: Perform tooth sectioning under continuous irrigation and wear appropriate personal protective equipment, including gloves, a mask, and protective eyewear.
3. Finishing of dentin discs
   1. Polish the dentin disc using sequentially decreasing grit finishing discs.
   2. Continue polishing until achieving a final dentin thickness of approximately 0.8 mm (±± 0.3 mm) and lateral dimensions of approximately 5 mm × 5 mm (±± 0.3 mm) (Figure 1D).
4. Storage of dentin discs
   1. Store the prepared dentin discs in 0.5% chloramine solution at 4°C until OoC device assembly.

Figure 1: Design and fabrication workflow of the dentin–pulp complex organ-on-a-chip (OoC) device. (A) Three-dimensional design of the device showing the three-layer architecture. (B) Top view of the device design with principal dimensions. (C) Lateral view of the three-layer device configuration. (D) Representative dentin disc prepared with an approximate thickness of 0.8 mm and lateral dimensions of approximately 5 mm × 5 mm. (E) Silicon master mold fabricated by photolithography and used for casting the top and bottom polydimethylsiloxane (PDMS) layers. (F) Removal of a cured PDMS layer from the silicon mold following thermal polymerization. (G) Plasma treatment of the PDMS surface prior to layer bonding. (H) Final three-layer dentin–pulp complex OoC device assembled by plasma bonding and incorporating the dentin interface. (I) Assembled device connected to tubing for leakage testing and preparation for microfluidic experiments. Please click here to view a larger version of this figure.

2. Fabrication of the dentin–pulp complex OoC device

1. Preparation of the workspace and materials
   1. Prepare materials and equipment
      1. Prepare all materials and equipment listed in the Table of Materials before beginning device fabrication.
      2. Prepare biopsy punches before device assembly.
      3. Perform all fabrication procedures in a clean, dust-free environment.
      4. Wear gloves, protective eyewear, and appropriate personal protective equipment throughout device fabrication and assembly.
         NOTE: Prepare the photolithography-produced mold before beginning the protocol (Figure 1E).
2. Fabrication of the microstructured mold
   1. Preparation of the silicon wafer
      1. Fabricate the mold on a 4-inch, single-side polished silicon wafer (N-type, As-doped, CZ growth method, <100> orientation, 500 ±± 25 µm thickness, TTV ≤ 10 µm, BOW ≤ 35 µm, resistivity 0.001–0.005 Ω·cm).
      2. Clean the wafer sequentially in acetone, isopropanol, and ultrapure water for 10 min each under ultrasonication.
      3. Dry the wafer with compressed air.
      4. Bake the wafer at 200°C for 10 min on a hot plate to remove residual moisture.
      5. Allow the wafer to cool to room temperature (RT; 20°C–25°C).
   2. Surface activation and photoresist coating
      1. Activate the silicon wafer surface using a vacuum plasma cleaner operated with air plasma at level 7 for 120 s.
      2. Apply GM 1060 epoxy-based negative photoresist by spin coating at 300 rpm for 10 s.
      3. Continue spin coating at 565 rpm for 100 s and 965 rpm for 1 s using an acceleration of 200 rpm/s.
      4. Allow the coated wafer to rest for 10 min.
      5. Perform a pre-bake at 65°C for 10 min.
      6. Continue the pre-bake at 95°C for 1 h.
      7. Allow the wafer to cool to RT.
   3. Exposure and development
      1. Expose the photoresist using a direct laser lithography system at a dose of 1,950 mJ/cm².
      2. Perform a post-bake at 65°C for 10 min.
      3. Continue the post-bake at 95°C for 30 min.
      4. Allow the wafer to cool to RT.
      5. Allow the wafer to rest for 10 min.
      6. Develop the pattern in PGMEA for 4 min.
      7. Extend development in 30-s intervals if residual photoresist remains.
      8. Stop development by rinsing with isopropanol.
      9. Dry the wafer using compressed air.
   4. Hard bake
      1. Perform a hard bake at 135°C for 2 h.
      2. Obtain a microstructured master mold containing microchannels measuring 1 mm in width and 50 µm in height.
3. Preparation and degassing of the PDMS mixture
   1. Preparation of the PDMS mixture
      1. Mix the PDMS base and curing agent at a 10:1 weight ratio.
      2. Stir the mixture manually with a glass rod in a plastic beaker for approximately 5 min.
      3. Maintain slow and continuous circular motions during mixing.
         NOTE: Avoid introducing air bubbles during mixing.
   2. Degassing
      1. Place the PDMS mixture in a vacuum desiccator connected to a vacuum pump.
      2. Degas the mixture at approximately 700 mmHg for 5 min.
      3. Release the vacuum briefly to allow surface bubbles to burst.
      4. Reapply the vacuum.
      5. Repeat the cycle until no visible bubbles remain.
         NOTE: Total degassing time is typically approximately 30 min.
4. Casting of PDMS on the mold
   1. Preparation of the mold
      1. Place the mold on a flat surface.
      2. Clean the mold using compressed air.
      3. Avoid direct contact with the patterned surface.
      4. Verify that the mold is free of dust and contaminants.
   2. Dispensing of PDMS
      1. Dispense approximately 8 g of PDMS around the dentin disc to fabricate the dentin embedding layer.
      2. Adjust the PDMS height to match the dentin disc thickness (0.8 ±± 0.3 mm).
      3. Prevent PDMS from covering the dentin surface.
      4. Dispense approximately 0.26 g of PDMS per mold unit to fabricate each microchannel layer.
      5. Gently tap the mold or apply vibration to facilitate filling of the microstructures.
      6. Allow the PDMS to settle into the smallest features.
      7. Cover the mold using a lid or flat cover without contacting the PDMS surface.
         NOTE: The dentin embedding layer is approximately 0.8 mm thick, and each microchannel layer is approximately 0.4 mm thick, yielding a total device thickness of approximately 1.6 mm.
         NOTE: Adjust layer thickness according to the working distance of the confocal objective when confocal imaging is planned.
5. Curing
   1. Place the molds in an oven at 50°C.
   2. Cure the PDMS for 4 h.
6. Removal of the PDMS layers
   1. Removal of cured PDMS
      1. Carefully peel the cured PDMS layers from the mold (Figure 1F).
      2. Handle the PDMS layers carefully to avoid damaging the microstructures.
7. Bonding of the PDMS layers and fabrication of inlet and outlet ports
   1. Fabrication of inlet and outlet ports
      1. Create perforations at the designated inlet and outlet locations using a 1 mm biopsy punch.
      2. Position the microchannel side facing upward during punching.
         NOTE: Place the PDMS layer on a cutting mat to avoid structural damage.
      3. Create two perforations in the top layer corresponding to the upper microchannel.
   2. Cleaning of PDMS layers
      1. Clean the PDMS layers using isopropyl alcohol or ethanol.
      2. Remove residual dust particles using sterile adhesive tape.
      3. Verify that all bonding surfaces are free of contaminants before assembly.
   3. Plasma activation of PDMS surfaces
      1. Position the top and middle PDMS layers for plasma treatment (Figure 1G).
      2. Activate the surfaces using a handheld corona plasma treater operating at approximately 25 W with ambient air as the working gas.
      3. Maintain a working distance of approximately 1 mm between the plasma source and the PDMS surface.
         CAUTION: Perform plasma treatment under appropriate ventilation and avoid direct exposure to plasma-generated ozone.
      4. Activate the plasma treatment device and allow stabilization for 5 s.
      5. Verify generation of a uniform plasma beam before treatment.
      6. Treat the top PDMS layer for 10 s using a continuous back-and-forth motion at approximately 5 mm/s.
      7. Treat the middle PDMS layer for 10 s using the same motion and distance.
      8. Apply an additional 1 s plasma treatment to the top layer immediately before bonding.
   4. Bonding of the top and middle layers
      1. Align the upper microchannel with the dentin disc immediately after plasma treatment.
      2. Bond the top and middle layers together.
      3. Place the bonded assembly in an oven at 75°C for 5 min.
      4. Consolidate the bond during incubation.
         NOTE: The dentin disc and microchannel are visible to the naked eye and can be aligned without additional alignment tools.
   5. Fabrication of lower microchannel ports
      1. Punch two perforations corresponding to the lower microchannel through the bonded top and middle layers.
      2. Perform punching through both layers simultaneously to ensure alignment.
         NOTE: This procedure generates four perforations in the top layer and two perforations in the middle layer.
      3. Place the bonded assembly on a firm cutting mat before punching.
      4. Apply steady downward pressure in a single motion.
         NOTE: Avoid rotational or lateral forces to minimize the risk of delamination.
      5. Remove debris from the perforation sites using lint-free sterile gauze moistened with isopropyl alcohol.
      6. Inspect the perforations and surrounding bonded areas under adequate lighting.
      7. Verify the absence of delamination, tears, or incomplete perforations before proceeding.
   6. Bonding of the bottom layer
      1. Plasma treat the bottom layer as described in Step 2.7.4.
      2. Align the inlet and outlet ports of the lower microchannel with the perforations in the top and middle layers.
      3. Bond the bottom layer to complete device assembly.
         NOTE: Mark the inlet and outlet locations on the external surface of the bottom layer before bonding to facilitate alignment.
         NOTE: Plasma chambers may provide stronger bonding than plasma pens, as reported in the literature³⁶.
         NOTE: The completed device consists of a three-layer microfluidic platform incorporating a dentin interface (Figure 1H). The lower dentin surface represents the pulp side of the model, whereas the upper dentin surface represents the coronal side for exposure to dental biomaterials or bacterial preparations.
8. Leak testing of the microfluidic device
   1. Inspection of the assembled device
      1. Inspect the assembled device for misalignment.
      2. Inspect the assembly for incomplete bonding and visible leakage.
   2. Leak testing
      1. Connect 19G Luer-lock needles to the inlet ports (Figure 1E).
      2. Connect the needles to microfluidic pump tubing.
      3. Press the needle hubs gently into the punched ports to obtain an airtight fit.
      4. Inject sterile PBS or culture medium at a flow rate of 10 µL/min.
      5. Maintain fluid flow for at least 5 min.
      6. Inspect all bonded interfaces, channel margins, and dentin disc borders during flow.
      7. Observe the device under adequate lighting during and after injection.
      8. Confirm the absence of fluid accumulation, wetness, or droplet formation.
      9. Classify the device as leak-free if no leakage is detected.
      10. Discard devices exhibiting leakage.
          NOTE: Perform leak testing on every fabricated device before cell seeding or experimental use.
9. Sterilization and storage of the microfluidic device
   1. Sterilization
      1. Expose the top surface of the assembled device to UV light (254 nm) for 30 min inside a laminar flow cabinet.
      2. Expose the bottom surface to UV light for an additional 30 min.
      3. Flush the microchannels with 70% ethanol for 20 min.
      4. Wash the microchannels three times with sterile PBS to remove residual ethanol.
   2. Storage
      1. Use the sterilized device immediately for cell culture experiments.
      2. Alternatively, store the device submerged in sterile PBS at 4°C inside a sterile container until use.
         ​CAUTION: Do not use chloramine or other disinfectant solutions after device assembly because PDMS may absorb these compounds and compromise cell viability.

3. Evaluation of dental pulp stem cell adhesion and viability

1. Preconditioning of the dentin surface
   1. Inject 10% EDTA into the device.
   2. Maintain contact between the EDTA solution and the dentin surface for 15 min.
   3. Inject 0.5 mL EDTA every 3 min to maintain fluid movement within the microchannel.
   4. Wash the dentin disc three times with ultrapure water.
   5. Maintain each wash for 5 min.
      NOTE: EDTA treatment removes the smear layer and exposes dentin matrix proteins.
      CAUTION: Handle EDTA solutions using gloves and protective eyewear to avoid skin and eye irritation.
2. Seeding of dental pulp stem cells
   1. Expansion of dental pulp stem cells
      1. Seed dental pulp stem cells into a T75 flask at a density of 1 × 10⁶ cells per flask.
      2. Incubate the cells at 37°C and 5% CO₂ until approximately 80% confluency is reached.
      3. Detach the cells using trypsin.
      4. Neutralize the trypsin using complete culture medium.
      5. Centrifuge the cell suspension to obtain a cell pellet.
      6. Assess cell viability using trypan blue.
      7. Use cell suspensions exhibiting viability ≥90%.
   2. Preparation of the cell suspension
      1. Prepare a pre-warmed cell suspension at 37°C.
      2. Adjust the cell concentration to 100,000 cells/mL in complete culture medium.
      3. Prepare a final suspension volume of 1–2 mL.
         NOTE: Cells at passage 5 were used in the present study.
   3. Cell seeding
      1. Inject the cell suspension into the inlet port of the lower microchannel.
      2. Invert the device after cell seeding
         NOTE: Device inversion promotes cell sedimentation toward the dentin surface and enhances cell adhesion.
   4. Incubation
      1. Place the inverted device inside a Petri dish.
      2. Add 1 mL culture medium to the Petri dish to minimize evaporation.
      3. Incubate the device at 37°C and 5% CO₂ for 24 h.
      4. Maintain static culture conditions throughout the incubation period.
         NOTE: Dynamic flow validation was not performed in the present study.
         NOTE: The present protocol describes static culture conditions for initial device validation and cell adhesion assessment.
   5. Initial device validation
      1. Evaluate cell adhesion using scanning electron microscopy.
      2. Retrieve the dentin disc by device disassembly before analysis.
3. Preparation of samples for scanning electron microscopy
   1. Sample retrieval and washing
      1. Disassemble the device after 24 h of incubation.
      2. Retrieve the dentin disc and transfer it to a culture plate or Petri dish.
      3. Position the cell-seeded dentin surface facing upward.
      4. Wash the dentin disc twice using 1 mL PBS.
   2. Fixation
      1. Add 2 mL of 2.5% glutaraldehyde in 0.1 M Sorensen’s phosphate buffer.
      2. Incubate the sample for 30 min at RT.
         CAUTION: Glutaraldehyde is toxic and may cause irritation to the skin, eyes, and respiratory tract. Perform fixation procedures inside a chemical fume hood while wearing appropriate personal protective equipment.
   3. Washing after fixation
      1. Remove the glutaraldehyde solution.
      2. Wash the dentin disc three times with ultrapure distilled water at RT.
   4. SEM analysis
      1. Analyze the samples using an environmental scanning electron microscope.
      2. Operate the microscope under low-vacuum conditions (25 Pa).
      3. Use an accelerating voltage of 10.0 kV and a working distance of 7.0 mm.
      4. Acquire images using backscattered electron 3D detection mode (BSE-3D).
      5. Capture images at ×200 magnification (Figure 2A).
4. Evaluation of cell viability through confocal microscopy
   1. Sample preparation
      1. Seed and culture cells as described in Step 3.2.
      2. Incubate the device for 24 h at 37°C and 5% CO₂.
      3. Disassemble the device and expose the dentin surface.
      4. Transfer the dentin disc to a well plate containing pre-warmed HBSS⁺/+ supplemented with Ca²⁺ and Mg²⁺.
   2. Sample washing
      1. Wash the dentin disc twice with HBSS⁺/⁺.
      2. Remove residual serum and phenol red before staining.
         NOTE: Residual serum and phenol red may increase background autofluorescence.
   3. Live-cell staining
      1. Prepare a fresh Calcein AM working solution at 2 µM in HBSS⁺/⁺.
      2. Add sufficient staining solution to fully cover the dentin surface.
      3. Incubate the sample for 20 min at 37°C in the dark.
      4. Remove the staining solution after incubation.
   4. Dead-cell staining
      1. Add Propidium Iodide working solution (1–2 µg/mL in HBSS⁺/⁺).
      2. Incubate the sample for 5–8 min at 37°C in the dark.
      3. Do not exceed 10 min of incubation.
      4. Wash the sample twice with HBSS⁺/⁺.
      5. Maintain the sample submerged in buffer until imaging.
   5. Confocal microscopy acquisition
      1. Acquire confocal images within 30 min of staining.
      2. Acquire images sequentially in the green and red channels.
      3. Use λ_ex 490 nm and λ_em 515 nm for live-cell imaging.
      4. Use λ_ex 535 nm and λ_em 617 nm for dead-cell imaging.
         ​NOTE: Protect fluorescent staining solutions and stained samples from light exposure during incubation and imaging procedures (Figure 2B).

Figure 2: Representative evaluation of dental pulp stem cell adhesion and viability within the dentin–pulp complex OoC device. (A) Scanning electron microscopy image showing adherent dental pulp stem cells (white asterisk) on the dentin surface after 24 h of static culture within the OoC device. After 24 h of culture, adherent cells covered 54.35 ± 0.92% of the dentin surface exposed to the microchannel. No cells are observed in the dentin region covered by PDMS (black asterisk), indicating effective sealing between device layers. Scale bar = 200 µm. (B) Confocal microscopy image showing viable dental pulp stem cells stained in green and adhered to the dentin surface (white arrows) after 24 h of static culture. Non-viable cells stained in red are also observed (blue arrows). Scale bar = 50 µm. Please click here to view a larger version of this figure.

Results

Successful fabrication of the tooth-on-a-chip device followed the workflow illustrated in Figure 1A–1I, including dentin disc preparation (Figure 1D), fabrication of the photolithography-derived mold (Figure 1E), removal of cured PDMS layers (Figure 1F), plasma activation prior to bonding (Figure 1G), assembly of the final device (Figure 1H), and connection of the device to the microfluidic setup (Figure 1I). The resulting device consisted of a three-layer PDMS assembly in which the dentin disc was fully integrated within the middle layer, with the upper and lower microchannels positioned on either side of the dentin surface (Figure 1H).

A positive outcome at this stage was characterized by the absence of visible air bubbles or delamination and by leak-free fluid passage following injection of PBS or culture medium through the inlet ports (Figure 1I). Misalignment of the microchannel ports, incomplete plasma bonding, or leakage at the dentin–PDMS interface were indicative of failed assembly and required repetition of the fabrication procedure.

A total of 30 devices were fabricated, of which approximately 20% failed during fabrication or testing. Failure was primarily associated with fluid leakage beyond the microchannel boundaries due to incomplete plasma bonding, whereas a smaller proportion of failures resulted from delamination in other regions of the device. No failures were attributed to leakage at the dentin–PDMS interface or to misalignment of the microchannel inlet and outlet ports.

Following 24 h of static culture, successful cell seeding was confirmed by the presence of adherent dental pulp stem cells exhibiting fibroblastic morphology on the dentin surface. Scanning electron microscopy provided assessment of both cell adhesion and device integrity (Figure 2A). Using the present protocol, adherent cells covered 54.35 ± 0.92% of the dentin surface exposed to the microchannel after 24 h of culture. A positive result was evidenced by the presence of adherent cells exclusively within the microchannel area, whereas the dentin surface covered by PDMS remained free of cell attachment (Figure 2A). This spatial distribution supported successful sealing between device layers and proper device assembly. Conversely, the presence of cells beyond the microchannel boundaries was indicative of leakage and should be considered a negative outcome.

Confocal microscopy using live/dead staining enabled non-destructive assessment of cell viability within the device. A successful outcome was characterized by a predominance of viable cells exhibiting green fluorescence and fibroblastic morphology, with only occasional non-viable cells stained in red (Figure 2B). In contrast, a high proportion of non-viable cells would indicate suboptimal dentin preconditioning, inadequate washing, or unsuitable culture conditions.

During protocol optimization, an initial fabrication strategy was evaluated in which a fixed mold incorporating a predefined recess was used to position the dentin disc within the middle PDMS layer (Figure 3A). Although this approach was intended to standardize disc placement and reduce leakage at the disc–PDMS interface, it proved inefficient in practice. Because the region corresponding to the dentin disc recess had to be manually cut from the cured PDMS layer, precise adaptation between the disc margins and the cut PDMS edges could not be reliably achieved, resulting in interfacial gaps that caused leakage in approximately 50% of the fabricated devices (Figure 3A).

Figure 3: Representative images illustrating unsuccessful fabrication approaches identified during optimization of the dentin–pulp complex OoC device. (A) Silicon mold containing fixed recesses designed for standardized dentin disc positioning (white arrows). Due to inherent variability in dentin disc dimensions resulting from manual specimen preparation, reliable adaptation to the fixed mold geometry could not be achieved, leading to interfacial gaps and leakage in approximately 50% of fabricated devices. (B) Scanning electron microscopy image of the dentin surface following application of a sealing material at the disc–PDMS interface. Exposed dentin (black asterisk) and sealing material deposits (white asterisk) are indicated. Unintended spreading of the sealing material onto the dentin surface compromised cell adhesion and reduced the available surface area for cell colonization. Scale bar = 100 µm. Please click here to view a larger version of this figure.

To address this limitation, a sealing material was applied at the disc–PDMS interface. However, this approach introduced an additional drawback, as the sealing material spread onto the dentin surface and prevented cell adhesion in the covered regions, as demonstrated by scanning electron microscopy (Figure 3B). These suboptimal outcomes highlighted two critical limitations of the initial fabrication strategy: inadequate adaptation between the manually cut PDMS recess and the dentin disc, and reduced cell adhesion associated with the sealing material.

Discussion

This study describes a step-by-step protocol for the fabrication of a vertically stacked tooth-on-a-chip device incorporating dentin and dental pulp stem cells and demonstrates representative cell adhesion and viability within the device under static culture conditions. The protocol was developed to provide a reproducible workflow for the fabrication and assembly of a dentin–pulp complex OoC platform suitable for in vitro studies of the dentin interface. The establishment of such a workflow is particularly significant because our recent systematic review highlighted that the field of microfluidic models for the dentin–pulp complex remains in an emerging stage, characterized by high methodological heterogeneity in device architecture and materials, which severely compromises the reproducibility and standardization of current tooth-on-a-chip models³⁴. By providing a detailed and reproducible protocol, this study addresses these limitations and contributes to the development of better-defined biological models. Within the broader context of OoC development applied to oral biology, França et al.¹⁵ pioneered the field with a microfluidic tooth-on-a-chip model incorporating dentin discs and odontoblast-like cells, demonstrating the feasibility of recreating the dentin–pulp interface in a dynamic microfluidic environment and evaluating the cytotoxic response to dental materials. Since this groundbreaking contribution, a growing number of oral OoC platforms have been reported in the literature³⁴, addressing diverse aspects of dental tissue biology, biomaterial interactions, and oral microbiology and establishing this approach as an active and rapidly expanding field of research. The present protocol contributes to this landscape by providing a detailed and reproducible fabrication workflow for a vertically stacked tooth-on-a-chip platform accessible to research groups with standard microfluidic capabilities.

Several steps in the protocol are critical for successful fabrication and performance of the device. In particular, pouring PDMS over the mold requires careful control to obtain uniform layer thickness and proper microchannel formation. The thickness of the PDMS layers should be selected according to the intended downstream application. For example, confocal microscopy requires sufficiently thin PDMS layers to permit optical visualization of cells adhered to the dentin surface. However, thinner layers may be more difficult to manipulate during fabrication and assembly procedures. Consequently, optimization of PDMS thickness represents an important balance between imaging accessibility and practical handling during device fabrication. Furthermore, in the present study, confocal microscopy of intact devices was not feasible because the PDMS layer overlying the cells had a thickness of 0.4 mm, which exceeded the penetration depth of the available confocal microscope (0.2 mm). As a result, confocal imaging was performed following device disassembly and direct exposure of the dentin surface. In addition, fluorescence microscopy was not applicable in this device because the inherent opacity of dentin precluded direct optical visualization of cells adhered to its surface.

This limitation is particularly relevant for vertically stacked OoC configurations, in which the microchannel containing the cells is positioned beneath the dentin disc, resulting in optical access being hindered by both the overlying PDMS layer and the opacity of the dentin itself. This limitation is not encountered in horizontally stacked chip designs, where the microchannels run laterally alongside the dentin rather than beneath it, thereby allowing direct optical access to the cell-seeded surface without interference from dentin opacity or PDMS thickness. The choice between vertical and horizontal stacking configurations should therefore consider not only fluidic and biological requirements but also the imaging modalities available in the laboratory. In an OoC system, device design plays a significant role, and the microenvironment must be maintained as close as possible to in vivo conditions. Like most tissues, the tooth is subjected to mechanical stimuli and exhibits anisotropic behaviour³⁵, meaning that its properties vary depending on the direction of the applied stimulus. Furthermore, histological differences exist between the coronal portion of the dentin–pulp complex, where dentin and odontoblasts are arranged vertically, and the lateral portion, where odontoblasts lie adjacent to dentin, with notable differences in cell number, organization, and dentinal tubule orientation. A vertically stacked chip design therefore provides a more faithful anatomical representation of carious lesions and restorative materials affecting the coronal aspect of the dentin–pulp complex.

Precise plasma surface treatment is essential for successful device assembly and remains one of the principal challenges in chip production³⁶. In this study, delamination occurred in 20% of the fabricated devices. This result is not unexpected because PDMS–PDMS bonding achieved with a handheld plasma pen, as used in the present study, generates a lower concentration of reactive oxygen species than vacuum-based plasma chamber systems, which allow oxygen concentrations above atmospheric levels and consequently produce a greater number of covalent surface groups available for bonding³⁶. As a result, the bonding strength achieved using a plasma pen is inherently lower and less reproducible than that obtained with a dedicated plasma chamber, which likely contributed to the observed delamination rate. Optimization of the dentin integration strategy also required multiple fabrication iterations to achieve intimate contact between the dentin disc and the middle PDMS layer while minimizing leakage. An initial approach involved fabrication of a predefined recess within the middle layer for dentin placement; however, manual sectioning of the recess prevented consistent adaptation between the dentin disc and PDMS margins, resulting in leakage during culture medium injection and yielding an overall fabrication success rate of approximately 50%.

A denture adhesive was subsequently evaluated to improve sealing at the dentin–PDMS interface, but scanning electron microscopy demonstrated interference with cell adhesion and reduced visualization of cells on the dentin surface. The final protocol was therefore optimized by positioning the dentin disc directly within the mold before carefully casting PDMS around the specimen, thereby improving integration of the dentin disc within the device and reducing the need for additional sealing materials. This approach effectively eliminated leakage at the dentin–PDMS interface by ensuring intimate and conformal contact between the dentin disc margins and the surrounding PDMS, without reliance on secondary sealing agents that could interfere with cell adhesion to the dentin surface. Several limitations of the present method should also be considered. Fabrication of microfluidic devices remains partially dependent on manual handling procedures, which may introduce variability between devices and operators. Although photolithography provides high precision and reproducibility for fabrication of microchannel geometries, the process may require specialized equipment and technical expertise. Alternative fabrication approaches, including three-dimensional printing techniques, may facilitate device prototyping and accessibility for some laboratories; however, factors such as printing resolution, residual monomers, and material biocompatibility require careful optimization³⁷.

The continuous supply of medium provides cells with a stable environment and protects them from waste accumulation or calcium/phosphate imbalance¹⁶. Several studies have shown that shear stress induced by constant flow can simulate mechanotransduction processes similar to in vivo signaling, which are key determinants of cell behaviour^38,39, contributing to cell differentiation and maturation and inducing significant changes in cell morphology, gene expression, and function³⁹. The present study reports the fabrication and initial validation of the device, demonstrating cell adhesion and viability after 24 h under static culture conditions; therefore, the effects of dynamic flow were not evaluated. The implementation of dynamic flow conditions represents an important avenue for future research and will be addressed in subsequent investigations. The present protocol provides a reproducible workflow for fabrication of a vertically stacked dentin–pulp complex OoC device incorporating dentin within a three-layer PDMS microfluidic platform. Potential applications of the model include investigation of dentin–biomaterial interactions, cellular responses to dental materials, and in vitro evaluation of the dentin–pulp interface under controlled experimental conditions.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Biological and chemical materials
Acetone	ExpertQ	SOLV 8	Silicon wafer cleaning
Alpha MEM culture medium	Cytiva	SH30265.02	Dental pulp stem cell culture medium
Chloramine (0.5%)	Merck	7080-50-4	Storage solution for dentin discs
Dental pulp stem cells	Not Applicable	Not Applicable	Primary cell culture. See Steps 3.2.1 and 3.2.2
EDTA (10%)	Winkler	N2222158E	Smear layer removal
Epoxy-based negative photoresist	Gersteltec	GM1060	Photolithography mold fabrication
Glutaraldehyde (2.5%) in 0.1 M Sorensen’s phosphate buffer	Sigma-Aldrich	210908-06	SEM fixation
Healthy human teeth (molars or premolars)	Not Applicable	Not Applicable	Human sample
Isopropanol	ExpertQ	SOLV 16	Wafer cleaning, development stopping, and PDMS cleaning
LIVE/DEAD Viability Kit	Sigma-Aldrich	451	Fluorescence viability staining.
PBS (phosphate-buffered saline)	Corning	46-013-CM	Washing and leakage testing
PDMS base and curing agent (polydimethylsiloxane)	Dow	Sylgard 184	Device fabrication
PGMEA	Sigma Aldrich	484431	Photoresist development
Trypan blue	Sigma Aldrich	T8154-100ml	Cell viability assessment
Trypsin	Corning	25-053-CI	Cell detachment
Ultrapure water	Cytiva	SH30529.01	Washing solution
Consumables and labware
19G syringe needles	Vmatic	VCB19050	Connection of device with micropump or syringe
Biopsy punch (diameter: 1 mm)	Harris Unicore	WHAWB100073	Inlet and outlet port fabrication
Cutting mat	Nicecool	Not Applicable	Surface protection during punching
Dental finishing discs (e.g., Sof-Lex)	3M	Sof-Lex	Dentin polishing
Female Luer-to-barbed adapters	Runze Fluidics	RH-M016	Tubing connection
Glass or plastic mixing stick	Duran	Not Applicable	PDMS mixing
Hypodermic syringes	Cranberry	pc-12246	Cell seeding and leakage testing
Lint-free sterile gauze	Linderson	Not Applicable	Removal of punching debris
Male Luer adapters	Runze Fluidics	RH-G016	Tubing connection
Mixing vessel	Duran	Not Applicable	PDMS preparation
Petri dish (140 mm × 20 mm)	Deltalab	2000219	Cell culture/incubation
Polytetrafluoroethylene base	CCTVal	Not Applicable	Intermediate dentin layer fabrication
Silicon wafer	Jiaozuo Commercial FineWin	Not Applicable	Mold fabrication
Silicone tube	Runze Fluidics	96421	Fluidic connection
T75 flask	Corning	353136	Dental pulp stem cell expansion
Well plate	SPL Life Sciences	30006	Confocal staining and imaging preparation
Equipment and instruments
Air dryer	METALPLAN	TITAN-040	Compressed air drying
Compressed air source	SCHULZ	CSD-18.1	Mold and surface cleaning
Confocal microscope	Olympus	Not Applicable	Model: FV1000
Direct laser lithography system	Heidelberg	uPG 101	Photoresist exposure
Environmental scanning electron microscope	Hitachi	Not Applicable	Model: Quanta 400
High-speed handpiece	Nakanishi	Not Applicable	Model: Pana-max2
Hot plate	LabTech	EH20A Plus	Wafer bake and photoresist bake
Laminar flow cabinet	BioBase	BBS8V1708313D	Device sterilization
Microfluidic pump	New Era	Not Applicable	Model: 1002X-ES. Leak testing at 10 µL/min
Microstructured mold (fabricated via photolithography)	CCTVal	Not Applicable	See Step 2.2 for mold fabrication specifications
Oven or incubator	Memmert	UN110	PDMS curing at 50°C and bond consolidation at 75°C
Vacuum Plasma Cleaning Chamber	Fari Plasma	GD-5	Wafer plasma activation
Handheld Corona Plasma	Aurora Pro Scientific	APS-CD-20AC	PDMS plasma activation
Ultrasonicator	Elma Sonic	S 10 (H)	Wafer cleaning
Vacuum chamber or desiccator (250 mm)	Winglass	Not Applicable	PDMS degassing
Vacuum pump	ROCKER	300	PDMS degassing