Bioprinting PDLSC-Laden Collagen Scaffolds for Periodontal Ligament Regeneration

Abstract

Periodontitis and severe trauma are major causes of damage to the periodontal ligament (PDL). Repairing the native conditions of the PDL is essential for the stability of the tissue and its interfaces. Bioprinting periodontal ligament stem cells (PDLSCs) is an interesting approach to guide the regeneration of PDL and interfacial integration. Herein, a collagen-based bioink mimicking the native extracellular matrix conditions and carrying PDLSCs was tested to guide the periodontal ligament organization. The bioink was tested at two different concentrations (10 and 15 mg/mL) and characterized by swelling and degradation, microstructural organization, and rheological properties. The biological properties were assessed after loading PDLSCs into bioinks for bioprinting. The characterization was performed through cell viability, alizarin red assay, and expression for ALP, COL1A1, RUNX2, and OCN. The in vivo biocompatibility of the PDLSC-laden bioinks was verified using subcutaneous implantation in mice. Later, the ability of the bioprinted PDLSC-laden bioinks on dental root fragments to form PDL was also investigated in vivo in mice for 4 and 10 weeks. The bioinks demonstrated typical shear-thinning behavior, a porous microstructure, and stable swelling and degradation characteristics. Both concentrations were printable and provided suitable conditions for a high cell survival, proliferation, and differentiation. PDLSC-laden bioinks demonstrated biocompatibility in vivo, and the bioprinted scaffolds on the root surface evidenced PDLSC alignment, organization, and PDLSC migration to the root surface. The versatility of collagen-based bioinks provides native ECM conditions for PDLSC proliferation, alignment, organization, and differentiation, with translational applications in bioprinting scaffolds for PDL regeneration.

Introduction

A healthy periodontal ligament (PDL) is crucial for the integrity of the supporting apparatus around the tooth because it links the alveolar bone closely to the cementum on the root surface. Furthermore, PDL harbors the cell source of cementoblasts which produce cementum.¹ Once PDL is lost, the integrity and regenerative capacity of periodontium are also severely impaired. Periodontitis and traumatic dental injuries are common conditions that compromise or cause PDL loss.^2,3 Conventional treatments for periodontal disease encompass debridement, root scaling, and surgical management of the tissue. Moreover, guided tissue regeneration (GTR) strategies via bone grafts and membranes that prevent epithelial tissue invagination into the area of the defect present some advantages in terms of repairing compromised periodontal sites.⁴ However, those cell-free strategies usually lead to disorganized tissue repair, which impairs the proper functional behavior of the periodontium, possibly related to the poor arrangement of the regenerated PDL fibers.⁵ Similarly, the clinical management of avulsed teeth has had little progress with cell-free approaches because the PDL is very vulnerable to dryness or improper storage. The success of replantation of the avulsed teeth is at the mercy of how long the tooth has been out of the socket or how it was stored.

In that regard, dental tissue engineering strategies and stem cell-based therapies could be advantageous to address PDL regeneration⁶⁻⁹ that would have a significant impact on treatments for avulsed teeth. Both cell-based and cell-free approaches have been tested and reported for periodontal regeneration through a myriad of mechanisms, which include the use of cell sheets containing periodontal ligament stem cells (PDLSCs), functionalized natural and synthetic polymers, and 3D printing technologies.¹⁰ For many years, cell-free strategies for periodontal regeneration were based on polymer-based electrospun membranes. However, the conventional electrospinning technique does not allow control over fibers deposition, which forms sheet-like scaffolds that limit cell organization and tissue vascularization.¹¹ In this sense, 3D-printed mono and multiphasic constructs resembling specific characteristics and compartments of the tissue are cutting-edge strategies to address full periodontal regeneration.¹²⁻¹⁵

Concurrently, cell-based therapies to treat periodontal defects that rely on periodontal tissue-derived cell sheets have been tested in animal models and clinical studies, and those models demonstrated to be safe and induce tissue repair.^6−8,16 Cell sheets are prepared by detaching and stacking a significant number of cell layers one over the other to have a high number of organized cells onto a self-supported extracellular matrix (ECM).^6,17 Despite the significant outcomes in vitro and in vivo, cell sheet preparation is complex and time-consuming, and the final structure presents poor mechanical stability that makes it hard to manipulate for implantation.¹⁷ Besides, combining cell sheets with periodontal membranes for the treatment of completely denuded roots in an orthotopic model did not result in satisfactory PDL regeneration, probably related to poor attachment of the cells to the root, limited attachment to the bone, and hypoxia conditions caused by the membrane.⁶ This scenario opens space for innovative methods such as 3D bioprinting as a more suitable alternative for cell-based periodontal ligament regeneration toward predictable management of avulsed teeth.

Bioprinting consists of 3D-printed cell-laden scaffolds for tissue regeneration, where it is possible to combine the biological and mechanical features of a material with cells from a specific source addressing the tissue to be regenerated. The bioprinting protocol involves mixing material solutions, the so-called bioinks, with the desired type of cells that can be printed through different methods such as extrusion-based, inkjet-based, and stereolithography (SLA)-based techniques to form 3D constructs mimicking the targeted tissue environment.¹⁰ Extrusion-based bioprinting is a relatively simple and affordable method that dispenses a continuous filament of bioink through a syringe and nozzle/needle system using controlled pneumatic or mechanical pressure.¹⁸ Meanwhile, inkjet-based systems use various mechanisms to produce droplets that can be dispensed in a picoliter scale in a precise manner.^19,20 The stereolithography or vat polymerization, on the other hand, uses a light source to precisely cure a photo cross-linkable bioink with the desired shape and high resolution with many applications in tissue engineering.²¹ Despite the valuable applications of SLA manufacturing techniques in bioprinting, the scenario of printing on the root of avulsed teeth demands continuous dispensing of the bioink directly on the surface to cover the root for subsequent replantation. Similarly, the inkjet method is a good approach to dispensing cells into a predesigned microenvironment but with limited application in the cases of full disruption of the PDL. In this sense, here, we adopted the extrusion-based method to bioprint our PDLSC-laden scaffolds in a controllable fashion to address cases of full PDL disruption in avulsed teeth. Using extrusion-based bioprinting also favors the storage of the bioinks in syringes for possible product commercialization and the necessary modifications on the equipment to integrate our further steps on bioprinting on the whole root surface.

Regarding the bioinks, they are biocompatible materials such as functionalized hydrogels or synthetic polymer-reinforced matrices. Although many synthetic polymers have been used as scaffolds for tissue regeneration strategies, they lack similarities with the native conditions of the ECM; therefore, hydrogels derived from collagen, gelatin, and hyaluronic acid are more frequently used for bioink preparation.²² Despite the reduced mechanical properties of natural components, natural polymers provide biological cues that favor cell survival, proliferation, nutrition, metabolic waste removal, and reduce hypoxia.²³ Noteworthy, collagen is the major protein contained in the ECM and can be obtained from different sources (e.g.: porcine, bovine, rodent),²⁴ which makes it a versatile alternative to various methods for tissue regeneration.²⁵⁻²⁷ Due to the abundance of collagen in connective and mineralized tissues such as bone, teeth, and the PDL, it is not unusual to see collagen-based scaffolds for biomedical applications, including resorbable GTR membranes⁴ and 3D-printed scaffolds for craniofacial regeneration.¹⁰ Nevertheless, it has been reported that the studies involving pure collagen as a bioink face the challenges of limited mechanical properties as they tend to use collagen at concentrations no higher than 5 mg/mL.^28,29 These limitations are critical for tissue engineering strategies in load-bearing areas, such as periodontal tissues, which might demand higher concentrations of collagen while still preserving the quality of the matrix for adequate cell survival, proliferation, and differentiation in cell-laden scaffolds.

Therefore, suitable collagen-based bioinks that mimic the ECM and allow specific cells to differentiate in their native environments are relevant for regenerative strategies. In the past years, some authors have worked on strategies of bioprinting to regenerate periodontal tissues either using commercially available self-assembling matrices⁹ or light-curable methacrylate hydrogels.³⁰⁻³² Those models tested different cell sources such as dental pulp stem cells and PDL cells that demonstrated osteogenic differentiation and regeneration after in vivo transplantation of the scaffold into the created periodontal defect.^9,32 Meanwhile, PDLSCs can differentiate toward PDL, cementum-, and bone-like lineages³³ and, combined with the appropriate scaffold, would facilitate PDL regeneration and integration with the alveolar bone and cementum in cases of fully disrupted PDL such as avulsed teeth, where the scaffold will be bioprinted directly on the root surface for tooth replantation.

Here, we investigated a collagen-based bioink mimicking the native ECM conditions and carrying PDLSCs to guide the PDL organization and regeneration. PDLSC-laden scaffolds were printed by an extrusion-based 3D bioprinter directly onto the root surface and subjected to in vitro and in vivo experimentation of their regenerative capacities.

Methods

Collagen Hydrogels Preparation

Clinical grade (sterilized) bovine type I lyophilized collagen (Revotek Co., Ltd. Lewes, DE, USA) was dissolved overnight in 0.1 M acetic acid under stirring. Acidic collagen solutions (10, 15, 20, and 25 mg/mL) were neutralized to pH 6.7–7.2 by adding 1 M NaOH in 10X DPBS (no calcium, no magnesium, 14200075, Gibco - Life Technologies, Grand Island, NY, USA) with thorough shaking. Each neutralized collagen solution was centrifuged at 2000xg for 1 min to remove air bubbles before any subsequent use.

Printing Parameters Optimization and Bioprinting

Collagen hydrogel printing was conducted using a BioX Bioprinter (Cellink, Gothenburg, SWE) through an extrusion-based bioprinting protocol with a temperature-controlled print-head set to 4 °C to avoid gelation of the gel into the cartridge. The temperature of the print-bed was set to 23 °C to accelerate the gelation after printing each layer and reduce the risk of collapsing the constructs. Different printing parameters were tested for all concentrations to optimize the printability of the bioinks. The 20 and 25 mg/mL concentrations did not present consistent printability and shape fidelity. The presence of bubbles trapped in the matrix was evidenced, and due to their viscosity, the printed layers were prone to collapse after printing. Therefore, only 10 and 15 mg/mL were used for the subsequent characterization and experimentation in this model (Figure S1). The following pressure/speed parameters were then adopted for all the studies involving bioprinting: 6–8 kPa/12–15 mm/s (10 mg/mL gels) and 10–15 kPa/12–15 mm/s (15 mg/mL gels) with 10 ms postflow using a 25G plastic nozzle (250 μm inner diameter). The parameters were further confirmed after mixing the cells to ensure that cell incorporation would not affect the printability.

Swelling Ratio and Enzymatic Degradation

The swelling ratio was derived by measuring the hydrated and dried weights of the collagen hydrogels. The neutralized hydrogels were added to cylindrical (10 mm diameter ×3 mm height) silicon molds and incubated for 30 min at 37 °C and 50% humidity to set (n = 4/group). The samples were then stored in PBS at 37 °C for 24 h, blot-dried, and weighed for the wet weight (Ww). After that, the samples were freeze-dried, and dry weights (Wd) were recorded. The swelling ratio was calculated using the following formula:

The enzymatic degradation analysis was quantified after immersion of the samples (n = 4/group) in 5 mL of PBS containing 1 U/mL collagenase type I (Gibco, Life Technologies) at 37 °C. Each sample was removed from the solution at specific time points, washed with deionized water, blot-dried, and weighed. The solutions were replaced every 3 days to preserve the enzymatic activity. The degradation ratio of the hydrogels was calculated using the following calculation:

W₀ represents the initial hydrated weight, while W_t is the weight at different time points.

Scanning Electron Microscopy

The microstructure of the collagen gels was evaluated by scanning electron microscopy (SEM). Round-shaped samples (2/group) were prepared as previously described for swelling and degradation and subsequently freeze-dried. The samples were cross-sectioned, mounted onto aluminum stubs, and gold-sputtered for SEM analysis (Evo HD LS15, Carl Zeiss, Oberkochen, Germany). Representative images of the surface and the cross-sectional area where the samples were cut were taken under high vacuum at different magnifications.

Rheological Properties

Rheological characterization of the hydrogels was performed using a Discovery HR-3 rheometer (TA Instruments, New Castle, DE, USA) with 1.2 mL of gel for each reading. The samples were tested at 4 °C and 180 s soaking time for homogeneous distribution of temperature through the whole specimen. Flow sweep readings were performed with a shear rate of 1 to 1000 s^–1 and 5 points per decade to calculate the apparent viscosity. The samples were also tested under oscillatory conditions to obtain storage and loss moduli under strain sweep conditions at an angular frequency of 10 rad/s. For each method, 3 samples per group were used, and the average rheological behavior was calculated.

PDLSC Isolation and Culture and GFP Labeling

Human PDLSCs were isolated from freshly extracted human teeth based on our established methods.³⁴⁻³⁶ The sample collection conformed to an exempt protocol approved by the Institutional Review Board of UTHSC (#19–06855 NHSR); no patient consents were needed. The standard medium for culturing PDLSC was α-MEM medium containing 10% FBS, 1% penicillin/streptomycin, and 2 mM of glutamine. Subsequently, the PDLSCs were labeled with a green fluorescent protein (GFP) by culturing the cells with regular medium added to a GFP-carrying lentiviral vector (8 μL/mL) and polybrene (0.5 μL/mL) for 24 h. The medium containing the lentiviral vector was removed after 24 h, and the wells were washed twice with sterile PBS before adding fresh medium. The cells were imaged using fluorescence microscopy to confirm GFP expression and were identified as PDLSC-GFP.

Cell Proliferation and Distribution

Bioprinting of grit-lattice constructs (10 × 10 × 0.3 mm) was conducted by extruding the cell-laden hydrogels (1 × 10⁶ cells/mL) into wells of a low-attachment six-well plate (Costar, Corning Incorporated, Kennebunk, ME, USA) following the established parameters according to the collagen concentration. After setting, 2 mL of α-MEM medium was added to each well, and the plates were incubated at 37 °C and 5% CO₂. Images were taken using fluorescence microscopy (Olympus SZX16, Olympus, Tokyo, JPN) and Nikon Eclipse Ti–S (Nikon Tokyo, JPN) at various time points for up to 14 days to observe cell growth.

Cell Survival after Bioprinting

As previously described, round-shaped samples were printed into wells of a low-attachment 6-well plate (Corning). After setting, the culture medium was added to the wells and the plates were incubated at 37 °C and 5% CO₂. The cell survival rate was tested 2 h and 3 days after printing as follows: the medium was aspirated, and the construct was washed with PBS and incubated for 30 min with 1.5 mM propidium iodide (PI) solution (Thermo Fisher Scientific, Waltham, MA, USA) to detect dead cells according to manufacturer’s recommendation. The PI solution was aspirated, and the wells were washed again with PBS to remove any unspecific staining before imaging the scaffolds. Images were taken under fluorescence microscopy (Nikon) at 4× magnification from three different regions of the construct using the red filter, and the experiment was performed in duplicate. Since the PDLSCs were GFP labeled, the green filter was also used to illustrate the live cells in the constructs. Cell survival in PDLSC-GFP-laden scaffolds without printing was also tested following the same protocol, to exclude any effects related solely to the nature of the bioinks and used for normalization of the cell viability values. Further quantification of the percentage of viable/dead cells was performed using ImageJ software (NIH, Bethesda, MD, USA) (n = 5/group).

Osteogenic Differentiation

The scaffolds were printed into wells of 6-well plates and incubated in a culture medium with osteogenic supplementation (10 mM β-glycerophosphate, 10 nM dexamethasone, 50 μg/mL ascorbic acid, and 10 nM vitamin D3) for 14 and 21 days. Bioprinted PDLSCs cultured in regular α-MEM medium with no osteogenic supplementation were used as a control. The medium was changed every 3 days. At each time point, the cells were washed with PBS and fixed with 10% formalin for 1 h. The formalin was removed, and the scaffolds were washed twice with PBS. Next, the scaffolds were stained with 1 mL of Alizarin Red S solution for 30 min under agitation. The excess staining was removed and washed four times with PBS, and pictures were taken under a microscope for qualitative evaluation. The quantification was performed using cetylpyridinium chloride to dissolve the crystals, and the supernatants’ absorbances were read at 562 nm.³⁶

The analyses of osteogenic marker expression were conducted after 7, 14, and 21 days of culturing the bioprinted PDLSC-laden hydrogels in an osteogenic medium. At each time point, TRIzol reagent (Thermo Fisher) was used for RNA extraction, the RNA purification was performed using RNeasy mini kit (Qiagen, Germantown, MD, USA), and the cDNA was obtained using Maxima First strand cDNA synthesis kit (Thermo Fisher). The expressions of RUNX2, ALP, OCN, and COL1A1 were investigated through quantitative polymerase chain reaction (RT-qPCR), and GAPDH was the housekeeping gene. The primers with each specific sequence for the tested genes are listed in Table S1. cDNA was detected using SYBR green master mix (Thermo Fisher), and the relative quantification of the gene expression was calculated using C_t fold change with GAPDH normalization.

Bioprinting on the Root Surface In Vitro

Root fragments (2/group) were obtained from the distal root of freshly extracted mandibular third molars. Under constant irrigation, the roots were cut into 8 mm × 6 mm × 2 mm using an Isomet low-speed saw (Buehler, Lake Bluff, IL, USA). After preparation, the root fragments were disinfected using sequential immersion into disinfectant solutions (17% EDTA, betadine, 3% sodium hypochlorite, and sterile PBS) under shaking for 10 min each. The samples were left to air dry and then incubated in serum-free α-MEM for 7 days to ensure disinfection. The root fragments were individually placed into wells of a 6-well plate, and the bioprinting succeeded on the external root surface. Two milliliters of medium were added to each well, the samples were incubated, and images were taken at specific time points up to 28 days. Additional root fragments were prepared using the same protocol for further in vivo implantation in an ectopic model in mice.

In Vivo Biocompatibility

Cylinder-shaped samples (Ø10 × 3 mm) were prepared by dispensing pristine and PDLSC-GFP-laden 10 and 15 mg/mL collagen hydrogels into custom-made polypropylene molds placed into the wells of a 6-well plate. After complete gelation, both cell-free and PDLSC-GFP-laden samples were incubated overnight immersed in culture medium and then subcutaneously implanted on the back of immunocompromised NSG mice for investigating the biocompatibility of the collagen-based bioink and the interaction of the PDLSC-GFP-laden scaffolds with the native tissue. All in vivo experiments were conducted under Institutional Animal Care and Use Committee (IACUC) approval (protocol #210261). Briefly, 8-week-old male NSG mice were anesthetized using 5% Isoflurane inhalation. Preemptive administration of anti-inflammatory was also performed (Meloxicam 2 mg/mL–100 μL per mouse). Bilateral subcutaneous pockets were created, and one sample was implanted into each side of the mouse. The wound was sutured, and a topical triple-antibiotic paste was applied. After recovering from anesthesia, the animals were single-housed and monitored according to the local LACU standards. After 1 week, the animals were euthanized and the tissue was retrieved and fixed in 10% formalin for histology. The samples were embedded in paraffin, sectioned (5 μm), and subjected to Hematoxylin/Eosin and Gomori’s Trichrome staining to evaluate the presence of any inflammatory response, the organization of the tissue, the degradation of the constructs, and the PDLSC morphology for the cell-laden groups. Additionally, immunohistochemistry staining was performed using anti-GFP (mouse monoclonal, GF28R, Invitrogen, Waltham, MA, USA) at 1:100 dilution by incubating the samples overnight at 4 °C. Secondary staining was conducted using Vectastain Elite ABC Universal kit, Peroxidase, R.T.U. (Horse Anti-Mouse/Rabbit IgG) (Vector Laboratories, Burlingame, CA, USA), according to the manufacturer’s recommendation. Nuclei counter-staining was done with hematoxylin.

In Vivo Root Fragments Implantation

Root fragments from third molars were prepared as described above and placed into the wells of 6-well plates for the printing process. The 10 and 15 mg/mL PDLSC/GFP-laden (2 × 10⁶ cells) and cell-free hydrogels were printed on the external surface of the root using a cylinder shape design (Ø5 × 0.5 mm) to cover the maximum area of the fragment. After setting, the samples were incubated overnight before implantation. An additional group of roots with no material on the surface was used as a blank control. The surgical protocol for implanting the root fragments containing the printed hydrogels was the same one used for biocompatibility. After 4 and 10 weeks, the animals were euthanized, and the tissue was retrieved and fixed in 10% formalin. The samples were embedded in methyl-methacrylate and sectioned using a microtome with a diamond blade (Indiana University School of Medicine, Histology Services Core, Indianapolis, IN, USA). The samples were stained for H&E, and images of the region of interest were taken at different magnifications. The thickness of the bioprinted scaffolds after 4 and 10 weeks was measured using ImageJ software, with ten measurements from the margin of the radicular dentin to the host soft tissue in different areas of the region of interest.

Statistical Analysis

The statistical analyses were performed using Prism 9 (GraphPad Software, San Diego, CA, USA). Data normality was verified using the Shapiro–Wilk test. Mass swelling was analyzed using independent t tests. Degradation rate, Alizarin Red S data, gene expression, and cell viability were analyzed using two-way ANOVA and Tukey post hoc. The thickness of the remaining bioprinted matrix in vivo was compared using one-way ANOVA and Tukey post hoc. For all the experiments, the significance level was established at α = 0.05.

Results

Physical Characterization of Hydrogels In Vitro

We first characterized some physical properties of the hydrogel. As illustrated in Figure 1A,B, the mass swelling for 10 mg/mL gels is significantly higher than for 15 mg/mL (p < 0.05). About 50% of the enzymatic degradation of the gels occurred in the first 3 days and continuously progressed up to 28 days, where complete degradation was evidenced. A higher degradation rate was evidenced for 10 mg/mL compared to 15 mg/mL at 3 and 7 days (p < 0.05). Rheological properties of the gels such as the viscosity and moduli were investigated after submitting the hydrogels to shear stresses under flow and oscillatory conditions. Figure 1C shows the flow behavior of the gels through their apparent viscosity under shear stress. The viscosity decreased as the shear rate increased with typical pseudoplastic behavior curves for both 10 and 15 mg/mL gels. Figure 1D presents the storage modulus (G′) and loss modulus (G″) for both gels under oscillatory shear stresses, where it is possible to evidence the higher moduli for 15 mg/mL hydrogels compared to 10 mg/mL and the linear region of the graph reaching up to 10% strain for both concentrations. The loss modulus values cross the storage modulus values for 10 mg/mL with about 100% strain, which occurred for 15 mg/mL with about 200% strain, indicating the plastic deformation and permanent changes in the material’s structure at those specific regions of the graph.

Figure 1.

Physical and biological properties of collagen hydrogels. A) Mean and standard deviation for mass swelling of 10 and 15 mg/mL collagen hydrogels. (B) Enzymatic degradation into collagenase solution over time. C) Flow behavior of the different concentrations of the collagen hydrogels. D) Storage and loss modulus of the hydrogels under oscillatory shear stresses. Asterisks (*) indicate statistical differences between groups * (p < 0.05); ** (p = 0.01). Swelling and degradation (n = 4/group/experiment). Rheological properties (n = 3/group/experiment).

SEM was used to observe the microstructure of the gels, where Figure 2 presents the SEM images with the surface and transversal views of the collagen hydrogels. The surfaces of both 10 and 15 mg/mL gels present similar wrinkled morphology with some pores appearing. The transversal views indicate the characteristic porous structure of the hydrogels with various pore sizes and shapes throughout the samples. Some pore deformation occurred in the cross-sectional view for the 15 mg/mL concentration, possibly as artifacts of sample preparation.

Figure 2.

Scanning electron microscopy images of the surface (top) and cross-sectional cut (middle and bottom) of the collagen hydrogels at 10 mg/mL (Left) and 15 mg/mL (Right). Dashed yellow squares indicate the magnified area of the cross-section. White arrows indicate the porous nature of the hydrogels with varying sizes.

Biological Characterization of the Bioprinted Constructs In Vitro

To characterize the cell behavior in the bioprinted scaffolds, we printed grid lattice (10 × 10 × 0.3 mm) constructs with GFP-labeled PDLSCs and cultured them in vitro for 14 days. Live and dead staining was used to test the survival of PDLSCs in the gel, and the representative fluorescence images of the bioprinted grid-lattice scaffolds laden with GFP-labeled PDLSCs are presented in Figure 3. It is possible to identify the spreading and proliferation of the cells in the matrix. Small changes in the shape of the grid and signs of degradation are evidenced on day 14 for the 10 mg constructs. Besides, the PDLSC alignment seems to follow the printing path as time passes (Figure S2). Representative merged images (green and red channels) of the cell survival assay for up to 3 days after bioprinting are shown in Figure 4A. A high cell survival rate was evidenced for both concentrations of the hydrogel, as indicated in the cell viability quantification (Figure 4B). Both concentrations of the collagen-based bioinks demonstrated high cell viability immediately after printing (>95%) which was maintained on day 3. The spreading and proliferation rate appears to be slower for 15 mg/mL gels than for 10 mg/mL. Additional magnified images of the red channels for better visualization of the dead cells are available in Figure S3.

Figure 3.

Fluorescence microscopy images of box-shaped bioprinted scaffolds (10 mm × 10 mm × 0.3 mm). Collagen gels, 10 and 15 mg/mL, laden with PDLSCs-GFP at day 0 (top) and day 14 of culture (bottom). Dashed yellow squares in the full-size images of the scaffolds (left) indicate the area from where the high-magnification images (right) were taken. White arrows in the high-magnification images on day 14 indicate signs of gradual degradation of the scaffold associated with the pulling forces of the cells. Scale bars: 1 mm (full-size images on the left); 500 μm (box-shaped areas on the right).

Figure 4.

A) Representative images of the cell viability assay immediately after printing (day 0) and on day 3 of culture for both 10 and 15 mg/mL gels. White arrows indicate the areas where cells are stained in red (dead cells). Magnified images of the red channel are available in the Supporting Information B) Bar graph with mean and standard deviation of the quantitative cell viability (%) for both groups on day 0 and day 3 after normalization to the nonprinted PDLSC-laden hydrogels. C) Mean and standard deviation for Alizarin Red assay and the representative images of the qualitative analysis for the nonosteogenic induced control, 10 and 15 mg/mL bioprinted cell-laden gels. D) Gene expression of mineralized tissue-related markers at different time points for both 10 and 15 mg/mL cell-laden bioprinted hydrogels. Asterisks (*) indicate statistical differences between groups *: p < 0.05; **: p < 0.001; ****: p < 0.0001. Scale bar: 500 μm

PDLSCs Showed In Vitro Osteogenic Differentiation Capacity in Hydrogel

The printed cell-laden hydrogels were cultured in osteogenic induction media for up to 21 days. Alizarin Red staining and RT-qPCR were used to assess the osteogenic differentiation of the PDLSCs. The graphs for the Alizarin Red staining and the gene expression of the osteogenic markers are presented in Figure 4C. For both time points, the osteogenic induction increased the mineral deposition, compared to the nonosteogenic control, regardless of the concentration of collagen (p < 0.05). On day 14, there was a slight difference in the osteogenic differentiation, where the 15 mg/mL cell-laden gels demonstrated more mineralization nodules (p < 0.05). In the meantime, on day 21, although both groups increased the mineral deposition, there was a shift in the pattern where 10 mg/mL concentration significantly presented the highest values of mineral deposition (p < 0.05). For the expression of the osteogenic markers, the difference between 10 and 15 mg of gel was minor despite being statistically significant for the upregulation of RUNX2 and OCN at day 14 in the 10 mg/mL gel compared to 15 mg/mL (Figure 4D). The expression levels of both RUNX2 and OCN significantly increased on day 14 for 10 mg/mL concentration, while the ALP expression spiked on day 21 for both groups with only intragroup differences. No statistical differences were evidenced for COL1A1 expression, even though the 10 mg/mL concentration presented higher mean values.

In Vitro Formation of Connective Tissue-Like Layer after PDLSC/GFP-Laden Gel Printed on the Root Surface

To observe the cell behavior after printing on the tooth root, we printed cell-laden collagen on the external surface of root fragments and cultured them in vitro for 28 days. The arrangement of the PDLSCs printed on the root after 28 days is presented in Figure 5. At low magnification, the location of printed cells shows a higher cell density as they proliferated over time in this 3D space. Cells also migrated away to populate the remaining root surface forming various levels of cell densities. Histological analysis of the transversal section from the middle of the PDLSC-laden scaffold (middle H&E panel) showed that a thin layer (150–200 μm) of soft connective tissue-like was well-attached on the root surface. Meanwhile, the layer of cells that migrated from the scaffold to the root (right H&E panel) was less than 50 μm thick. No newly formed cementum was observed, even though some cells that migrated into the matrix and the monolayer of cells that migrated to the root were attached to the surface.

Figure 5.

In vitro PDL regeneration on a tooth root. A) Schematic of the bioprinting process onto the external surface of the root fragment. Created with BioRender.com. B) Representative images of the PDLSC-laden hydrogels bioprinted on the root fragment on day 0 and after 28 days of culture in vitro. Cells almost cover the whole root surface after migrating from the originally printed area (dashed blue area) to the root. C–D) High-magnification fluorescence microscopy showing the orientation of the cells attached to the root surface beneath the original bioprinted area and the PDLSC alignment on the root surface after 28 days of culture (left), where the cells aligned and attached to the root forming a transition zone between the original area of the printed gel and the root surface. Dashed area in blue indicates the margin of the printed PDLSC-laden scaffold. Red asterisks indicate the layer of cells that migrated out of the original spot and covered the root. H&E staining of the fixed samples after 28 days of culture showing the cross-sectional view of the center of the PDLSC-laden scaffold printed on the root (middle) showing the PDLSC distribution (white arrows) into the matrix of the scaffold, and the monolayer of cells that migrated and attached to the root surface (right). RD (radicular dentin).

PDLSC-GFP-Laden Gel In Vivo

The in vivo tissue integration of the gel was tested by implanting the cell-laden or cell-free hydrogels subcutaneously in NSG mice. The experimental design and protocol of implantation are listed in Figure 6A. The H&E, Gomori’s Trichrome staining, and IHC staining for GFP are presented in Figure 6B. After 7 days, 15 mg/mL gels were more stable and degraded less when compared to 10 mg/mL. PDLSCs distributed in the cell-laden constructs are evidenced in both H&E and Trichrome staining, while no cells were present in the gel area of the cell-free gel samples, suggesting that mouse cells did not migrate into the gel after 7 days. The elongation and alignment characteristics of the PDLSCs are depicted in the high-magnification images stained in red for the Gomori’s staining and purple for the H&E. Besides, the cells inside the matrix of the scaffold were positively stained for GFP, which confirms those are the PDLSC-GFP and not cells from the host that invaginated into the matrix. There is a higher number of cells close to the periphery of the constructs in contact with the native tissue suggesting their out-migrating behaviors, likely to seek nutrition. No remaining piece of gel was identified in the histological analyses for the cell-free 10 mg/mL gels, which indicates complete degradation of the construct by the host.

Figure 6.

In vivo biocompatibility of cell-free and cell-laden gels. A) Schematic of cell-free and PDLSC-GFP-laden hydrogels preparation for the in vivo ectopic implantation on the back of the NSG mice. Created using BioRender.com. B) Representative images of the H&E staining, Gomori’s Trichrome staining, and IHC staining for GFP of the implanted samples 1 week postimplantation. PDLSC-GFP were stained and brown precipitates in the matrix of the PDLSC-GFP-laden scaffolds as indicated by the white arrows. Cell-free samples stained negative for GFP. Yellow stars indicate the bulk of the collagen matrix for the samples. Cell-free 10 mg/mL hydrogel was degraded entirely after 1 week. Dashed lines in both H&E and trichrome staining mark the margins of the collagen-based scaffold.

In Vivo PDL Formation on the Root Surface

The regeneration of PDL by our bioprinted cell-laden gels (5 × 0.5 mm) in an established mouse model of subcutaneous implantation was investigated, as shown in Figure 7A. The tooth roots were collected at 4 weeks and 10 weeks postimplantation, and the images for the H&E staining are displayed in Figure 7B–C. The cell-free hydrogel at 10 mg/mL concentration of collagen almost completely degraded, while the cell-laden matrix at the same concentration remained in place and attached to the root surface apparently forming an integrated zone with the root. It is also possible to notice the PDSLC aligning parallel to the root into the matrix for both time points. Regarding the 15 mg/mL concentration, the bioprinted matrix was preserved regardless of the presence of PDLSCs. Nevertheless, the tissue in the cell-laden strategy is integrated to the root surface where it is also possible to see PDLSC alignment and distribution into the bulk of the bioprinted scaffold and on the root surface. The thickness of the remaining bioprinted collagen scaffold varied according to the concentration and strategy where cell-free 10 mg/mL was the thinnest (less than 100 μm) and PDLSC-laden 15 mg/mL the thickest (∼400 μm) for both time points (p < 0.05) (Figure S4).

Figure 7.

In vivo PDL formation on the root surface. A) Schematic of the scaffolds printed onto the root fragment and ectopic implantation on the back of the NSG mice. Created using BioRender.com. B) Representative images of the H&E staining at different magnifications of the printed area for the cell-free and PDLSC-laden bioprinted collagen on the root surface after 4 weeks of implantation. Matrix is largely degraded in the 10 mg/mL cell-free strategy. Collagen matrix is fully attached to the surface of the root. PDLSCs are organized and aligned into the bioprinted scaffold for the cell-laden strategy. C) Representative images of the H&E staining at different magnifications of the printed area for the cell-free and PDLSC-laden bioprinted collagen on the root surface after 10 weeks of implantation. Bioprinted scaffold is still preserved for the cell-laden strategy and for the 15 mg cell-free. PDLSCs are more organized into the matrix. Interface between the bioprinted scaffold and the root has cells attached to it. Red asterisks indicate the root surface. White arrows indicate PDLSCs into the matrix of the bioprinted collagen scaffold.

Discussion

This study tested the use of 3D bioprinting technology for cell-based PDL regeneration. Our data indicated the first step of these possibilities with our evidence in the printing parameters, hydrogel properties, and PDLSC growth for PDL regeneration in in vitro and in vivo settings.

Critical injuries to the PDL originating from periodontitis or dental trauma result in periodontal defects and a loss of tooth stability. Cell-based strategies hold a promising potential for PDL regeneration.⁶ The advent of bioprinting using ECM-like matrices loaded with cells has created novel perspectives on the development of 3D-printed scaffolds for tissue regeneration.³⁷ The herein proposed model evidenced the potential benefits of bioprinting PDLSC-laden collagen bioinks for periodontal ligament regeneration due to the favorable conditions provided by the collagen matrix with controllable printability, and the ability of PDLSCs to differentiate toward the three key compartments of the periodontium–cementum, PDL, and surrounding bone.³³ As previously discussed, different factors influence the proper selection of materials for bioprinting and, despite the fact that collagen is the major component of the ECM and good for cell survival and proliferation, not all concentrations of the material are suitable for extrusion-based bioprinting without additional cross-linking strategies to reduce printing inaccuracy.^24,28 Even though four different concentrations of the collagen-based bioink were tested for calibration of the printing parameters, only 10 and 15 mg/mL presented enough accuracy and printing fidelity for conducting further experimentation, and the properties of both were subsequently investigated in vitro and in vivo.

Bioprinting scaffolds for tissue regeneration demands attention to the rheological properties of the bioinks, which impacts the survival rate of the cells during printing.^9,38 The tested concentrations of collagen in this study evidenced a typical shear-thinning behavior in the rheological characterization, which promotes favorable conditions for cell survival during the printing process because of the minimized shear stresses when the cell-laden bioinks are pushed through the nozzle with reduced viscosity.³⁹ Furthermore, the elastic modulus and the loss modulus were mostly determined by the concentration of collagen with a viscoelastic behavior illustrated by the linear stage until 10% strain, with a progressive decline after passing that range.

Apart from that, the stability in the environment and the ability of the matrix to allow cell proliferation and nutrition are of major interest while designing bioinks.³⁷ The swelling ratios of the collagen-based bioinks indicate up to 50% medium uptake, while the enzymatic activity led to significant degradation in the first week and complete degradation occurred between 3 and 4 weeks. Although the degradation profiles in vitro were similar, due to the nature of collagen, the 10 mg/mL bioink degraded significantly more than 15 mg/mL on days 3 and 7. Overall, these properties were dictated by the concentration of collagen, which is expected since that was the only variable in terms of processing the bioinks. Complementary, even though the freeze-drying method results in some damage to the collagen structure compared to protocols like critical point drying and cryo-SEM,^40,41 the SEM analysis confirmed the interconnected porous network of the collagen matrix, a primordial condition for cell proliferation and nutrition in scaffold-based regenerative strategies.⁴²

After the optimization of the printing parameters and confirmation that 20 and 25 mg/mL concentrations presented limitations regarding printing fidelity and stability, the printability of the PDLSC-laden scaffolds and cell behavior in the matrix of the printed scaffolds was investigated only for 10 and 15 mg/mL bioinks. The high cell survival rates in vitro for all the groups corroborate with previous findings when using 25G tips to print PDLC-laden hydrogels.³⁰ Besides, the degree of PDLSCs spreading and proliferation was dictated by the stiffness of the matrix as has been reported before by Lee et al.⁴³ Even though the biological mechanisms driving cell orientation into bioprinted matrices are not fully understood, it has been demonstrated that micropatterning and scaffold arrangement play important roles in guiding periodontal ligament cells’ commitment and orientation.^13,44−46 The evidenced straight or curved patterns of PDLSC alignment according to the area (middle or corners) of the grid-shaped bioprinted scaffolds presented here corroborate previous reports. That phenomenon occurred regardless of the concentration of collagen and despite the stiffness of the bioink impacting the timing for cell spreading and proliferation in vitro, where the 10 mg/mL bioink allowed faster proliferation and typical morphological organization compared to the 15 mg/mL group. Therefore, one could infer that the resulting shear forces from the printing path and the consequent mechanotransduction⁴⁷ drive specific molecular mechanisms to guide cell orientation into the matrix of the scaffolds, but further clarification on that matter is still necessary.

The PDLSCs’ ability to differentiate after the bioprinting process was subsequently validated in vitro, and both concentrations of bioink allowed PDLSC differentiation as evidenced in the Alizarin Red and gene expression assays. The stiffer matrix (15 mg/mL) significantly stimulated the mineral deposition for 14 days, while the softer matrix upregulated RUNX2 and OCN for the same time point and later (21 days) significantly increased the mineral deposition. Those characteristics are most likely related to the fact that the stiffness of the ECM regulates stem cells’ fate and differentiation.^48,49 The peak of upregulation for RUNX2 and OCN may have influenced the late mineral deposition at 10 mg/mL, where the soft matrix favors cell-to-cell interaction to optimize the differentiation response. Even though other models of periodontal regeneration using different bioinks to mimic the ECM did not find significant differences for Col1A1,(32) the fact that we used a collagen-based bioink may also explain the absence of statistical differences in our model. Regarding ALP, the changes in the gene expression were only time-dependent, with a gradient similar to those of both groups, which suggests a pure relation to the ability of PDLSCs to express alkaline phosphatase over time. In this sense, the in vitro osteogenic differentiation suggests 10 mg/mL as a more favorable microenvironment for PDLSC. However, balancing the properties and working on the versatility of collagen matrices to build multiphasic scaffolds may be an important model to address the regeneration of different compartments of the periodontium and their interfaces.⁵⁰

To further confirm the suitability of the model for PDL regeneration in cases of avulsed teeth denuded of PDL, the PDLSC-laden bioinks were bioprinted on the external surface of root fragments and cultured in vitro, where we evidenced the cells forming a transition zone from the bulk of the scaffold and migrating to cover the whole surface of the root after 28 days of culture. The cells attached to the root demonstrated an aligned pattern, while the ones close to the surface of the scaffold are more scattered which would suggest different commitments for tissue regeneration as a result of the interaction with the substrate. The histological analysis of the cross-sectional samples also confirmed the distribution of the PDLSCs inside the matrix and their outmigration to form a monolayer covering the root surface for potential differentiation in case of additional external stimuli. Moreover, despite the occurrence of some degradation after 28 days, the thickness of the bioprinted scaffolds cultured in vitro was within the range of the human PDL thickness,⁵¹ which reinforces the stability of the ECM-mimicking bioink.

After the in vitro experimentation that supported the applicability of the collagen-based bioink for the model of PDL regeneration, the biocompatibility of the material in both cell-free and PDLSC-laden strategies was confirmed in vivo using an ectopic model in mice, where the scaffolds are well tolerated by the host tissues. Additional bioprinted scaffolds on the root surface were also tested in vivo for 4 and 10 weeks to validate these PDLSC-laden bioinks’ applications for the treatment of dental avulsion. Notably, the bioinks were fully attached to the root surface at both time points, and the PDLSC-laden strategy evidenced cell survival, PDLSC proliferation inside the matrix of the scaffold, and a well-defined parallel alignment. Also, as it occurred in our in vitro model, the thickness of the bioprinted scaffolds on the root surface in the PDLSC-laden strategy varied according to the concentration of the gel but remained in the range of the human PDL that goes from ∼150 to 380 μm⁵¹ for both concentrations up to 10 weeks in vivo. In this sense, the tested concentrations of the collagen-based bioink indicate suitability for cell-based strategies guiding PDL regeneration, with the potential to be printed directly on the root surface of avulsed teeth. Even though these outcomes are promising, the parallel alignment of the PDLSCs suggests limited tissue maturation regarding the functional orientation of the PDL fibers toward the root surface.⁵² This is related to the absence of mechanical stimuli in this ectopic model since it is well understood that the mechanical forces of occlusion and mastication are an important mechanism driving the progenitor cells’ organization and the PDL arrangement for proper load distribution.^53,54 Furthermore, the thickness of the layer of material deposited on the surface of the root was not homogeneous due to the use of a preset design to print on an irregular surface such as a dental root, which is a limitation of the current model.

Therefore, our future steps include using image-based 3D designs obtained from the root surface and integrating a dynamic collector to compensate for the irregularities of the root during the printing process to create a homogeneous layer of bioprinted material for replantation of the tooth in the alveolar socket after dental avulsion. Moreover, further experimentation using an orthotopic model to investigate the integration between the three compartments of the periodontium in large animals would provide more insights into the efficacy of the model for complete regeneration of the damaged tissue in vivo for critical defects, such as the cases of avulsed teeth where the whole PDL is disrupted.

Conclusions

Bioprinting collagen-based PDLSC-laden scaffolds has great potential for PDL regeneration as the scaffolds preserve cell viability, stimulate PDLSC differentiation, and demonstrate biocompatibility and suitable PDLSC organization and attachment to the root surface both in vitro and in vivo. Moreover, through the evidence of the shear forces of bioprinting influencing cell alignment, it is possible to target areas involving different PDL fibers’ organization, moving toward image-based and defect-specific bioprinted scaffolds to regenerate PDL in critical periodontal defects.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.4c13830.

• Sequences of primers used for the gene expression analysis, representative images of the printed grid-shaped gel, illustration of printing path and representative images of PDLSC orientation, representative images of the live and dead assay, and representative images of the H&E staining for bioprinted scaffolds on the root surface with the thickness measurements (PDF)

The authors declare no competing financial interest.