https://orcid.org/0000-0001-6817-008X
Abstract
The use of dental implants to replace lost or damaged teeth has become increasingly widespread due to their reported high survival and success rates. In reality, the long-term survival of dental implants remains a health concern, based on their short-term predicted survival of ~15 years, significant potential for jawbone resorption, and risk of peri-implantitis. The ability to create functional bioengineered teeth, composed of living tissues with properties similar to those of natural teeth, would be a significant improvement over currently used synthetic titanium implants. To address this possibility, our research has focused on creating biological tooth substitutes. The study presented here validates a potentially clinically relevant bioengineered tooth replacement therapy for eventual use in humans. We created bioengineered tooth buds by seeding decellularized tooth bud (dTB) extracellular matrix (ECM) scaffolds with human dental pulp cells, porcine tooth bud-derived dental epithelial cells, and human umbilical vein endothelial cells. The resulting bioengineered tooth bud constructs were implanted in the mandibles of adult Yucatan minipigs and grown for 2 or 4 months. We observed the formation of tooth-like tissues, including tooth-supporting periodontal ligament tissues, in cell-seeded dTB ECM constructs. This preclinical translational study validates this approach as a potential clinically relevant alternative to currently used dental implants.
The primary objective of our research is to develop improved therapies for tooth loss. Toward this end, we have developed a novel method to create bioengineered tooth buds by seeding dental stem cells onto decellularized tooth bud (dTB) extracellular matrix (ECM) scaffolds, and subsequently implanted them into fresh mini pig tooth extraction sockets. Our results demonstrate the formation of mature tooth structures, supporting the potential use of dTB scaffolds for applications in whole tooth tissue engineering.
Introduction
Treatment of edentulism, congenital tooth agenesis, and tooth damage using dental implant placement has become increasingly popular in recent years.1,2 Dental implants, fabricated from titanium or titanium alloy, exhibit good mechanical strength, are biocompatible, and securely osseointegrate with the surrounding jawbone.3,4 Unfortunately, the direct transmission of mechanical forces of chewing from the implant to the supporting jaw bone can result in bone resorption over time and implant failure.5 In contrast, natural teeth are tethered to the jawbone via periodontal ligament (PDL) tissues, which absorb and modulate the forces of mastication, thereby promoting healthy bone maintenance.6
Tissue engineering technologies have shown significant potential for treating a wide variety of human congenital and acquired tissue defects. Selecting an appropriate 3-dimensional (3D) scaffold is essential for effective tissue engineering strategies, to provide proper physical support and biological cues to promote cell adhesion, migration, proliferation, and differentiation.7 Ideally, biodegradable scaffolds will gradually be replaced by natural extracellular matrix (ECM) produced by seeded or infiltrating cells, facilitating bioengineered tissue formation. Therefore, natural ECM and ECM-mimetic scaffolds are ideal for tissue engineering approaches, as they contain the necessary signals to direct tissue-specific cell function and differentiation.8
The composition of natural tissue ECMs varies considerably and can exhibit dynamic remodeling during development, creating challenges for biomimetic tissue engineering strategies.9 Therefore, methods have been devised to create decellularized natural ECM (dECM) scaffolds, by gently removing immune-responsive cellular components from natural tissues or organs, leaving behind the tissue-specific dECM. Published reports validate the use of dECM scaffolds for a variety of regenerative tissue applications including heart valves,10 nerves,11 blood vessels,12 whole hearts,13 livers,14 lungs15 and bladders.16 dECM scaffolds have been FDA approved and are currently available for use in clinical applications.17,18 Our lab has optimized a decellularized protocol to efficiently remove the cells from the porcine tooth buds.19
Our research focuses on creating bioengineered teeth. To date, embryonic tooth bud cells and tissues have been used to create small, functional teeth in rodents.20,21 But human embryonic tissues are difficult if not impossible to obtain and are therefore not a viable tissue source for applications in regenerative medicine. Therefore, our efforts focus on the use of post-natal dental cells as clinically relevant sources for bioengineered tooth formation. Adult dental pulp stem cells have been shown to generate osteodentin or dentin-pulp complex-like tissues,22,23 and our previous work has shown that the combined use of adult dental mesenchymal (DM) and dental epithelial (DE) stem cells can be used to bioengineer small teeth containing dentin, enamel, cementum, pulp, and periodontal ligament.24 Although these results are quite promising, the ability to engineer full-sized, functional replacement teeth is far more challenging.
We next developed bioengineered tooth bud constructs using porcine tooth bud derived dECM scaffolds, which we implanted into young, 6-month-old mini-pig mandibles.25 While this study demonstrated the formation of bioengineered teeth containing mature enamel, pulp, dentin, periodontal ligament and immature tooth roots, bioengineered teeth formation was observed in less than one-third of the implants. We believe that the major issue with this study was that successive permanent tooth eruptions in these young minipigs damaged and/or dislocated our implanted tooth constructs. Therefore, to avoid this issue, in this study described here we revised our experimental design to include only mature, 2-year-old minipig hosts with fully mature dentition and jawbone.
The results of this significantly revised experimental approach demonstrated that post-natal DE and DM cells seeded onto porcine decellularized tooth bud (dTB) ECM scaffolds much more consistently formed organized, human sized bioengineered teeth. We propose dTB-ECM as a promising and instructive scaffold for eventual therapeutic applications for human tooth regeneration.
Materials and Methods
Porcine dTB ECM scaffold preparation
Porcine second molar (M2) tooth buds were decellularized as previously published by us.19 The resulting dTB ECM scaffolds were dissected into 5 parts (Figure 1), and one part was paraffin-embedded, sectioned, and stained with 4’,6-diamidino-2-phenylindole to validate full decellularization. The remaining four parts were used for implantations. Only fully decellularized tb-ECM scaffolds were used in this study.
Study design. (A) Detailed schematic of implant study. Each decellularized porcine tooth bud was split into 5 parts, each of which were implanted or used for histological/immunohistochemical analyses. (B) Right mandible implant surgeries. (B1) Prior to tooth extraction. (B2) Immediately after tooth extraction. (B3) Implant placement. (B4) Sutured tooth socket. Arrows indicate where the implants were placed. Abbreviations: hDPCs: Human dental pulp cells; pDEs: porcine tooth bud derived dental epithelial cells; HUVEC: Human umbilical vein endothelial cells.
Cell isolation and in vitro expansion
Human dental pulp cells (hDPCs) and porcine tooth bud derived dental epithelial cells (pDEs) were harvested and characterized as previously published by us.25,26 The Human umbilical vein endothelial cells (HUVEC) cell line was purchased from ATCC (PSC100010, Manassas, VA), expanded in vascular basal media (PCS100030, ATCC) with VEGF growth kit (PCS100041, ATCC) in 5% CO2 at 37 °C, and cryopreserved at passage 3. All 3 types of cells were expanded in vitro to the required cell numbers immediately prior to use.
Constructing bioengineered tooth buds
Cultured cells were labeled as follows. pDEs were labeled with DiO (V22886, Thermo Fisher Scientific, Waltham, MA), hDPCs were labeled with DiD (V22887, Thermo Fisher Scientific), and HUVECs were labeled with DiR (D12731, Thermo Fisher Scientific). For each dTB-ECM scaffold, a mixture of 0.5 × 106 pDEs and 0.5 × 106 HUVECs (50 µl volume) was injected into the enamel organ (EO), a mixture of 1 × 106 hDPCs and 1 × 106 HUVECs (50 µl volume) was injected into the pulp organ (PO), and the 2 were then securely sutured together. Acellular dTB-ECM constructs consisted of sutured dEO and dPO with no cells. All constructs were then cultured in vitro for 1 week in a bioreactor in Osteogenic Media, separately for cell-seeded and acellular dTB-ECM constructs, in a mixture of 1:1:1 (dental mesenchymal: dental epithelial: endothelial) medium with osteogenic supplements (10 mM β-glycerol phosphate, 100 nM dexamethansone, and 50 μM ascorbic acid).
On the day of implantation, replicate in vitro cultured replicate scaffolds (n = 3) were fixed and later used to confirm proper cell localization and survival via immunofluorescent (IF) analyses for the DM cell marker Vimentin (sc-6260, Santa Cruz Biotechnology, Dallas, TX), the DE cell marker E-Cadherin (ABIN1858334, antibodies-online Inc. Atlanta, GA), and the HUVEC marker Factor VIII (ab20721, Abcam, Cambridge, MA). Quantification of DE, DM, and HUVEC cells at 5 locations in replicate paraffin-embedded and sectioned constructs throughout each construct was performed using ImageJ software. Differences among cell types were evaluated using the one-way ANOVA and the T-tests. P-value of <0.05 was considered to be statistically significant. Freshly isolated second molar tooth buds harvested from a 6-month-old pig sacrificed for an unrelated study at Tufts CMS animal facility were used as natural tooth bud (nTB) control implants.
Implanting constructs in the Mini Pig jaw
All animal experiments were performed using Tufts University IACUC approved protocols, including systemic anesthesia (1–2% Isoflurane) and electrocardiogram and heart rate monitors. Two-year-old mini pig hosts were used due to their similar size and anatomy to human mandibles, and for the fact that successional tooth eruption was completed. Briefly, the mandibular third incisor and first premolar teeth were extracted, and the extraction sockets were carefully curettaged and slightly enlarged to ensure full removal of all natural tooth tissues (Figure 1). Individual randomized constructs were then placed in each fresh extraction socket and sutured closed. For each time point (2 and 4 months), 6 pigs received 4 implants each, and a negative control pig received tooth extractions and no implants. In summary, Recell-dTB (n = 8), dTB (n = 8), and nTB (n = 8) were transplanted for each time point. Animals received antibiotics for 10 days post-surgery and were fed a gel diet for 1 week followed by a regular diet. Mini pigs were monitored 3 times a day for 3 days post-op, and then daily until sacrifice. All mucosal wounds healed within 7 days and no w8 loss or adverse reactions were observed in any of the host animals.
To visualize mineralized tissue formation over time, we performed i.v. injection of Xylenol orange (90 mg/kg, 398187, Sigma, St. Louis, MO) at 1 and 9 weeks, and Calcein green (20 mg/kg, C0875, Sigma) at 5 and 13 weeks. At 2 or 4 months of implantation, anesthetized hosts underwent perfusion fixation with formalin and mandibles were harvested and fixed in formalin for an additional week.
Microcomputed tomography imaging
Harvested, fixed 2-month mandibular implants were scanned using Xradia MicroXCT-200 (Carl Zeiss X-ray Microscopy, Inc. Pleasanton, CA), and 4-month implants were scanned using a newly purchased SkyScan 1176 mCT (Bruker Micro-CT, Kartuizersweg 3B, Belgium). Data reconstruction and analysis were performed using Avizo software (FEI, Inc. Hillsboro, OR).
Histological and immunohistochemical analysis
Each harvested and fixed jaw-implant segment was split sagittally through the center using a band saw. One half was prepared for mineralized tissue analysis via polymethylmethacrylate (PMMA) embedding and sectioning, and the other half was demineralized, paraffin embedded, and sectioned.
PMMA embedding and sectioning was performed at the University of Minnesota Hard Tissue Core Facility. Under the direction of Drs. Hari Prasad and Michael Rohrer, 100 µm sections were prepared and analyzed by fluorescent microscopy to detect cell tracker labeling, and XO and Calcein stained mineralized tissues. PMMA sections were then stained with Stevenel’s blue and Van Gieson’s Picro-Fuchsin for histological analysis.
The other implant halves were decalcified in formic acid/sodium citrate for 6 months, paraffin embedded and serially sectioned at 6 μm for histochemical and IHC analyses. Antibodies included: α-ECad (1:100, Antibodies-Online Inc., Atlanta, GA), α-VM (1:200, Santa Cruz Biotechnology, Dallas, TX), α-AM (1:50, Millipore, Burlington, MA), and α-DSPP (1:200, Santa Cruz Biotechnology, Dallas, TX).
Statistical analyses
Differences among cell types were evaluated using the one-way ANOVA and the T-tests. P–value of <0.05 was considered to be statistically significant.
Results
Fabrication and characterization of in vitro cultured bioengineered tooth bud constructs
A schematic, timeline, and image of the experimental and surgical approach used in this study is shown in Figure 1. All the cells used in this study—hDPCs, pDEs, and HUVECs—were easily expanded after cryopreservation and appeared healthy after fluorescent cell tracker staining (Figure 2A-C). A portion of each type of labeled cell was cultured in vitro for additional time to validate that the labeled cells maintained their typical morphology, proliferation rate, and viability. The dTB-ECM scaffolds, created as previously described,19,27 exhibited an opaque appearance, with dEO-ECM scaffolds appearing more transparent and softer as compared to the stiffer dPO-ECM scaffolds (Figure 2D). dTB-ECM scaffold sections were randomly selected for cell seeding or acellular controls. All dTB-ECM constructs were similar in size, ~0.4 cm3 (Figure 2E).
Recell-dTB construct fabrication. The 3 types of cells used in these constructs—pDE (A), hDPCs (B), and HUVECs (C)—all appeared healthy before and after being stained with fluorescent cell tracker DiO, DiD, and DiR, respectively (see insets). (D) Dissected dTB ECM EO and PO scaffolds before cell seeding (arrow indicates cusp). (E) Cell seeded dTB ECM scaffold (Recell-dTB) ready for implantation (arrow indicates sutures). (F) Implants being cultured in a perfusion bioreactor. Abbreviations: EO, enamel organ; PO, pulp organ. Scale bars = 100 µm (A-C), 200 µm (A-C insets), and 2 mm (D, E).
The dEO- and dPO-ECM scaffolds were seeded with pDE/HUVEC and hDPSC/HUVEC cell mixtures, respectively, and then tightly sutured together (Figure 2E). For optimized nutrient and oxygen exchange, cell viability and proliferation, a bioreactor system was used to in vitro culture the constructs for 1 week prior to implantation (Figure 2F). One-week in vitro bioreactor cultured paraffin sectioned replicate constructs exhibited high cellularity at the periphery of the re-seeded dTB-ECM constructs (Recell-dTB), especially at the dEO- and dPO-ECM scaffold interface (Figure 3). Fewer but sufficient numbers of cells were present throughout the scaffolds, although comparatively less than in natural tooth buds (nTBs), which also showed a more even cell distribution (Figure 3A). No nuclei were detectable in acellular dTB-ECM constructs (Figure 3A). All of the Recell-dTB appeared smaller in size after 1-week in vitro bioreactor culture, indicating compaction of the scaffold by the seeded cells.
Characterization of in vitro cultured tooth bud constructs. (A) H&E stained Recell-dTB, acellular dTB and natural TB (TB) constructs. Recell-dTBs exhibited good cellularity (arrows indicate some seeded cells), while no cells were detected in unseeded dTB constructs. nTB samples exhibited a more even cell distribution as compared to Recell-dTB constructs. (B) IF analyses showed E-cadherin-positive DE cells in the EO ECM, VM-positive DM cells in PO ECM, and Factor 8-positive HUVECs in both EO and PO ECMs. (C) Cell count result indicated that 40% of cells expressed E-cadherin, 46% expressed VM, and 20% expressed Factor 8. Abbreviations: EO, enamel organ; PO, pulp organ. Scale bars = 1 mm (A, B, low mag images); 50 µm (higher mag boxed areas and negative control). White-dotted lines indicate the EO and PO border.
IF histochemistry using cell type-specific antibodies was used to quantify all 3 cell types in 1 week bioreactor cultured Recell-dTB-ECM constructs prior to implantation (Figure 3B). Recell-dTB-ECM constructs consisted of approximately 40% E-Cad-positive pDEs, 45% VM-positive hDPCs and 15% Factor8-positive HUVECs (Figure 3C). No significant statistical difference was observed between the cell numbers of pDEs and hDPCs, but HUVECs showed way lower numbers (P ≤ .05). As expected, pDEs were enriched in dEO-ECM, hDPCs were enriched in dPO-ECM, and HUVECs were present throughout. Recell-dTB-ECM scaffolds exhibited higher expression of the basal membrane markers Col IV and Laminin, and Fibrillin 1 and 2, ECM markers for periodontal ligament and tooth root formation, respectively, when compared with acellular dTB-ECM scaffolds (Figure 4A). Recell-dTB-ECM constructs also showed robust expression of Dentin sialophosphoprotein (DSPP), a specific marker for odontoblasts and dentin, which was not detected in acellular constructs (Figure 4B, arrows).
Characterization of ECM components of tooth bud constructs. (A) IF analyses of CoIV, Laminin, Fibrillin 1, and Fibrillin 2 showed higher expression in 1 week in vitro cultured recell-dTB when compared with acellular dTB constructs. (B) DSPP was detected in Recell-dTB and nTB constructs, but not in acellular dTB constructs. Scale bar = 200 µm (A, B low mag images), and 50 µm (B, high mag images of boxed areas and negative control). Abbreviations: EO, enamel organ; PO, pulp organ. Dotted lines indicate the EO and PO border.
Hard tissue analyses demonstrate bioengineered tooth formation
In vitro cultured bioengineered Recell-dTB and dTB constructs and the nTB-positive controls were individually implanted in the mandibles of 2-year-old mini pigs. No contamination was observed in any of the in vitro cultured constructs prior to surgery, and none of the minipigs exhibited any adverse reactions to any of the implants. For analyses of hard tissue formation, 2 methods were used—(1) micro-CT and (2) plastic embedded and sectioned implants. XO and Calcein fluorescent dye injections, 4 weeks apart, were also used to monitor newly formed mineralized tissue formation.
Micro-CT showed mineralized tooth-like structures formed in Recell-dTB-ECM constructs after 2 and 4 months of implantation (Supplementary Videos S1 and S2, Figure S1). In total, 5/8 (62.5%) Recell-dTB-ECM constructs formed recognizable tooth-like structures after 2 months of implantation, and 4/8 (50%) formed tooth-like structures after 4 months (Supplementary Figure S1). In contrast, 2/8 (25%) acellular dTB-ECM and 1/8 (12.5%) nTB implants formed tooth-like structures after 2 or 4 months of implantation (Supplementary Figure S1). Micro-CT images correlated well with hard-tissue sections (Figure 5). Fluorescent dyes were incorporated into bioengineered alveolar bone and tooth-like hard tissues (Figure 5, A6-C6). Polarized light microscopy (POL) showed that Recell-dTB-ECM implants exhibited more mature collagen organization as compared to acellular dTB-ECM and nTB implants (Figure 5, A5-C5). The volumes of bioengineered teeth were generally smaller than natural porcine teeth, but contained a variety of natural tooth tissues (Supplementary Figure S2).
Hard tissue sectioned 4-month samples. Tooth-like structures were present in Recell-dTB-ECM, Acellular dTB-ECM, and nTB samples. Recell-dTB-ECM constructs appeared more mature than dTB-ECM and nTB implants. (A) Recell-dTB-ECM; (B) Acellular dTB-ECM; (C) nTB sample. (A1-C1) µCT cross-section through implant center. (A2-C2) Unstained hard tissue section. (A3-C3) Stevenel’s blue and Van Gieson’s picro fuchsin stained hard tissue sections. Boxed areas indicate implant. (A4-C4) High mag images of boxed areas in A3-C3. (A5-C5) Polarized light images of A4-C4. (A6-C6) Fluorescent imaging of XO and Calcein stained mineralized tissues in the same areas shown in A-C, 4, 5. Regenerated tooth-like structures were indicated by arrows (panel 1), boxes (panels 2 and 3), and circles (panels 5~6). Abbreviations: I: incisor, PM: premolar. Scale bar = 2 mm (upper panels 1), 1 cm (upper panels 2-3), and 1 mm (upper panels 4-6).
Demineralized sectioned implants demonstrated bioengineered tooth formation
H&E stained demineralized, paraffin embedded, and sectioned implants showed the cellular and tissue organization of bioengineered tooth-like structures (Figure 6). As per hard-tissue sections, low magnification imaging revealed grossly similar tissues in all 3 types of implants after 2 or 4 months implantation (Figure 6; Supplementary Figures S3 and S4). In contrast, high magnification imaging showed a variety of mineralized tissue types, including those typical of dentin and cementum (Figure 6). Bioengineered PDL-like tissues with distinctive Sharpey’s fibers were oriented perpendicular to the surface of bioengineered cementum (Figure 6, arrows). At 2 months, IHC analyses revealed robust DSPP expression in Recell-dTB-ECM derived bioengineered teeth, while dTB-ECM and nTB teeth did not (Figure 7A-C). In acellular dTB-ECM constructs, strong DSPP expression was detected in tissues surrounding the newly formed calcified tooth-like structures, but not in the construct itself (Figure 7D-F). Together, these results showed that the DSPP antibody used in this study detected DSPP in nascent bioengineered dentin/cementum-like tissues, but not in mature porcine dentin or bone. Finally, hDPCs were detected in both 2- and 4-month Recell-dTB-ECM implants, using a human mitochondria antibody (Supplementary Figure S5).
Histological analyses of demineralized sectioned 2 and 4 month implants. H&E stained bright field (panels 1-3) and polarized light (panels 1’-3’) images of constructs as indicated. High magnification images revealed the formation of dentin (D), cementum (Ce) and PDL. Bioengineered Recell-dTB-ECM derived tooth tissues closely resembled those of nTBs including the presence of PDL tissue with Sharpey’s fibers oriented perpendicular to the surface of the cementum. Panels 1-3 are high mag images of boxed areas of low mag panels. Abbreviations: C, canine; Ce, cementum; D, dentin; I, Incisor; PDL, periodontal ligament; PM, premolar. Scale bar = 2 mm (low mag images), and 50 µm (high mag images).
DSPP expression in bioengineered dental tissues. (A-C) 2-month implants. (D-F) 4-month implants. (Panels 1-3) High magnification images of boxed areas (1-3) in A-E, respectively. (A) 2-month Recell-dTB-ECM implants exhibited robust DSPP expression (brown stain, arrows). (B) 2-month acellular dTB-ECM derived bioengineered teeth showed limited DSPP expression. (C) 2-month nTB implant showed no detectable DSPP expression. (D) 4-month Recell-dTB-ECM teeth exhibited no detectable DSPP expression. (E) Strong DSPP expression was detected in bone surrounding 4-month dTB-ECM implants. (F) Positive control natural minipig tooth bud. (G) Negative control. (H) Mature porcine dentin (I) bone on the same slide of 4 month porcine jaw showed no DSPP expression. Scale bar = 200 µm (A-E); 20 µm (1-3); and 50 µm (F-I).
Discussion
The use of titanium implants to replace missing teeth has become increasingly widespread due to their reported success (~10 years) and highly lucrative clinical market.28,29 The global dental implant market is a valued at USD 6.6 billion in 2023 and was predicted a compound annual growth rate of 9.6% to reach a USD 10.53 billion market in 2029.30 Unfortunately, the corresponding increased numbers of implant failures will require alternative tooth replacement therapies, in addition to those individuals not wanting synthetic implants to begin with. It is therefore essential to devise effective alternative tooth replacement therapies to ensure healthy long-term patient outcomes, and to avoid a dental healthcare crisis. We propose that one reasonable approach is to create functional, bioengineered living teeth, complete with functional bioengineered PDL tissues. An effective bioengineered replacement tooth therapy could provide better patient satisfaction, oral and overall health, and reduced medical and dental care costs.
Previously, our lab defined methods to decellularize natural porcine tooth buds harvested from discarded 6-month-old pig jaws, using 5 cycles of 1% SDS/1%Triton X-100 treatment.19 This process preserved the major components of the tooth bud ECM, including collagen, laminin, and collagen networks,31 indicating the potential use of dTB-ECMs for bioengineered tooth regeneration.19 We used these constructs to conduct a pilot study in 6-month-old mini pig hosts,27 testing 4 types of replicate implants: (1) dTB-ECM scaffold alone, (2) Recell-dTB-ECM scaffolds, (3) dTB-ECM scaffold with BMP-2, and (4) freshly isolated nTB.27 All constructs were implanted and grown for 3 or 6 months. Subsequent analyses revealed the formation of small, bioengineered tooth-like structures, particularly in the Recell-dTB group. Certain concerns with this study included the fact that successive tooth formation in young porcine hosts displaced some of the implanted constructs, and that necrotic and improperly fixed tissues were found in all 4 types of implants at both 3 and 6 months, indicating the need for improved vascularization of bioengineered tooth bud constructs, and perfusion of the jaws prior to harvest. This pilot study also showed that the addition of BMP did not improve bioengineered tooth formation.
Based on these prior contraindicative results, here in this study we have significantly revised and improved our model as follows. First, we have standardized our methods to ensure reliable and consistent bulk production of dTB-ECM scaffolds. Next, here we used Yucatan mini pigs at least 2 years of age that have completed the formation of mature dentition (no successional tooth formation), and grew the implants for 2 and 4 months rather than 3 and 6 months. The larger sized jaws in 2-year-old pigs also allowed us to implant 2 constructs in each hemi mandible, on either side of a permanent canine that demarcated the implant sites. These significant modifications to our experimental approach allowed us to achieve a significantly higher rate of successful tooth regeneration—in 50% of the implants as compared to less than one-third—and the formation of bioengineered teeth approaching full human size. Finally, another major modification included in this study was to use of fluorescent dyes to label the implanted cells,32 and to label newly formed bone and tooth tissues,33 which allowed us to identify newly formed calcified tissues. Our sections of PMMA embedding samples showed a clear layer pattern of fluorescent staining around the newly formed tooth-like structures, similar to a thicker layer of fluorescent staining that can be observed around the trabeculae of the surrounding alveolar bone, which suggests bone formation was faster than the dentin formation in this model (Figure 5). Our approach was to label the 3 types of cells with different fluorescent cell trackers in order to be able to identify and track individual cell types in harvested bioengineered tooth buds. Unfortunately, the fluorescent cell labeling showed only weak signals after 1 week in vitro culture and was not detectable after 2 months of implantation. We therefore conducted IF staining of paraffin-embedded and sectioned specimens using cell-type specific vimentin antibody for hDPCs, E-cadherin for pDE cells, and factor VIII for HUVECs. Positive signals were then analyzed to quantify and assess the distribution of each cell type (Figure 3B, C). For in vivo implanted and harvested constructs, anti-human mitochondrial antibody was used to identify donor human versus porcine host cells (Fig. S5). Identifying the implanted cells within the newly regenerated tissue is crucial for understanding the mechanisms underlying cell-based therapies by illustrating the cell location and the interaction of seeded cells with the host microenvironment.34 Two types of cell labeling, direct and indirect cell labeling have been used for long-term cell monitoring.35 Direct cell labeling using radiotracers, quantum dots, or nanoparticles, does not require genetic modification of the cells but is susceptible to reduced signal with each cell division.36 Indirect labeling can be achieved via genetic modification such as a lentiviral reporter gene.37 The implanted cells can then be visualized by an appropriate probe or substrate. Although our IF analyses confirmed the presence of human donor cells in regenerated tooth tissues it does not fully reveal the interactions of host and donor contributions over time to the regenerated dental tissues. Future studies will consider improved cell labeling techniques to better analyze donor and host cell interactions in tooth regeneration.
In summary, the extensive modifications introduced into this study significantly improved our ability to definitively monitor bioengineered tooth development. We observed the formation of human-sized mineralized tooth-like structures, most frequently in the Recell-dTB implant group, which were significantly larger than those created in any of our previous studies27,38 or other published reports.22,39
Cell-seeded dTB-ECM scaffolds exhibited increased Fibrillin 1 and 2 expression, essential ECM markers for natural periodontal ligament and tooth root formation.40 Moreover, we observed the formation of mature PDL tissue around the bioengineered teeth after only 2 months in vivo growth. These results show that bioengineered teeth exhibit key properties of natural teeth that are lacking in titanium implants, and also indicate the potential ability to bioengineer PDL tissues onto titanium implants. We are currently working to test the feasibility of this approach.
The clinical success of any tissue-engineered construct is dependent on its long-term survival and integration with the host, in particular with the host vasculature.41 In addition, since the denuded vasculature of organ scaffolds is highly thrombogenic even when anti-coagulation drugs are used,42 dECM scaffolds must be recellularized in order to achieve successful transplantation. For these reasons, we maintained the use of HUVECs in our bioengineered tooth bud constructs, and HUVEC plus dental cell-seeded constructs showed increased calcified dental tissue formation in this study. We also used perfusion fixation of host jaws prior to harvest, and observed much reduced necrotic tissue formation and improved fixation and histology of implants as compared to our previous study.
Conclusion
During development, the ECM mediates biophysical stimuli and biochemical and molecular signals that direct the temporospatial organization of developing organs.43 The constant crosstalk between cells and ECM, called dynamic reciprocity, directs proper cell differentiation and tissue formation.44,45 Most excitingly, we conclude that the results presented here demonstrate, for the first time, the potential use of dTB-ECM scaffolds to engineer full human-sized replacement teeth.
Supplementary material
Supplementary material is available at Stem Cells Translational Medicine online.
Acknowledgments
We thank all members of the Yelick Lab for their expert advice and review. We kindly thank Drs. Hari Prasad and Michael Rohrer, University of Minnesota Hard Tissue Core Facility, for PMMA embedding and sectioning of mineralized implants.
Author contributions
W.Z.: experimental design, experimental execution, data collection, manuscript preparation, manuscript edits. P.C.Y.: experimental design, manuscript preparation, manuscript edits.
Funding
This research was supported by NIH/NIDCR R01 DE016132 (P.C.Y.) and NIH/NIBIB/NIDCR R01 DE026731 (P.C.Y.).
Conflict of interest
Authors declare no competing interest.
Data availability
All published data will be made available upon request.
