Orthodontic tooth movement alters cementocyte ultrastructure and cellular cementum proteome signature

Abstract

Cementum is a mineralized tissue that covers tooth roots and functions in the periodontal attachment complex. Cementocytes, resident cells of cellular cementum, share many characteristics with osteocytes, are mechanoresponsive cells that direct bone remodeling based on changes in loading. We hypothesized that cementocytes play a key role during orthodontic tooth movement (OTM). To test this hypothesis, we used 8-week-old male Wistar rats in a model of OTM for 2, 7, or 14 days (0.5N), whereas unloaded contralateral teeth served as controls. Tissue and cell responses were analyzed by high-resolution micro-computed tomography, histology, tartrate-resistant acid phosphatase staining for odontoclasts/osteoclasts, and transmission electron microscopy. In addition, laser capture microdissection was used to collect cellular cementum, and extracted proteins were identified by liquid chromatography coupled to tandem mass spectrometry. The OTM model successfully moved first molars mesially more than 250 μm by 14 days introducing apoptosis in a small number of cementocytes and areas of root resorption on mesial and distal aspects. Cementocytes showed increased nuclear size and proportion of euchromatin suggesting cellular activity. Proteomic analysis identified 168 proteins in cellular cementum with 21 proteins found only in OTM sites and 54 proteins only present in control samples. OTM-down-regulated several extracellular matrix proteins, including decorin, biglycan, asporin, and periostin, localized to cementum and PDL by immunostaining. Furthermore, type IV collagen (COL14A1) was the protein most down-regulated (−45-fold) by OTM and immunolocalized to cells at the cementum-dentin junction. Eleven keratins were significantly increased by OTM, and a pan-keratin antibody indicated keratin localization primarily in epithelial remnants of Hertwig’s epithelial root sheath. These experiments provide new insights into biological responses of cementocytes and cellular cementum to OTM.

1. INTRODUCTION

The periodontal attachment complex of the tooth includes cementum, periodontal ligament (PDL), and alveolar bone [1]. Cementum is present in two main varieties that cover the root dentin. Acellular cementum is found on cervical root surfaces and is important for tooth attachment. Cellular cementum covers apical regions of root dentin and plays a role in adjustment of occlusion throughout life. Cementum and bone feature similar extracellular matrix (ECM) composition; however unlike bone, cementum is non-innervated and avascular. Cellular cementum includes cementocytes, osteocyte-like cells embedded in lacunae (spaces within the mineralized ECM) that radiate cellular processes through a network of canaliculi (interconnected tunnels) [2,3]. Osteocytes are mechanoresponsive cells and regulators of bone remodeling by signaling to osteoblasts (bone building cells) and osteoclasts (bone resorbing cells), remodeling their own perilacunar ECM, and directing mineral metabolism that controls bone mineralization [4,5]. It has been hypothesized that cementocytes may regulate aspects of cellular cementum biology in parallel fashion to osteocytes in bone. This proposition is supported by morphology and expression profiles of cementocytes in vivo and in vitro [2,3], parallel cementocyte and osteocyte defects in genetic models, and a study showing that experimentally-induced apposition in mice promoted accelerated cellular cementum formation in conjunction with changes in cell ultrastructure and in the proteomic profile of cementum [6,7]. However, conclusive evidence for roles of cementocytes has not yet been reported.

Animal models introducing challenges have been instrumental in demonstrating the functional importance of osteocytes, e.g. resistance of mice to unloading-induced bone loss following ablation of osteocytes provided robust evidence that osteocytes direct bone remodeling in response to changes in loading [8]. Orthodontic tooth movement (OTM) is a challenge to alveolar bone; application of orthodontic appliances creates loading-induced regions of compression and tension on PDL and bone by stimulating bone remodeling through osteoclast and osteoblast activities and allowing teeth to move within the alveolar bone. Ablation of osteocytes in mice reduced osteoclastic bone resorption and tooth movement signifying a role for these cells in induction of osteoclast activity [9]. OTM, particularly when large, prolonged forces are introduced, is associated with pathological tooth root resorption that targets apical cellular cementum [10,11]. OTM alters mechanical loading on cementum; therefore if cementocytes parallel some regulatory roles of osteocytes, they would be expected to show cellular alterations in response to OTM.

In this study, we aimed to analyze cementocytes and cellular cementum in a rat model of OTM. We hypothesized that OTM would induce changes in cementocyte activity reflected by altered cell characteristics and cellular cementum ECM proteome profile. To test our hypothesis, we used a multimodal approach including high resolution micro-computed tomography, transmission electron microscopy, laser capture microdissection of cellular cementum coupled with global proteomic analysis, and immunohistochemistry.

2. MATERIAL AND METHODS

2.1. Animals

Animal procedures were performed in compliance with the guidelines of the University Committee for Ethics in Animal Research n°4910–1/2018. A total of fifty 8-week-old male Wistar rats (Rattus norvegicus albinus) were used in this study and housed with access to water and rodent chow ad libitum. Animals were anesthetized with ketamine (80 mg/kg; Dopalen, Vetbrands, São Paulo, Brazil) and xylazine (8 mg/kg; Rompun, Bayer, São Paulo, Brazil), and maxillary first molars of rats were randomly ligated to maxillary incisors with 7 mm nickel-titanium closed-coil springs to deliver 0.5 N of force in the mesial direction for 2, 7, or 14 (n=4–6/group) days [12]. In this split-mouth study design, unloaded contralateral teeth served as controls. Tissues for histology and micro-computed tomography (described below) were harvested after 2, 7, and 14 days of OTM. Rats were perfused with 4% paraformaldehyde and 0.5% glutaraldehyde in 0.01 M phosphate buffer (pH 7.2) and euthanized. Maxillae were dissected and placed in 4% paraformaldehyde and 0.5% glutaraldehyde in 0.01 M phosphate buffer pH 7.2. Tissues for proteomic analysis (described below) were fixed in 10% Protocol buffered formalin (Fisher Diagnostics, USA) at 4°C for 2 hours, rinsed in phosphate-buffered saline (PBS), decalcified in 20% EDTA, and prepared for sectioning.

2.2. Macroscopic analysis

To document first molar movement, maxillae were imaged on a Stemi 305 (Carl Zeiss Microscopy, LLC, Thornwood, NY, USA) coupled to a computer running Labscope software version 2.8.0.

2.3. Micro-computed Tomography (Micro-CT)

To measure tooth movement, hemi-maxillae (n=5–6/group) were scanned in a SkyScan 1172-D scanner (Kontich, Belgium) at 70kV, 200μA, 0.5 mm Al filter, 0.4°rotation per projection, 10 frames averaged per projection, and 40 ms exposure time with 13.64 μm³ voxel size. To determine pore spaces/cementocyte lacunae, hemi-maxillae (n=3/group) were scanned under higher resolution in a μCT 50 scanner (Scanco Medical, Brüttisellen Switzerland) at 70 kVp, 76 μA, 0.5 mm Al filter, 900-ms integration time, and 2 μm³ voxel size. DICOMs were loaded into AnalyzePro 1.0 software (AnalyzeDirect, Overland Park, KS) for analysis. Scans were calibrated to a standard curve of five known hydroxyapatite densities (mg/cm³). All scans were oriented to a standard anatomical orientation. Cementum was traced on the 2 μm scans as previously described [13]. In brief, a median filter with a kernel size of 11 in x-, y-, and z-axes was performed. On this median filtered image, a mask of cementum was traced with a density range between 450 −1050 mg HA/cm³ with manual corrections. This mask was then placed over the original scan and cementum was identified as any object under the mask above 650 mg HA/cm³. This was further subdivided into cellular vs. acellular cementum manually by visual analysis for presence or absence of lacunae. Mesial vs. distal halves of the mesial molar root were divided by the most central axis of the mesial root’s pulp chamber in the sagittal orientation.

For detailed analysis of pores consistent in size with cementocyte lacunae, two cubical regions of interest (ROI) were selected from each tooth’s mesial root. The cubes were 100 μm³ and were selected from the most outer edge of cementum nearest to the apical end of the root as possible to be fully enclosed in cellular cementum on both the mesial and distal sides. For each of these ROI, all void spaces were traced as discrete objects with a density less than 450 mg HA/cm³. Based on area measurements of rat cementocyte lacunae by TEM, the estimated volumes of cementocyte lacunae were calculated by an approach previously described [14]. Using a simplified approach as cementocyte lacunae are erratic, spherical volume was estimated by measuring areas of 61 lacunae from 3 samples (ImageJ version 1.51; National Institutes of Health, Bethesda, MD, USA). Volumes were divided into 3 groups: upper 25%, middle 50%, and lower 25%. For the upper and lower groups, a 95% confidence interval was generated for each to make the upper and lower bounds for acceptable volumes for lacunae. Based on this approach, the range consistent with cementocyte lacunar size was determined to be 26–1,507μm³. Void spaces outside this range were excluded. Based on this estimate, cementocyte densities (pores/lacunae per mm³) and average pore/lacunar volumes (μm³) were obtained.

2.4. Histological analysis

Tissues for histology were decalcified in 10% EDTA solution for 4 weeks and paraffin-embedded for 6 μm serial sagittal sections (n=3–6). Hematoxylin and eosin (H&E) staining was performed as previously described [15]. Cementocytes on mesial and distal aspects of the mesial root of the first maxillary molar were counted and normalized to cellular cementum area (n=2–3 slides/each). Tartrate-resistant acid phosphatase (TRAP) staining (Takara Bio Inc., Japan) was performed to identify odontoclast/osteoclast-like TRAP⁺ cells on root surfaces as previously described [16]. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was performed to identify cells undergoing apoptosis with Alexa Fluor 488 dye, according to manufacturer’s instructions (ThermoFisher Scientific, Waltham, MA, USA).

Immunohistochemistry (IHC) was used to localize target proteins in cellular cementum. IHC was performed on paraffin sections (n=3/group) using an avidin-biotinylated peroxidase enzyme complex (ABC) based kit (Vector Labs, Burlingame, CA, USA) with 3-amino-9-ethylcarbazole (AEC) chromogenic substrate (Vector Labs) to generate a red-brown final product. Primary antibodies included: Goat polyclonal IgG anti-asporin (ASPN; Abcam Inc., Cambridge, MA, USA), rabbit polyclonal IgG anti-biglycan (BGN; Courtesy Dr. Larry Fisher, NIDCR/NIH, Bethesda, MD, USA), rabbit polyclonal IgG anti-decorin (DCN; Courtesy Dr. Larry Fisher, NIDCR/NIH, Bethesda, MD, USA), rabbit polyclonal IgG anti-periostin (POSTN; Abcam Inc., Cambridge, MA, USA), rabbit polyclonal IgG anti-collagen alpha-1(XIV) chain (COL14A1; Novus Biologicals, Centennial, CO, USA), and rabbit polyclonal IgG anti-wide spectrum cytokeratin (KER; Abcam Inc.). Rabbit polyclonal IgG anti-active caspase 3 (R&D Systems, Minneapolis, MN, USA) was used to detect apoptosis. With the exception of COL14A1, these antibodies were previously verified in mouse periodontal tissues [17–19]. Antibodies were verified for expected localization patterns in rat periodontal tissues (Appendix Figure 1).

2.5. Transmission Electron Microscopy (TEM)

Formalin-fixed, decalcified maxillae were post-fixed in 1% osmium tetroxide, dehydrated, and embedded in LR White resin (n=3/group). Sagittal sections (70–90 nm) were obtained from mesial roots, transferred to Formvar carbon grids, stained with uranyl acetate followed by lead citrate, and analyzed on a JEOL JEM 1400 (JEOL, Akishima, Tokyo, Japan). Similar regions in cellular cementum were imaged, and lacunae, cell nuclear sizes, and chromatin/euchromatin were measured using ImageJ software.

2.6. Proteomic Analysis

Tissues were collected by laser capture microdissection (LCM) as previously described for formalin-fixed paraffin-embedded (FFPE) samples (n=5/group) [19]. Longitudinal 6 μm thick sections of maxillary first molars were placed on PEN membrane glass slides (Applied Biosystems, USA). Sections were deparaffinized, dried, and microdissected [7]. Microdissected tissues were digested, and peptide samples were loaded on a mass spectrometer LTQ Velos Orbitrap (Thermo Fisher Scientific, Waltham, MA, USA) connected to a nanoflow LC (NLC-MS / DM) for the EASY-NLC system (Proxeon Biosystems, West Palm Beach, FL, USA) through a Proxeon nanoelectrospray ion source. The MS/MS spectra (MSF) were generated from the raw data files using Proteome Discover version 1.3 (Thermo Fisher Scientific) with Sequest (version 1.3.0.339; Thermo Finnigan, San Jose, CA, USA). Resulting spectrum count values were used to analyze the distribution of identified proteins in control and OTM groups. Detailed methods for LCM, peptide preparation, LC-MS/MS, and proteomic analysis are provided in the Appendix.

2.7. Statistical Analyses

Mean ± standard deviation is shown in graphs. For pairwise comparisons, quantitative data were analyzed by independent samples (Student’s) t-test (α=0.05), whereas multiple comparisons were performed by one-way ANOVA followed by post-hoc Tukey test for multiple comparisons (α=0.05).

3. RESULTS

3.1. Tissue Responses to Orthodontic Tooth Movement

Orthodontic appliances were placed in 32 dpn Wistar rats (n=50). Mesial movement of first maxillary molars was achieved by ligating NiTi coil springs between first maxillary molars and incisors to deliver 0.5 N of force for 2, 7, or 14 days (Figure 1A). In a split-mouth design, one side underwent orthodontic tooth movement (OTM), and the other served as a contralateral control. All rats completed the study except for four animals that were excluded for loss of greater than 20% body weight (n=1), cutaneous infection (n=1), or premature removal of the appliance (n=2).

Figure 1. Tissue responses to orthodontic tooth movement.

(A) NiTi coil springs were ligated between first molars (M1) and maxillary incisors of 32 dpn Wistar rats (n=50). In a split-mouth design, one side underwent orthodontic tooth movement (OTM) and the other served as a contralateral control. A red arrow indicates the mesial direction M1 will move under orthodontic loading. (B, C) While there is no visible space between M1 and second molar (M2) on the control side, OTM results in gaps (red stars in C) between M1 and M2 on days 2, 7, and 14. (D-F) 3D microCT renderings show the coil spring on the OTM side and non-ligated control side. (G) MicroCT measurements reveal increasing gaps at days 2, 7, and 14. Statistical analyses were made by ANOVA and post-hoc Tukey test (*P<0.05; ***P<0.001; **** P<0.0001). (H-J) MicroCT shows that by day 14, cellular cementum volume decreases on both the mesial and distal aspects of the M1 mesial root on OTM vs. control sides. (K, L) Histology reveals that OTM sides exhibit features consistent with root resorption (red star in L). Inset shows multinucleated TRAP⁺ odontoclast/osteoclast-like cells associated resorbing roots in OTM (M) Counts of TRAP⁺, multinucleated odontoclast/osteoclast-like cells increase 3-fold on OTM vs. control sides, including both mesial and distal aspects of M1 mesial root. Statistical analyses were made by independent samples t-test (*P<0.05; **P<0.01; ***P <0.001). M3=third molar; AB=alveolar bone; CC=cellular cementum; DE=dentin.

After euthanasia and tissue harvest, stereomicroscopy showed mesial displacement of the first molar on OTM sides of all animals. Compared to non-ligated control (Figure 1B), OTM introduced gaps between first and second molars at all three time points (Figure 1C). 3D micro-CT analyses revealed average gaps of 143 μm at day 2, 156 μm at day 7, and 255 μm by day 14 (Figure 1D-G).

Next, we focused on the mesial root of the first maxillary molar at 14 days for further analyses. Volumetric analysis by micro-CT indicated that cellular cementum volume on the OTM vs. control sides decreased over time on both the mesial aspect (by about 10%) and distal aspect (by about 25%) (Figure 1H-J). Histology revealed that OTM-treated teeth exhibited features consistent with root resorption and counts of TRAP⁺, multinucleated odontoclast/osteoclast-like cells increased 3-fold on OTM vs. control sides (P=0.0005) (Figure 1K-M). Numbers of TRAP⁺ odontoclast/osteoclast-like cells increased to a similar extent on mesial and distal aspects of molar mesial roots undergoing OTM (P=0.037 and 0.006, respectively). Collectively, these analyses validated the OTM model and showed that root resorption may be induced in association with OTM.

3.2. Altered Cementocyte Ultrastructure Associated with OTM

We next evaluated effects of OTM on cementocytes at 14 days; the time point marked by increases in: tooth movement, numbers of odontoclast/osteoclast-like cells, and cellular cementum resorption. We noted a small number of empty lacunae in cellular cementum on OTM-treated teeth near hyalinized zones of PDL and/or root resorption (Figure 2A, B). Compared to control sides, both mesial and distal aspects of roots undergoing OTM featured greater numbers of cells positive for the TUNEL assay for apoptosis (Figure 2C, D). Some cells on OTM-treated teeth were immunopositive for caspase 3, a marker for apoptosis (Figure 2E, F). By TEM, some cells exhibited ultrastructural features consistent with apoptosis, including cell shrinkage, pyknotic nuclei, and other characteristics (data not shown).

Figure 2. Altered Cementocyte Ultrastructure Associated with OTM.

(A, B) Empty lacunae (red arrow in B) are evident in cellular cementum (CC) on OTM sides near hyalinized zones of PDL. (C, D) In OTM vs. control side CC, more numerous cementocytes are positive (green) by TUNEL staining for apoptosis. (E, F) Cementocytes immunopositive for caspase 3 (red) are observed on OTM sides. (G-J) Cementocyte (Ccy) cell densities are not different on mesial or distal OTM vs. control sides. (K) The range of sizes for rat Ccy were estimated from area measurements in TEM images. Based on area measurements, the average volume of rat Ccy lacunae is 372 μm³. (L-N) High resolution 3D micro-CT analyses of pores consistent in size with Ccy lacunae (multicolor objects shown in representative cubical regions of interest in K) indicates no differences in pore/lacuna density or pore/lacuna sizes in OTM vs. control sides. Pores on mesial aspects are significantly larger than those on distal aspects in both experimental groups (40% larger; P=0.03–0.04). (O-R) TEM measurements of Ccy suggest no difference in Ccy lacunar (Lac) or cell areas in OTM vs. control sides, though a 20% increase in nuclear size associated with OTM (P=0.015). (S) TEM shows the area of heterochromatin is similar in OTM vs. control groups, but euchromatin increased 40% in OTM vs. control samples (P=0.046). DE=dentin.

Cementocyte cell densities, counted from H&E stained tissue sections, were not different on OTM vs. control sides on mesial or distal aspects of cellular cementum (Figure 2G-J). Using TEM images, we estimated the average area for rat cementocyte lacunae as done previously for mouse cementocyte lacunae [14]. Volumes were divided into 3 groups: upper 25%, middle 50%, and lower 25% (Figure 2K). For the upper and lower groups, a 95% confidence interval was generated for each to make the upper and lower bounds for acceptable volumes for lacunae. Based on this approach, the range consistent with cementocyte lacunar size was determined to be 26–1,507μm³ with the average volume of rat cementocyte lacunae estimated at 372 μm³. High resolution 3D micro-CT analyses of pores consistent in size with cementocyte lacunae indicated no differences in pore/lacuna density or pore/lacuna sizes in OTM vs. control sides, though pores on mesial aspects were significantly larger than those on distal aspects in both experimental groups (40% larger; P=0.3–0.4) (Figure 2L-N).

2D TEM measurements also suggested no difference in cementocyte lacunar or cell areas in OTM vs. control sides, though measurements revealed 20% increased nuclear size associated with OTM (P=0.015) (Figure 2O-R). TEM further indicated that while the area of heterochromatin was similar in OTM vs. control groups, euchromatin area was increased 40% in OTM samples (P=0.046) (Figure 2S).

3.3. Altered Proteomic Profile of Cellular Cementum Associated with OTM

Increased nuclear sizes and proportion of euchromatin suggested cementocyte activation in response to OTM. To identify altered protein secretion and cellular cementum ECM composition accompanying cementocyte activation, we used laser capture microdissection (LCM) to carefully select only cellular cementum and resident cementocytes from molar root tissues on OTM and control sides at 14 days. The proteomes of collected tissues and cells were analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). From this approach, a total of 168 proteins were identified, whose distribution is shown in the Venn diagram in Figure 3A. Among these, 93 proteins (55.3%) were shared between groups, 54 proteins (32.1%) were exclusively identified in control samples, and 21 proteins (12.5%) were detected only in the OTM group.

Figure 3. Altered Proteomic Profile of Cellular Cementum Associated with OTM.

(A) Venn diagram summarizing 168 proteins identified at 14 days in cellular cementum of only control (blue; 54) or OTM (orange; 21) groups, or shared between the two groups (overlap; 93). (B) Stacked bar chart demonstrating proportion of protein class distribution (% of whole) identified in control and OTM groups. (C) Direction of regulation (fold-change) of differentially abundant proteins in control (red) vs. OTM (blue) groups.

As a first step to understand these distinct proteomic profiles in OTM vs. control cellular cementum, we characterized the protein class distribution in each experimental group (Figure 3B). While most protein classes were represented in both groups, OTM was associated with increased proportion of calcium-binding proteins, chromatin/chromatin-binding proteins, cytoskeletal proteins, and protein modifying enzymes, and decreased proportion of chaperone proteins, defense/immunity proteins, and transporters.

We next examined the identities of exclusively and differentially identified proteins in OTM vs. control groups at day 14. Exclusively identified proteins in either group are listed in Appendix Table 1. Those identified only in control samples included some collagens (COL5A1, COL5A2, and COL6A1), the small leucine-rich proteoglycan (SLRP), osteomodulin (OMD), and several enzymes with varying functions. Those proteins exclusively identified in OTM samples included several enzymes and receptor-associated factors, many described as intracellular factors that would implicate resident cementocytes as their origin.

Proteins differentially abundant in OTM vs. control samples at day 14 are ranked in Figure 3C with factors decreased by OTM (i.e. more abundant in controls) in red at the top of the list, and factors increased by OTM in blue at the bottom of the list. Several proteins previously identified in cellular cementum were down-regulated by OTM. These included periostin (POSTN; −3-fold), several SLRPs e.g. decorin (DCN; −2-fold), biglycan (BGN, −3-fold), and asporin (ASPN; −4-fold), and several collagens e.g. COL1A1 (−1-fold), COL1A2 (−1-fold), and COL12A1 (−1.4-fold). COL14A1, a member of the FACIT (fibril-associated collagens with interrupted triple helices) collagen family, was the most down-regulated protein by OTM (−45-fold). Proteins increased by OTM included COL11A1 (1.5-fold) and 11 different keratins (ranging from 1 to 3.4-fold) with KRT83 being the most up-regulated (7.6-fold) factor.

To put these differentially identified factors into context and identify concerted effects on functional pathways, we performed gene ontology (GO) analysis using PANTHER to classify molecular function ontology into common classes, including molecular function, biological process, and pathway distribution (Figure 4). In terms of molecular function, control and OTM groups were remarkably similar, though molecular transducer activity (GO: 0060089) was not identified in the OTM group (Figure 4A). Five biological processes were also not identified in the OTM group: Biological phase (GO:0044848), immune system process (GO: 0002376), interspecies interaction between organisms (GO:0044419), reproduction (GO: 0000003), and reproductive process (GO: 0022414) (Figure 4B). Pathways showed more substantial differences between groups with control samples expressing 7 exclusive pathways (Axon guidance mediated by Slit/Robo, P00008; Axon guidance mediated by netrin, P00009; CKR signaling map, P06959; Glutamine glutamate conversion, P02745; Plasminogen activating cascade, P00050; Ras Pathway, P04393; and p38 MAPK pathway, P05918) and the OTM group expressing 6 exclusive pathways (Heterotrimeric G-protein signaling pathway-Gi alpha and Gs alpha mediated pathway, P00026; Heterotrimeric G-protein signaling pathway-rod outer segment phototransduction, P00028; Parkinson disease, P00049; Angiogenesis, P00005; Apoptosis signaling pathway, P00006; and B cell activation, P00010) (Figure 4C). While the functional interpretations for these proteomic differences in molecular function, biological process, and pathway distribution remain unclear, they do signify OTM-associated expression changes that likely include cementocytes.

Figure 4. Annotation of Proteins Differentially Abundant in OTM Versus Control Cellular Cementum.

Gene Ontology (GO) analysis using PANTHER was performed to annotate differentially abundant proteins into molecular function, biological process, and pathway groups. Values in pie charts are demonstrated in percentage (%). Exploded slices in pie charts indicate GO groups exclusively found in control or OTM groups. (A) Pie charts of molecular function GO groups in control and OTM groups. (B) Pie charts of biological process GO groups in control and OTM groups. (C) Pie charts of pathway distribution GO groups for control and OTM.

We used immunohistochemistry (IHC) to confirm localization of several factors differentially abundant in OTM vs. control samples (Figure 5). The SLRPS, BGN, and DCN showed strong localization in the PDL, outer cellular cementum ECM, and at the cementum-PDL interface (Figure 5A, B). Another SLRP, ASPN, showed more abundance in cervical vs. apical PDL but was also present in the outer cellular cementum ECM (Figure 5C). POSTN showed intense presence throughout the PDL with substantial staining within cellular cementum that appeared to be associated with embedded Sharpey’s fibers from the PDL (Figure 5D). COL14A1 has been previously identified in alveolar bone by proteomic analysis [20]. However, to our knowledge this is its first identification in cementum. IHC showed a wide distribution of COL14A1 within PDL, as well as pericellular localization within cellular cementum ECM, particularly in interior regions at the cementum-dentin junction (Figure 5E). Several keratins were increased by OTM, and a pan-keratin antibody marked a relatively small number of interior cells at the cementum-dentin junction that were likely associated with epithelial remnants of Hertwig’s epithelial root sheath (HERS) (Figure 5F).

Figure 5. Immunolocalization of Proteins Differentially Abundant in Cementum Undergoing OTM.

Immunohistochemistry (IHC) was used to validate (by red or brown-red reaction color) elected protein markers differentially abundant in cellular cementum (CC) of OTM vs. control groups. The blue dotted line in A indicates the cementum-dentin junction (CDJ). (A, B) Biglycan (BGN) and decorin (DCN) localize to the PDL, CC, and strongly at the CC-PDL interface. (C) Asporin (ASPN) localizes to cervical PDL and outer CC. (D) POSTN shows intense throughout the PDL and staining within CC associated embedded Sharpey’s fibers (red stars). (E) Type XIV collagen (COL14A1) shows a wide distribution within PDL and pericellular localization within CC, particularly at the CDJ. (F) Pan-keratin immunostaining marks cells likely associated with epithelial remnants of Hertwig’s epithelial root sheath near the CDJ.

4. DISCUSSION

Cellular cementum covers apical regions of tooth roots and functions as part of the periodontal attachment complex. Cementum and bone feature similar extracellular matrix (ECM) composition, and cellular cementum includes cementocytes, cells that reside within the mineralized ECM. It has been hypothesized that cementocytes regulate some aspects of cellular cementum biology in parallel fashion to osteocytes in bone, though little is known about functions of these cells. We challenged cementocytes in rat molars by applying a model of orthodontic tooth movement (OTM) where maxillary first molars were ligated to incisors with a coil spring appliance. This model promoted mesial movement of molars and induced apical root resorption; therefore, we analyzed cementocyte response through a multimodal approach. Compared to contralateral controls, cementocytes under OTM exhibited increased nuclear size and proportion of euchromatin. Laser capture microdissection (LCM) of cellular cementum paired with proteomic analysis revealed proteins exclusively or differentially abundant in cementum under OTM, implicating alterations in several biological processes and signaling pathways. These included proteins known to localize intracellularly that may signify the cementocyte response. Several proteoglycans, collagens, and collagen-associated proteins were down-regulated by OTM, and keratins were increased by OTM. These experiments provide new insights into biological responses of cementocytes and cellular cementum to challenge.

4.1. Effects of Orthodontic Tooth Movement on Periodontal Cells and Tissues

We aimed to provide new insights into cementocyte biology by challenging cellular cementum and its associated cells with OTM. A NiTi coil spring ligated to maxillary incisors was used to mesially move maxillary first molars, an approach used in several previous studies that delivers consistent results in rodents [21–23]. Stereomicroscopy and micro-CT analysis confirmed average movement of 255 μm for molars by 14 days, validating the model. Movement appeared to occur in two stages as previously described with immediate translation noted by 2 days resulting from deformation of soft tissues of the PDL and additional movement by 14 days coinciding with alveolar bone resorption [24,25]. Contrary to expectations, this model did not simply promote lateral movement but also tipped molars in a more complex 3D way, making compression and tension zones vary between the three maxillary molar roots in different rats. Importantly, this outcome would not have been evident without the 3D analysis capabilities of micro-CT but was additionally supported by observations such as localization of TRAP⁺ odontoclasts/osteoclasts and evidence of resorption on both mesial and distal aspects of the molar root. One factor possibly contributing to this unexpected relocation of the molar may be attachment to the continuously erupting rat incisors, leading to a more complex loading regimen. This limitation should be considered when analyzing effects of OTM in studies in rodents.

Periodontal tissue reactions to OTM have been well studied and described. OTM creates regions of compression and tension based on mechanical loading induced by the orthodontic appliance. Compression results in decreased blood flow (ischemia) and reduction of oxygen content (hypoxia) in the PDL tissues, inducing cell death and degradation of the PDL ECM (hyalinization) [26,27]. This promotes osteoclast-mediated alveolar bone resorption in compression areas allowing the tooth to move as the bone ECM is cleared. Root resorption is also associated with hyalinized tissue in compressed areas [28] and is considered one of the most frequent adverse occurrences of orthodontic treatment [29]. We demonstrated root resorption in our rat model of OTM by quantitative micro-CT analysis and presence of TRAP⁺ odontoclast/osteoclast-like cells. We noted TRAP⁺ odontoclasts/osteoclasts and resorption on both mesial and distal aspects of the molar mesial root, possibly due to variable compression/tension zones as discussed above. Effects of OTM on periodontal cells are understood to varying degrees and associated with particular tissue responses.

Apoptosis is associated with some cells during OTM. Apoptosis is a mechanism of programmed cell death characterized by chromatin condensation, nuclear fragmentation, and nuclear membrane disruption, while the integrity of cytoplasmic organelles is maintained [30,31]. Caspases, a family of proteases, play important roles in the apoptotic process and are often used as markers of programmed cell death [32]. Alveolar bone osteocytes respond to OTM and play an important role in the bone remodeling process [9]. In response to OTM, osteocytes express greater levels of cytokines, and those closer to hyalinized regions undergo apoptosis [9,27]. Osteocytes in the process of cell death release immune-stimulating molecules through their lacunar-canalicular network in the direction of the bone surface, promoting the releasing of proinflammatory cytokines by macrophages and activation of osteoclastogenesis [33]. When osteocytes were ablated in vivo in transgenic mice, tooth movement was significantly reduced pointing to an important role for these cells in promoting alveolar bone resorption in OTM [9]. Responses of cementocytes to mechanical forces such as OTM remain poorly understood. An in vivo study of OTM in rats demonstrated that some cementocytes close to the hyalinized tissue region undergo apoptosis [34]. We noted some empty cementocyte lacunae in our study. TUNEL staining, IHC for caspase 3, and ultrastructural observations by TEM supported presence of some cementocytes undergoing apoptosis, and proteomic analysis identified apoptosis pathway activation in OTM vs. control cementum. While cell death might contribute to or prevent cellular cementum resorption, the effects of cementocyte apoptosis remains unclear.

4.2. Altered Cellular Cementum Proteome under OTM

Even though a small number of cementocytes under OTM presented signs of apoptosis, the most striking morphological change we observed in cementocytes was increased nuclear size and proportion of euchromatin suggesting cells became activated in response to OTM. We adapted previously validated approaches to pair LCM with proteomic analysis of cellular cementum ECM [7,19,20,35,36]. We identified many proteins up- or down-regulated in OTM vs. control samples and further put these into a functional context using gene ontology (GO) groups.

Factors reduced by OTM included several members of the small leucine-rich proteoglycan (SLRP) family common components of ECM. Decorin (DCN), like many SLRPs, binds to and regulates collagen fibrillogenesis and is proposed to regulate collagen mineralization [37]. DCN is identified in PDL and cellular cementum, and genetic ablation of DCN in mice resulted in abnormal PDL collagen fibrils [38,39]. Biglycan (BGN) is a multifunctional SLRP widely abundant in connective tissues that interacts with numerous collagens and is proposed to act as a signaling molecule [40,41]. BGN is identified in PDL and cellular cementum, and its knockout in mice results in defective bone formation and altered tooth root-periodontal structures [19,42]. Asporin (ASPN) is widely abundant in dental and periodontal tissues and competes with DCN for collagen binding sites [19,43]. ASPN is reported to negatively regulate PDL progenitor cell differentiation and modulate mineralization [44], though its functional effects on periodontal tissues in vivo have not yet been reported. Periostin (POSTN), though not a SLRP, is also widely abundant in connective tissues and functions in collagen assembly in development and healing [45], and like the SLRPs reviewed above was down-regulated by OTM. POSTN is highly abundant in PDL, and its genetic deletion in mice resulted in severe breakdown of the periodontium [46,47]. Collectively, altered levels of these collagen-interacting factors in the cellular cementum proteome suggest a dynamic response to OTM involving multiple ECM modifications. While functions of these factors are partially understood, their altered abundance in cementum should be explored further, perhaps using genetically edited mice in challenge models like OTM or super-eruption of molars, to better understand their functional importance.

Collagen type XIV (COL14A1) was the most down-regulated marker associated with OTM, though several collagens were localized to cementum in this and previous proteomic studies [7,20]. To our knowledge, type XIV collagen has not been previously identified in cementum, and we noted specific localization in cells adjacent to the cementum-dentin junction (CDJ) by immunostaining. Type XIV collagen is a fibril-associated collagen with an interrupted triple helix (FACIT) associated with early stages of collagen fibrillogenesis, and genetic deletion in mice led to premature fibril growth and large fibril diameters in tendons [48]. Intriguingly, type XIV collagen has been identified in high mechanical stress tissues such as heart, tendons, and skin, leading to the hypothesis for its role in biomechanical processes [49,50]. The altered response of COL14A1 to OTM and its localization at the CDJ, a mechanically important tissue interface [51], should prompt additional studies into the functions of type XIV collagen in the periodontium.

Surprisingly, 11 keratins were increased in association with OTM. Pan-keratin immunostaining showed specific localization in a small number of cells and are likely epithelial remnants of the Hertwig’s epithelial root sheath (HERS) that directs tooth root formation and then disintegrates prior to cementogenesis [52]. As no keratin localization was identified in cementocytes or more broadly in cellular cementum ECM, and keratins have primarily been shown in HERS and epithelial remnants of Malassez (ERM) in association with root formation [53,54], this raises the intriguing possibility that epithelial cells were responsive to OTM; a hypothesis requiring further studies.

5. Conclusions

Cementum is an essential component of the periodontal attachment apparatus. We provide new insights into the rodent OTM model and associated response of cementocytes and proteome profile of cementum. These results do not yet directly identify cementocytes as actors in cellular cementum biology but provide additional evidence supporting this hypothesis by employing a challenge model. Discovery of type XIV collagen provides a novel candidate for further studies in cementum. Additional studies testing the roles of cementocytes are necessary for further insights and may eventually add to understanding of processes of root resorption and repair cementum, contributing to development of therapies that promote predictable repair and regeneration.

Supplementary Material

Supplemental Fig. 1. Supplemental Figure 1. Verification of Antibodies Used of Immunostaining.

Immunohistochemistry (IHC) was used to verify (by red or brown-red reaction color) immunostaining of selected candidate proteins in rat dental and periodontal tissues. (A) Negative control including no primary antibody shows no staining in dentin (DE), cellular cementum (CC), periodontal ligament (PDL), or alveolar bone (AB). A black dotted line shows the cementum-dentin junction (CDJ). Protein targets include: (B) Biglycan (BGN); (C) Decorin (DCN); (D) Asporin (ASPN); (E) Periostin (POSTN); (F) Collagen type XIV (COL14A1); and (G) Pan-keratin (KER). Staining in control vs. OTM tissues is shown in Figure 5.