Molecular and circadian controls of ameloblasts

Abstract

Stage-specific expression of ameloblast-specific genes is controlled by differential expression of transcription factors. In addition, ameloblasts follow daily rhythms in their main activities i.e. enamel protein secretion and enamel mineralization. This time related control is orchestrated by oscillations of clock proteins involved in circadian rhythms regulation. Our aim was to identify the potential links between daily rhythms and developmental controls of ameloblast differentiation. The effects of selected transcriptional factors Distal-less homeobox 3 (Dlx3) and Runt related transcription factor 2 (Runx2) and clock gene Nuclear receptor subfamily 1, group D, member 1 (Nr1d1) on secretory and maturation ameloblasts [using stage-specific markers amelogenin (Amel), enamelin (Enam) and kallikrein related-peptidase 4 (Klk4)] were evaluated in HAT-7 ameloblast cell line. Amel and Enam steady-state RNA expression levels were down-regulated in Runx2 over-expressing cells and up-regulated in Dlx3 over-expressing cells. In contrast, Klk4 was up-regulated by both Dlx3 and Runx2. Furthermore, a temporal and spatial relationship between clock genes and ameloblast differentiation markers was detected. Of interest, clock genes not only affected rhythmic expression of ameloblast specific genes but also influenced the expression of Runx2. Multi-scale mathematical modeling is being explored to further understand the temporal and developmental controls of ameloblast differentiation. Our study provides novel insights into the regulatory mechanisms sustaining ameloblast differentiation.

Dental enamel is formed mainly during two distinct developmental stages (1, 2) i.e. the secretory stage (at the end of which the full thickness of enamel is completed) and the maturation stage (during which residual organic material is removed and the tissue is eventually occluded by hydroxyapatite crystals). Ameloblasts, the cells responsible for making enamel, are specialized epithelial cells with distinct morphological features that change during ameloblast differentiation (2). Secretory and maturation ameloblasts are characterized by restricted expression of enamel stage-specific genes and by stage-specific functions (3–8). However, the control of ameloblast gene expression resulting in specialized functions that direct enamel secretion and maturation is unclear (9–13).

Mineralized tissue development results from a complex temporo-spatial expression of adhesion molecules and growth and transcription factors. Different mineralized tissues share common signaling pathways. Runx2 and Dlx3 are both key regulatory transcription factors that control bone formation (14). Runx2 is also expressed by maturation stage ameloblasts during enamel formation and Runx2 mutations result in enamel abnormalities (15). Similar to Runx2, Dlx3 is strongly expressed in ameloblasts (16) and Dlx3 mutations are linked to amelogenesis imperfecta (17). This study aims to elucidate transcriptional targets of Runx2 and Dlx3 during amelogenesis.

In addition to the stage-specific regulation by transcription factors, it has been long suggested that gene expression and dental tissue formation are under circadian control both in rodents and humans. Previous studies demonstrated that the formation of incremental lines in rat dentin reflect a circadian rhythm in the collagen-synthetic and secretory activities of odontoblasts (18). Similar to rat dentin, human enamel is formed by appositional growth leaving growth marks on the enamel surface every 24 h during the secretory stage (19). In addition, amelogenin secretion shows clearly daily oscillations (20). At a later stage of development, the maturation stage, ameloblasts oscillate between smooth-ended and ruffle-ended morphologies every 8 h in rat and express a different set of proteins at each part of their cycle (2). Therefore, ameloblast differentiation is directly correlated with cyclical gene expression and specialized cell functions. However, no direct evidence for a “dental” circadian clock exists. It is also unclear if ameloblast-expressed genes are under circadian control and how circadian control affects ameloblast differentiation and enamel formation. This is the first study that aims to elucidate how clock genes regulate formation and maturation of mineralized tissues and how stage-specific regulation is linked to daily circadian controls.

Material and Methods

Cell culture and study of circadian effects

Ameloblast-like cells HAT-7 (21) were cultured in DMEM F12 (1:1) + L-Glutamine and 15 mM HEPES and 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA). Cells were passaged just before confluence and plated in 6-well plates. For measuring the circadian effects of clock genes HAT-7 cells were allowed to reach 80% confluence and then the medium was supplemented with 0.1 mM forskolin. Forskolin is known to induce cell cycle synchronization of cultured cells (22). Total RNA was harvested every 4 h for 28 h using Trizol (Invitrogen). Two μg of RNA was reverse transcribed with TaqMan reverse transcription reagents (Applied Biosystems, Branchbury, NJ, USA), following the manufacturer’s recommendations. cDNA was then quantified, and used for real-time quantitative RT-PCR (qRT-PCR). qRT-PCR was conducted using SYBR Green (Invitrogen) and specific primers (Table 1) for Beta-actin (Actb), Amelogenin (Amel), Enamelin (Enam), Ameloblastin (Ambn), Matrix metalloproteinase 20 (Mmp20), kallikrein related-peptidase 4 (Klk4), Aryl hydrocarbon receptor nuclear translocator-like (Bmlα1), Nuclear receptor subfamily 1, group D, member 1 (Nr1d1), Distal-less homeobox 3 (Dlx3), and Runt related transcription factor 2 (Runx2).

Table 1.

RT-PCR primers

Detailed sequence information is provided here for all the primers used in this study. The following gene symbols are used in this table: Beta-actin (Actb), Amelogenin (Amel), Enamelin (Enam), Ameloblastin (Ambn), Matrix metalloproteinase 20 (Mmp20), kallikrein related-peptidase 4 (Klk4), Aryl hydrocarbon receptor nuclear translocator-like (Bmlα1), Nuclear receptor subfamily 1, group D, member 1 (Nr1d1), Distal-less homeobox 3 (Dlx3), and Runt related transcription factor 2 (Runx2).

Gene Name	Sequence 5′ – 3′
Ambn	Forward: GTCCAGAAGGCTCTCCACTG Reverse: GTCATTGGGGAAAGCAAGAA
Amelx	Forward: TACCACCTCATCCTGGAAGC Reverse: CTGTTGAGACAGCACAGGGA
Dlx3	Forward: ACCCAGTGTCGGTGAAAGAG Reverse: GCCAGATACTGGGCTTTCTG
Enam	Forward: GATGCCCATGTGGCCTCCACCA Reverse: GCCAAATGGTGGGAATGGCTGA
Klk4	Forward: ACAAGGGCTCGTGTCTATGG Reverse: GTCTCAGGTTCCCTCAGCAG
Mmp20	Forward: AGCTCGTCCTTTGATGCAGT Reverse: TGGACATTAGCTGGGGAAAG
Nr1d1	Forward: CTTCCGTGACCTTTCTCAGCA Reverse: TGTGCGGCTCAGGAACATCAC
Runx2	Forward: CCGTCCATCCACTCTACC Reverse: TGCCTGGCTCTTCTTACTG
Actb	Forward: AAGTACCCCATTGAACACGG Reverse: ATCACAATGCCAGTGGTACG
Bmlα1	Forward: CCAAGAAAGTATGGACACAGACAAA Reverse: GCATTCTTGATCCTTCCTTGGT

Transfection and real-time quantitative RT-PCR

HAT-7 ameloblast-like cells were cultured with DMEM F12 (1:1) + L-Glut and 15 mM HEPES and 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA). Cells were passaged and plated in 6-well plates. Cells were then transfected at 80% confluence with 2 mg Nr1d1 or Dlx3 (gift of Dr. Maria M Morasso, Developmental Skin Biology Section, NIAMS-NIH, USA) or Runx2 (gift of Dr. Renny Franceschi, University of Michigan, Ann Arbor, USA) or control (empty pCDNA) expression vectors using Lipofectamine LTX and Plus Reagent (Invitrogen). Total RNA was isolated 24 or 48 h later from HAT-7 ameloblast cells using TRIzol (Invitrogen), and 2 μg of RNA was reverse transcribed with (Applied Biosystems) following the manufacturer’s recommendations. The resulting cDNA was then amplified by qRT-PCR. RT-PCR amplifications were performed at 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s using specific primers (Table 1). Relative expression levels for each gene was calculated based on Actb expression levels and differences are represented in graphs using the 2^−ΔΔC_T method. The statistical method to calculate the p-values was a two-sample t test. RT-PCR products were also sub-cloned into pGEM-T Easy vector (Promega, Madison, WI, USA) and mRNA expression was confirmed by direct sequencing.

Results

Effects of Runx2 and Dlx3 on ameloblast-specific gene expression

HAT-7 cells were transfected with Runx2 expression vector and the changes on mRNA expression levels of stage-specific ameloblast genes (i.e. Amel, Enam and Klk4) were evaluated by qRT-PCR. Our data showed that Runx2 down-regulates Enam RNA levels (Fig. 1A) as well as Amel RNA levels (not shown) and up-regulates Klk4 RNA levels (Fig. 1A). HAT-7 were also transfected with the Dlx3 expression vector. Ameloblast specific genes Amel and Enam (markers of secretory ameloblasts) and Klk4 (marker of maturation stage ameloblasts) were all up-regulated upon Dlx3 over-expression in HAT-7 cells (Fig. 1B). Cells transfected with a control vector (pCDNA) showed no significant changes.

Figure 1.

A: Runx2 effects on HAT-7 ameloblast cells. Runx2 over-expression resulted in down-regulation of Enam (p<0.05) and up-regulation of Klk4 (p<0.05). Experiments are done in triplicate. B: Dlx3 effects on HAT-7 ameloblast cells. Over-expression of Dlx3 encoded for full-length protein (1–287aa) resulted in up-regulation of amelogenin, enamelin and Klk4 (p<0.05 for all three genes). All data was evaluated 24 h after transfection. C: Circadian oscillations at RNA level were found for clock genes in ameloblasts after cell cycle synchronization using forskolin. The results are shown here for Nr1d1, which showed a pic of expression 12 h after cell cycle synchronization. D: Transfection of HAT-7 cells with Nr1d1resulted in statistically significant up-regulation of amelogenin (p<0.05) and down-regulation of Klk4 (p<0.056) and Mmp20 (p<0.05). In addition, Runx2 mRNA levels were also up-regulated 24 h after Nr1d1 transfection of HAT-7 cells (not statistically significant). In contrast, Enam steady state mRNA levels remained unchanged upon Nr1d1 over-expression. * reports statistically significant; changes; + reports very close to be statistically significant changes.

Effects of Nr1d1 on ameloblast-specific gene expression

Cell cycle synchronized HAT-7 cells were used to evaluate the expression levels of clock genes in ameloblasts in regular daily intervals. Several clock genes were detected and found to oscillate at the RNA level. One of the most regularly oscillated clock genes in ameloblasts was Nr1d1 (Fig. 1C). We then decided to evaluate if over-expression of Nr1d1changes the expression of mRNA levels of ameloblast-specific genes. Nr1d1 over-expression resulted in up-regulation of Amel and down-regulation of Mmp20 and Klk4 (Fig. 1D). In contrast, Enam (Fig. 1D) and Ambn (not shown) RNA expression was unchanged upon Nr1d1 over-expression. Furthermore, we have also evaluated changes in the expression levels of Runx2, a key regulator of ameloblast-specific genes (Fig. 1D).

Multi-level and time dependent control of ameloblast-specific gene expression and cell functions

We are also being analyzed how these two networks (stage-specific regulation and circadian control) govern ameloblast differentiation and enamel formation using a multi-scale modeling approach. In our computational approach, cells are modeled as discrete entities that respond to intracellular and extracellular signals, which are modeled continuously with differential equations. Key circadian clock genes involved in amelogenesis are being integrated into a Boolean gene network. In their simplest form, Boolean models are interaction networks where each biochemical species is represented as a node in one of two possible states: expressed (“on” or 1) or non-expressed (“off” or 0) (23). Transfer functions between states are derived from biochemical interactions using logical operators (e.g., AND, OR, and NOT). The response to signals from the intracellular gene network determined whether each cell differentiate, proliferate or die and, therefore, directly influences the cellular and the extracellular tissue scales. The spatial distribution of cells is computed using a continuous macroscopic tissue model based on viscous liquid theory of tissue dynamics. Finally, the number and spatial configuration of cells are used to activate tissue signals, which in turn were input into the Boolean model (Fig. 2). This combination of discrete and continuous modeling of several steps of amelogenesis will be used to analyze key cellular events (such as ameloblast extension of differentiation) and to predict the most important regulatory networks necessary for enamel formation. This multi-scale modeling approach provides a powerful tool for addressing questions of how cells interact with each other and their environment and how these interactions affect in their turn gene expression.

Figure 2.

A multi-scale Boolean model is being designed to predict the complex interactions between circadian controls and stage-specific regulators such as Runx2 and Dlx3 that control ameloblast gene expression and ultimately orchestrate ameloblast differentiation and enamel formation. Schematic view of the multi-scale nature of our model composed of four different levels. At the genetic level we integrate the main genes involved in the regulation of amelogenesis within a Boolean network and that results in regulatory signals that control differentiation. The response to these signals occurs at the cellular level, determining whether each cell progress through differentiation or dies. Given this information, at the macroscopic model the new spatial distribution of the cells is computed at the tissue level. The number and spatial configuration of cells determine the activation of the regulatory signals, which in turn input to the genetic level. Clock genes induce daily oscillations of key transcription factors, which, in the model, activate stage-specific ameloblast genes at the genetic level.

Discussion

Our lab focuses on the study of gene expression during ameloblast differentiation. Enamel formation depends largely on a complex temporo-spatial expression of adhesion molecules and growth and transcription factors, described in early tooth development (24) that continues during cell differentiation and enamel formation. During the first stage of amelogenesis, secretory ameloblasts delineate the enamel space, and Amel, Enam and Ambn proteins are secreted and assembled to form an extracellular framework (25). Ameloblasts then transport calcium and phosphate ions into this framework, forming hydroxyapatite crystallites (4). During enamel formation, the organic materials in the matrix are degraded by two proteases, Mmp20 and Klk4, leaving behind a fluid filled porous tissue where secondary crystal growth and mineral accretion can occur to go on to produce the final mature enamel. Defects in the formation of enamel are seen in patients with Amelogenesis Imperfecta (AI) and mutations in AMELX, ENAM, KLK4, MMP20) have been implicated in the etiology of AI (26). Of all these genes expression of enamelin is exclusive to secretory ameloblasts (27). Amel and Mmp20 are expressed in secretory ameloblasts as well as odontoblasts (5), while Klk4 is mainly expressed in the maturation stage ameloblasts (28).

The aim of this study was to test the hypothesis that Runx2 and Dlx3 are involved in ameloblast stage-specific gene regulation. We found that Runx2 down regulates Enam and up-regulates Klk4. This is consistent with Enam and Klk4 developmental expression patterns. Runx2 expression is initiated at the end of the secretory stage when enamelin expression is suppressed. Runx2 expression continues during the maturation stage when Klk4 is exclusively expressed. Therefore, we propose that Runx2 is a key regulator of ameloblast differentiation with a role to suppress genes expressed in the secretory stage such as amelogenin and enamelin and to up-regulate genes of the maturation stage such as Klk4. Further studies in vivo are needed to confirm these preliminary indications and to identify Runx2 partners that may be involved in the down- or up-regulation of ameloblast genes studied here.

Dlx3, is another major player of amelogenesis. Dlx3 is strongly expressed by ameloblasts (16) and Dlx3 mutations result in amelogenesis imperfecta (17). In this study, we showed that Dlx3 up-regulates Amel, Enam and Klk4 expression. These findings are consistent with suggested roles of Dlx3 during both the secretory and the maturation stages of amelogenesis. Our data is also in accordance with previous studies that reported amelogenin regulation by Dlx2 based on gel shift assays and promoter Dlx2 binding sites predictions (29). It may be possible that synergistic and/or competitive relationships between DLX proteins take place during amelogenesis. More studies are needed to clarify the precise roles of the Dlx family of transcription factors in amelogenesis. Nevertheless, our data support a key role of Dlx3 in ameloblast differentiation.

Circadian rhythms are self-sustained endogenous oscillations occurring over a 24 h period. They correspond to the environmental light-dark cycles of an organism but persist even after the light-dark stimulus is removed. These biological rhythms are involved with most physiological processes. Though there is a site in the suprachiasmatic nucleus of the brain that is considered the “master clock”, peripheral clocks have been found in several tissues in the body. The relationship between these two kinds of circadian biological clocks is not very clear yet (30). Several genes have been identified as core maintainers of the circadian rhythm. The main mammalian genes include Clock, Bmal1, Per1, Per2, Per3, Cry1 and Cry2. The genes Nr1d1, Nr1d2, Rorα, and Dbp also play a key role modifying the expression of the main clock genes (31). Transcription of these “clock genes” oscillates over a 24 h period, and their output signals induce rhythms of target gene expression that create patterns in physiological processes. Inducing a rhythm involves a clock gene transcription factor binding to the promoter region of a clock-controlled gene (32).

We have recently showed that clock genes and clock proteins are expressed during ameloblast differentiation (20, 33). This present study further supports the concept that clock genes are expressed in ameloblasts and that their expression is oscillating in regular 24 h intervals. Furthermore, we showed evidence that clock genes regulate several ameloblast stage-specific genes supporting the idea that clock genes are key regulators of ameloblast differentiation. These data are consistent with our previous discoveries reporting that the amounts of secreted amelogenin vary during different daily intervals (20). In addition, we showed that over-expression of Nr1d1 results in up-regulation of Runx2 a key transcription factor strongly expressed in maturation ameloblasts. Our data also showed that Runx2 regulates the expression of enamelin and Klk4. We therefore, hypothesize that clock genes may regulate the daily variations of gene expression in ameloblasts by directly regulating their transcriptional rates or indirectly by regulating the expression of key transcription factors (Runx2 in our case) that regulate ameloblast gene expression. More studies are needed to understand the precise roles of clock genes in enamel formation. Nevertheless, we suggest that in addition to the stage-specific controls, amelogenesis is subject to very precise rhythmical daily controls of gene expression levels and cell activity (Fig. 3).

Figure 3.

This cartoon summarizes our findings reported here. Ameloblast gene regulation is orchestrated by stage-specific controls as we showed for amelogenin, enamelin, and Klk4 regulation by Runx2 and Dlx3. Ameloblast gene expression is also subject to circadian controls as we showed for amelogenin, Mmp20, Klk4, and Runx2 regulation by Nr1d1. Accordingly, we postulate that circadian control of ameloblast-genes can be direct but also indirect (e.g. through Runx2). At the moment is unknown if Dlx3 also is subject to circadian regulation. It is also unknown if Runx2 and/or Dlx3 can also regulate the expression of clock genes in ameloblasts in a negative feedback loop. Another complexity is that clock genes regulate each other resulting in complex network interactions. We are currently analyzing these networks and their effects on ameloblast differentiation and enamel formation using mathematical modeling.

In conclusion, our study offers novel insights on the role of clock genes in ameloblast differentiation and explores the potential links between circadian control and stage-specific regulation of ameloblast-genes. Our hypothesis that an ameloblast peripheral clock regulates enamel formation orchestrating the expression of ameloblast-specific genes is further strengthened. It remains to show direct links between clock genes expression alterations and dental diseases using in vivo models. Nevertheless this initial study using an ameloblast cell line lays the foundation for more research in the chronobiology of tooth development and diseases.