Background

Medaka (Oryzias latipes) is a small, egg-laying freshwater fish [1] that has a smaller genome size than zebrafish [2]. Medaka have oral and pharyngeal dentition, and adult fish have more than a thousand teeth [3]. The pharyngeal teeth of medaka are a conical in shape and are connected to the pharyngeal bone via a tooth-supporting bone called the pedicle or attachment bone [4, 5]. The functional teeth are attached to the pedicle either by ankylosis in the lower teeth or by fibrous connections in the upper teeth [6]. A previous study identified tooth families and tooth-forming units in medaka and determined the tooth replacement cycles and the localization of odontogenic stem cell niches in the pharyngeal dentition, providing insights into these mechanisms and clearly indicating that continuous tooth replacement is maintained [4]. Furthermore, tooth replacement in medaka occurs due to the turnover of pedicles by osteoblasts and osteoclasts [5].

The secretory calcium-binding phosphoprotein (scpp) gene cluster functions in mineralized tissues such as dental and bone tissues, and the genes encoding multiple SCPPs are reportedly conserved in fishes, amphibians, reptiles, and mammals [7,8,9]. SCPPs can be classified into two subclasses: acidic SCPPs and Pro/Gln (P/Q)-rich SCPPs [8]. Acidic SCPPs function in bone, dentin, and cementum formation, whereas P/Q-rich SCPPs are involved in enamel and enameloid formation [8,9,10]. Since enameloid is derived partly or mainly from mesenchymal cells, it differs from enamel formation. In enameloid, odontoblasts provide an organic matrix containing collagen that is then dissolved and degenerated by epithelial cells, which supply inorganic ions during advanced crystal growth [11]. SPP1, odontogenic ameloblast-associated protein (ODAM), and their common ancestor, secreted protein, acidic, cysteine-rich like 1 (SPARCL1), are conserved in jawed vertebrates [12]. In teleost fishes such as fugu and zebrafish, sparcl1 and other scpp genes (except for spp1) constitute the sparcl1-scpp gene cluster [12]. The sparcl1-scpp gene cluster of zebrafish and fugu is located on chromosome 1 and chromosome 17, respectively, whereas in medaka, it is also located on chromosome 1. Other members of the scpp gene cluster are likely more specific between lineages, indicating that they arose with lineage-specific duplications and deletions, which possibly resulted in the specialization of certain scpp family genes in vertebrates [12, 13]. A recent study showed that the formation of ganoid scales in ray-finned fish involves the largest known repertoire of scpp genes, which includes many genes previously unknown in teleosts [13, 14]. In addition, that study confirmed that scpp6 and fa93e10 are teleost orthologues of ameloblastin (ambn) and enamelin (enam), respectively [13].

Previous studies characterized the expression patterns of scpp genes during the tooth germ stage of fugu [7], zebrafish [8, 15], and cichlid [16]. In the tooth germ stage of fugu, a previous study reported that the expression patterns of scpp genes, scpp2, scpp4, and scpp5 in the ide, scpp1, and scpp5 in odontoblasts, and scpp3a and scpp3b in the ide [7]. The expression patterns of scpp1, scpp2, and scpp5 in zebrafish were similar to those of fugu. In addition, scpp9 was strongly expressed in the ide, and spp1 was strongly expressed in the cells of the pedicle and jawbone [8]. In zebrafish, expression of ambn and enam was detected in ide cells [15]. Scpp2 and scpp5 in the cichlid were also similar to the expression patterns of fugu and zebrafish [16]. Recently, it was reported that scpp5 in the pharyngeal tissues of zebrafish was expressed in the odontoblasts, but not pedicles (dentinous bone) [17]. In the present study, we examined the expression patterns of several scpp genes in the pharyngeal teeth of medaka to elucidate their roles during tooth formation and replacement.

Methods

Medaka

Himedaka (Japanese medaka, Oryzias latipes) of both sexes (n = 22, body length: 28 to 33 mm) were used in this study. The fish were maintained in the laboratory at a water temperature of 23–25 °C under a 14-h/10-h light/dark cycle. All animal experiments were performed in accordance with policies and protocols approved by the guidelines of the Animal Welfare Committee of Tokushima University. In addition, we confirmed that all of the methods are in accordance with ARRIVE guidelines 2.0 for reporting of animal experiments.

Cloning of scpp genes in medaka

Genome Data Viewer (GDV) (https://www.ncbi.nlm.nih.gov/genome/gdv/) [18] and Ensembl Genome Browser (EGB) 105 (https://www.ensembl.org/index.html) were used to search gene sequence databases to identify gene orthologues and examine the expression of scpp genes in medaka (ASM223467v1). Genomic synteny analysis was also conducted using GDV and EGB105. Each scpp gene assembly number in medaka was referenced to GDV and EGB105, as listed in availability of data. For scpp4, scpp7, scpp9, and enam, similar sequences were identified in the medaka genome by performing genomic alignment with zebrafish (GRCz11) and fugu (fTakRub1.2) in EGB105. Those sequences were then obtained by searches using BLAST (blastin) (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Genomic locations of the scpp genes were extracted from publicly available annotations in EGB105. Specific primers were designed using Primer3Plus (https://primer3plus.com), and the primer sequences are listed in Table 1. Brain, liver, fin ray bone, and pharyngeal tissues (teeth and bone) were dissected from the medaka, and total RNA was extracted from each tissue using TRI REAGENT (Molecular Research Center Inc., Cincinnati, OH, USA). Total RNA from the fin ray bone was extracted after dissecting the fin ray from the body without removing the epithelium. cDNAs were generated using a 1st Strand cDNA Synthesis kit for RT-PCR® (Roche, Basel, Switzerland).

Table 1 Specific primers for scpp genes and mineralized tissue-related genes used in qPCR and ISH analyses
Full size table

Real-time quantitative reverse transcription-polymerase chain reaction (PCR) (qPCR)

qPCR was performed using PowerUp™ SYBR® Green Master Mix (Applied Biosystems™, Thermo Fisher Scientific k.k., Tokyo, Japan) and a MyGo mini system (IT-IS Life Science Ltd., Dublin, Ireland). The qPCR conditions for amplification were as follows: initial denaturation at 95 °C for 120 s, followed by 40 cycles of 10 s at 95 °C and 30 s at 60 °C. The length of each amplicon was approximately 50–150 base pairs, and the confirmed efficacy of each specific primer pair was 2.0. Gene expression levels were normalized to those of β-actin. The expression levels in each sample were compared to those from the fin ray bone (at level 1). Three different samples were analyzed, and reactions were repeated three times. Relative mRNA expression levels were calculated by the ΔΔCt method. All data were calculated as the mean ± standard deviation (SD). Data were analyzed using one-way analysis of variance for multi-group comparisons, and the significance was determined by Tukey’s comparison test using SPSS Statistics software (version 24, IBM Inc., Chicago, IL, USA). A p-value of < 0.05 or < 0.01 was considered statistically significant.

Histological analysis

The heads of medaka were fixed in a 4% paraformaldehyde solution (0.1 M phosphate buffer, pH 7.4) at 4 °C overnight, decalcified in 8% ethylenediaminetetraacetic acid (EDTA) solution at 4 °C for 3 days, then embedded in paraffin according to conventional methods. The prepared paraffin blocks were cut into sagittal and horizontal sections at a thickness of 5 µm.

Pharyngeal tissues of medaka were fixed in 4% paraformaldehyde solution (0.1 M phosphate buffer, pH 7.4) at 4 °C overnight, and then the samples were embedded in Technovit 8100 (Heraeus Kulzer, Wehrheim, Germany) with a coagulant (Heraeus Kulzer). The resin blocks were cut into horizontal sections at a thickness of 1 µm. Toluidine blue staining and tartrate-resistant acid phosphatase (TRAP) reaction were performed to identify osteoclasts [3].

In situ hybridization (ISH)

ISH was performed as described previously [19]. Briefly, template cDNAs for mRNA probes were generated using gene-specific primer sets encoding T3 or T7 RNA polymerase promoter sequences. The primer sequences are listed in Table 1. The RT-PCR products for ISH were generated using Ampdirect® Plus (Shimazu, Japan) with a Mastercycler® nexus gradient (Eppendorf, Japan). The PCR conditions were as follows: initial denaturation at 94 °C for 120 s, followed by 35 cycles of 30 s at 94 °C, 30 s at 60 °C, and 60 s at 72 °C. Digoxigenin (DIG)-labeled, single-stranded antisense and sense strands were then generated from the PCR products using a DIG RNA Labeling kit® (Roche). After de-waxing, histological sections were treated in the following solutions: Proteinase K (1.0 µg/ml, Sigma) in 10 mM Tris–HCl buffer (pH 8.0) containing 1.0 mM EDTA at 37 °C for 10 min; 0.2 M HCl at 25 °C for 10 min; and 0.1 M triethanolamine containing 0.5% acetic anhydride at 25 °C for 15 min for acetylation. Hybridization was performed by incubating the samples with diluted RNA probes (1.0 µg/ml) at 50–60 °C overnight. The sections were washed according to conventional protocols. Samples were visualized using an anti-DIG antibody as follows: sections were incubated with 1.5% blocking buffer at 25 °C for 1 h and then incubated with alkaline phosphatase–labeled anti-DIG antibody (Roche) at 25 °C for 2 h. Finally, the sections were reacted with BCIP/NBT (Roche) solutions and developed at 25 °C overnight. The sections were counterstained with methyl green.

Results

Synteny analysis of scpp in medaka

The GDV was used to search for orthologues of medaka scpp1, scpp2, scpp5, and sparc in zebrafish (GRCz11) and fugu (fTakRub1.2). For scpp4, scpp7, scpp9, and enam, similar sequences were identified in the medaka genome by performing genomic alignment with zebrafish and fugu using EGB105. The resulting sequences were then searched for using BLAST (blastin). Genomic locations of the scpp genes were extracted from publicly available annotations in EGB105. The location and transcriptional orientation of the scpp genes are schematically illustrated in Fig. 1. Of the various scpp genes, expression of scpp3 or ambn orthologues was not observed in the present study. However, genomic alignment analysis showed that fugu scpp3a and scpp3b are similar to a non-coding RNA in medaka. Fugu scpp3a was similar to uncharacterized LOC101161295 ncRNA or uncharacterized LOC110015140 ncRNA, whereas fugu scpp3b was similar to uncharacterized LOC101160534 ncRNA or uncharacterized LOC101161034 ncRNA. As these are non-coding RNAs, they were not considered in the present study. Fugu scpp3c did not show sequence similarity to that gene in medaka.

Fig. 1
figure 1

Synteny analysis of secretory calcium-binding phosphoprotein (scpp) in medaka, fugu, and zebrafish. Schematic illustration of the location and transcriptional orientation of the sparcl1 (yellow box), Pro/Gln-rich scpp (orange box), and acidic scpp (blue box) genes. Teleost orthologues (gray vertical bars) obtained from GDV and EGB105 are shown (orthologous were). In addition, further putative orthologues supported by low sequence similarities (?) connect scpp4, scpp7, scpp9, and enam, respectively. ambn: ameloblastin; enam: enamelin; Chr: chromosome

Full size image

Expression levels of scpp and mineralized tissue-related genes in medaka

The expression levels of transcripts for scpp1, scpp9, sparc, and spp1 were analyzed in the pharyngeal tissues, fin ray bone, brain, and liver (Fig. 2a). The expression levels of these genes were significantly higher in the pharyngeal tissues than in other tissues (Fig. 2a). Transcripts for scpp2 were detected only in the pharyngeal tissues (Fig. 2c). Transcripts for scpp4, scpp5, and enam were detected in mineralized tissues, but not in soft tissues, and the expression levels of the scpp4 and scpp5 transcripts were significantly higher in pharyngeal tissues than in fin ray bones (Fig. 2a). However, enam expression levels did not differ significantly between the mineralized tissues, such as bone and teeth (Fig. 2a). Transcripts for scpp7 were detected in each tissue, and the expression levels were significantly higher in fin ray bone than in other tissues (Fig. 2a). Levels of transcripts for collagen 1a1 (col1a1) and bone gla protein (bglap) were higher in mineralized tissues than in other tissues (Fig. 2b). The levels of transcripts for runt-related transcription factor 2 (runx2) were higher in mineralized tissues than in soft tissues (Fig. 2b). Transcripts for sp7 were detected only in mineralized tissues, and no significant difference in expression level was found between the pharyngeal tissues and fin ray bone (Fig. 2b). In addition, there was no significant difference between the expression levels of transcripts for runx2 and sp7 in the tissues examined.

Fig. 2
figure 2

Results of qPCR indicating expression levels of scpps and mineralized tissue-related genes in medaka. The qPCR results are obtained from an experiment in which the expression levels of target genes and an endogenous control (β-actin) are evaluated. The expression levels in each sample were compared to those in fin ray bone. a Transcripts for scpp1, scpp9, sparc, and spp1 were detected in the pharyngeal tissues, fin ray bone, brain, and liver. Transcripts for scpp4, scpp5, and enam were detected in the mineralized tissues but not in the soft tissues, and the expression levels of scpp4 and scpp5 were significantly higher in the pharyngeal tissues than in fin ray bone. Transcripts for scpp7 were detected in each tissue, and the expression levels were significantly higher in fin ray bone than in other tissues. b Transcripts for collagen 1a1 (col1a1) and bone gla protein (bglap) were detected at higher levels in mineralized tissues than in other tissues. Transcripts for runt-related transcription factor 2 (runx2) were found to be expressed at higher levels in mineralized tissues than in soft tissues. Transcripts for sp7 were only detected in mineralized tissues, and no significant difference in expression level was found between the pharyngeal tissues and fin ray bone. c Results of PCR analysis of the expression of scpp2 in medaka. Gene expression levels were normalized to those of β-actin. Data were presented as the mean ± SD (n = 3). *p < 0.05, **p < 0.01

Full size image

Expression patterns of scpp and mineralized tissue-related genes during tooth germ stages

In the enameloid organ of the tooth germ, the ide and outer dental epithelium (ode) are clearly distinguishable (Fig. 3a, b). Cap enameloid (acrodin) is shown as the dotted yellow area after decalcification, and dentin was observed adjacent to the cap enameloid. In addition, odontoblasts were observed adjacent to dentin, and dental papilla under the odontoblasts (Fig. 3a, b). Erupted functional teeth were connected to pedicles by collagenous fibers (Fig. 3c, d). In horizontal sections at the height of the pedicles under the erupted teeth (Fig. 3e), two pedicles and two tooth germs were typically observed in a row (Fig. 3 f, g, h, i). In this study, we classified these structures as the first pedicle, second pedicle, third tooth germ, and fourth tooth germ based on their developmental stages (Fig. 3h). Numerous TRAP-positive cells were observed in the anterior side of the first pedicles but not in the second pedicles (Fig. 3h).

Fig. 3
figure 3

Histological image of tooth families in the pharyngeal tissues. a, c Hematoxylin–eosin staining of upper tooth germ and the functional tooth in the pharyngeal tissues. b, d Schematic diagram of tooth germ and the functional tooth. e, f Alizarin red staining of the upper left of pharyngeal tissue. g Toluidine blue staining of the pharyngeal tissue was cut at the level of the pedicles under the erupted functional teeth for Fig. 3e (yellow dotted line). h High-magnification views of the boxed areas in (g), showing two pedicles and two tooth germs in a row. These structures were identified as the first pedicle (1), second pedicle (2), third tooth germ (3), and fourth tooth germ (4) based on their developmental stages. i TRAP staining of the horizontal section of the pharyngeal teeth. E: enameloid; D: dentin; ode: outer dental epithelium; ide: inner dental epithelium; dp: dental papilla; ped: pedicle. *: epithelial cells adjacent to the surface of the tooth shaft. A: anterior; P: posterior. Scale bars: 20 µm (a, c), 200 µm (e, f), 100 µm (g), and 50 µm (h, i)

Full size image

When tooth germ stages were examined, transcripts for scpp1 were detected in odontoblasts, but not in the ide and dental papilla (Fig. 4a). Transcripts for scpp2, scpp4, scpp9, and enam were detected only in the ide adjacent to enameloid, but not in ide adjacent to the cervical loop (Fig. 4b, c, f, g). Transcripts for scpp5 and sparc were detected in both the ide adjacent to the cervical loop and in odontoblasts during dentin formation, but no transcripts were detected in either the ode or dental papilla (Fig. 4d). Transcripts for scpp7 were weakly detected in the ide adjacent to enameloid (Fig. 4e). No spp1 or bglap transcripts were detected in the tooth germ (Fig. 4i, k). Transcripts for col1a1 were weakly detected in the ide adjacent to enameloid and strongly expressed in odontoblasts (Fig. 4j). Transcripts for runx2 were detected in the ide adjacent to enameloid but not in ide adjacent to the cervical loop or dental papilla (Fig. 4l). Transcripts for sp7 were weakly detected in the ide adjacent to enameloid and dental papilla, but not in odontoblasts (Fig. 4m).

Fig. 4
figure 4

In situ hybridization (ISH) analysis of sagittal sections showing the expression of scpps and mineralized tissue-related genes in the upper tooth germs of pharyngeal tissues. Broken line indicates the area of the enameloid organ. Transcripts for scpp1, -5, sparc, and col1a1 were detected in odontoblasts (orange). Transcripts for scpp2, -4, -7, -9, enam, col1a1, runx2, and sp7 were detected only in the ide adjacent to the enameloid (yellow). Transcripts for scpp5 and sparc were detected in ide adjacent to the cervical loop (green). No spp1 or bglap transcripts were detected in the tooth germ. Transcripts for sp7 were weakly detected in the dental papilla (white). Scale bars: 20 µm

Full size image

Expression patterns of scpp and mineralized tissue-related genes in functional teeth

When functional teeth were examined, no scpp1, scpp4, scpp7, or scpp9 transcripts were detected in the cells of pharyngeal teeth and pedicles (Fig. 5a, c, e, f). Transcripts for scpp2 were detected in the epithelial cells adjacent to the surface of the tooth shaft (Fig. 5b). Transcripts for scpp5 were detected within the odontoblasts above the border between the teeth and pedicles, but not in the cells of the pedicles (Fig. 5d). Transcripts for enam were detected in the mesenchymal cells of pedicles, but not in odontoblasts of the pharyngeal teeth (Fig. 5g). In horizontal sections, enam was more strongly expressed in the mesenchymal cells of the second pedicles than in those of the first pedicles (Fig. 6g). Transcripts for sparc were detected in the odontoblasts and mesenchymal cells of pedicles (Fig. 5h). Sparc was not detected in the mesenchymal cells around the first pedicles; however, it was expressed in mesenchymal cells around the second pedicles (Fig. 6h). Transcripts for spp1 were weakly detected in the pulp area of the pharyngeal teeth near the pedicles and in the surrounding mesenchymal cells of the pedicles (Figs. 5i, 6i). Transcripts for col1a1 were detected in the mesenchymal cells located on posterior side of the pedicles but not in cells within the pedicles or pharyngeal teeth (Fig. 5j). In horizontal sections, col1a1 transcripts were detected in mesenchymal cells on the posterior side of the second pedicles but not in the first pedicles (Figs. 3g, 6j). No bglap transcripts were detected in the cells of pharyngeal teeth and pedicles, but transcripts were detected in osteoblasts of pharyngeal bone (Fig. 5k). Transcripts for runx2 were detected in mesenchymal cells of the pedicles and weakly detected in the cells surrounding the pedicles, but they were not detected in pharyngeal teeth (Figs. 5l, 6k). Transcripts for sp7 were detected in the cells on the posterior side of the second pedicles, but not in the mesenchymal cells of the pharyngeal teeth and pedicles (Figs. 5m, 6l).

Fig. 5
figure 5

Results of ISH analysis show the expression of scpps and mineralized tissue-related genes in the upper functional teeth of pharyngeal tissues. Broken gray line indicates the area of dentin in the pharyngeal teeth. Black line indicates the region of the pedicles. No transcripts for scpp1, -4, -7, and -9 were detected in the cells of pharyngeal teeth and pedicles. Transcripts for scpp2 were detected in the epithelial cells adjacent to the surface of the tooth shaft (yellow). Transcripts for scpp5, sparc, and spp1 were detected within the odontoblasts (orange). Transcripts for enam, sparc, and spp1 were detected within the mesenchymal cells of pedicles (white). Transcripts for col1a1 and sp7 were detected in the mesenchymal cells located on the posterior side of the pedicles (blue). Transcripts for bglap were detected in osteoblasts of pharyngeal bone (green). Transcripts for runx2 were detected in mesenchymal cells of the pedicles and weakly detected in the cells surrounding the pedicles (white). A: anterior; P: posterior. Scale bars: 50 µm

Full size image
Fig. 6
figure 6

ISH analysis of horizontal sections showing the expression of scpps and mineralized tissue-related genes. Transcripts for scpp1, -5, sparc, col1a1, and sp7 were detected in odontoblasts (orange). Transcripts for scpp2, -4, -5, -7, -9, enam, sparc, col1a1, runx2, and sp7 were detected in the ide (yellow). Transcripts for enam, sparc, spp1, and runx2 were detected in the mesenchymal cells of pedicles (white). Transcripts for enam, spp1, and runx2 were detected in the mesenchymal cells located on the anterior side of the pedicles (dark blue). Transcripts for enam, sparc, spp1, col1a1, and sp7 were detected in the mesenchymal cells located on the posterior side of the pedicles (blue). A: anterior; P: posterior. Scale bars: 50 µm

Full size image

Figure 7 shows a schematic representation from the tooth germ stage to the functional teeth of medaka based on ISH data. Figure 7a was prepared with reference to the sections shown in Figs. 3h and 6. It shows two pedicles and two successional tooth germs in a row. The cross-section of same pedicles and tooth germ are shown in Fig. 7b. The areas of scpp and mineralized tissue-related gene expression were classified into epithelial and mesenchymal components (Fig. 7c, d). The specific expression patterns of the scpp genes were in each stage (Fig. 7c). In summary, scpp1, scpp4, scpp7, and scpp9 were expressed in tooth germ, but not in the cells of teeth and pedicle (Fig. 7c), while transcripts of spp1 exhibited the opposite expression pattern (Fig. 7c). The expression of enam was detected in the ide of tooth germ and the cells of the second pedicles (Fig. 7c). In addition, scpp2 was consistently expressed in the epithelium (Fig. 7c). Other genes of the epithelial components that were detected in the tooth germ were not detected in the cells of functional teeth and pedicles (Fig. 7c). Scpp5 expression was only detected in the cells that formed enameloid and dentin (Fig. 7c).

Fig. 7
figure 7

Summary of scpp and mineralized tissue-related genes expression patterns in the pharyngeal tissues. a, b Schematic representation of a section at the level of the pedicles under the erupted functional teeth. The schematic representation shows two pedicles and two successional tooth germs in a row. The cross-section was cut so that the pedicle and the tooth germ are continuous (orange dotted line), and the cross-section shown in the schematic is obtained on the right. c, d Schematic representation of expression patterns of scpp and mineralized tissue-related genes. Expression in the epithelium is shown in orange, and expression in the mesenchyme is shown in blue. The specific expression patterns of the scpp genes were in each stage (Fig. 7c). Scpp1, scpp4, scpp7, and scpp9 were expressed in tooth germ, but not in the cells of teeth and pedicle (Fig. 7c), while transcripts of spp1 exhibited the opposite expression pattern (Fig. 7c). The expression of enam was detected in the ide of tooth germ and the cells of the second pedicles (Fig. 7c). In addition, scpp2 was consistently expressed in the epithelium (Fig. 7c). Other genes of the epithelial components that were detected in the tooth germ were not detected in the cells of functional teeth and pedicles (Fig. 7c). Scpp5 expression was only detected in the cells that formed enameloid and dentin (Fig. 7c)

Full size image

Discussion

To investigate the genetic basis of the formation of tooth and periodontal tissues in medaka, we examined the expression of several scpp genes encoded in the medaka genome. The present study is the first to describe the expression patterns of scpp genes during pharyngeal tooth formation in medaka.

The role of scpp genes in dentin formation

The expression levels of scpp1 and scpp5 in medaka were higher in pharyngeal tissues (Fig. 2a). Expression of scpp1 was detected only in odontoblasts in tooth germ, and scpp5 was expressed in both ide and odontoblasts in tooth germ (Fig. 4a, d). These tooth germ expression patterns are similar to those of fugu, zebrafish, and cichlids [7, 8, 15, 16]. Interestingly, expression of scpp5 was detected only in odontoblasts of functional teeth, and no other scpp gene was expressed in these cells (Fig. 5d). In addition, seahorses, and pipefish (family Syngnathidae) are toothless, a phenomenon known as edentulism, and in seahorses, scpp5 is a pseudogene [20]. Furthermore, the pharyngeal teeth in scpp5-deficient zebrafish exhibit a distinct phenotype characterized by fewer functional teeth [21]. In addition, the expression patterns of scpp1 and scpp5 were similar to those of the dentin sialo phosphoprotein (Dspp) gene in rats and mice [22, 23]. In the rat, the Dspp gene is expressed in odontoblasts at all stages and transiently in preameloblasts during tooth development [22]. Although the genomic position of scpp1 in zebrafish is similar to that of Dspp and dentin matrix protein 1 (Dmp1), the amino acid sequence of scpp1 differs from that of Dspp and Dmp1 [8]. Furthermore, no Dspp or Dmp1 orthologues have been found in teleosts fish [8]. However, Mikami et al. reported that Dspp or Dmp1 orthologues have been found in ray-finned fish [14]. These results suggest that both scpp1 and scpp5 in medaka play a critical role in dentin formation.

The role of scpp genes in enameloid formation

Among the scpp genes in medaka, the expression of scpp2, -4, -5, -7, -9, enam, and sparc was detected in the ide (Figs. 4b, c, d, e, f, and 7). These expression patterns were consistent with those in fugu, zebrafish, and cichlid [7, 8, 15, 16]. Kawasaki et al. reported that scpp2, -4, -5, and -9 play important roles during enameloid formation in fish [7, 8, 15]. The genes scpp4, -5, -7, and -9 are fish specific. The scpp2 gene encodes ODAM and is an orthologue of a gene found in mice. In mice, ODAM is expressed in the inner enamel epithelium and junctional epithelium [24, 25] where it modulates the mineralization of enamel via the regulation of matrix metalloproteinase-20 in conjunction with Runx2 [26]. The above results suggest that the expression of scpp genes is also involved in the formation of highly mineralized tissues such as cap enameloid (acrodin).

In fish, enameloid is usually found at the tooth cap, and it is called cap enameloid (acrodin) [11]. Moreover, in Polypterus senegalus, collar enamel forms at the surface of the tooth shaft [11], while only cap enameloid (acrodin) is observed in medaka. Kawasaki et al. reported that scpp2 and scpp9 are involved in the hyper-mineralization of collar enameloid in addition to cap enameloid (acrodin) [8]. A recent study reported that scpp5, ameloblastin, and enam are expressed in the ide cells adjacent to the surface of the tooth shaft in gar and zebrafish [15]. Interestingly, in the present study on medaka, no expression of scpp5, scpp9, or enam was detected in the cells adjacent to the surface of the tooth shaft; only scpp2 was detected in those cells (Figs. 4b, 7). Thus, the expression of scpp2 in medaka appears to be related to another role for ODAM in mediating active adhesion mechanisms at the junctional epithelium [27]. These data suggest that expression of scpp2 adjacent to the surface of the tooth shaft in medaka contributes to active adhesion to the epithelial tissue adjacent to the tooth shaft rather than playing a role in hyper-mineralization.

Other scpp genes

In channel catfish, scpp7 is highly expressed in the skin [28]. In the present study, total RNA was extracted from fin ray bones without removing the epithelium, and analyses revealed that scpp7 was strongly expressed in fin ray bone (Fig. 2a). In ISH, transcripts for scpp7 were detected in the epithelial cells adjacent to the fin ray bone, but not in osteoblasts (Additional file 1). In addition, scpp5, scpp7, spp1, and sparc are expressed over time during scale regeneration in zebrafish [29]. These results suggest that scpp7 expression might be critical during skin and scale formation and regeneration.

Spp1 encodes osteopontin, a sialic acid–rich, phosphorylated, integrin-binding extracellular matrix glycoprotein that enhances osteoclast activity [30, 31]. The inhibitory effect of osteopontin on calcification depends on the degree of phosphorylation of serine [30]. In medaka, pharyngeal teeth replacement is accompanied by the remodeling of pedicles, which occurs under the first functional teeth [5]. Such dynamic tooth replacement is associated with the dynamic resorption of pedicles by the large population of TRAP-positive osteoclasts [6]. Bone matrix–resorbing osteoclasts are anchored by osteopontin bound both to the mineral of the bone matrix and to a vitronectin receptor on the osteoclast plasma membrane [32]. In addition, osteopontin is involved in cell attachment during cementogenesis in mice [10]. In the present study, strong spp1 expression was detected in mesenchymal cells on the anterior side of pedicles (Fig. 5i). TRAP-positive cells were also observed in the anterior side of the first pedicles (Fig. 2h). In contrast, spp1 expression was detected in mesenchymal cells on the posterior side of pedicles (Fig. 5i). These results suggest that the expression pattern of spp1 reflects the process of tooth replacement.

Sparc, also known as osteonectin or basement membrane protein 40, is one of the most abundant non-collagenous proteins in bone [33]. Sparc is expressed in both mineralized and non-mineralized tissues [33]. Osteonectin-null mice show reduced bone formation and decreased osteoblast and osteoclast surface area and number, which leads to a decrease in bone remodeling with a negative bone balance and profound osteopenia [34]. In the present study, the expression pattern of sparc in the tooth germ was similar to that in fugu, zebrafish, and cichlids [7, 8, 35]. In catsharks, sparc is expressed in odontoblasts, and in thornback rays, it is weakly expressed in the mesenchyme cells of tooth buds [36]. These results suggest that expression of other scpp genes occurs during tooth development in teleosts and chondrichthyans. In the functional teeth of medaka, sparc was expressed continuously in the mesenchymal cells during formation of the pharyngeal teeth and pedicles (Fig. 4h). The expression pattern of sparc during pharyngeal teeth formation in this study was consistent with that in mice, suggesting that sparc has a similar function in teleosts and mammals.

The expression patterns of mineralized tissue-related genes change during the tooth replacement process

The findings of our previous study showed that individual functional teeth and their successional teeth are organized into families [4]. The pharyngeal teeth of medaka are replaced with the next group of teeth within 4 weeks [4]. Tooth replacement in medaka involves pedicle remodeling by osteoblasts and osteoclasts [5]. In the present study, sparc and col1a1, which are osteoblast markers, were expressed in the posterior side of the first pedicles but not in the anterior side (Fig. 6h, j), and sp7 expression was observed in the posterior side of the second pedicles (Fig. 6l). In addition, while enam was expressed in the second pedicle during the process of bone formation, no enam expression was observed in the already-formed first pedicles (Figs. 5g, 6g). In ST2 cells, exogenous ameloblastin and enam suppressed RANKL expression, which attenuated multinucleated osteoclast formation [37]. In contrast, numerous TRAP-positive cells and spp1 expression were detected in the cells on anterior side of the first pedicles (Fig. 3h). Considered collectively, these expression patterns are consistent with the cell dynamics of pedicle remodeling during tooth replacement.

Is the pedicle a bone-like tissue or a dentin-like tissue?

Two major points must be considered to answer this question. The first point is that scpp genes are specifically expressed in the cells that form dentin and pedicles. In this study, the expression of only scpp1 and scpp5 was detected in odontoblasts. Scpp5 transcripts were expressed in odontoblasts during the stages from tooth germ formation to the formation of functional teeth. Previous studies have shown that scpp5 exhibits a similar expression pattern in fugu, zebrafish, and gar [7, 8, 15]. A recent study reported that the pharyngeal teeth in scpp5-deficient zebrafish exhibit a distinct phenotype, particularly a decreased number of functional teeth [20], Rosa et al. reported that scpp5 in zebrafish was expressed in the odontoblasts during the tooth germ stage, but not in the cells depositing the “bone of attachment” [17]. The expression patterns of scpp5 in zebrafish were similar to those in medaka, implying that, scpp5 plays a role in dentin formation in teleosts fish. In addition, the findings showed that scpp1, and scpp5 were not expressed in the cells of pedicles, whereas spp1, enam, runx2, and sp7 were expressed. Interestingly, these factors cause morphological abnormalities in the tooth roots of mammals. In mice, Spp1 is strongly expressed by osteoblasts in alveolar bone but rarely expressed in odontoblasts [38]. In enam-deficient mice, excessive absorption of dentin and cementum of the tooth root was observed [39]. In runx2-overexpressing mice, the tooth roots form osteodentin-like structures [40]. In addition, the jaw teeth of sp7-mutant medaka and zebrafish fail to form pedicles [41, 42]. Micro-computed tomography analysis revealed that Osterix conditional knock-out mice have short roots and a thin dentin matrix [43]. Alterations, either through deletion or overexpression of these genes, results in abnormal tooth root morphology in mammals. These abnormalities were specifically observed in the cells responsible for forming the pedicles. Notably, similar root and pedicles phenotypes associated with sp7 have been observed in both mice and teleosts fish.

The other point that must be considered is that the epithelial elongation of the enameloid organs terminates at a certain level during tooth formation in teleost fish such as medaka. Tooth development is controlled by interactions between the epithelium and mesenchyme. In tooth development in mammals, the cervical loop grows after the crown formation stage, becoming Hertwig’s epithelial root sheath. However, the extent to which Hertwig’s epithelial root sheath descends toward the root apex varies among species [44]. The teeth of medaka are also formed by interactions between the epithelium and mesenchyme. However, no epithelial tissue adjacent to the pedicle is observed during pedicle formation.

Rosa et al. reported that given the presence of dentinal tubule-like cell extensions and the near absence of osteocytes, the characteristics of the pedicles in zebrafish were intermediate between bone and dentin, and that this material was better termed “dentinous bone” [17]. However, the cells responsible for depositing the “bone of attachment” are more closely associated with osteoblasts, as evidenced by the presence of the osteoblast marker Zns-5 and the lack of a covering epithelium, rather than with odontoblasts [17]. While osteocytes have been confirmed in the pedicles of the jaw teeth of Pagrus auratus [45], medaka does not osteocytes its pedicles. This suggests that pedicles of medaka might be similar to those of zebrafish in this context.

In summary, the pedicle cannot be characterized as dentin for the following reasons: 1) the genes expressed in the teeth are not detected in the cells that form pedicles; and 2) no close spatial association with epithelial cells is detected in pedicle-forming cells. Thus, the pedicle is considered to be a bone-like tissue with characteristics that are similar to dentin.

Conclusions

We characterized changes in the expression patterns of scpp genes in medaka during the formation of pharyngeal teeth and pedicles. Among the scpp genes, scpp5 in medaka is considered to be a crucial gene in the tooth formation.