A Systematic Analysis for Localization of Predominant Growth Factors And

Home , FGF4, Inner enamel epithelium, Outer enamel epithelium, PDGFA, Stellate reticulum, Stratum intermedium

Biomedical Research (Tokyo) 36 (3) 205-217, 2015

A systematic analysis for localization of predominant growth factors and their receptors involved in murine tooth germ differentiation using in situ hybridization technique

1 1 2 3 3 Meri HISAMOTO , Marie GOTO , Mami MUTO , Junko NIO-KOBAYASHI , Toshihiko IWANAGA , and Atsuro 1 YOKOYAMA 1 Department of Oral Functional Prosthodontics, Division of Oral Functional Science, Graduate School of Dental Medicine, Hokkaido University, Sapporo, Japan; 2 Department of Orthodontics, Division of Oral Functional Science, Graduate School of Dental Medicine, Hokkaido University, Sapporo, Japan; and 3 Department of Anatomy, Graduate School of Medicine, Hokkaido University, Sapporo, Japan (Received 22 April 2015; and accepted 2 May 2015)

ABSTRACT Tooth development is regulated by various growth factors and their receptors. However, the over- all mechanism of growth factor-mediated odontogenesis remains to be elucidated. The present study examined expression sites and intensities of major growth factors and receptors in the tooth germ of murine fetuses and neonates. Signals of TGF-β and CTGF in fetuses were released from the enamel epithelium, while their neonatal signals arose in odontoblasts. Moreover, BMP/Smad signaling may affect the differentiation of ameloblasts, in contrast to PDGFα whose signals may cause odontoblast differentiation. Growth factors associated with the formation of the periodontium were IGF1, IGF2, IGFBP3, CTGF, and PDGFα. Concerning cusp formation, the enamel knot selectively expressed FGF4, BMP2, and BMP4 with an expression of PDGFα in the enamel-free area. It is concluded that many molecules play critical roles in the epithelium-mesenchyme interaction of tooth germ differentiation, and their expressions are precisely controlled.

Tooth germ differentiation commonly progresses in genesis, partially due to insufficient in vivo system- the following order: the initiation, bud, cap, bell, and atic analyses of the expression of such factors. root formation stages. According to the stages, pro- The transforming growth factor-β (TGF-β) super- genitor constituents of the tooth germ proliferate and family may take a leading role in the regulation of are differentiated to form specialized parts of teeth cell proliferation, differentiation, and apoptosis, also and associated tissues. Tooth development is pre- being one of the major signals for tooth and peri- cisely regulated by cross-talk between adjacent tis- odontium morphogenesis. Immunohistochemical sues and various growth factors secreted in autocrine studies of the mouse tooth germ detected TGF-β1 in and paracrine modes. Recent studies in this research the inner enamel epithelium, ameloblasts, and odon- field have identified many molecules which play im- toblasts; the expression of TGF-β receptor-1 (TGF- portant roles in signaling associated with tooth germ βR1) was weak in the enamel epithelium but differentiation, but it is largely unknown which increased in intensity in differentiated ameloblasts growth factors are the main contributors to odonto- (6). An in situ hybridization study using mouse embryos identified an intense mRNA expression of TGF-β2/3 in the odontoblast layer (20). Among Address correspondence to: Dr. Meri Hisamoto, Depart- bone morphogenetic proteins (BMPs) which belong ment of Oral Functional Prosthodontics, Division of Oral Functional Science, Graduate School of Dental to the TGF-β family, it has been reported that the Medicine, Hokkaido University, North 13, West 7, Kita- expression sites of BMP2 and BMP4 mRNA in the ku, Sapporo 060-8586, Japan mouse tooth germ shift from the enamel knot to Tel: +81-11-706-4270, Fax: +81-11-706-4270 dental papilla and odontoblasts, and their signalings E-mail: [email protected] are required for the epithelium-mesenchyme interac- 206 M. Hisamoto et al. tion during tooth development (1). Another immu- factor) show characteristic expressions in develop- nohistochemical study using the rat tooth germ ing tooth germs. CTGF is a member of the CCN showed that the immunoreactivity of BMP receptor- family involved in cell proliferation, differentiation, 1b (BMPR1b) increased in intensity with the differ- and angiogenesis (2, 8, 12). Shimo et al. demon- entiation of ameloblasts and odontoblasts (15). strated by an in situ hybridization analysis of mice Therefore, BMPs may also act as key regulators in that CTGF and Fisp12 (fibroblast-inducible secreted the differentiation of tooth germs (1, 15). However, protein) are expressed in the dental follicle as well the precise direction and comparative intensity of as inner/outer enamel epithelium (22). Furthermore, signaling by TGF-βs and BMPs have not been fully FGF4 displays a unique expression pattern in the elucidated. enamel knot which regulates the patterning of the Platelet-derived growth factors (PDGFs) mainly tooth crown, including cusp formation, and releases regulate the migration and multiplication of mesen- many signaling molecules such as BMP2/4, Shh, chymal cell lineages. PDGF signaling consists of and Wnt as well as FGF4 (9, 10, 14, 24). The pri- four ligands, PDGF-A to D, and two different recep- mary enamel knot induced at the late bud stage un- tors, PDGF receptor-α (PDGFRα) and PDGFRβ. dergoes apoptosis during the cap stage, and then the PDGFRα binds to PDGF-A/B/C, and PDGFRβ has secondary enamel knot occurs at the future cusps to affinity for PDGF-B/C/D (5, 27). Concerning tooth trigger the folding of the inner enamel epithelium, development, a recent immunohistochemical study leading to cusp formation (14). In relation to cusp using a tooth germ organ culture system reported formation, rodent molars exhibit specific regions at that expressions of PDGFA and PDGFB were in- the tips of cusps, called the enamel-free area (23). tense and extensive in the oral and dental epithelia Although various morphological studies have dealt and underlying mesenchyme, while their receptors with the enamel-free area, it is unknown what mole- (PDGFRα and PDGFRβ) were expressed mainly in cules are specifically expressed there and related to the mesenchyme (27). Moreover, this organ culture signaling for tooth development. Among FGFs, study suggested that PDGFAA, a disulfide-linked FGF2 is detected immunohistochemically in the homodimer of PDGFA, accelerates cusp formation, dental mesenchyme and stellate reticulum of the while PDGFBB induces mesenchyme proliferation mouse enamel organ (3, 21). Thus, we examined the (27). PDGFA and PDGFRα may regulate the size expression profiles of FGF2/4 and CTGF in associa- and stage of tooth development via an autocrine tion with other growth factors in the present study. mechanism (5). However, it is still unclear where Information from systematic analyses of predomi- the PDGF signals arise and are directed. nant growth factors is insufficient to fully understand Insulin-like growth factors (IGFs) are involved in tooth germ differentiation. To comprehend its differ- the differentiation, proliferation, and morphogenesis entiation, we need to elucidate the stage-dependent of various tissues, including odontogenesis (4, 30). expressions of growth factors with respect to interac- The IGF-1 receptor (IGF-1R) and high-affinity IGF- tion with associated factors and signaling directions. binding proteins (IGFBPs) mediate most of the ac- The present study, using an in situ hybridization tions of IGFs. IGFBPs, primarily IGFBP3, inhibit technique, investigated cellular expressions of vari- the bioactivity of IGFs by preventing interaction ous growth factors and their receptors in the fetal with IGF-1R. On the other hand, IGFBPs function and neonatal tooth germs of mice. Our method over- as carrier proteins to direct IGFs from the circula- comes the shortage of specific antibodies and may tion to target tissues and extend their half-lives by provide useful information on growth factors in protecting them from enzymatic degradation (7, 13, tooth development. 16). Morphologically, an in situ hybridization and immunohistochemical study using rat incisors showed METHODS that IGF-1/2 and IGF-1R/2R are expressed in the enamel organ and ameloblasts, suggesting an impor- Animals and tissue sampling. Pregnant ddY mice tant role of the IGF family in the differentiation and were supplied by Japan SLC (Shizuoka, Japan). physiological activity of ameloblasts (30). Another Mice were sacrificed by the intraperitoneal injection immunohistochemical study using human third mo- of an overdose of pentobarbital sodium (Schering- lars reported that IGFBP3 was localized in the den- Plough Animal Health, the Netherlands). Heads of tal follicle (13). E16.5 embryos and day-1 neonates were used in the Among other growth factors, CTGF (connective present study. The tissues were directly embedded tissue growth factor) and FGF (fibroblast growth into a freezing medium (OCT compound; Sakura Fi- Growth factors in tooth germ 207

Table 1 Targeted nucleotide residues of probes used in this study Gene code Accession no. Residues Gene code Accession no. Residues Tgfb1 NM_011577 1182–1226 Smad1 NM_008539 601–645 1839–1883 1561–1605 Tgfb2 NM_009367 1566–1610 Smad2 NM_010754 841–885 2877–2921 1401–1445 Tgfb3 NM_009368 1546–1590 Smad3 AF016189 481–525 2311–2355 1261–1305 Tgfbr1 MM_009370 373–417 Smad5 NM_008541 431–475 1218–1262 1451–1495 Tgfbr2 NM_00937 685–729 Igf1 NM_001111276 341–375 1585–1629 574–618 Tgfbr3 NM_011578 691–735 Igf2 NM_010514 621–665 1921–1965 869–913 Bmp2 NM_007553 1221–1265 Igf1r NM_010513 1521–1565 2261–2305 3601–3645 Bmp4 NM_ 007554 581–625 Igfbp3 NM_008343 571–615 1401–1445 951–995 Bmpr1a NM_009758 850–894 Fgf2 NM_008006 271–315 1611–1655 543–587 Bmpr1b NM_007560 721–765 Fgf4 NM_010202 121–165 1741–1785 481–525 Bmpr2 NM_007561 1381–1425 Pdgfa NM_008808 271–315 2878–2922 551–595 Ctgf NM_010217 481–525 Pdgfra NM_011058 1741–1785 961–1005 2621–2665 netechnical Co., Ltd., Tokyo, Japan), quickly frozen 0.1% sarkosyl, dehydrated through a graded series in liquid nitrogen, and stored −80°C until use. All of ethanol, and air-dried. Sections were dipped into experiments were performed based on protocols fol- an autoradiographic emulsion (NTB-2; Kodak) and lowing the Guidelines for Animal Experimentation, exposed at 4°C for 7–8 weeks. The hybridization Graduate School of Medicine, Hokkaido University, sections were counterstained with hematoxylin after Japan. development. The same sections were observed under bright and dark fields using a light microscope In situ hybridization. Two non-overlapping 45-mer (BX51; Olympus, Tokyo, Japan). antisense oligonucleotide probes for mRNAs were In situ hybridization analysis using two non-over- synthesized. The targeted nucleotide residues of lapping antisense probes showed consistent labeling probes used in this study are shown in Table 1. in all tissues examined. The specificity of hybridiza- These oligonucleotides were labeled with 33P-dATP tion was also confirmed by the disappearance of the using terminal deoxynucleotidyl transferase (Invitro- signals with an excess dose of unlabeled antisense gen, Carlsbad, CA, USA). Fourteen-micrometer- probes (twenty times more than labeled probes). thick fresh frozen sections were prepared and mounted on glass slides precoated with 3-aminopro- RESULTS pyltriethoxysilane. They were fixed with 4% para- formaldehyde in 0.1 M phosphate buffer for 15 min Comparative expression profiles for mRNAs of and then acetylated with 0.25% acetic anhydride in growth factors and their receptors are summarized 0.1 M triethanolamine-HCl (pH 8.0) for 10 min. Hy- in Tables 2 and 3. bridization was performed at 42°C for 10 h by adding 10,000 cpm/μL of 33P-labeled oligonucleotide probes. TGF-β Slides were rinsed at room temperature for 30 min Among the three TGF-β subtypes, mRNA expres- in 2 × SSC (1× SSC: 150 mM sodium chloride, sion of TGF-β1 and β2 was significant in the devel- 15 mM sodium citrate) containing 0.1% sarkosyl, oping tooth germ. Signals for both TGF-β1 and twice at 55°C for 40 min in 0.1× SSC containing TGF-β2 in the fetal tooth germ were localized in the 208 M. Hisamoto et al.

Fig. 1 TGF-β1/2 and TGF-βR2 gene expressions in fetal (A, B) and neonatal tooth germs (C, D) determined by in situ hybridization. Fig. 1A2 and B2 are dark field images of Fig. 1A1 and B1, respectively. An intense expression of TGF-β2 in a fetus (E16.5) is seen in a part of the enamel epithelium, particularly the cervical loop (arrows in A). TGF-βR2 is expressed mainly in the fetal dental papilla (asterisk in B), but not in the epithelium of the enamel organ (B). In adjacent sections of the neonatal (day-1) tooth germ, signals of TGF-β1 are intense in odontoblasts (C), while the expression of TGF-βR2 is more intense in the ameloblast layer (D). inner/outer enamel epithelium, where an intense in ameloblasts (Fig. 1C); its expression in amelo- TGF-β2 signal condensed in the cervical loop, being blasts tended to decrease in intensity according to important for root formation (arrows in Fig. 1A). dental hard tissue formation. TGF-β2 was expressed TGF-βR1 and TGF-βR3 in the tooth germ of fetuses in both ameloblasts and odontoblasts, where the ex- were expressed faintly, while a moderately intense pression was more intense in ameloblasts than odon- expression of TGF-βR2 was observed throughout toblasts (Table 2). Significant signals for TGF-βR1 the dental papilla (asterisk in Fig. 1B). Therefore, a and TGF-βR2 in neonates were localized in both predominant TGF-β signaling pathway in fetuses ameloblasts and odontoblasts, where TGF-βR2 was may be directed from the enamel epithelium toward predominant and more intensely expressed in amelo- the dental papilla if autocrine regulation by TGF-β blasts than odontoblasts (Fig. 1D). Therefore, TGF-β1 is excluded. is a potent signal in neonates and may be directed Generally, expressions of TGF-β and TGF-βRs in from odontoblasts toward ameloblasts with a more the tooth germ increased in intensity in neonates intense expression for all receptor subtypes (Table 2). (Table 2). Unlike fetuses, neonatal TGF-β1 expres- These findings were partially supported by the sion was the most intense in odontoblasts, but weak mRNA expression of Smad3 which is involved in Growth factors in tooth germ 209

Table 2 Expression of members of TGFβ and Smad families in fetuses/neonates TGF-β1 TGF-β2 TGF-βR1 TGF-βR2 TGF-βR3 Smad3 BMP2 BMP4 Smad5 fetus inner enamel epithelium + +* ± ± ± + − − ± outer enamel epithelium − +* ± ± ± + − − ± cervical loop − ++ ± ± ± + − − ± stellate reticulum − − − − − + − − ± enamel knot − − ± − − ± +++* +++*** ± dental papilla + − − ++ ± + +++ +++ ± neonate ameloblasts + ++ ++ +++ + ++ − ++ +++ odontoblasts +++ + + + − + +++ + + pulp − − ± ± − + ++ − + +++: strong, ++: moderate, +: weak, ±: faint, −: negative *: limited in cervical loop, **: primary enamel knot, ***: secondary enamel knot

Table 3 Expression of signaling molecules FGF2 FGF4 PDGFα PDGFRα IGF-1 IGF-2 IGFBP3 IGF-1R CTGF fetus inner enamel epithelium − − +++ − − + − ++ +++ outer enamel epithelium − − +++ − − + − ++ +++ cervical loop − − +++ − − − − − − stellate reticulum − − − − − +++ − − ++ stratum intermedium − − − ++ − − − +++ − enamel knot − +++* − − − − − − − dental papilla − − − +++ − + − − − dental follicle − − +++ − +++ +++ +++ − + neonate ameloblasts − − +++ − − + − + − stratum intermedium − − − +++ − + − +++ − enamel-free area − − +++ − − − − − − odontoblasts − − − +++ − + − + ++ pulp − − − + − + + − − dental follicle − − +++ − +++ +++ +++ − ++ +++: strong, ++: moderate, +: weak, ±: faint, −: negative *: primary enamel knot the intracellular transduction of TGF-β signaling: mesenchyme interaction from the initiation stage to Smad3 expression occurred mainly in ameloblasts of crown stage (1, 15, 17). Thus, we examined mRNA neonates (Table 2). It is generally accepted that expression of BMP2 and BMP4 together with three TGF-β ligands first bind to TGF-βR2 serine/threo- BMP receptor subtypes (BMPR1a, 1b, and 2). Ex- nine kinase receptors, where TGF-βR1 is linked pression of BMP2 in the fetal tooth germ was found with TGF-βR2 under activating conditions and is mainly in the dental papilla, with more intense ex- phosphorylated by it. In turn, the phosphorylated pression in the lower (deeper) region (asterisk in TGF-βR1 stimulates its downstream targets such as Fig. 2A). Signals of BMP2 in the enamel organ Smad2/3 (31, 32). were restricted to the primary enamel knot (Fig. 2B) and a part of the inner enamel epithelium which BMP may correspond to the secondary enamel knot (arrow Although previous studies documented several BMP in Fig. 2A), as reported previously (14). Neonatal ex- subtypes (BMP2 to –7) expressed in the tooth germ pression of BMP2 increased in intensity and accumu- (1), in particular, expression of BMP2 and BMP4 lated in odontoblasts (Fig. 2C). However, the entire may be marked and associated with the epithelium- region of the dental pulp still maintained a moderate- 210 M. Hisamoto et al.

Fig. 2 Expression of BMP2/4 in fetuses (A, B, D) and neonates (C). In fetuses, both BMP2 and BMP4 are localized in the dental papilla, where BMP2 is intense in the lower dental papilla (asterisk in A), and BMP4 is intense in the upper dental papilla (asterisk in D). Other signals for BMP2 are seen in the primary enamel knot (B) and a restricted area of the inner enamel epithelium, probably the secondary enamel knot (arrow in A). Expression of BMP2 in neonates is very intense in odontoblasts and moderately intense throughout the entire dental pulp (C). Fig. 2A (A1/A2: bright and dark field images) and Fig. 2D are taken from serial sections. ly intense BMP2 expression, while ameloblasts lacked tuses; neonatal BMP4 signals were found weakly any significant expression of BMP2 in neonates as to moderately in odontoblasts and ameloblasts, and well as fetuses (Fig. 2C). According to the expres- the latter which displayed more intense signals (Ta- sion profile, BMP2 signals generally arise from the ble 2). BMP4 expression in the dental pulp was al- fetal dental papilla and neonatal odontoblasts. most undetectable. Like BMP2, BMP4 in the fetuses was selectively Although figures are not shown, three types of expressed in the dental papilla; however, more in- BMP receptor (BMPR1a/1b and BMPR2) in fetuses tense expression was found in the upper dental pa- were expressed faintly in the entire area of the tooth pilla close to the inner enamel epithelium (asterisk germ, while moderately intense expression of BMPR2 in Fig. 2D), being different from the localization of was found in both ameloblasts and odontoblasts of BMP2 (Fig. 2A). Besides the dental papilla, the sec- neonates. When we examined mRNA expression of ondary enamel knot, but not the primary enamel Smad1 and Smad5 which mediate BMP signals, knot, intensely expressed BMP4 (Fig. 3A). BMP4 Smad5 was expressed faintly in the fetal tooth germ expression in neonates was less intense than in fe- over its entire length, and more intensely in amelo- Growth factors in tooth germ 211

Fig. 3 Figures showing the localization of growth factors involved in cusp formation. BMP4 is expressed in the secondary enamel knot (arrow) and adjacent mesenchyme of the dental papilla (A1/A2: bright and dark field images). In Fig. 3B, restricted expression of FGF4 in fetuses corresponds to the primary enamel knot also seen in Fig. 2B for BMP2. blasts of neonates than in other areas (Table 2). On Expression of PDGFα, a gene coding for PDGFA, the other hand, a weak expression of Smad1 was lo- was localized in the inner and outer enamel epitheli- calized throughout the tooth germ in both fetuses um including the cervical loop in fetuses (Fig. 4A). and neonates (data not shown in Table 2). The neonatal tooth germ showed a restricted expression of PDGFα in ameloblasts (Fig. 4C) and the FGFs and PDGFA enamel-free area, which was occupied by small It is well-known that among FGFs, FGF4 shows a round cells (arrow in Fig. 4E). While PDGFRα was characteristic expression pattern in the enamel knot intensely expressed throughout the dental papilla of which regulates patterning of the cusps and, hence, fetuses (Fig. 4B), its selective expression in neonates the shape of the tooth crown (9, 10, 14, 24). Fur- was found in odontoblasts and the stratum interme- thermore, an immunohistochemical study demon- dium of the enamel organ (arrows in Fig. 4D). strated the existence of FGF2 in the tooth germ of Therefore, PDGFα signaling may be directed from mice (3). To investigate signaling molecules associ- the enamel epithelium/ameloblasts toward dental pa- ated with cusp formation, we examined the expres- pilla/odontoblasts. sion of FGF4 and FGF2. FGF4 expression in fetuses was restricted to the primary enamel knot (Fig. 3B), Growth factors in the dental follicles but not found in the secondary enamel knot express- The dental follicle in fetuses and neonates consis- ing BMP4 (Fig. 3A). On the other hand, no signals tently expressed IGF-related molecules except IGF- of FGF4 in neonates occurred throughout the tooth 1R (Table 3). The expressions of IGF-1, IGF-2, and germ. Therefore, the tooth germ uses FGF4 only in IGFBP3 were intense in the dental follicles through- the early stage of development, associated with cusp out the developing stages (asterisks in Fig. 5A–E). formation. Also, no significant signals for FGF2 Another intense expression of IGF-2 in fetuses was were detectable anywhere in fetuses and neonates, found in the tips of the enamel organs (arrow in in contrast to previous immunohistochemical studies Fig. 5C), and a weak expression of IGF-2 was rec- showing the localization of FGF2 in the dental mes- ognizable in the dental papilla and inner/outer enamel enchyme, dental papilla, and stellate reticulum of epithelium of fetuses while neonates weakly ex- the tooth germ (3, 21). pressed IGF-2 in the entire area of the tooth germ 212 M. Hisamoto et al.

Fig. 4 PDGFα and PDGFRα expression patterns in fetuses (A, B) and neonates (C, D, E): A–B and C–D are serial sections. PDGFα in fetuses is expressed in the inner/outer enamel epithelium including the cervical loop (arrow) (A). The predominant expression of PDGFRα is restricted to the dental papilla (B). In neonates, PDGFα is expressed in ameloblasts and the dental follicle (asterisk) (C), and PDGFRα is expressed in odontoblasts and the stratum intermedium (arrows) (D). In some neonates, PDGFα expression is concentrated in the enamel-free area (arrow) as well as ameloblasts (E1/ E2: bright and dark field images).

(Table 3). IGF-1R was undetectable in the dental DISCUSSION follicle but expressed intensely in the stratum intermedium of the enamel organs in both fetuses and Members of the TGF-β family (TGF-βs and BMPs) neonates (Table 3). Furthermore, IGFBP3 was selec- are well-known as the most important signals in the tively expressed in blood vessels distributed in the epithelium-mesenchyme interaction during tooth de- pulp of fetuses and neonates (arrowheads in Fig. 5B velopment (19, 25). When the TGF-β and BMP li- and E) as well as the dental follicle. gands bind to specific receptors on the cell surface, The predominant expression sites of CTGF were the intracellular transduction for TGF-βs is initiated the enamel epithelia, stellate reticulum, and dental by Smad2/3, while BMP signaling is mediated by follicle in fetuses (Fig. 5F), but the predominant ex- Smad1/5/8 (26, 31, 32). Some studies have docu- pression sites in the tooth germ of neonates shifted mented the involvement of Smad molecules in to odontoblasts, maintaining intense signals in the odontogenesis (28). Our present analysis of the ex- dental follicle (Fig. 5G). pressions of TGF-βs and TGF-βRs in the tooth germ of mouse fetuses indicates that the intense signals of Growth factors in tooth germ 213

Fig. 5 Expression of IGF-1, IGF-2, IGFBP3, and CTGF in the dental follicle. In fetuses, IGF-1(A) and IGFBP3 (B) are intensely localized in the dental follicles (asterisks). Their expressions in the dental follicle (asterisks) are maintained in neonates (D, E). Another expression site of IGFBP3 is found in blood vessels in the dental papilla (arrowhead) of the fetus (B) and neonate (E). IGF-2 is localized in the dental follicle (asterisk in C) and stellate reticulum (arrow in C). CTGF in fetuses is intensely expressed in most of the inner/outer epithelium, a part of the stellate reticulum, and dental follicle (asterisks) (F).

Expression of CTGF in neonates is restricted to odontoblasts (arrows) and the dental follicle (asterisk) (G1/G2: bright and dark field images).

TGF-βs are released from the inner enamel epitheli- tion. In neonates, TGF-βs and TGF-βRs were ex- um toward the mesenchyme or dental papilla, and pressed in both odontoblasts and ameloblasts with that TGF-β2 is specifically involved in root forma- different intensities. Therefore, it is difficult in neo- 214 M. Hisamoto et al.

Fig. 6 The signaling pathways of various growth factors in the epithelium-mesenchyme interaction of the tooth germ. In fetuses, TGF-β1/2 and PDGFA affect the development of the mesenchyme, while BMP2/4 signals arise from the mesenchyme (A). TGF-β1 and BMP2 in neonates modulate the differentiation of ameloblasts, on the contrary neonatal PDGFA affects odontoblasts (B). Signals of CTGF originate from the enamel epithelium in fetuses but from the mesenchyme in neonates (A, B). nates to decide on the direction of the signaling ings largely correspond to those of Aberg et al. (1), pathway of TGF-βs working in autocrine/paracrine who performed an in situ hybridization analysis for manners. Nevertheless, TGF-β1, a predominant sig- the expression of BMP2–7 in the mouse tooth germ. nal in neonates, was intensely expressed in odonto- The present study showed that both BMP2 and blasts with a weak expression in ameloblasts. For BMP4 in fetuses were localized in the epithelium, the receptors in neonates, TGF-βR2 was predomi- but, in neonates, BMP2 was expressed in the mes- nant and localized mainly in ameloblasts together enchyme in contrast to epithelial BMP4; conse- with other receptor subtypes, suggesting that signals quently, the expression patterns of BMP2 and BMP4 of TGF-β1 in neonates may be directed from odon- were different during tooth development. Concern- toblasts toward ameloblasts. This finding was sup- ing Smads, we identified a significant mRNA - ex ported by the mRNA expression of Smad3, which pression of Smad5, which transmits signals of BMP was localized mainly in ameloblasts of neonates. members, only in ameloblasts of neonates. Taken to- Based on these findings, we can conclude that TGF- gether, we consider that BMP/Smad signaling arises βs are essential for differentiation of the dental pa- mainly from the dental mesenchyme to affect the pilla into odontoblasts in fetuses and, in turn, as differentiation of ameloblasts (Fig. 6). tooth development advances, are critical for the dif- PDGFA is also a representative growth factor ferentiation of ameloblasts (Fig. 6). with epithelium-mesenchyme interaction. Our pres- Among the TGF-β family involved in the epithe- ent data revealed an intense expression of PDGFα lium-mesenchyme interaction, the signals of BMP2 in the enamel epithelium of fetuses and ameloblasts clearly arose from the dental papilla (fetus) and of neonates, while the expression of PDGFRα al- odontoblasts (neonate). BMPR2, a predominant re- ways appeared in the dental papilla and odontoblasts, ceptor subtype for BMP in the tooth germ, was in- indicating that epithelial PDGFα signals induce the tensely expressed in neonatal ameloblasts, although development of mesenchyme and subsequent odon- BMPR2 in fetuses was extensively distributed in the toblast differentiation (Fig. 6). However, a recent ex tooth germ. Consequently, BMP2 signaling is basi- vivo study reported that PDGFA is expressed in the cally directed from the dental mesenchyme toward enamel epithelium and PDGFRα is localized in the ameloblasts to regulate their differentiation (Fig. 6). inner enamel epithelium as well as mesenchyme, The dental mesenchyme of fetuses intensely ex- and so they concluded that PDGFA may have ef- pressed BMP4 as well as BMP2. On the other hand, fects on the development of the inner enamel epi- the expression of BMP4 in neonates differed from thelium (27). This idea is in contrast to ours that that of BMP2: BMP4 was expressed moderately in epithelial PDGFA/PDGFα signals always regulate ameloblasts and weakly in odontoblasts, but not in the differentiation of the mesenchyme (Fig. 6). the dental pulp, while BMP2 was expressed only in CTGF is another growth factor which may medi- the dental pulp including odontoblasts. These find- ate the epithelium-mesenchyme interaction. It was Growth factors in tooth germ 215 reported by a study employing in situ hybridization by analysis methods with different sensitivities. analysis that the predominant expression site of The expression of IGF-1R in fetuses and neonates CTGF and Fisp12 was preameloblasts at the bell indicated a unique feature of the stratum intermedi- stage, and their expression decreased in intensity in um of the enamel organ. The stratum intermedium secreting ameloblasts (22). Another in situ hybrid- is an epithelial cell layer located between the inner ization study of the mouse tooth germ reported that enamel epithelium and stellate reticulum, but its CTGF was expressed equally in ameloblasts and function remains largely unknown. It has been re- odontoblasts in the late developmental stage (29). ported that the stratum intermedium produces Shh However, in the present study, no significant expres- (11) and participates in the differentiation of amelo- sion of CTGF was found in either preameloblasts or blasts (18). Our data further revealed that signaling ameloblasts. Instead, our data revealed that the defi- via IGF-1/IGF-1R is present there and it may affect nite expression site of CTGF was localized in the the differentiation of the adjacent ameloblast layer. enamel epithelium in fetuses and shifted to the den- Moreover, PDGFRα was also expressed selectively in tal follicle and odontoblasts in neonates (Fig. 6), the stratum intermedium. Therefore, PDGFα/PDGFRα presenting another feature that the expression sites may possess the same functions as IGFs/IGF-1R, of growth factors move from the enamel epithelium and PDGFA and IGF signalings may have a mutual to the mesenchyme during tooth development, like relationship in promoting the functional significance TGF-β1. of the stratum intermedium. The expressions of growth factors in the dental The present study draws special attention to follicle which gives rise to the periodontium includ- growth factors associated with cusp formation. It is ing the periodontal ligament, osteoblasts, and ce- well-known that the differentiation of odontoblasts mentbolasts were distinctive in both fetuses and starts from the tip of the cusp and that the signals neonates. Our data revealed the intense expression released from the enamel knot regulate the pattern- of three IGF family members (IGF-1, IGF-2, and ing of cusps and, hence, the shape of the tooth IGFBP3) in the dental follicle of fetuses. Marked crown, and subsequently participate in differentia- expression of PDGFα and CTGF was also detected tion of the dental papilla and odontoblasts (10, 24). in the dental follicle of neonates. An in vitro study The enamel knot expresses many signaling mole- using a mouse periodontal ligament cell line also in- cules, such as FGFs, BMPs, Shh, and Wnt families dicated the involvement of CTGF (or CCN2) in the (9, 10, 14, 24). In particular, FGF4 functions as a differentiation of periodontal tissues (2). Therefore, cusp activator while BMPs function as inhibitors to these growth factors may play an important role in regulate the distance between adjacent cusps (10). constituting the periodontium. In addition to the Our data strengthen the evidence for the significance dental follicle, we noted an intense expression of of the enamel knot by showing that the primary IGFBP3 in blood vessels of the dental pulp. This re- enamel knot expressed FGF4 and BMP2, and the sult corresponds to the idea that IGFBP3 functions secondary enamel knot expressed BMP4. In addi- as a carrier protein to direct IGFs from the circula- tion, the expression sites of BMP2/4 in fetuses shift- tion to target tissues (7, 13, 16). On the other hand, ed from the primary/secondary enamel knots to the Gabriel et al. (13) reported based on the immuno- dental papilla, being in agreement with findings pre- histochemistry of human third molars that IGF-1 sented by Aberg and colleagues (1). As another and IGFBP3 were broadly expressed in preodonto- growth factor of cusp formation, the present study blasts, odontoblasts, cementoblasts, dental papilla demonstrated for the first time a restricted expression mesenchyme, and the pulp as well as expression of of PDGFα to a special part, the enamel-free area in IGFBP3 in the dental follicle. According to an im- the subsequent stage. The enamel-free area is unique munohistochemical and in situ hybridization study in rodent molars, and is not covered by enamel by Yamamoto et al. (30), IGF family members such when the teeth erupt (23). Although Vaahtokari et as IGF-1, IGF-2, IGF-1R, and IGF-2R were ex- al. (25) reported based on an in situ hybridization pressed in ameloblasts of rat incisors. However, our technique that TGF-β1 was expressed in non-secret- careful observation failed to detect any significant ing ameloblasts of the enamel-free area in day-4 expression of IGF-1 and IGFBP3 in the enamel epi- postnatal mice, we failed to observe TGF-β1 in the thelium and ameloblasts, although IGF-2 and IGF-1R enamel-free area of day-1 and day-3 postnatal mice were weakly detected in those areas. These discrep- (for day-3, our unpublished data). Such varying re- ancies in the expression profiles may be caused by sults may be caused by differences in the stage of different tooth types and maturation stages, and also tooth development and sensitivity of probes used. 216 M. Hisamoto et al.

Since the inner enamel epithelium in the enamel- 6. Gao Y, Li D, Han T, Sun Y and Zhang J (2009) TGF-beta1 free area does not differentiate into ameloblasts (23), and TGFBR1 are expressed in ameloblasts and promote MMP20 expression. Anat Rec 292, 885–890. PDGFα in rodent molars may inhibit ameloblast dif- 7. Gotz W, Heinen M, Lossdorfer S and Jager A (2006) Immu- ferentiation in this area and regulate cusp formation. nohistochemical localization of components of the insulin- The present observation failed to identify cell types like growth factor system in human permanent teeth. Arch of the small round cells which occupied the enamel- Oral Biol 51, 387–395. free area. Further studies are required to define the 8. Igarashi A, Okochi H, Bradham DM and Grotendorst GR (1993) Regulation of connective tissue growth factor gene characteristic and function of the enamel-free area. expression in human skin fibroblasts and during wound re- pair. Mol Biol Cell 4, 637–645. 9. Jernvall J, Kettunen P, Karavanova I, Martin LB and Thesleff CONCLUSIONS I (1994) Evidence for the role of the enamel knot as a con- trol center in mammalian tooth cusp formation: non-dividing The regulatory mechanism of tooth formation is com- cells express growth stimulating Fgf-4 gene. Int J Dev Biol plex and controlled by intricate cross-talk between 38, 463–469. adjacent tissues and by released growth factors. Sys- 10. Jernvall J and Thesleff I (2000) Reiterative signaling and tematic morphological analyses of the growth factors patterning during mammalian tooth morphogenesis. Mech are indispensable for the comprehensive understand- Dev 92, 19–29. 11. Koyama E, Wu C, Shimo T, Iwamoto M, Ohmori T, Kurisu K, ing of tooth formation. The present data may provide Ookura T, Bashir MM, Abramas WR, Tucker T and Pacifici valuable information concerning signaling pathways M (2001) Development of stratum intermedium and its role of growth factors in the epithelium-mesenchyme in- as a Sonic Hedgehog-signaling structure during odontogene- teraction during odontogenesis. Our in situ hybrid- sis. Dev Dyn 222, 178–191. ization method using many probes with the same 12. Leask A and Abraham DJ (2006) All in the CCN family: essential matricellular signaling modulators emerge from the base length is useful for the cellular localization of bunker. J Cell Sci 119, 4803–4810. growth factors and comparison of the signal intensi- 13. Magnucki G, Schenk U, Ahrens S, Navarrete SA, Gernhardt ties. Interestingly, the expression strength and direc- RC, Schaller HS and Cuong Hoang VC (2013) Expression of tion of signals are regularly changeable during tooth the IGF-1, IGFBP-3 and IGF-1 receptors in dental pulp stem cells and impacted third molars. J Oral Sci 55, 319–327. development, suggesting that various growth factors 14. Matalova E, Antonarakis GS, Sharpe PT and Tucker AS (2005) play stage-dependent roles in odontogenesis. Cell lineage of primary and secondary enamel knots. Dev Dyn 233, 754–759. 15. Moriguchi M, Yamada M, Miake Y and Yanagisawa T (2011) DECLARATION OF INTEREST Immunolocalization of TAK1, TAB1, and p38 in the developing rat molar. Anat Sci Int 86, 69–77. The authors have no conflicts of interest to declare. 16. Morimoto LM, Newcomb PA, White E, Bigler J and Potter JD (2005) Variation in plasma insulin-like growth factor-1 REFERENCES and insulin-like growth factor binding protein-3: Genetic factors. Cancer Epidemiol Biomarkers Prev 14, 1394–1401. 1. Aberg T, Wozney J and Thesleff I (1997) Expression patterns 17. Nadiri A, Kuchler-Bopp S, Haikel Y and Lesot H (2004) Im- of bone morphogenetic proteins (Bmps) in the developing munolocalization of BMP-2/-4, FGF-4, and WNT10b in the mouse tooth suggest roles in morphogenesis and cell differ- developing mouse first lower molar. J Histochem Cytochem entiation. Dev Dyn 210, 383–396. 52, 103–112. 2. Asano M, Kubota S, Nakanishi T, Nishida T, Yamaai T, 18. Nakamura H and Ozawa H (1990) Lectin cytochemistry on Yosimichi G, Ohyama K, Sugimoto T, Murayama Y and the stratum intermedium and the papillary layer in the rat in- Takigawa M (2005) Effect of connective tissue growth factor cisor enamel organ. Arch Histol Cytol 53, 351–369. (CCN2/CTGF) on proliferation and differentiation of mouse 19. Oka S, Oka K, Xu X, Sasaki T, Bringas P Jr and Chai Y (2007) periodontal ligament-derived cells. Cell Commun Signal 3, Cell autonomous requirement for TGF-β signaling and dentin 11. matrix formation. Mech Dev 124, 409–415. 3. Cam Y, Neumann MR, Oliver L, Raulais D, Janet T and 20. Pelton RW, Dickinson ME, Moses HL and Hogan BL (1990) Ruch JV (1992) Immunolocalization of acidic and basic fi- In situ hybridization analysis of TGF beta 3 RNA expression broblast growth factors during mouse odontogenesis. Int J during mouse development: comparative studies with TGF Dev Biol 36, 381–389. beta 1 and beta 2. Development 110, 609–620. 4. Caviedes BJ, Canales-SP, Castrillon SN, Jovel-GJ, Alvarez 21. Russo LG, Maharajan P and Maharajan V (1997) Basic fi- VJ, Rivero C, Azuero HMM, Diaz E and Munoz HR (2009) broblast growth factor (FGF-2) in mouse tooth morphogene- Expression of insulin-like growth factor-1 and proliferating sis. Growth Factors 15, 125–133. cell antigen in human pulp cells of teeth with complete and 22. Shimo T, Wu C, Billings PC, Piddington R, Rosenbloom J, incomplete root development. Int Endo J 42, 686–693. Pacifici M and Koyama E (2002) Expression, gene regula- 5. Chai Y, Bringas P Jr, Mogharei A, Shuler CF and Slavkin tion, and roles of Fisp12/CTGF in developing tooth germs. HC (1998) PDGF-A and PDGFR-α regulate tooth formation Dev Dyn 224, 267–278. via autocrine mechanism during mandibular morphogenesis 23. Sutcliffe JE and Owens PD (1980) A light and scanning elec- in vitro. Dev Dyn 213, 500–511. tron microscopic study of the development of enamel-free ar- Growth factors in tooth germ 217

eas on the molar teeth of the rat. Arch Oral Biol 25, 263–268. 29. Yamaai T, Nakanishi T, Asano M, Nawachi K, Yoshimichi G, 24. Thesleff I, Keranen S and Jernvall J (2011) Enamel knots as Ohyama K, Komori T, Sugimoto T and Takigawa M (2005) signaling centers linking tooth morphogenesis and odonto- Gene expression of connective tissue growth factor (CTGF/ blast differentiation. Adv Dent Res 15, 14–18. CCN2) in calcifying tissues of normal and cbfa-1-null mu- 25. Vaahtokari A, Vainio S and Thesleff I (1991) Associations tant mice in late stage of embryonic development. J Bone between transforming growth factor β1 RNA expression and Miner Metab 23, 280–288. epithelial-mesenchymal interactions during tooth morphogen- 30. Yamamoto T, Oida S and Inaga T (2006) Gene expression esis. Development 113, 985–994. and localization of insulin-like growth factors and their re- 26. Wang Y, Li L, Zheng Y, Yuan G, Yang G, He F and Chen Y ceptors throughout amelogenesis in rat incisors. J Histochem (2012) BMP activity is required for tooth development from Cytochem 54, 243–252. the lamina to bud stage. J Dent Res 91, 690–695. 31. Yokozeki M, Afanador E, Nishi M, Kaneko K, Shimokawa H, 27. Wu N, Iwamoto T, Sugawara Y, Futaki M, Yoshizaki K, Yokote K, Deng C and Tsuchida K (2007) Smad3 is required Yamamoto S, Yamada A, Nakamura T, Nonaka K and for enamel biomineralization. Biochem Biophys Res Commun Fukumoto S (2012) PDGFs regulate tooth germ proliferation 305, 684–690. and ameloblast differentiation. Arch Oral Biol 55, 426–434. 32. Zhao H, Li S, Han D, Kaartinen V and Chai Y (2011) Alk5- 28. Xu X, Jeong L, Han J, Ito Y, Bringas P Jr and Chai Y (2003) mediated transforming growth factor β signaling acts upstream Development expression of Smad1–7 suggests critical func- of fibroblast growth factor 10 to regulate the proliferation tion of TGF-β/BMP signaling in regulating epithelial-mesen- and maintenance of dental epithelial stem cells. Mol Cel Biol chymal interaction during tooth morphogenesis. Int J Dev 31, 2079–2089. Biol 47, 31–39.