Lrp4/Wise regulates palatal rugae development through Turing-type reaction-diffusion mechanisms

Periodic patterning of iterative structures is diverse across the animal kingdom. Clarifying the molecular mechanisms involved in the formation of these structure helps to elucidate the process of organogenesis. Turing-type reaction-diffusion mechanisms have been shown to play a critical role in regulating periodic patterning in organogenesis. Palatal rugae are periodically patterned ridges situated on the hard palate of mammals. We have previously shown that the palatal rugae develop by a Turing-type reaction-diffusion mechanism, which is reliant upon Shh (as an inhibitor) and Fgf (as an activator) signaling for appropriate organization of these structures. The disturbance of Shh and Fgf signaling lead to disorganized palatal rugae. However, the mechanism itself is not fully understood. Here we found that Lrp4 (transmembrane protein) was expressed in a complementary pattern to Wise (a secreted BMP antagonist and Wnt modulator) expression in palatal rugae development, representing Lrp4 expression in developing rugae and Wise in the inter-rugal epithelium. Highly disorganized palatal rugae was observed in both Wise and Lrp4 mutant mice, and these mutants also showed the downregulation of Shh signaling, which was accompanied with upregulation of Fgf signaling. Wise and Lrp4 are thus likely to control palatal rugae development by regulating reaction-diffusion mechanisms through Shh and Fgf signaling. We also found that Bmp and Wnt signaling were partially involved in this mechanism.


Introduction
The genetic commonality of developmental processes is found in many organs. It is believed that they share the same molecular mechanisms and fundamental processes. Periodic patterning is one of the common features observed in many organs, which is believed to develop through general molecular mechanisms. In 1952, Turing proposed a simple model that two morphogens diffusing through a tissue could create self-regulating periodic patterns, the reaction-diffusion model [1]. Simulation of these mechanisms replicates many biological pattern types, such as fish stripes, digits, and feather and hair spacing [2][3][4][5][6]. Turing-type reaction-diffusion mechanisms have thus been shown to play a critical role in regulating periodic patterning in organogenesis [3,5]. Clarifying the detail of Turing-type reaction-diffusion mechanisms during organogenesis helps to elucidate many other biological processes.
Palatal rugae are corrugated structures on the hard palate and are conserved in all mammals, including humans, mice and pigs [7][8][9] They are believed to function in the tactile sensing of objects or food, assisting in holding and crushing food between the tongue and the palate, and aiding in tongue placement during the production of certain speech sounds. There are eight or nine palatal rugae in mice. Three transverse ridges (antemolar rugae) are formed just behind the incisor teeth. The most anterior ruga is fused to the incisive papilla. Five or six rugae (intermolar rugae) are observed between the molar teeth. These rugae are shorter, more oblique and do not cross the midline (Fig 1A). Localized thickening of the palatal epithelium to form placodes is observed as the first morphological sign of rugae development, while the underlying mesenchymal cells condense (Fig 1B) [8]. The outer surface of the epithelium exhibits flat and the thickened epithelium is confined to the inner mesenchymal surface. Then, slight protrusions can be observed on the surface of the developing palatal shelf as the thickened epithelium protrudes over the surface by the flattening of the basement membrane. Subsequently, the placode regions bulge toward the oral cavity to form an overall corrugated appearance [10]. Palatal rugae are sequentially added on the growing palate, their interposition appears to be dependent on activation-inhibition mechanisms, and rugae development has been proposed as a simple tool to study regulation of patterning of serial structures [11]. Although, the number and the patterns of the palatal rugae are species specific, palatal rugae are consistent in mice. Therefore murine palatal rugae development is believed to be under strict genetic control [11][12][13][14][15][16][17][18][19][20]. We previously found that Turing-type reaction-diffusion mechanisms are involved in murine palatal rugae development acting through Fgf and Shh signaling [21]; however, the process itself is not fully understood.
The low-density lipoprotein (LDL) receptor family is a large evolutionarily conserved group of transmembrane proteins (for reviews, see [22,23]). The LDL receptor mainly regulates the concentration of lipoproteins in the extracellular fluids and delivers them to cells (i.e. for the uptake of cholesterol). LDL receptor family members have also been shown to function as direct signal transducers or modulators for a broad range of cellular signaling pathways. Lrp5 and Lrp6 function as co-receptors in the Wnt signaling cascade [24][25][26]. Lrp4 (also called Megf7) belongs to the LDL receptor family and ENU-induced Lrp4 null mutants die at birth with defects in the formation of multiple embryonic tissues [27]. We have previously reported that Lrp4 is involved in regulating tooth development [28,29]. Wise (also known as USAG-1, Sostdc1 and Ectodin) is a cystine-knot secreted BMP-antagonist, that shows relatively high homology with members of the DAN and CCN peptide families [30]. Wise has also been identified as a context-dependent activator or inhibitor of Wnt signaling by directly binding to Lrp6 [31][32][33]. Mutation in either of Wise or Lrp4 in mice produces multiple, but identical abnormalities in tooth development linked to alterations in BMP and Wnt signaling [28,32,[34][35][36][37][38][39][40]. We previously reported that physical interaction occurs between Wise and Lrp4 [28].
We report here that Lrp4 is expressed in a complementary manner to Wise expression in murine palatal rugae development, and Lrp4 and Wise mutants display disorganized palatal rugae, which was accompanied by disturbance in Shh and Fgf signaling. Lrp4/Wise is thus involved in murine rugae development by regulating reaction-diffusion mechanisms.

Materials and methods
All animal experiments were conducted in compliance with the protocol, which was reviewed by the Institutional Animal Care and Use Committee and approved by the President of Niigata University (Permit Number: #25 Niigata Univ. Res. 255-6).

Production and analysis of transgenic mice
Production and analysis of transgenic mice Lrp4 and Wise mutant mice were performed as described by Johnson et al. [41] and Kassai et al. [34], respectively. K14-Bmp4 and K14-Noggin mice were produced as described by Guha et al. [42]. K14-Shh mice were produced as described by Cobourne et al. [43]. Mouse heads were fixed in 4% paraformaldehyde (PFA), wax embedded and serially sectioned at 7μm. Sections were split over 5-10 slides and prepared for histology or radioactive in situ hybridization. All mice were sacrificed by cervical dislocation.

In situ hybridization
Radioactive in situ hybridization with [35S]UTP-labeled riboprobes and whole-mount in situ hybridization with DIG-labeled riboprobes was carried out as described previously by Ohazama et al. [28,44]. In situ hybridization was performed in triplicate.

Immunohistochemistry analysis
After deparaffinization of sections, tissues were treated with proteinase K and then incubated with an antibody to Phosphorylated-Smad 1/5/9 (Cell signaling Technology). As a negative control, normal rabbit serum or normal goat serum were used instead of primary antibody. Tyramide signal amplification system was performed (Parkin Elmer Life Science) for detecting phosphorylated-Smad 1/5/9. Slides were mounted with Aquamount. Pictures were taken with the same exposure between control, wild-type, Wise and Lrp4 mutant mice. Immunohistochemistry analysis was performed in triplicate.

Whole-mount nuclear fluorescent imaging
Detailed morphology of palatal rugae was analyzed by nuclear fluorescent imaging technique, called "Pseudo SEM" as previously described [45]).

Lrp4 and Wise expression in palatal rugae development
In order to clarify the detail of Turing-type reaction-diffusion mechanisms during palatal rugae development, we were interested in Lrp4 and Wise. It is known that both molecules interacted each other in organogenesis such as tooth and mammary development [28,29,40]. Disorganized patterns of palatal rugae have previously reported in Wise mutant mice [15]. We found that Lrp4 was expressed in a complementary manner to Wise expression in palatal rugae development from days of gestation (E) 12.5 to E16.5 (Fig 2). Radioactive in situ hybridization analysis exhibited that Wise and Lrp4 were expressed in the epithelium during all the examined developmental stages. Lrp4 expression was observed in the developing rugae, whereas Wise was expressed in the inter-rugal epithelium during palatal rugae development. It is known that palatal rugae are sequentially added to the developing palate [11,14]. To understand how Lrp4 and Wise were expressed in the sequential addition of rugae in murine rugae development, we performed the whole mount in situ hybridization of Lrp4 or Wise and Shh (as the marker of palatal rugae) on the left and right palatal shelves from same embryo, respectively. While Shh expression were observed in the regions corresponding to ruga 2 and 8 (E11.5-E12.0), Lrp4 was only weakly expressed in the same region (Fig 3A and 3B). Timing and localization of Lrp4 expression were similar to those of Shh during the sequential appearance of rugae ( Fig  3C-3F). While Shh expression was observed in the regions corresponding to ruga 2 and 8 (E11.5-E12.0), Wise was expressed in the region anterior to ruga 2 alone (Figs 4A, 4B and 5A). Wise expression between rugae then expanded. Wise was, therefore, expressed through two mechanisms: expansion from a single-expression domain and fusion of two expression domains. Interestingly, the former was observed in the intermolar rugae region and the latter in the antemolar rugae.

Palatal rugae phenotypes in Lrp4 and Wise mutant mice
In order to investigate the role of Lrp4 and Wise in palatal rugae development, palatal rugae were examined in Lrp4 and Wise mutant mice. Both mutants showed disorganized pattern of palatal rugae as full penetrance, which differed in each mutant mouse (Fig 6B and 6C, Lrp4 mutant mice: N = 20/20, Wise mutant mice: N = 35/35). In both mutants, antemolar rugae exhibited minor disorganized pattern, whereas intermolar rugae was highly disorganized. These suggested that Lrp4 and Wise were essential molecules for palatal rugae development, especially for intemolar rugae. After the initiation of rugae, ruga growth extended toward the midline in wild-type mice (Fig 6D) [8]. The arrest of rugae extension, ectopic initiation of rugae and the changing of the direction of ruga growth were observed during mutant rugae development (Fig 6E, data not shown). To determine whether there is any interaction between Lrp4 and Wise during the palatal rugae development, we examined Lrp4 and Wise expression in Wise and Lrp4 mutant mice, respectively. Lrp4 expression was significantly reduced in Wise mutant mice, particularly in the intermolar region, while Wise expression was disorganized in Lrp4 mutants (Fig 7A-7F). In addition to other ectoderm-derived organs including mammary gland and tooth [28,29], Lrp4 and Wise were thus likely to be interact each other to regulate palatal ruage development.

Shh and Fgf signaling in palatal rugae of Lrp4 and Wise mutant mice
Shh has been shown to act as an inhibitor in Turing-type reaction-diffusion mechanisms during palatal rugae development [21]. In fact, overexpression of Shh in palatal epithelium resulted in reduced palatal rugae (Fig 8B and 8D). Shh expression was found to be significantly downregulated in the intermolar rugae in Wise mutant mice at E14.5 (Fig 8F). A slight alteration in Shh expression was observed in Wise mutant mice at E12.5, when ruga2 and 8 were   formed ( Fig 8H). Subsequently, the downergulation of Shh expression was evident in mutant palates (Fig 8J and 8L). Shh expression was also downregulated in Lrp4 mutants (Fig 8N). In Turing-type reaction-diffusion mechanisms during palatal rugae development, Fgf signaling was previously identified as an activator [21]. Conversely to Shh signaling, expression of Erm, a marker of Fgf signaling, was expanded in developing palatal rugae of both Wise and Lrp4 mutants at E13.5 (Fig 8P and 8R). The abnormal palatal rugae found in Wise and Lrp4 mutants was thus likely to be involved in the disturbance of Turing-type reaction-diffusion mechanisms. These also suggested that Wise and Lrp4 mutants were excellent experimental models for investigating the detail of Turing-type reaction-diffusion mechanisms during palatal rugae development. Lrp4 and Wise are involved in Bmp signaling in tooth development [28,34,[36][37][38][39][40]. In order to investigate whether Bmp signaling is involved in palatal rugae development, we firstly examined the expression of major Bmp ligands in the rugae development. Bmp2, Bmp4 and Bmp5 were expressed in mesenchyme underneath the developing rugae, whilst Bmp7 expression was  Lrp4/Wise in palatal rugae development found in the developing rugae at E14.5 and E16.5 (Fig 9A-9H). In addition to ligands, expression of the Bmp inhibitor Noggin, was found in the inter-rugal epithelium (Fig 9I and 9J). To identify where Bmp signaling is activated in rugae development, immunohistochemistry for phosphorylated Smad1/5/9 (p-Smad1/5/9) was performed in the developing palatal rugae. P-Smad1/5/9-positive cells were observed in developing rugae, suggesting that Bmp signaling is activated in developing rugae, but not in the inter-rugal region where Noggin was expressed ( Fig 9K). Bmp signaling is thus likely to be controlled by the balance between ligands and inhibitors ( Fig 9M). P-Smad1/5/9 immunolocalization showed downregulation in Wise and Lrp4 mutants (Fig 9L, data not shown). In order to investigate the role of Bmp signaling in palatal rugae development, we also examined the palatal rugae of mice overexpressing Bmp4 (K14-Bmp4) or Noggin (K14-Noggin) under K14 promoter. Both K14-Bmp4 and K14-Noggin mice showed only minor anomalies in the palatal rugae. K14-Noggin mice showed only five intermolar rugae, while six are usually observed in wild-type at this stage (Figs 6A and 9N). Only slightly disorganized palatal rugae were observed in K14-Bmp4 mice (Fig 9O, K14-Bmp4 mice:N = 7/7, K14-Noggin mice: N = 9/9). Lrp4 and Wise are also known to be involved in Wnt signaling in tooth development [28,34,36,37,39]. Lef1 and Axin2 have been shown to be expressed in palatal rugae development [18]. In order to examine the changes of canonical Wnt signaling in Lrp4 and Wise mutant embryos, we examined the expression of Lef1 and Axin2 (a marker of canonical Wnt signalling). Both mutants showed downregulation of Axin2 expression at E14.5 (Fig 10A-10C). Lef1 expression was also reduced in mutants, particularly in the intermolar rugae ( Fig  10E).

Discussion
Wise can physically bind to Lrp4, and Lrp4 is expressed in a complementary pattern to Wise expression in tooth development. Furthermore, tooth phenotypes are identical between Lrp4 and Wise mutants, suggesting that Lrp4 interacts with Wise during tooth development [28,29,34,36,37,39]. Similarly, Lrp4 and Wise expression was found to be complementary during palatal rugae development (Figs 2-5), and the phenotypes of palatal rugae were comparable between Lrp4 and Wise mutants (Fig 6). Lrp4 is thus highly likely to interact with Wise during palatal rugae development.
We found that palatal rugae development is controlled by a Turing-type reaction-diffusion mechanism through Fgf and Shh signaling morphogens. Disturbance of Shh and Fgf signaling Lrp4/Wise in palatal rugae development has been shown to lead to disorganized palatal rugae, which were identical to those of mice with altered Shh and Fgf signaling [21]. We have previously shown that the disturbance of the reaction-diffusion mechanism by applying inhibitor of Shh signaling led to the expansion of Fgf response makers in palatal rugae development [21]. In Wise and Lrp4 mutants, the downregulation of Shh signaling was accompanied by the expansion of Fgf signaling (Fig 8). This suggested that disorganized palatal rugae in the Wise and Lrp4 mutants is likely to be caused by the disturbance of Turing-type reaction-diffusion mechanisms through Fgf and Shh signaling. Wise and Lrp4 mutant mice are thus excellent experimental model for investigating the detail of Turing-type reaction-diffusion mechanisms during palatal rugae development.
Lrp4 and Wise expression are found in a complementary pattern in both tooth and rugae development. Wise expression is observed in tooth mesenchyme, while Lrp4 is expressed in tooth epithelium, suggesting that correlation between Lrp4 and Wise is through epithelial-mesenchymal interaction. However, unlike their expression in tooth development, both Wise and Lrp4 were expressed in palatal epithelium, with Lrp4 in the developing rugae and Wise in the inter-rugal epithelium. Our results also showed that Bmp signaling activity and Bmp inhibitor expression are found in a complementary pattern in palatal rugae development, with activity in the epithelial placode, and inhibition in the inter-placode epithelium. We found that palatal rugae development is under the control of Turing-type reaction-diffusion mechanisms through Fgf and Shh signaling morphogens [21]. Shh and Erm were also expressed in palatal rugae epithelium [15,16]. Additionally, hair follicle spacing has been shown to be determined by reaction-diffusion mechanisms using morphogens expressed in the epithelium [5]. Interaction between developing rugae and inter-rugal epithelium thus plays a critical role in palatal rugae development. No significant changes in K14-Bmp4 and K14-Noggin mice, suggesting that Bmp and Noggin might be able to be compensated by other molecules during palatal rugae development. It is possible that there is an interaction between Wise and Noggin in the interrugal epithelium, and between Lrp4, Shh and Bmp in developing rugae.
Our work, in addition to observations by other groups, indicates that Lrp4 and Wise are strongly linked to Bmp signaling in tooth development [28,34,36,37,39]. We found that Bmp signaling was downregulated in developing palatal rugae of both mutants (Fig 9L). However, the disturbance of Bmp signaling resulted in only minor anomalies during palatal rugae development. The alteration of Bmp signaling due to Lrp4/Wise mutation thus weakly influenced palatal rugae development, suggesting that the down-regulation of Bmp signaling is not a major cause of abnormal palatal rugae formation in Wise and Lrp4 mutants. We could not exclude the possibility that the minor rugae phenotypes observed in K14-Bmp and K14-Noggin mice are related to inter-strain variability. In addition to Bmp signaling, Lrp4 and Wise are strongly related to Wnt signaling in tooth development [28,34,36,37,39]. Downregulation of Wnt signaling was observed in the palatal rugae of both Lrp4 and Wise mice (Fig 10). It has been shown that Fgf and Shh pathways are major downstream targets of Wise-regulated Wnt signaling in tooth development [38]. In addition, Shh signaling has been shown to be downstream of Wnt signaling in palatal rugae development [18]. It is conceivable that there is feedback loop between Shh and Wnt signaling in palatal rugae (i.e. the lack of Shh lead to the downregulation of Wnt signaling). It is also possibile that Wise/Lrp4 is directly interacted with Shh and Fgf signaling and the alteration of Bmp and Wnt signaling was consequence of the changes of Shh and Fgf signaling. On the other hand, it has been shown that none of Wnt Lrp4/Wise in palatal rugae development signaling results in no palatal rugae formation, indicating that Wnt signaling determines the initiation of palatal rugae which contain regulating palatal rugae patterning [18]. Therefore, we could not exclude the possibility that reduced Wnt signaling allow rugae to form, but affect patterning of rugae in Wise and Lrp4 mutants.
Rugae disorganization was more severe in the intermolar rugae subset, compared to those of the antemolar rugae in Lrp4 and Wise mutants (Fig 6B and 6C). Changes of Shh signaling in mutants were also found to be different between antemolar and intermolar rugae (Fig 8). K14-Bmp4 and K14-Noggin also showed different phenotypas of palatal rugae between antemolar and intermolar rugae (Fig 9N and 9O). We also found that the processes of establishment in Wise expression also differed between antemolar and intermolar rugae (Figs 4 and 5). The palatal rugae show periodic patterning, but the pattern is slightly different between the anterior and posterior region; the antemolar rugae are transverse ridges, and the intermolar rugae are shorter, more oblique and do not cross the midline. It is likely that molecular mechanisms in rugae development is slightly difference between regions.
Periodic patterning of iterative structures, the palatal rugae, develops by Turing-type reaction-diffusion mechanisms. Our results also indicate that the intricate molecular network is involved in the palatal rugae development, although two morphogens diffusing through a tissue finally create self-regulating periodic patterns.