ABSTRACT
Hertwig's epithelial root sheath (HERS) interacts with dental apical mesenchyme and guides development of the tooth root, which is integral to the function of the whole tooth. However, the key genes in HERS essential for root development are understudied. Here, we show that Axin1, a scaffold protein that negatively regulates canonical Wnt signaling, is strongly expressed in the HERS. Axin1 ablation in the HERS of mice leads to defective root development, but in a manner independent of canonical Wnt signaling. Further studies reveal that Axin1 in the HERS negatively regulates the AKT1-mTORC1 pathway through binding to AKT1, leading to inhibition of ribosomal biogenesis and mRNA translation. Sonic hedgehog (Shh) protein, a morphogen essential for root development, is over-synthesized by upregulated mTORC1 activity upon Axin1 inactivation. Importantly, either haploinsufficiency of the mTORC1 subunit Rptor or pharmacological inhibition of Shh signaling can rescue the root defects in Axin1 mutant mice. Collectively, our data suggest that, independently of canonical Wnt signaling, Axin1 controls ribosomal biogenesis and selective mRNA translation programs via AKT1-mTORC1 signaling during tooth root development.
INTRODUCTION
Organs such as teeth, hair follicles, and mammary glands share a similar initial developmental pattern, whereby epithelial cells undergo complex reciprocal interactions with underlying mesenchymal cells to form multiple specialized structures (Thesleff and Sharpe, 1997; Mikkola and Millar, 2006; Fuchs, 2007). The tooth organ is composed of crown and root parts (Chen et al., 1996; Huang et al., 2010). Despite a large number of studies focused on tooth crown formation (Zhang et al., 2005; Jussila and Thesleff, 2012), studies regarding tooth root development are rather limited.
The tooth root develops through dynamic interactions between the epithelial structure (Hertwig's epithelial root sheath, HERS) and the mesenchyme (apical dental papilla). During this process, HERS cells undergo proliferation, downward migration, apoptosis, and epithelial-mesenchymal transition (EMT). HERS guides the root development, providing positional information for the establishment of root shape as well as inducing the adjacent mesenchymal cells to differentiate into odontoblasts. Subsequent secretion of extracellular components, including collagen I, dentin matrix protein 1 (Dmp1) and dentin sialophosphoprotien (Dspp) by odontoblasts ensures the formation of root dentin, the major hard tissue in tooth roots (Huang and Chai, 2012; Huang et al., 2009; Liu et al., 2015). However, the key gene(s) in HERS and its regulatory mechanisms in tooth root development are rather understudied.
Axin1, a scaffold protein promoting the assembly of β-catenin destruction complex, negatively regulates canonical Wnt signaling (Nong et al., 2021; Qiu et al., 2024). Deletion of Axin1 in mice causes early embryonic lethality owing to defects in Wnt-mediated anterior-posterior axis determination (Clevers and Nusse, 2012; Zeng et al., 1997). A large number of studies have underscored the pivotal role of Axin1 in the modulation of many physiological and pathological processes through regulatory control of canonical Wnt signaling (Figeac and Zammit, 2015; Xie et al., 2022; Xu et al., 2022). Additionally, Axin1 also mediates other non-Wnt pathways, including p53 and transforming growth factor β (TGFβ) pathways (Li et al., 2009; Liu et al., 2006). To date, the association between Axin1 and tooth development is poorly understood.
Ribosomes are the key cellular machines for mRNA translation, a fundamental process in all forms of life. mTORC1, a well-established downstream component of PI3K-AKT signaling, regulates ribosomal biogenesis (production of different ribosomal components) and ribosomal function (mRNA translation) through two mediators, namely ribosomal protein S6 kinase (S6K) and eukaryotic translation initiation factor 4E (eIF4E)-binding protein 1 (4EBP1) (Ma and Blenis, 2009; Pelletier et al., 2018). mTORC1-mediated ribosomal biogenesis is known to have a role in translational control of selective mRNAs (Thoreen et al., 2012; Truitt et al., 2015). This set of translated mRNAs determines precise spatial-temporal expression pattern of crucial proteins in developmental settings such as angiogenesis, myogenesis and erythropoiesis (Oberkersch et al., 2022; Yan et al., 2021; Folgado-Marco et al., 2023). Ribosomopathies, caused by dysregulation of ribosomal production, manifest in a tissue-specific manner, and ribosomal dysfunction has a clear connection with craniofacial developmental defects (Kadakia et al., 2014; Lipton and Ellis, 2009; Weaver et al., 2015). Additionally, it has been suggested that ribosomal S6 kinase 2 (RSK2) and a key ribosomal biogenesis regulator, PAX9, are required for tooth development (Koehne et al., 2016; Farley-Barnes et al., 2020). Despite this, changes in ribosomal function have been largely unexplored within developmental contexts associated with teeth.
Here, we present the requirement and underlying mechanisms of Axin1 in the tooth root development of mice. We discover that Axin1 is highly expressed in HERS but absent from the apical dental mesenchyme during root development, and conditional inactivation of Axin1 in HERS of mice leads to defective root development, characterized by truncated roots and thin root dentin. Mechanistically, independently of canonical Wnt signaling, Axin1 in HERS binds to AKT1 and suppresses the phosphorylation level of AKT1 to modulate the appropriate translation of Shh via mTORC1-mediated selective translational programs of transcripts. AKT1-mTORC1 activity and Shh translation are enhanced upon Axin1 deficiency. Either haploinsufficiency of mTORC1 subunit Rptor or pharmacological inhibition of Shh can apparently rescue the root phenotype in Axin1 mutant mice. Our findings therefore elucidate how Axin1 regulates tooth root development via mTORC1-mediated translational regulation and aid better understanding of the precise mechanism involved in tooth root formation.
RESULTS
Axin1 is strongly expressed in the HERS during tooth root development and is required for tooth root development in mice
In the beginning of tooth root development, the growth of cervical loop cells into the deeper tissues results in the formation of HERS, which functions as an essential structure for the guidance of root formation (Li et al., 2017) (Fig. 1A). To screen the essential genes in HERS that are responsible for regulating tooth root development, microarray data reflecting the global expression profile of primary HERS cells (Li et al., 2019; retrieved from Gene Expression Omnibus, accession number GSE109622) were analyzed and the expression of Axin1 was found to be the highest among all factors associated with tooth or bone development (Fig. S1A).
To examine Axin1 expression during root development, immunofluorescence staining was performed in the mouse lower molars at postnatal day (PN) 4 and PN9. The results showed that Axin1 was highly expressed in the HERS but completely excluded from the adjacent mesenchyme (Fig. 1B; Fig. S1B). Given that the HERS is crucial for root development, the enriched expression of Axin1 in HERS provides a clue that Axin1 may play a specific role in root development.
To verify the functional significance of Axin1 in tooth root development, K14-Cre;Axin1fl/fl (Axin1cKO) mice were generated with specific ablation of Axin1 in HERS. Successful deletion of Axin1 in HERS was confirmed by immunofluorescence (Fig. S1C). X-ray of mouse mandibles showed that tooth root elongation in Axin1cKO mice was delayed obviously at PN14 and PN21, although the morphology of tooth crowns was similar to that in control mice (Fig. 1C). Micro-computed tomography (micro-CT), Hematoxylin and Eosin (H&E) staining, and quantitative analysis of the first mandibular molars were subsequently conducted. At PN14, the root furcation had commenced formation in the mandibular first molars of control mice, but was absent in those of Axin1cKO mice (Fig. 1D), which was also confirmed by histological observation showing that root dentin formation in the furcation region was barely seen in Axin1cKO mice compared with controls (Fig. 1E). The root lengths and root/crown ratios assessed by three-dimensional reconstruction of micro-CT images and measurement of the dentin widths on histological sections were significantly reduced in Axin1cKO mice at PN14 and PN21, although the crown lengths were not altered (Fig. 1F-I). Collectively, these data demonstrate that Axin1 expression in HERS is required for normal tooth root development.
Ablation of Axin1 in HERS disrupts cell behaviors
We next examined the effects of Axin1 on HERS cell proliferation, apoptosis and EMT in mice, which are important for establishment of root shape. Reverse transcription-qPCR (RT-qPCR) results showed elevated mRNA of Ki67 (Mki67) in the root apical tissues (Fig. S1D), and Ki67 immunofluorescence confirmed a significant increase of cell proliferation in the HERS of Axin1cKO mice at PN9 and PN12 (Fig. 2A,B). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) analysis revealed no significant difference in the number of apoptotic HERS cells between Axin1cKO and control mice (Fig. 2C,D). These results suggest that Axin1 has an inhibitory effect on proliferation but not apoptosis of HERS cells. EMT is a process through which cells lose some epithelial characteristics, such as tight junctions, and gain mesenchymal features (Leathers and Rogers, 2022). E-cadherin (cadherin 1; hereafter, Ecad) is a key protein of intercellular adhesion as well as an epithelial marker. Immunohistochemical staining of Ecad was used to reflect the biological process of EMT and outline the HERS structure at PN10 and PN14. The results showed that, compared with littermate control mice, HERS of Axin1cKO mice was larger and more bulging with a delayed dissociation, as well as an increased number of Ecad-positive epithelial cells (Fig. 2E,F). These results suggest that Axin1 contributes to the EMT process of HERS cells.
HERS-H1 cells, a HERS cell line established with rat tooth germs (Li et al., 2019), were used to validate the effects of Axin1 on cell proliferation, apoptosis and migration by small interfering RNA (siRNA) knockdown, which gave rise to ∼75% reduction of Axin1 mRNA as assessed by RT-qPCR (Fig. S2A). Cell proliferation was measured by 5-ethynyl-2'-deoxyuridine (EdU) assay, which showed significantly increased EdU-positive cells after Axin1 knockdown (Fig. S2B,C). TUNEL analysis revealed no obvious difference in levels of apoptosis between the knockdown and control groups (Fig. S2D,E), consistent with the unaltered expression of the apoptotic markers Bax, Fas and Bcl2, as assayed by RT-qPCR (Fig. S2F). HERS-H1 cells undergo EMT when treated with TGFβ1 (Zhang et al., 2020; Li et al., 2019; Huang and Chai, 2012). To investigate whether Axin1 affects the EMT of HERS-H1, the expression of several genes related to tight junctions and EMT was quantified by RT-qPCR. The results showed that mRNA levels of the tight junction-related genes Ocln and Tjp1 and the epithelial marker gene Cdh1 were increased, whereas mRNA levels of the mesenchymal markers Vim and Fn1 were decreased in Axin1-knockdown HERS-H1 cells in the presence of TGFβ1 (Fig. S2G). Cell migration assays demonstrated a significantly decreased migratory distance as a result of Axin1 knockdown (Fig. S2H,I). These results further confirmed that Axin1 inhibits proliferation and adhesion but promotes the migration of HERS cells.
Ablation of Axin1 in HERS disturbs the differentiation of odontoblasts through a paracrine mode
Tooth root dentin is formed by odontoblasts, a specific type of differentiated dental mesenchyme cells (Li et al., 2017). The impaired dentin formation observed in Axin1cKO mice appeared to be the consequence of the deformed HERS as it has been reported that aberrant epithelium-derived signaling disturbs normal epithelial-mesenchymal crosstalk, thereby affecting mesenchymal cell behaviors (Li et al., 2017; Huang et al., 2009; Liu et al., 2015). We started with examining the expression levels of the odontoblast differentiation markers Dspp, Dmp1 and collagen I by RT-qPCR and immunofluorescence. The results showed a dramatic decrease of both mRNA and protein levels of these genes in Axin1cKO mice compared with control mice (Fig. 2G,H; Fig. S1E). These results indicate that Axin1 ablation in HERS hampers the differentiation of neighboring odontoblasts, leading to decreased root dentin deposition.
To confirm whether the impact on the differentiation of mesenchymal cells by the impaired HERS can be attributed to a paracrine mechanism, HERS-H1-conditioned medium (CM) was used to culture the primary mesenchymal cells from root apical region of mouse molars. After 5 days of culture in the CM from Axin1-knockdown HERS-H1 cells, the expression levels of the odontoblast marker genes Dspp, Dmp1 and Col1a1 were significantly decreased in the mesenchymal cells compared with those cultured with CM from the scramble siRNA-transfected cells (Fig. S2J). Alizarin Red staining showed that CM from Axin1-knockdown HERS-H1 cells impeded the mineralization ability of mesenchymal cells (Fig. S2K,L). Thus, we conclude that Axin1 deficiency in HERS impairs the differentiation of odontoblasts through epithelial–mesenchymal interactions via a paracrine mode.
Axin1 regulates root development independently of canonical Wnt signaling
Previous studies have emphasized the importance of Axin1 as a rate-limiting factor in the regulation of canonical Wnt signaling (Lee et al., 2003). We therefore investigated whether the activity of canonical Wnt signaling is involved in the regulation of root development by Axin1. The mRNA levels, detected by RT-qPCR, of the Wnt/β-catenin target genes Axin2, Myc and Ccnd1 were increased in root apical tissues from Axin1cKO mice (Fig. S3A). The protein levels of β-catenin and Axin2 displayed an apparent increase in HERS of Axin1cKO mice by immunofluorescence (Fig. S3B,C) and in Axin1-knockdown HERS-H1 cells by western blot (Fig. S4A). These results reveal that the canonical Wnt signaling is indeed strengthened in HERS cells with Axin1 deficiency or knockdown.
To determine whether the enhanced canonical Wnt signaling is responsible for the root defects in Axin1cKO mice, MSAB, a small-molecular inhibitor of Wnt signaling that promotes degradation of β-catenin, was used to treat Axin1cKO mice and their control littermates. Intraperitoneal injection of MSAB successfully reversed the increased mRNA levels of Axin2, Myc and Ccnd1 in the root apical tissues (Fig. S3A), but failed to rescue the developmental root defects (Fig. S3D-J). Although MSAB treatment partially reversed the increased numbers of Ki67-positive and Ecad-positive cells in the HERS of Axin1cKO mice (Fig. S3K-N), it failed to rescue the expression levels of the odontoblast markers Dspp and Dmp1 in Axin1cKO mice (Fig. S3O,P). Similar experiments with iCRT14, another Wnt inhibitor that inhibits β-catenin–TCF binding, confirmed that tooth root developmental defects in Axin1cKO mice are not associated with increased canonical Wnt signaling (Fig. S3Q-V).
To exclude a possible correlation between the tooth root developmental defects and the upregulated canonical Wnt signaling, HERS-H1 cells with Axin1 knockdown were treated with MSAB. Although cell proliferation was partially rescued, as shown by EdU immunofluorescence (Fig. S4B,C), cell migration remained at a relatively slower speed (Fig. S4F,G). TUNEL analysis revealed no significant alteration in the number of apoptotic HERS cells after MSAB treatment (Fig. S4D,E). The differentially expressed mRNA levels of the tight junction- and EMT-related genes were not rescued either (Fig. S4H). Moreover, similar to CM from Axin1-knockdown HERS-H1 cells, CM from Axin1-knockdown HERS-H1 cells with additional MSAB treatment retained the negative impact on the differentiation and mineralization ability of mesenchymal cells shown by mRNA levels of the odontoblast marker genes Dspp, Dmp1 and Col1a1 and formation of mineralized nodules, respectively (Fig. S2J-L).
Taken together, the above in vivo and in vitro experiments strongly support the suggestion that Axin1 regulates tooth root development independently of canonical Wnt signaling.
Axin1 negatively regulates ribosomal biogenesis
To further unveil the mechanism by which Axin1 regulates tooth root development, RNA sequencing (RNA-seq) was performed using the root apical tissues from Axin1cKO and control mice. Differential gene expression profiles as well as Gene Ontology (GO) and KEGG enrichment analyses were carried out. We identified 1124 differently expressed genes (DEGs) in Axin1cKO versus control mice, among which Dspp, Dmp1 and Col1a1 were all significantly downregulated, consistent with the phenotype of root defects (Fig. 3A; Table S1). GO analysis of DEGs demonstrated extensive alterations in translation processes, rRNA processing, and formation of translation initiation complex, and KEGG enrichment analysis and gene set enrichment analysis (GSEA) showed that genes related to ribosomes were significantly enriched in the upregulated gene set (Fig. 3B-D).
Ribosomes are complex molecular machines made of ribosomal RNAs (rRNAs) and ribosomal proteins (RPs), ensuring protein synthesis in living cells (Pelletier et al., 2018). Many ribosomal genes in Axin1cKO mice showed an upward expression in the sequencing data (Fig. 3E), of which the upregulated expression of Rps6 and Rpl5 as two examples were verified by immunofluorescence (Fig. 3F,G). In HERS-H1 cells with Axin1 knockdown, RT-qPCR verified the significant upregulation of multiple ribosomal genes including Rps6, Rps10, Rps12, Rpl5 and Rpl12 (Fig. 3H). Moreover, western blot confirmed that the protein levels of several RPs were upregulated, including Rpl5, Rps6 and Rps12 (Fig. 3I; Fig. 4C). All these results indicate enhanced ribosomal biogenesis and implicate a possible mechanism related to mRNA translation in the root developmental defects of Axin1cKO mice.
Axin1 negatively regulates ribosomal biogenesis and protein synthesis through inhibition of AKT1-mTORC1 signaling
The serine/threonine kinase mTOR is a master regulator of mRNA translation (Ma and Blenis, 2009). The kinase AKT is the main upstream activator of the mTORC1 (Glaviano et al., 2023), and KEGG analyses showed that upregulated genes in Axin1cKO mice were enriched in the PI3K-AKT signaling pathway (Fig. 3C). Considering that the AKT-mTORC1 signaling pathway regulates mRNA translation initiation and ribosomal biogenesis mainly through S6K and 4EBP1 (Pelletier et al., 2018; Neasta et al., 2011; Wu et al., 2018), the levels of phosphorylated AKT1 (p-AKT1), p-4EBP1 and p-S6K were examined in Axin1cKO mice by immunofluorescence. As a substrate of p-S6K (Jenö et al., 1988), the level of p-Rps6 was also examined. The results showed that p-AKT1, p-4EBP1, p-S6K and p-Rps6 were significantly increased in the HERS of Axin1cKO mice compared with control mice (Fig. 4A,B). In addition, when Axin1 was knocked down, western blot results showed an apparent upregulation of p-AKT1, p-4EBP1, p-S6K and p-Rps6 levels in HERS-H1 cells (Fig. 4C). These results suggest that Axin1 has an inhibitory effect on the phosphorylation of AKT1 and mTORC1 activity (shown by protein levels of p-4EBP1, p-S6K and p-Rps6). When Axin1-silenced cells were treated with MK2206 (a pan-AKT inhibitor), the upregulated p-AKT1 as well as mTORC1 activity were reversed (Fig. 4H). Inhibition of mTORC1 activity with two mTORC1 inhibitors, rapamycin (Rapa) or Torin1, failed to attenuate increased AKT1 phosphorylation attributed to Axin1 knockdown, confirming that AKT1 acts upstream of mTORC1 in HERS-H1 cells. Therefore, we conclude that Axin1 negatively regulates the AKT1-mTORC1 signaling axis in the HERS.
To examine whether Axin1 regulates mRNA translation via mTORC1, HERS-H1 cells were treated with Torin1 following Axin1 knockdown and then tested by protein synthesis assays and polysome profiling. Puromycin (Puro) was used to label nascent peptides for investigating protein synthesis. The results showed that Axin1 knockdown enhanced global protein synthesis, which was reversed by Torin1 treatment (Fig. 4D). Polysome profiling experiments showed that polysomal distribution of the mRNAs, reflecting the translation efficiency, was significantly increased after Axin1 knockdown, and mTORC1 inhibition with Torin1 brought the upregulated proportion back to a relatively normal level (Fig. 4E). These results indicate that Axin1 represses mRNA translation by inhibiting mTORC1 activity.
It has been reported that mTORC1 is involved in regulating ribosomal biogenesis, including the synthesis of many ribosomal proteins (Ma and Blenis, 2009; Bouyahya et al., 2022). To investigate whether Axin1 regulates ribosomal biogenesis at translational levels via mTORC1, the distribution of some ribosomal-specific mRNAs in the polysome fractions was assessed. With Axin1 knockdown, the distribution of mRNAs for ribosome proteins, including Rps6, Rps10, Rps12, Rpl5 and Rpl12, rose significantly in the polysome fractions and reverted to normal levels upon Torin1 treatment (Fig. 4F). Meanwhile, Puro was incorporated into nascent peptides and nascent RPs were examined via co-IP with an anti-Puro antibody. We found that all of the investigated nascent RPs including Rps6, Rps12 and Rpl5 were obviously increased upon Axin1 knockdown, which was resumed by Torin1 treatment (Fig. 4G). Additionally, the upregulated levels of RPs including Rps6, Rps12 and Rpl5 upon Axin1 knockdown were restored by pan-AKT or mTORC1 inhibitor treatment (Fig. 4H). Therefore, Axin1 regulates mRNA translation of RPs and thereby ribosome biogenesis through AKT1-mTORC1 signaling.
Axin1 as a scaffold protein can bind to multiple proteins through direct interactions and modulate signaling pathways besides the canonical Wnt pathway (Li et al., 2009; Liu et al., 2006; Luo et al., 2003; Zhou et al., 2019; Xie et al., 2022). To explain the change in AKT1 phosphorylation level after Axin1 ablation or knockdown, we attempted to predict protein–protein interactions between Axin1 and AKT1 on the STRING website and found that, in addition to other members of the β-catenin destruction complex, AKT1 has a possible direct connection with Axin1 (Fig. S5A). To test this prediction, co-immunoprecipitation (co-IP) and in situ proximity ligation assay (PLA) were performed to identify and visualize the interaction between Axin1 and AKT1. The results revealed that Axin1 bound to AKT1 in both HERS-H1 cells and HERS cells in mouse molars, but this physical interaction disappeared upon Axin1 knockdown or ablation (Fig. 4I,J). According to previous reports, there are many steps involved in fully activating AKT1, one of which is relocating AKT1 and 3-phosphoinositide dependent protein kinase-1 (PDPK1, also known as PDK1) to membrane sites, followed by PDPK1 phosphorylating AKT1 (Manning and Toker, 2017; Glaviano et al., 2023). Therefore, we first checked the expression of AKT1 and PDPK1 in Axin1cKO mice by immunofluorescence. The results showed that the colocalization between AKT1 and PDPK1 was obviously increased in the HERS of Axin1cKO mice compared with control mice (Fig. S5B). Furthermore, when Axin1 was knocked down, the PDPK1 inhibitor GSK2334470 could resume the heightened interaction between AKT1 and PDPK1 as well as the increased level of p-AKT1 (Fig. S5C). These results indicate that Axin1 inhibits AKT1 phosphorylation by PDPK1 through physical interaction with AKT1.
Rptor haploinsufficiency in HERS rescues the tooth root defects of Axin1cKO mice
To test whether elevated mTORC1 activity is responsible for tooth root defects in Axin1cKO mice, we performed rescue experiments using K14-Cre;Axin1fl/fl;Rptorfl/+ (Axin1cKO;Rptorfl/+) mice with deficiency of one allele of Rptor, which encodes a key component of mTORC1. Immunofluorescence confirmed haploinsufficiency of Rptor in the HERS of Axin1cKO;Rptorfl/+ mice compared with Axin1cKO and control littermates (Fig. S6H). Elevated p-S6K and p-4EBP1 levels in the HERS of Axin1cKO mice were evidently inhibited with Rptor haploinsufficiency (Fig. 5A,B). Moreover, the increased expression of Rpl5 and Rps6 proteins in the HERS of Axin1cKO mice were also reversed, reflecting the rescue of ribosomal biogenesis owing to Rptor haploinsufficiency (Fig. 5C,D). Therefore, the elevated mTORC1 activity and mTORC1-mediated ribosomal biogenesis in Axin1cKO mice are reversed by Rptor haploinsufficiency. Next, X-ray, micro-CT and H&E staining were performed. The results showed that tooth roots in Axin1cKO;Rptorfl/+ mice exhibited similar morphology and size to those in control mice, revealing that haploinsufficiency of Rptor rescued the tooth root defects in Axin1cKO mice (Fig. 5E-K). Immunostaining showed that the enhanced cell proliferation and the increased number of Ecad-positive cells in HERS of Axin1cKO mice were partially restored by haploinsufficiency of Rptor (Fig. 5L,M; Fig. S6I,J). Ki67 immunostaining, cell migration experiments and RT-qPCR illustrated that mTORC1 inhibition significantly rescued the increased proliferation and reduced migration of HERS-H1 cells caused by Axin1 knockdown (Fig. S6A,B,E-G). TUNEL analysis showed no significant difference in the number of apoptotic HERS cells after mTORC1 inhibition (Fig. S6C,D). Notably, the expression of the odontoblast differentiation markers Dspp and collagen I also returned to relatively normal levels with Rptor haploinsufficiency (Fig. 5N,O). When wild-type mice were intraperitoneally injected with the mTORC1 inhibitor Rapa, which successfully inhibited the activity of mTORC1 signaling in vivo, they developed root developmental defects similar to Axin1cKO mice shown by X-ray, micro-CT and H&E staining (Fig. S8), which further indicated a role of mTORC1 in the tooth root defects of Axin1cKO mice.
Axin1 inhibits Shh translation through suppression of mTORC1 activity
mTORC1 is a master regulator of translation and mTORC1 activation is known to promote the translation initiation of selective group of mRNAs (Ma and Blenis, 2009). KEGG enrichment analysis showed that the Hedgehog (Hh) signaling pathway was significantly enriched in the upregulated genes in Axin1cKO mice (Fig. 3C). Recent studies have reported that an optimal level of sonic hedgehog (Shh) signaling is important for tooth root development, as either inhibition or activation of Shh signaling can hinder root elongation (Huang et al., 2010; Liu et al., 2015). We firstly investigated the expression of Shh and Axin1 in the whole tooth germ by immunofluorescence and found they were colocalized in HERS (Fig. S7A). We next examined the changes in Shh signaling pathway in the HERS of Axin1cKO mice and in HERS-H1 cells with Axin1 knockdown. Immunofluorescence, western blot and RT-qPCR experiments revealed an upregulation of both protein and mRNA levels of Gli1 and Ptch1 (Fig. 6A-G), two well-established readouts of Shh signaling activity. However, the mRNA level of Shh exhibited no significant change upon Axin1 deletion or knockdown (Fig. 6F,G), despite the Shh protein level being upregulated (Fig. 4H; Fig. 6E). Therefore, it is reasonable to infer that, upon Axin1 deletion or knockdown, translation of Shh mRNA may be increased leading to increased Shh protein and thus Shh signaling activity.
To confirm the effect of Axin1 on Shh mRNA translation, we tested Shh mRNA distribution in the polysome fractions. RT-qPCR results showed that polysome-incorporated Shh transcripts were higher upon Axin1 knockdown, but Gapdh and Bmp4 transcripts (as controls) were not (Fig. 6H). Puro-labeled nascent Shh protein was also analyzed by immunoprecipitation with an anti-Puro antibody. The results showed a significant increase of nascent Shh protein upon Axin1 knockdown (Fig. 6I). These results confirm that Axin1 inhibits Shh mRNA translation. However, the increased protein level of Shh after Axin1 knockdown returned to a normal level after the inhibition of either mTORC1 or pan-AKT activity (Fig. 4H), suggesting that elevated AKT1-mTORC1 signaling axis is probably responsible for the upregulated Shh mRNA translation. Accordingly, mTORC1 inhibition restored the increased Shh mRNA incorporation and nascent Shh protein to a relatively normal level (Fig. 6H,I), confirming that Axin1 mediates the regulation of Shh mRNA translation by mTORC1.
mTORC1 has been reported to regulate translation of a subset of transcripts that carry a specific stretch of pyrimidines known as the 5′ terminal oligopyrimidine (TOP) motif or TOP-like motif (Thoreen et al., 2012). A relaxed TOP-like motif (5′-ttaaaatcaggctctttttgt...) was found in Shh 5′UTR. We then generated constructs with wild-type (WT) Shh 5′UTR or Shh 5′UTR with presumed TOP-like motif deleted (DEL) as upstream of a firefly luciferase reporter (GenBank: U47298). Renilla luciferase reporter (GenBank: AF025844) was used as an internal control (Fig. 6J). After these reporters were transfected into HERS-H1 cells, Axin1 knockdown caused a significant increase in the relative luciferase activity in a mTORC1-dependent manner, which was eliminated by deletion of the TOP-like motif in the Shh 5′UTR (Fig. 6K).
All these above in vitro results indicate that Axin1 negatively regulates translation of a set of proteins including Shh through mTORC1 signaling, and the in vivo evidence that Rptor haploinsufficiency in mice inhibited the elevated Shh signaling activity in Axin1cKO mice also supports this conclusion (Fig. 6L,M). Furthermore, the influence of the canonical Wnt pathway on Shh was excluded (Fig. S7B).
Axin1 regulates root development through Shh signaling
To further test whether root developmental defects in Axin1cKO mice was associated with upregulated Shh signaling, a small-molecular Hh inhibitor, vismodegib, was used, which reduced the root length of wild-type mice (Fig. S8), as previously reported (Liu et al., 2015). The activity of Shh signaling in Axin1cKO mice and in HERS-H1 cells with Axin1 knockdown was inhibited to a normal level by vismodegib treatment (Fig. S7C,K).
The Axin1cKO mice intraperitoneally injected with vismodegib developed normal tooth roots as shown by the results of X-ray, micro-CT and H&E staining (Fig. 7A-G). Both in vivo and in vitro experiments indicated that the increased cell proliferation in HERS cells was not apparently reversed (Fig. S7D,E,L,M). TUNEL analysis showed no significant difference in the number of apoptotic HERS cells after vismodegib treatment (Fig. S7F,G). The increased Ecad-positive HERS cells in Axin1cKO mice were partially rescued by treatment with Hh inhibitor (Fig. S7N,O). Similarly, cell migration was partially rescued in vitro (Fig. S7H,I). The differential mRNA levels of tight junction- and EMT-related genes with Axin1 knockdown was completely reversed by vismodegib in HERS-H1 cells (Fig. S7J).
Notably, the expression levels of odontoblast differentiation genes returned to a relatively normal level with vismodegib treatment in vivo (Fig. 7H,I). To validate whether ablation of Axin1 in HERS cells influence the differentiation of mesenchymal cells through Shh signaling via a paracrine mode, CM from Axin1 knockdown HERS-H1 cells was supplemented with vismodegib to culture the primary mesenchymal cells from the apical papilla region. Compared with mesenchymal cells cultured in CM from Axin1 knockdown HERS-H1 cells, addition of vismodegib significantly rescued the impeded mineralization as well as the expression levels of the odontoblast marker genes Dspp, Dmp1 and Col1a1, as shown by Alizarin Red staining assay and RT-qPCR (Fig. 7J-L).
Furthermore, we also overactivated Shh signaling using Shh recombinant protein or SAG (a small molecular Shh-Smoothened agonist) in HERS-H1 cells and apical dental mesenchymal cells. The results showed that the cell behaviors of HERS-H1 cells exhibited similar changes to those with Axin1 knockdown (Fig. S9A-E). The differentiation and mineralization ability of mesenchymal cells treated by Shh protein or SAG were reduced (Fig. S9F-H), similar to mesenchymal cells cultured with CM from HERS-H1 cells with Axin1 knockdown. Therefore, we conclude that Axin1 in HERS modulates epithelial–mesenchymal interactions through Shh signaling.
DISCUSSION
Tooth root development serves as an excellent model to study how epithelial–mesenchymal interactions contribute to organogenesis (Li et al., 2017; Kumakami-Sakano et al., 2014). Although HERS is considered to be responsible for guiding the differentiation of underlying mesenchymal cells and subsequent root formation and elongation, the key gene in HERS during tooth root development is rather understudied. Here, we show that HERS-specific inactivation of Axin1 causes obvious root developmental defects. Our in vivo and in vitro studies demonstrate that Axin1 inhibits proliferation and adhesion, as well as promotes migration of HERS cells. Ablation of Axin1 in HERS disrupts the odontoblastic differentiation of adjacent mesenchymal cells through a paracrine mode, which strengthens the theory that HERS plays a guiding role during tooth root development (Fig. 7M).
It is known that HERS guides root development; therefore, we screened expression of Axin1 in HERS and confirmed high expression in this region. Studies on the transcriptional regulation of the Axin1 gene have identified several transcription factors actively involved in promoting its transcription in different contexts, including GATA4 (Suthon et al., 2022), RUNX1 (Chimge et al., 2016), C/EBPβ (Park et al., 2018) and EGR1 (Zhang et al., 2016). Further experiments are needed to test whether these or other factors participate in the regulation of Axin1 expression in HERS.
To explore the exact role of Axin1 in HERS, we examined the phenotype of tooth root development in Axin1cKO mice. It has been reported that HERS induces the adjacent mesenchymal cells to differentiate into odontoblasts and contributes to the pool of cementoblasts in the root via EMT (Huang et al., 2009, 2010). Because previous studies have shown that direct contact is required for HERS in regulating odontoblast differentiation and abnormal EMT of HERS disturb the interruption of epithelial root sheath, resulting in abnormal differentiation of odontoblasts (Cerri and Katchburian, 2005; Huang and Chai, 2012), here, in addition to cell proliferation and apoptosis, we also characterized the effect of Axin1 on the EMT of HERS cells. We found that ablation of Axin1 in HERS disrupted the EMT process and odontoblast differentiation.
It is well established that Axin1 negatively regulates the canonical Wnt signaling pathway, which controls many physiological and pathological processes (Qiu et al., 2024; Albrecht et al., 2021). Canonical Wnt signaling activity is indeed elevated in Axin1 deficient or knockdown HERS cells. However, such enhanced canonical Wnt signaling does not account for the tooth root defects in Axin1cKO mice as both Wnt signaling inhibitors, MSAB and iCRT14, failed to reverse the phenotype. However, it has been observed in vivo that stabilization of β-catenin in HERS leads to interrupted root elongation (Yang et al., 2021). This phenotypic discrepancy may be due to the expression pattern of Axin1 in the HERS and different degrees of Wnt signaling activation. Considering that Axin2 is also expressed in HERS and it is a functionally identical isoform of Axin1 (Clevers and Nusse, 2012), Axin2 may partially compensate for regulation of Wnt signaling activity in the absence of Axin1. Similarly, it has been reported that knockdown of Axin1 in RKO cells (human colorectal adenocarcinoma cells) is buffered by Axin2 (Moshkovsky and Kirschner, 2022).
In the present study, it was tentatively inferred by analyses of our sequencing results that Axin1 negatively regulates ribosomal biogenesis. Ribosomal biogenesis requires the production of different ribosomal components and plays pivotal roles in many physiological and pathological processes, including cell fate determination, stem cell maintenance, organ regeneration, and tumorigenesis (Saba et al., 2021). We next experimentally confirmed the upregulation of ribosomal components at both the mRNA and protein levels. Considering that the cellular processes essential for growth and development are highly dependent on protein synthesis mediated by ribosomes, we assumed that Axin1 is linked to certain signaling pathways involved in translational regulation.
Axin1 has been reported to function in non-Wnt pathways, including the c-Jun N-terminal kinase (c-JNK), TGFβ, p53, and AMP-activated protein kinase (AMPK) pathways (Li et al., 2009; Liu et al., 2006; Luo et al., 2003; Salahshor and Woodgett, 2005; Lin and Hardie, 2018). Recently, Axin1 was proposed to regulate catabolic-anabolic transition by coordinating AMPK activation and inactivation of mTORC1 (Zhang et al., 2013, 2014). Combined with analyses of sequencing results, we subsequently identified the involvement of the AKT1-mTORC1 signaling axis. This is because mTOR regulates the synthesis of ribosomal components (ribosomal biogenesis) at both the transcriptional and translational levels (Ma and Blenis, 2009). A previous report showed that overexpression of Axin1 in embryonal carcinoma cells inhibits AKT-mTOR signaling pathway (Xu et al., 2017), which is consistent with our observation that the phosphorylation of AKT1 and mTORC1 pathway downstream target proteins were increased significantly with Axin1 knockdown or ablation. The amount of nascent proteins labeled by Puro and the distribution of mRNAs in the polysome fractions detected by polysome profiling reflect the promotion of ribosomal function by Axin1 knockdown in HERS-H1 cells. It was further proved that Axin1 regulates ribosomal function via mTORC1 activity as inhibition of the hyperactive mTORC1 activity rescued the enhanced ribosomal function attributed to Axin1 knockdown. The interaction between Axin1 and AKT1 was then verified both in vitro and in situ of tissues. We observed that Axin1 inhibited the phosphorylation of AKT1. Thus, our studies reveal the function of Axin1 in regulating protein synthesis, beyond its role as a Wnt signaling regulator.
The root developmental defects in Axin1cKO mice are significantly rescued by the deletion of one allele of Rptor. Our study uncovers that an optimal level of mTORC1 signaling activity is required for root formation. Because mTORC1 signaling controls the synthesis of a set of proteins, we attempted to identify the key regulatory protein involved in the tooth root defects of Axin1cKO mice that was regulated by mTORC1 signaling in HERS. The enrichment of the Hedgehog signaling pathway in KEGG analysis attracted our attention. We confirmed the enhancement of Shh signaling activity in the HERS of Axin1cKO mice, as well as a relatively normal activity of Shh signaling pathway in the HERS of Axin1cKO;Rptorfl/+ mice. Furthermore, we provided evidence that Axin1 inhibits Shh protein expression at the translational level through mTORC1 signaling. Inhibition of Hh signaling rescued the decreased root length of Axin1cKO mice, indicating that Axin1 regulates root development through Shh signaling. Notably, odontoblastic differentiation of the neighboring mesenchymal cells was also rescued. Therefore, Axin1 regulates mTORC1 signaling, and hence controls Shh translation and epithelial–mesenchymal crosstalk. Shh has been reported to have a dual effect on retinal ganglion cell axonal growth, acting as a positive factor at low concentrations and a negative factor at high concentrations (Kolpak et al., 2005). Likewise, we propose that the hyperactive Shh activity in the root region hampers the odontoblast differentiation of mesenchymal cells.
Collectively, our findings demonstrate that Axin1 regulates tooth root development by suppressing AKT1-mediated mTORC1 activation and Shh translation in HERS cells in a canonical Wnt signaling-independent fashion, revealing a regulatory mechanism underlying tooth root formation and regeneration. This epithelium-derived spatiotemporal regulation of signaling events by Axin1 may also be utilized during other organ development and contribute to a better understanding of physiological and pathological ectodermal organogenesis.
MATERIALS AND METHODS
Animal experiments
The animal studies were approved by the Animal Welfare and Ethics Committee of the School and Hospital of Stomatology at Wuhan University. All animal studies complied with relevant guidelines and ethical regulations. Conditional Axin1 loss-of-function mutant mice were generated by intercrossing double heterozygous for a floxed Axin1 allele and the K14-Cre transgenic allele (Axin1flox/+; K14-Cre+/−) with homozygous floxed Axin1 (Axinflox/flox) or homozygous floxed Axin1 and heterozygous floxed Rptor (Axinflox/flox;Rptorflox/+) mice. The Axin1flox/flox mice were kindly provided by Professor Di Chen (Xie et al., 2022). The Rptorflox/flox mice were kindly provided by Prof. Weiguo Zou (Han et al., 2019). Genomic DNA extracted from mouse tails was used for genotyping. Mice were housed under specific pathogen-free conditions at room temperature (20-23°C, ∼50% humidity) with a 12 h light–dark cycle set with access to water and food. The Hh inhibitor GDC-0449 (MedChemExpress, HY-10440; 24 mg kg−1; once a day), canonical Wnt inhibitor MSAB (MedChemExpress, HY-120697; 20 mg kg−1; once a day), iCRT14 (MedChemExpress, HY-16665; 20 mg kg−1; once a day) and Rapa (MedChemExpress, HY-10219; 2 mg kg−1; once a day) were dissolved in dimethylsulfoxide (DMSO; Sigma-Aldrich, D2650) and injected intraperitoneally from PN3 to PN9 or PN13 for the inhibitor assay in vivo. Control pups were injected with the same dosage of DMSO alone.
Tissue processing, radiographic assay, histological staining and immunostaining
For histological analysis, mice at defined ages were euthanized, and their mandibles were carefully dissected. The dissected mandibles were fixed in neutral buffered formalin for 24 h and then transformed into PBS (pH 7.4), before an initial X-ray analysis using a Bruker X-ray cabinet. Dissected tissues were scanned at a spatial resolution of 5 μm and a medium resolution using an in vivo micro-CT system (SkyScan 1276, Bruker). The 3D images of the first molars were reconstructed from the scan slice data. The dissected mandibles were then decalcified in 10% ethylenediaminetetraacetic acid (EDTA) for 20-30 days, embedded in paraffin after being dehydrated through a series of graded ethanol, and cut into 5-μm-thick sections using a microtome (Leica RM2235). For H&E staining, the sections were rehydrated in graded ethanol after being deparaffinized in xylene and stained with Hematoxylin for 20 s, followed by Eosin for 45 s. After staining, sections were dehydrated through 100% ethanol and cleared in xylene.
For immunohistochemistry staining, the rehydrated sections were first treated with 3% hydrogen peroxide for 15 min to reduce endogenous peroxidase activity. For antigen retrieval, the slices were immersed in sodium citrate buffer for 15 min at 99°C and blocked with 5% bovine serum albumin for 30 min afterwards. The primary antibodies were incubated overnight at 4°C and secondary antibody was applied for 1 h at room temperature the following day and then the signal was detected using a DAB Staining Kit (Beyotime, P0203). The TUNEL assay was performed using an assay kit (G3250; Promega) according to the manufacturer's instructions. For immunofluorescence staining, the rehydrated sections were processed for antigen retrieval and then incubated with the indicated primary antibodies overnight at 4°C and secondary antibodies for 1 h at room temperature the following day. Nuclei were counterstained with DAPI then the sections were observed under an Olympus BX61 fluorescence microscope. Images were quantified using ImageJ (version 1.53a).
Antibodies
The following primary antibodies were used for immunostaining and western blot in this study: rabbit anti-Axin1 (Cell Signaling Technology, C76H11, 2087; 1:50 for immunostaining, 1:1000 for western blot), rabbit anti-Actin (Abclonal, AC038; 1:10,000 for western blot), rabbit anti-Ki67 (Novus Biologicals, NB110-89717; 1:100 for immunostaining), rabbit anti-Dspp (Novus Biologicals, NBP2-92546; 1:100 for immunostaining), rabbit anti-Dmp1 (Novus Biologicals, NBP1-89484; 1:100 for immunostaining), rabbit anti-collagen I (Abclonal, A1352; 1:100 for immunostaining), rabbit anti-p-Rps6 S240/S244 (Abcam, ab215241; 1:100 for immunostaining, 1:5000 for western blot), rabbit anti-Rps6 (Abcam, ab225676; 1:200 for immunostaining, 1:1000 for western blot), rabbit anti-p-Akt1 S473 (Abclonal, AP0637; 1:100 for immunostaining, 1:1000 for western blot), rabbit anti-Akt1 (Abclonal, A11016; 1:1000 for western blot), rabbit anti-p-4EBP1 T37/46 (Abclonal, AP0030; 1:100 for immunostaining, 1:1000 for western blot), rabbit anti-4EBP1 (Abclonal, A19045; 1:1000 for western blot), rabbit anti-p-S6K T421/S424 (Abclonal, AP0502; 1:100 for immunostaining, 1:1000 for western blot), rabbit anti-S6K (Abclonal, A16658; 1:1000 for western blot), rabbit anti-Rpl5 (Abclonal, A1977; 1:50 for immunostaining, 1:500 for western blot), rabbit anti-Rps12 (Abclonal, A5890; 1:1000 for western blot), rabbit anti-Shh (Abclonal, A12695; 1:100 for immunostaining, 1:2000 for western blot), rabbit anti-Gli1 (Genetex, GTX106207; 1:100 for immunostaining, 1:1000 for western blot), rabbit anti-Ptch1 (Abclonal, A0826; 1:100 for immunostaining, 1:1000 for western blot), rabbit anti-Axin2 (Proteintech, 20540-1-AP; 1:100 for immunostaining), rabbit anti-β-catenin (Genetex, GTX101435; 1:100 for immunostaining), rabbit anti-E-cadherin (Cell Signaling Technology, 24E10, #3195; 1:200 for immunostaining), mouse anti-PDPK1 (ZenBio, 221339; 1:500 for western blot).
Secondary antibodies used were purchased from Antgene (goat anti-mouse-HRP, ANT019, 1:5000; goat anti-rabbit-HRP, ANT020, 1:5000).
Cell cultures, siRNA transfection and cell staining
HERS-H1 cells were cultured in epithelial cell medium (ScienCell, 4101SC) as described previously (Li et al., 2019). Recombinant human TGFβ1 (Yeasen Biotechnology) was added to the medium at 10 ng ml−1 to induce EMT of HERS-H1 cells. For siRNA experiments, HERS-H1cells were transfected with Axin1 siRNA (GenePharma) with the recommended amounts of Lipofectamine 3000 (Thermo Fisher Scientific). The efficiency of the gene knockdown was assessed. The following Axin1 siRNA oligos were used: (#1) sense sequence 5′-CCUGCUGGACUUCUGGUUUTT-3′ and anti-sense sequence 5′-AAACCAGAAGUCCAGCAGGTT-3′; (#2) sense sequence 5′-CCAUCUACCGAAAGUACAUTT-3′ and anti-sense sequence 5′-AUGUACUUUCGGUAGAUGGTT-3′. Treatment with MSAB (MedChemExpress, HY-120697; 10 μM, 16 h), Rapa (MedChemExpress, HY-10219; 25 nM, 12 h), Torin1 (MedChemExpress, HY-13003; 250 nM, 4 h), MK2206 (MedChemExpress, HY-10358; 10 μM, 12 h), vismodegib (MedChemExpress, HY-10440; 20 μM, 12 h), GSK2334470 (MedChemExpress, HY-14981; 1 μM, 12 h) or DMSO as control were performed after transfected with Axin1 siRNA or the negative control (Scramble; GenePharma) for 36 h. Treatment with SAG (5 μM, 48 h) or Shh (5 μg ml−1, 48 h) was performed with their respective vehicles (DMSO or PBS). EdU immunolabeling was performed using the Click-iT EdU Imaging Kit (Beyotime). The TUNEL assay was performed using an assay kit (G3250; Promega) according to the manufacturer's instructions. Cells were fixed in 4% paraformaldehyde for 40 min for cell staining or collected for protein or RNA extractions.
For primary mesenchymal cells, apical papilla tissues of first mandibular molars from rat at PN1 were isolated under an Olympus SXZ-16 stereomicroscope and shredded into tissue fractions. Then, the tissues were incubated with 3% collagenase I (Sigma-Aldrich) at 37°C for 1 h. After that, the digested tissues were transferred to high-glucose DMEM medium (Hyclone Laboratories) containing 10% fetal bovine serum (Gibco), 100 U ml−1 penicillin and 100 μg ml−1 streptomycin (Gibco) at 37°C with a humidified atmosphere of 5% CO2. CM was harvested from the supernatants of Axin1-knockdown HERS-H1 cells for 36 h (or then treated with 10 μM MSAB for 16 h or 20 μM vismodegib for 12 h) after removing the possible cell debris by centrifugation (800 g for 10 min) and were mixed with fresh culture medium supplemented with 5 mmol l−1 β-glycerophosphate (Sigma-Aldrich), 50 μg ml−1 ascorbic acid (Sigma-Aldrich) and 100 nmol l−1 dexamethasone (Sigma-Aldrich) to produce a conditioned differentiation medium (siAxin1#1-CM or siAxin1#1+MSAB-CM). For odontoblast-like differentiation induction, mesenchymal cells were cultured in the conditioned differentiation medium for 5 days. The culture solution was half exchanged at 24 h and fully exchanged at 60 h. For the mineralization assay, the fixed cells were stained with 2% Alizarin Red S (pH 4.2, Sigma-Aldrich) for 30 min with gentle shaking. Alizarin Red staining was extracted with cetylpyridinium chloride and measured with a spectrophotometer at 562 nm for quantification analysis.
RNA extraction and RT-qPCR
RNA isolation was performed using TRIzol (Yeasen Biotechnology) according to the manufacturer's recommendations. Up to 1 μg of RNA was subjected to cDNA synthesis using a qScript cDNA Synthesis kit (Yeasen Biotechnology). qPCR with reverse transcription analyses were performed using SYBR Green Master Mix (Yeasen Biotechnology) and amplification was performed on a CFX instrument (Bio-Rad), with gene-specific primers (listed in Table S2).
RNA-seq and data analysis
RNA from the developing apical complex tissues was extracted using TRIzol (Yeasen Biotechnology), followed by purification using a RNeasy Mini Kit (QIAGEN). RNA-seq was performed using a NovaSeq6000 and High Output v2 kit (Illumina), and the clean data were analyzed by Anoroad Gene Technology (Beijing, China). Three biological replicates were sequenced for each sample. A difference was considered significant if the P-value was less than 0.05. Volcano plots and heatmaps were generated by an online platform for data analysis and visualization (https://www.bioinformatics.com.cn). GO and KEGG enrichment analysis was conducted with DAVID. GSEA was conducted using GSEA version 4.0.1. RNA-seq data are presented in Table S1.
Protein extraction, western blot and co-IP
To prepare cell lysate, HERS-H1 cells were washed with ice-cold PBS, lysed in NP40 lysis buffer supplemented with the protease inhibitor cocktail (1:100; Abcam) and then collected on ice followed by centrifugation (13,000 g for 10 min). Protein concentration was quantified using the Enhanced BCA Protein Assay Kit (Beyotime).
For western blot, cell lysate was mixed with SDS loading buffer and denatured at 95°C. Next, samples were resolved on 12.5% SDS–PAGE and electrotransferred to a polyvinylidene difluoride (PVDF) membrane. Membranes were blocked with 5% dry skimmed milk in PBST (PBS containing 0.05% Tween-20) for 2 h at room temperature and incubated with primary antibodies overnight at 4°C. Membranes were washed with TBST (TBS containing 0.05% Tween-20) the next day and then incubated with secondary antibodies (goat anti-mouse-HRP, 1:1000, Antgene, ANT019; goat anti-rabbit-HRP, 1:1000, Antgene, ANT020). Detection was performed using the High-sig ECL Western Blotting Substrate (Tanon). Images were captured using an Amersham Imager 600, and densitometry was measured using Image Studio.
For co-IP, cell lysate was incubated with indicated antibodies at 4°C overnight with mixing. Protein A/G magnetic beads (TJ273976A; Thermo Fisher Scientific) were then added and incubated for another 1 h at 4°C. Then protein samples were washed in IP buffer five times (5 min each time), resuspended in SDS loading buffer and denatured at 95°C. Equivalent amounts of lysate was used as a positive control. Next, samples were then analyzed by western blot as described above.
Protein synthesis assay
Puro (10 μg ml−1; MedChemExpress, HY-B1743) was added to the cell incubation media and further incubated for 30 min. During this incubation time, newly synthesized proteins were labeled with Puro. Then, protein lysates were prepared and blotted. Puro-labeled proteins were identified on blots using a rabbit anti-Puro antibody (Abclonal, A23031).
Polysome profiling
Polysome profiling was conducted as previously described (Young-Baird et al., 2020). In brief, HERS-H1 cells were treated with 250 nM Torin1 or DMSO as control for 4 h after transfected with Axin1 siRNA or the negative control (Scramble; GenePharma) for 36 h as mentioned above. Cells were treated with 100 µg ml−1 cyclohexamide (CHX; MedChemExpress, HY-12320) 10 min before lysis, washed in ice-cold PBS plus 100 µg ml−1 CHX at 4°C and then lysed in polysome lysis buffer [0.3 M NaCl, 15 mM MgCl2, 15 mM Tris-HCl (pH 7.6), 1% Triton X-100, 0.1 U µl−1 RNAsin (Yeasen Biotechnology), 100 µg ml−1 CHX, 1 µg ml−1 heparin] at 4°C. Cells were centrifuged at 9500 g for 10 min at 4°C, and the resulting supernatant was layered onto a 5 ml continuous sucrose gradient [15-50% sucrose in 0.3 M NaCl, 15 mM MgCl2, 15 mM TrisHCl (pH 7.6), 100 µg ml−1 CHX, 1 µg ml−1 heparin, 0.1 U µl−1 RNAsin]. Gradients were centrifuged at 260,000 g for 2 h in an SW41-Ti rotor at 4°C, and then sampled using a Teledyne ISCO density gradient fractionation system with continuous monitoring at 254 nm. Fractions were collected (ten fractions per sample) and RNA was extracted with the RNeasy isolation kit (QIAGEN). PCR reactions were carried out as mentioned above. mRNA amounts in each fraction were plotted as a percentage of the total mRNA in the sample.
Dual luciferase reporter assay
A 314 bp portion of the Shh 5′UTR was synthesized by Miaolingbio and cloned into the KpnI/NheI site of the pGL3 promoter vector. The pGL3-basic plasmid containing the Shh 5′UTR region, the pGL3-basic luciferase plasmid, and the pRL-TK plasmid were purchased from Miaolingbio (Wuhan, China). Deletion of the putative TOP-like motif in Shh 5′UTR was conducted using Mut Express II Fast Mutagenesis Kit V2 (Vazyme Biotech) according to the manufacturer's instructions. The dual luciferase reporter assay was performed as described (Doyle et al., 2015). Briefly, HERS-H1 cells were treated with 250 nM Torin1 or DMSO as control for 4 h after transfected with Axin1 siRNA or the negative control (Scramble; GenePharma) for 36 h as mentioned above. Then the cells were co-transfected with the indicated plasmids into cells with Lipofectamine™ 3000 (Thermo Fisher Scientific). pRL-TK was applied as internal reference. After being transfected for 48 h, the relative fluorescence ratio of the cells was detected using the kit of Dual-Luciferase Reporter Assay System (Promega) and the Glomax Luminometer (Promega) according to the manufacturer's instructions.
In situ PLA
To visualize histologically the interaction between Axin1 and AKT1 in HERS-H1 cells and mouse molars, in situ PLA was performed using the Duolink kit (Sigma-Aldrich) as described in our previous study (Yuan et al., 2015). In brief, 5-µm-thick slices were incubated with rabbit anti-Axin1 antibody (Cell Signaling Technology, C76H11, 2087; 1:50) together with mouse anti-AKT1 antibody (Abclonal, A10605; 1:50) at 4°C overnight. Afterward, the slices were washed and incubated with the probe solution following the manufacturer's instructions. Slices subsequently went through a ligation reaction as well as an amplification reaction. Finally, slices were mounted with DAPI and observed under an Olympus BX51 fluorescence microscope with an Olympus DP70 digital camera.
Statistical analysis
Experiments were independently repeated and information on the number of biologically independent samples analyzed and the number of times experiments were performed is included in the figure legends. All statistical analyses were performed using GraphPad Prism software (version 8.0). Statistical significance in this study was determined by unpaired, two-tailed Student's t-test or one-way analysis of variance (ANOVA) with Bonferroni's post-test at *P<0.05, **P<0.01, ***P<0.001, as indicated in the figures. The results are represented as mean±s.d.
Acknowledgements
We thank Professor Hongbing Shu from Medical Research Institute, Wuhan University, and Professor Xiaorong Zhang from Institute of Biophysics, Chinese Academy of Sciences, for their helpful discussions and suggestions in the study. We acknowledge the Research Center for Medicine and Structural Biology, Wuhan University, for their assistance with micro-CT analysis.
Footnotes
Author contributions
Conceptualization: Z.Z., W.G., G. Yang, Z.C., D.C., Y.C., G. Yuan; Methodology: Z.Z., G. Yang, Z.C., D.C., Y.C., G. Yuan; Formal analysis: X.Z., H.H., Z.Z., Z.C., D.C., Y.C.; Investigation: X.Z., H.H.; Resources: Z.Z., W.G., G. Yang, Z.C., D.C.; Writing - original draft: X.Z., H.H.; Writing - review & editing: X.Z., H.H., Z.Z., W.G., G. Yang, Z.C., D.C., Y.C., G. Yuan; Supervision: G. Yang, Z.C., G. Yuan; Project administration: G. Yang, Z.C., G. Yuan; Funding acquisition: G. Yang, Z.C., G. Yuan.
Funding
This research was supported by grants from the National Natural Science Foundation of China (82370913, 82170914 to G. Yuan; 81570942 to G. Yang), the National Natural Science Foundation of China Key Program (82230029 to Z.C.), the Fundamental Research Funds for the Central Universities (2042022dx0003) and the Innovation Project of Wuhan Municipal Science and Technology Bureau (2023020201010170). Open Access funding provided by Wuhan University. Deposited in PMC for immediate release.
Data availability
RNA-seq data have been deposited in the Sequence Read Archive (SRA) database under accession number PRJNA1171776.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.202899.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.