ClC-7 Deficiency Impairs Tooth Development and Eruption

CLCN7 gene encodes the voltage gated chloride channel 7 (ClC-7) in humans. The mutations in CLCN7 have been associated with osteopetrosis in connection to the abnormal osteoclasts functions. Previously, we found that some osteopetrosis patients with CLCN7 mutations suffered from impacted teeth and root dysplasia. Here we set up two in vivo models under a normal or an osteoclast-poor environment to investigate how ClC-7 affects tooth development and tooth eruption. Firstly, chitosan-Clcn7-siRNA nanoparticles were injected around the first maxillary molar germ of newborn mice and caused the delay of tooth eruption and deformed tooth with root dysplasia. Secondly, E13.5 molar germs infected with Clcn7 shRNA lentivirus were transplanted under the kidney capsule and presented the abnormal changes in dentin structure, periodontal tissue and cementum. All these teeth changes have been reported in the patients with CLCN7 mutation. In vitro studies of ameloblasts, odontoblasts and dental follicle cells (DFCs) were conducted to explore the involved mechanism. We found that Clcn7 deficiency affect the differentiation of these cells, as well as the interaction between DFCs and osteoclasts through RANKL/OPG pathway. We conclude that ClC-7 may affect tooth development by directly targeting tooth cells, and regulate tooth eruption through DFC mediated osteoclast pathway.


Results
Local Clcn7 silencing impairs tooth morphogenesis and tooth eruption. Comparing to the control ( Fig. 1A,D,G,H), the enamel and dentin at day 7 post-injection became discontinuous and had uneven thickness in chitosan (CS)-Clcn7-siRNA group (Fig. 1B,C,E,F,I,J). On P17, the tooth with well developed crown and root was normally erupted in control group ( Fig. 1K-M,Q,R); while CS-Clcn7-siRNA treated group showed abnormal dentin, irregular tooth roots (Fig. 1N,O,P,S,T) and impacted tooth (Fig. 1N,T). Micro-CT scanning showed that the left maxillary first molar treated with CS-Clcn7-siRNA nanoparticles became impacted and its eruption was delayed in P17 group (Fig. 1U~W).

ClC-7 deficiency disrupts morphogenesis of tooth in kidney capsule. After 3 days of lentivirus
infection, obvious GFP fluorescence could be detected in the infected tooth germs, meanwhile, the Clcn7 mRNA level was decreased about 70% in the Clcn7-shRNA treated group, compared to the negative-shRNA group (P = 0.0001, data not shown). The teeth were well developed with clear profiles including normal cusps and roots in both blank control groups ( Fig. 2A,D,G,J) and negative-shRNA group (Fig. 2B,E,H,K,M). The abnormal morphological alterations were found in Clcn7-shRNA group, such as short and/or crooked roots, abnormal cusps and apical foramen (Fig. 2C,I,N). Part of predentin became thicker and the odontoblasts lost their regular column shape (Fig. 2F). The imposing characteristics of AZON staining in Clcn7-shRNA group were demonstrated as Dentin became disrupted and uneven (J, black arrow) and the boundary between dentin and predentin was indiscernible (F, asterisk), and the disorganized odontoblasts were found (J, asterisk). In P17 CS-Clcn7-siRNA group, the tooth was not fully erupted (N,T). The malformed tooth showed the abnormal cusps and short roots (N, black arrow; T) or disappeared root (N, asterisk). Abnormal and irregular fibrous tissues surrounded tooth instead of bone tissue (N, black triangle). The disorganized dentin and pulp cavity were observed (O,P asterisk). The shape of the tooth crown was irregular and the root dentin was disrupted (S, black arrow) and uneven. The continuity and integrity of cervical loop and Hertwig's epithelial root sheath was disappeared (S, black arrow). Figure  Impact of ClC-7 on the differentiation of DFCs. We confirmed the general mesenchymal characteristics of DFCs by positive immunocytochemical staining for anti-vimentin and negative staining for anti-pan cytokeratin ( Supplementary Fig. 1).
The osteogenic induction with treated dentin matrix medium (TDMM) upregulated the mRNA expression of Alp, Bsp, Opn, Col1, and Tgfb1 (P < 0.01) with various ratios; while these induced upregulations of Alp, Bsp, Opn, and Tgfb1 were inhibited after Clcn7 gene level was lowered by Clcn7 shRNA lentivirus infection (Fig. 4A~C). However, the lowered Clcn7 level did not have any effect on calcified node formation with alizarin red staining ( Supplementary Fig. 2A~H).
Effect of ClC-7 on osteoclast formation through DFCs. DFCs became round and lucent, and looked like multi-dendrites cells under the condition of DFCs/MNCs co-culture (Fig. 5A,B). The monocytes survived no more than three days in LS8 cells/MNCs co-culture system (Fig. 5C), but formed TRAP-positive mulitinucleated cells at day 7 post DFCs/MNCs co-culture, and lowered ClC-7 expression level resulted in less TRAP-positive mulitinucleated staining cells (Fig. 5D~F). These TRAP+ cells showed small and shallow pits on the dentin slices, and the numbers and area of resorption lacunae were decreased in ClC-7 deficiency group (Fig. 5G~H, Supplementary Fig. 3). qRT-PCR demonstrated the upregulation of Opg and downregulation of Rankl after Clcn7 gene was silenced in DFCs (Fig. 5I).

Discussion
Resorption of alveolar bones by osteoclasts is essential for tooth eruption 7 . The osteopetrosis patients caused by osteoclasts dysfunction usually presented impacted and malformed teeth and the involved genes include RANKL 15,16 , TCIRG1 17 , etc. ClC-7 is highly expressed in the ruffled membrane of osteoclasts, and responsible for acidifying resorption lacuna. Osteoclasts isolated from ClC-7 deficient mice failed to resorb calcified bone 6,18 , which might contribute to the tooth malformations and eruption abnormalities in osteopetrosis patients with CLCN7 mutation 2 and Clcn7 gene knockout mice 3,5 . In Clcn7-shRNA, the tooth roots became curved and lost the regular shape (C,I). The predentin was enlarged and the odontoblasts were irregular arranged (F, asterisk) comparing to the other two groups (D,E, asterisk). The periodontal tissue of Clcn7-shRNA group was filled with poor-arranged fibroblasts (L, asterisk) other than normal periodontal ligament, cementum and alveolar bone (J,K, asterisk). Crooked tooth root (N, arrow), abnormal cusps (N, asterisk) and enlarged apical foramen (N, triangle) were observed in micro-CT scanning image. Figure  In the present study, we formulated siRNA with chitosan to protect siRNA from degradation, extended its in vivo effects and mimicked the gene knockout effect of Clcn7 around tooth germs 19 . We chose the first upper molar of P1 mouse as the siRNA target site, because the cusps of P1 molar germ begin to form while the development of root has not started yet, which is a good time point to observe tooth eruption and root development. We found that CS-Clcn7 siRNA nanoparticles directly targeted on some tooth tissues or cells. The early responding tissues to Clcn7 siRNA located in enamel and dentin, later in root and periodontal tissue. The local injection of CS-Clcn7 siRNA eventually resulted in impacted or delayed tooth eruption as we expected. We also tested the effect of local injection of chitosan formulated siRNA of Atp6v0a3, a gene involved in osteopetrosis 17 , and found that Atp6v0a3 siRNA may also target tooth cells and cause tooth malformation and root dysplasia after 17 injection days ( Supplementary Fig. 4). Thus the local injection of chitosan formulated siRNA of certain genes around tooth germ enables us to test a local gene silence or knock out effect during tooth development and this method should be widely used in the future. In order to identify the above tooth malformations caused by the effect of ClC-7 on osteoclasts or on tooth cells, the tooth germs infected with Clcn7 shRNA viruses were transplanted under kidney capsule which provides an osteoclasts-poor environment. The malformations in the transplanted tooth germs infected with Clcn7 shRNA viruses provided the direct evidence of ClC-7's contribution to the tooth development through tooth cells. Meanwhile the milder changes in the transplanted tooth germs infected with Clcn7 shRNA viruses comparing to that in situ CS-Clcn7 siRNA injection further confirmed that the contribution of ClC-7 deficient osteoclasts to the abnormal tooth phenotypes in situ.
ClC-7 deficiency caused the lower expressed DSP protein level in dentin in our study. Enamel and dentin changes had also been found in other osteopetrosis mouse models, such as Src(− /− ) mice 20 . Comparing to the dentin, less effect was found on enamel of Src(− /− ) mouse, which quite similar to our results. Thus ClC-7 may affect enamel and dentin development by regulating various key molecules of ameloblasts and odontoblasts. We did not find the lowered Dspp mRNA level in MDPC-23 cell line, which might be because of the different responses of Dspp to lowered Clcn7 level in vivo and in vitro, and the lower Δ CT value of Dspp in MDPC-23 cells.
A recent study in 3-week old ClC-7 mutant mice found that the molars did not form roots and the incisors were smaller than their age-matched controls. But the authors did not find obvious changes in the enamel of incisors and they regarded that ClC-7 deficiency might not significantly disrupt amelogenesis 5 . The different enamel's reaction to ClC-7 deficiency in the reference 5  and Clcn7 gene knockout mice 5 . On the other hand, the incisors were covered by enamel only in one side and the incisor growth is quite different from molar, especially during their root development. The deeper mechanism involved in the above different enamel changes under ClC-7 deficiency condition needs further exploring.
The deformed roots and abnormal periodontal tissues in ClC-7 deficient group switched our attentions to dental follicle cells. As the heterogeneous cell lineage, DFCs may differentiate into several types of cells, such as periodontal-type, cementoblastic-type and osteoblastic-type 21−24 . The upregulations of osteoblast related genes such as Alp, Opn, and Bsp in DFCs with TDMM induction confirmed the ability of DFCs to differentiate into osteoblasts; however, the above changes could be inhibited by Clcn7 gene deficiency. Thus ClC-7 might take part in the regulating of osteogeneic differentiation of DFCs.
One interesting finding is that the mRNA level of Tgfb1 was downregulated in this osteogenic process after lowering the ClC-7 level. TGF-β 1 has been reported to be involved in the regulation of tooth development 25,26 . The release and activation of TGF-β 1 stimulates the synthesis of various extracellular matrix proteins and inhibits their degradation which possibly results in tissue fibrosis 27,28 . Our previous work found that the deficient ClC-5 and ClC-3 were related with the upregulation of TGF-β 1 29,30 . Here again we found TGF-β 1 also being involved in the regulation of ClC-7 but with different mechanisms.
The interaction between DFCs and osteoclasts has also been studied previously 23 . Early in eruption, the coronal part of dental follicle accumulates monocytes which may differentiate into osteoclasts 31 . TRAP positive monocytes were present in the dental follicle prior to the onset of eruption and then declined in number during eruption 23,32 . Here we found that DFCs/monocytes co-culture system could induce the monocytes to differentiate into osteoclasts. The co-cultures of human/rat DFCs and osteoclast precursors have been applied in contact or non-contact systems before 33,34 . Due to the inhibition function of OPG secreted by DFC, both systems turned out a lower efficiency of TRAP+ cell formation 33,34 and fewer and shallow resorption pits 33 . Those TRAP+ cells in co-culture of DFC and monocytes only formed few and shallow pits in dentin slices, which might also be foreign body giant cells 35 . Our contact co-culture of mouse DFCs and monocytes confirmed the above findings. Meanwhile, we found that the lowered expression of ClC-7 in DFCs resulted in less TRAP positive multinuclear cells, less dentin resorption pits and smaller resorption area. Thus ClC-7 deficiency in DFCs might inhibit osteoclastgenesis.
RANK-RANKL-OPG signaling axis and downstream transcription factors play essential roles in the regulation of osteoclastogenesis 36 . DFCs are assumed to secret RANKL/M-CSF through a paracrine way, whilst OPG secreted by DFCs may inhibit osteoclastogenesis 8,37 . Consistent with the previous data 8, 38 , we found the mRNA expression of Rankl and Opg in DFCs, and the lowered Clcn7 level downregulated Rankl and upregulated Opg, which cause the inhibition of osteoclastogenesis. We suggest that ClC-7 is involved in the process of osteoclastogenesis through DFCs mediated RANK-RANKL-OPG signaling pathway. Meanwhile, we admitted that the mechanisms by which osteoclast precursors interact with DFCs and differentiate into osteoclasts in vitro and in vivo are more complicated than we expected, and our findings indicated that multiple factors might take part in the interaction between DFCs and osteoclast precursors under ClC-7 deficiency conditions.
In conclusion, our experiments demonstrate that ClC-7 deficiency affects tooth development and eruption in several pathways. The dysfunction of ClC-7 may directly influence the formation and calcification of enamel and dentin, or impair the differentiations of DFCs and DFCs mediated osteoclastogenesis through RANK-RANKL-OPG signaling pathway during tooth eruption. All the above changes contribute to the malformed teeth and altered tooth eruption. Tooth germs kidney transplantation. pGCSIL-GFP lentiviral vector (GeneChem, China) was used to generate shRNA against Clcn7 gene with the same siRNA targeting sequences. The plasmids were co-transfected into HEK 293T cells with lentiviral packaging plasmids to generate Clcn7 shRNA lentivirus (Clcn7-shRNA group) or negative shRNA lentivirus (negative-shRNA group). The mandibular first molar germs dissected from E13.5 Balb/c mice were infected with Clcn7-shRNA lentivirus (5000MOI/tooth germ) with the medium containing Dulbecco Modified Eagle Medium (DMEM), 10% fetal bovine serum (FBS, Gibco, USA) and 5 μ g/ml polybrene (GeneChem, China). After 24 hours' infection, the tooth germs were transferred to the transwell system, and the medium was switched to complete medium and cultured for about18 hours 40,41 . Then the tooth germs were transplanted into the renal capsule of adult male Balb/c mice. After 2 and 4 weeks of transplantation, micro CT scanning was performed to observe the tooth formation. The kidneys containing tooth germs were harvested for histological analysis and immunohistochemical staining after 4 weeks of transplantation.

Materials and Methods
Micro-CT scanning and analysis. Siemens Inveon system (Siemens, Germany) was used for the live animal scanning and three dimension images reconstruction. For CS-siRNA nanoparticles injection model, the heads of P17 mice were separated and scanned with the parameters as 80kV, 500 μ A, 500 ms exposure time, and scan angle of 360°. For the tooth germ and kidney subcapsular transplantation, after 2 weeks and 4 weeks of the transplantation, the mice and the kidneys with tooth germs were scanned with the same parameters as above, respectively. The experiment group and the controls were subjected to the same radiation.
Histological analysis and immunocytochemical staining. All the tooth samples and the kidneys con-  42,43 . Using modified Wise's method 21 , the dental follicle tissues were dissected from the tooth germs of 3-5 postnatal days Balb/c mouse, cut into pieces and digested in a solution of 2 g/L type I collagenase for 1 hour at 37 °C. The cells were cultured in α -MEM containing 20% FBS, sodium pyruvate (0.1 g/l), streptomycin (100 μ g/ml) and penicillin (100 U/ml). The 3 rd passage cells were used for all the experiments.
Lentivirus infection of cells or siRNA transfection. Cells (LS8 cells, MDPC-23 cell, and DFCs) grew up to 30% to 40% confluence were transfected with Clcn7 siRNA or control siRNA using lipofectin 2000, or infected with Clcn7 (or control) shRNA lentivivus at an MOI of 100. The mRNA were harvested after 72 hours for qRT-PCR analysis.
Quantitative realtime polymerase chain reaction (qRT-PCR). Quantitative RCR was performed as previously described methods 39 . The specific primers for amplification of Gapdh, Clcn7, M-csf (macrophage colony stimulating factor), Trap (acid phosphatase 5, tartrate resistant), Ctsk (cathepsin K) and Rank (receptor activator of nuclear factor-κ B), Rankl (receptor activator of nuclear factor-κ B ligand), Opg (osteoprotegerin), Alp (alkaline phosphatase), Bsp (bone sialoprotein), Tgfb1(transforming growth factor-β 1), Cap (cementum attachment protein), Opn (osteopontin), and Col1 (collagen type І) are shown in Supplementary Table 1. Osteogenic assay. Treated dentin matrix medium (TDMM) was prepared as the following methods: After being soaked in deionized water for 5 hours with an interval of 20 minutes ultrasonic treatment per hour, the cattle dentin matrices were serially soaked in 17% EDTA, 10% EDTA, and 5% EDTA for 5~10 minutes, with a deionized water wash for 10 minutes in an ultrasonic cleaner during each intervals. The treated dentin matrices were ground into powder with the treatment of liquid nitrogen, and sterilized with Co 60 irradiation. TDMM was made with a ratio of 20 g dentin powder per 100 ml α -MEM medium and incubated for 5 days at 37 °C, then the medium were filtered using a 0.22 μ m filter and stored at − 20 °C. DFCs were treated with TDMM for 3 or 21 days, then performed with qRT-PCR or Alizarin red staining.
Osteoclast function assay. Co-culture of DFCs and monocytes (MNCs). The monocytes were isolated using a Mouse Bone Marrow Monocyte Cell Isolation Kit (TBD science, China) from the femur and tibia of 6-weeks-old male Balb/c mice according to the manufacturer's protocol. For the establishment of DFCs/MNCs co-culture system, cell density of DFCs and MNCs were adjusted to 10 5 /ml and 10 6 /ml, respectively. LS8 cells/ MNCs co-culture, DFCs alone, MNCs alone were set as the controls.
TRAP staining and dentin resorption assay. After being co-cultured for 7 days, the DFCs/MNCs were stained for TRAP using Acid Phosphatase, Leucocyte (TRAP) Kit (Sigma, USA). The number of TRAP-positive multinucleated cells (≥ 3 multinuclear) was counted and compared among Clcn7-shRNA group, negative-shRNA group and control group. After 14 days co-culture of DFCs/MNCs on dentin slices (4 mm × 4 mm × 200 μ m), the number of resorptive lacunae of each dentine slice was counted under the scanning electron microscopy (SEM). The areas of the lacunas were measured with NIH Image J software (NIH, USA) for at least 50 lacunas each group. Statistical analysis. All the experiments were repeated at least three times and the data were presented as mean ± SD (standard deviation). Comparison between two different groups was performed using Independent-Samples T Test. Comparison among three different groups were performed using a One-Way ANOVO, and multiple comparisons were performed using LSD test and SNK test when data meet the homogeneity of variance, otherwise Dunnett T3 test and Dunnett C test were performed. The difference was considered to be statistically significant when the P value was < 0.05 or 0.01.