CLEC-2 suppresses calcification in cultured osteoblasts

Podoplanin is the only counter-receptor of platelet CLEC-2 and is expressing on mature osteoblast, but there is no report on the role of podoplanin and CLEC-2 in calcification. This study aimed to investigate the role of podoplanin binding to CLEC-2 in the calcification of osteoblasts carrying homozygously deleted Pdpn alleles (PdpnΔ/Δ) by heterozygously expressing collagen type I alpha 1 promoter (Col1a)-driven Cre recombinase. There were no macroscopic abnormalities in the bone and dentin of Col1a11-Cre;PdpnΔ/Δ mice but the coccygeal bone medullary cavity was very narrow. In the quantitative analysis for alizarin red-stained products and alkaline phosphatase activities on the cultured calvarial osteoblasts, the amounts of calcified products and alkaline phosphatase activity of calvarial osteoblasts of both Pdpnfl/fl and Col1a11-Cre;PdpnΔ/Δ mice were significantly higher in the calcification medium than in the α-mem. Both the amounts of calcified products and alkaline phosphatase activity of calvarial osteoblasts from Pdpnfl/fl mice were significantly lower in the calcification medium with CLEC-2 than without CLEC-2 while there were no significant differences in the amounts of calcified products and alkaline phosphatase activities of calvarial osteoblasts from Col1a11-Cre;PdpnΔ/Δ mice with CLEC-2. Platelet CLEC-2 may play a role in regulating the calcification via binding to podoplanin on mature osteoblasts expressing podoplanin in the medullary cavity of a part of the bone.


Introduction
Podoplanin is a mucin-type O-glycosylated type I transmembrane protein with high sialic acid content which is expressed on the ventricular choroidal plexus epithelium, salivary gland myoepithelial, tooth germ, craniofacial bone, nerve sheath, and Meckel's cartilage in head and neck organs. In the tooth germ, podoplanin is expressed on the enamel cord, cervical loop, internal and external enamel epithelium, and odontoblasts (1,2). In 1997, podoplanin was identified as a component playing a major role in glomerular podocyte atrophy in puromycin-induced nephrosis in rats (3); in rat type I alveolar epithelial cell antigen T1alpha (4); in 1998 in rat type I alveolar epithelial cell antigen RT 140 (5), and in 1992 identified in mouse peripheral lymphoid tissue stromal cell glycoprotein (gp) 38 (6,7); in 1999 identified in human lymphoid tissue gp36 (8), and in the skin tumor cell line antigen PA 2.26 reported (9); in 2003 identified in the AGGRUS, that is overexpressed on tumor cell surface and causes platelet aggregation because of the binding activity to the lectin-like receptor c-type lectin like receptor-2 (CLEC-2) on platelets (10,11). Further, podoplanin has been generally accepted to play an important role in cell process elongation and contraction by actin cytoskeleton rearrangement dependent on the binding activity of the cytoplasmic portion of podoplanin with the cytoplasmic linker protein ezrin via RhoA family signaling (12,13). The expression of podoplanin in the cell membrane promotes membrane-bound Rho-GTPase and results in the phosphorylation of ezrin, thereby facilitating mediation of the binding of podoplanin to F-actin by phosphorylated ezrin, resulting the formation of a cell membrane-actin structure and cell process extension. It has also been reported that CLEC-2 binding to podoplanin induces cell process extension by the dissociation of the actin cytoskeleton and relaxation of cells (12,13).
The earliest reported podoplanin gene for bone-related cells is OTS-8, a partial cDNA cloned from mouse osteoblast-like MC3T3-E1 (14). The OTS-8 was cloned from the early response protein cDNA library of mouse osteoblast-like MC3T3-E1 cells treated with the tumor promoter 12-O-tetradecanoylphorbol-13-acetate, and Farr et al. reported that the 38 amino acid sequence epitope gp38 of mouse thymic epithelium recognized by a hamster monoclonal antibody derived from the clone 8.1.1 closely resembles OTS-8 (6). Furthermore, podoplanin was established as an E11 antigen that is recognized by a monoclonal antibody for the rat osteoblastic osteosarcoma cell line ROS17/2.8 cells (15,16). Currently, E11/podoplanin is generally recognized as a bone cell marker. Mature late osteoblasts and osteocytes express E11/podoplanin, and MC3T3-E1 and human osteoblast-like cells MG63 increase the expression of podoplanin in calcification medium (16)(17)(18)(19)(20). As mature osteoblasts and osteocytes express podoplanin at the plasma membrane of cell processes and bone extract contains a large amount of podoplanin, there may be soluble podoplanin which participates in crystal sheath to the maturation of the mineralized nodule. The E11/podoplanin has increased expression in osteocyte dendrites more than the mature osteoblast/osteocyte marker dentin matrix protein 1 (DMP-1) and sclerostin (19)(20)(21). Mouse long bone osteocyte-like cells MLO-Y4 and IDG-SW3 express podoplanin more strongly than MC3T3-E1 (20)(21)(22)(23). Differentiation of cultured mouse calvarial osteoblasts and pre-osteoblastoid cell MLO-A5 into osteocytes in calcification medium containing βglycerophosphate and ascorbic acid, and biomaterial-induced calcification are simultaneously accompanied by an increase in podoplanin (24)(25)(26)(27). Also, in the mature osteoblast MLO-A5 cell line, an increase in podoplanin coincides with dendrite formation, calcification, and RhoA activation (28).
Since podoplanin plays an important role in cell process formation via RhoA signaling (12,13), podoplanin may be inducing osteocyte process extension with osteoblast maturation.
The podoplanin KO mice were established by Schacht et al. in 2003 (24), but there were no reports of podoplanin conditional KO (cKO) mice until our report in 2018 (29). Podoplanin KO mice are lethal at birth due to respiratory failure (24,29). The podoplanin KO mouse fetus do not have alveolar sacs at the end of the alveolar duct (29), and alveoli consisting of alveolar ducts with TTF-1 positive type II alveolar epithelial cells but lack alveolar sacs with type I alveolar epithelial cells. The flat and thin type I alveolar cells function as an air-blood barrier and type II alveolar cells differentiate into type I alveolar cells. The podoplanin-deficient type II alveolar cells appear unable to differentiate into type I lung cells, resulting in lung atrophy (29). All our attempts to generate cKO mice by embryos obtained from TIGM and EUCCOM have failed, it is believed that it is very difficult to become integrated in the germ line transmission of the podoplanin-targeted ES cells in the chimeric mice. Therefore, little is known of any details of the function of podoplanin in somatic tissues applying cKO mice. This study aimed to investigate the role of podoplanin and the receptor CLEC-2 in the calcification of osteoblasts from podoplanin-targeted mouse carrying the homozygously deleted Pdpn alleles (Pdpn Δ/Δ ) by heterozygously expressing 2.3-kb collagen type I alpha 1 (Col1a) promoter-driven Cre recombinase..

Animals
This study aimed to investigate medullary cavity formation in conditionally podoplanin-deficient mice. All of the specimens were collected from euthanized mice and the manuscript was prepared following the ARRIVE guidelines (29). The protocol of experiments for animal use was approved by the Animal Experiment Committee of Fukuoka Dental College in accordance with the principles of the Helsinki Declaration. Breeding and experiments were performed in a room with a 100% controlled atmosphere which had passed an examination for bacteria and is located in the Fukuoka Dental College Animal Center. Mice grew normally and lived healthily under conventional atmosphere conditions with normal feeding in cages. The mice were housed with an inverse 12-hour day-night cycle with lights on from 7:00pm in a room where the temperature (22˚C) and humidity (55%) were completely controlled. Humane endpoints were used in the experiments as a rapid and accurate method for assessing the health status of the mice, that is, mice which have lost the ability to ambulate and to access food or water were euthanized by induction anesthesia (1 l/min of 2% isoflurane mixed with 30% oxygen and 70% nitrous oxide with an anaesthetic apparatus) followed by cervical dislocation and intraperitoneal injections with 3.5% chloral hydrate (10 ml/kg, trichloroacetaldehyde monohydrate, Kanto Chemical, Tokyo, Japan) in the saline.
All of the subjects were collected from mice euthanized as described above.

Generation of Col1a1-dependent Pdpn cKO mice
We adopted a modular strategy for the construction of targeting vectors (Supplementary 1) [29]. The podoplanin gene (Pdpn) targeting vector HTGR03003_Z_2_G05 (European Conditional Mouse Mutagenesis Program, EUCOMM) was constructed by recombineering with a modular strategy of Gateway systems, assembling the vector backbone and Pdpn from the C57BL/6J BAC libraries, which allow reporter-tagging conditional mutation of the Pdpn (EUCOMM project ID: 36335) [30].
The targeting vector has the promoter-driven targeting cassette which consists of a gene-trap and selection cassettes as described elsewhere (29). In the mice with the Pdpn gene conventional knockout-first allele, referred to as Pdpn KO1st , the targeting vector insertion disrupts the targeted Pdpn

Genotyping
A 1-mm length of the tails were collected from mice under the anesthesia for genotyping. Genomic DNA from the tails was isolated with a QIAamp DNA Blood and Tissue Kit (Qiagen, Hilden, Germany), with all procedures as described in our previous report (29). The PCR was performed with 30 cycles for amplification using the Ex Taq hot start version (Takara Bio Inc., Otsu, Japan) with 50 pM of primer sets: Pdpn without the loxP site (wild, 133bp) and Pdpn including the third loxP site μg/ml, ascorbic acid, and 10mM β-glycerophosphate for the osteogenic differentiation to test the calcification. Cells (10,000 cells/well) were seeded on collagen I-coated 6-well Culture Plates (25mm diameter, 9.6-cm 2 wells; Iwaki).

Immunostaining
The morphology of the bone and teeth were investigated on tissue sections of 4 week old male mouse heads including the upper incisors, femurs, and coccyxes of Col1a-Cre;Pdpn Δ mice. The sections were made by Kawamoto's film method and tested by immunostaining as described elsewhere (29).
The specimens were embedded in super cryo-embedding medium (Leica Microsystems Japan, Tokyo, Japan) and frozen in liquid N 2 , and the undecalcified frozen sections (4 μm) were cut in a cryostat

Test for alkaline phosphatase activity and calcification
The calvaria osteoblasts from Pdpn fl/fl and Col1a1-Cre;Pdpn Δ/Δ mice, and osteoblasts from bone marrow (Cosmo Bio) were cultured in mouse osteogenesis culture kits (Cosmo Bio) in 6-well plates.
The 80% confluent cells were stained by an Alkaline phosphatase (ALP) staining kit (Cosmo Bio).
The relative ALP activities were determined by the absorbance at 405nm after treatment with 1.0 mg/ml pNPP diluted in 20 ml of 2M Tris-buffered saline (pH9.1). The 80% confluent cells were also fixed by 10% formalin-PBS and treated with 40mM alizarin red S (pH 4.2) for 30 min at room temperature. Calcified nodules stained bright red in the culture were solubilized in 5% formic acid and the calcification amounts were determined by the absorbance at 405nm for the formic acid solution colored yellow.

Enzyme-linked immunosorbent assay (ELISA)
Mouse calvaria osteoblasts originating from 80% confluent monolayers in the 24-well plates were

Reverse transcription (RT)-PCR and real-time PCR
Total RNA extraction from the cultured cells was performed with a QIAshredder column and an RNeasy kit (Qiagen, Inc., Tokyo, Japan). Contaminating genomic DNA was removed using DNAfree (Ambion, Huntingdon, UK), and the RT was performed on 30 ng of total RNA, followed by 30 cycles of PCR for amplification using the Ex Taq hot start version (Takara Bio Inc., Otsu, Japan) with 50 pM of primer sets for mouse mRNA of β-actin, podoplanin, osteopontin, and osteocalcin, where the specificities had been confirmed by the manufacturer (Sigma-Aldrich Corp., Tokyo, Japan). The RT-PCR products were separated on 2% agarose gel (NuSieve; FMC, Rockland, ME, USA) and visualized by Syber Green (Takara). The correct size of the amplified PCR products was confirmed by gel electrophoresis and amplification of accurate targets was confirmed by a sequence analysis. To

Statistics
All experiments were repeated five times, and the data was expressed as mean + SD. Statistically significant differences (p< 0.01) were determined by one-way ANOVA and the unpaired two-tailed Student's t test with STATVIEW 4.51 software (Abacus concepts, Calabasas, CA, USA).

Development of bone and teeth in Col1a1-Cre;Pdpn Δ/Δ mice
In the macroscopic observation and in the tissue sections there were no abnormalities in the femur bone formation of 4 week old Col1a11-Cre;Pdpn Δ/Δ mice, and osteocytes showed no expression of osteocyte marker podoplanin (Fig. 1). In the macroscopic observations and in the head sections there were no abnormalities in the dentin formation of the 4 week old Col1a11-Cre;Pdpn Δ/Δ mice, and odontoblasts showed no expression of odontoblast marker podoplanin (Fig. 2). In the macroscopic observation and in the tissue sections there were no abnormalities in the coccygeal bone formation of the 4 week old Col1a11-Cre;Pdpn Δ/Δ mice but the medullary cavity was very narrow (Fig. 3). The relative size of the coccygeal medullary cavity versus the bone was significantly smaller in Col1a11-Cre;Pdpn Δ/Δ mice than in the Col1a11-Cre;Pdpn fl/f mice, without differences in the femurs (Data not shown).

Expressions of alkaline phosphatase and mineralized nodule-associated proteins in cultured calvarial osteoblasts with CLEC-2
The calvarial osteoblasts from Col1a1-Cre;Pdpn Δ/Δ mice cultured in the calcification medium were stained with alkaline phosphatase substrate and immunostained by anti-podoplanin, anti-osteopontin, and andi-osteocalcin (Fig. 4). The gene expressions of podoplanin, osteopontin, and osteocalcin were also detected from the osteoblasts cultured in α-mem, in the calcification medium, and in the calcification medium with CLEC-2 (0.1 and 1μg/ml) (Fig. 4). In the Cell ELISA analysis, the produced amounts of mineralized nodule-associated protein: podoplanin, osteocalcin, and osteopontin, increased with the calcification medium and remained unchanged by the CLEC-2 administrations (0.1 and 1μg/ml) in the cultured calvarial osteoblasts from Col1a1-Cre;Pdpn fl/fl mice and in the cells from Col1a1-Cre;Pdpn Δ/Δ mice (not shown) (Fig. 4).

Production of calcified products in cultured calvarial osteoblasts with CLEC-2
In the analysis for calcified products in the cultured calvarial osteoblasts, alizarin red-stained products were seen in cells from Pdpn fl/fl mice which were cultured in the calcification medium for 20 days (Fig. 5).

Development of bone and teeth in Col1a1-Cre;Pdpn Δ/Δ mice
Tooth formation is controlled by the inner enamel epithelium. When pre-odontoblasts which have lost their proliferating ability after the terminal differentiation of the dental papilla cells mature into odontoblasts and begin secreting pre-dentin, the inner enamel epithelium subsequently terminally differentiates into pre-ameloblasts by interaction with the dentin matrix, and loses its proliferative ability, and the pre-ameloblasts mature into ameloblasts and secrete enamel matrix. In the tooth germ, podoplanin is expressed on the enamel cord, cervical loop, internal and external enamel epithelium, and odontoblasts, but not in the dental pulp cells, and the podoplaninn expression on odontoblasts disappears after dentinogenesis (1,2). In our previous study systematic interference for Pdpn allele splicing results in no abnormalities in the development of bone and teeth (29). In the study here, there were no abnormalities in the femur bone formation at 4 weeks of age, and the osteocytes showed no expression of osteocyte marker podoplanin (Fig. 1). There were also no abnormalities in the dentin formation of 4 week old Col1a11-Cre;Pdpn Δ mice, where odontoblasts showed no expression of the odontoblast marker podoplanin (Fig. 2). However in the coccygeal bone of the 4 week old Col1a11-Cre;Pdpn Δ/Δ mice, the medullary cavity was very narrow (Fig. 3). The relative size of the coccygeal medullary cavity to the bone was significantly smaller in the Col1a11-Cre;Pdpn Δ/Δ mice than in the Col1a11-Cre;Pdpn fl/f mice, while these mice showed no differences in the femurs (data not shown).
Taken together, this would suggest that podoplanin is not a critical regulator in the development and homeostasis of systemic bone and teeth, however, podoplanin would seem to play a so far not reported role to regulate the calcification in the medullary cavity.

Effect of CLEC-2 on the calcification of cultured calvarial osteoblasts
Mature osteoblasts secrete matrix vesicles into the bone matrix to induce calcification. Alkaline phosphatase (ALP) in the matrix vesicles hydrolyzes pyrophosphate to form phosphate and takes up calcium to form calcium phosphate crystals. Apatite ribbons on the calcium phosphate crystal core forms the mineralized nodules and are covered by a crystal sheath composed of non-collagen protein like OPN, OCN, and others (35). Osteoblasts produce ALP and OPN in the middle stage of maturation, and OCN and podoplanin in late maturation (19)(20)(21)(36)(37)(38)(39)(40). The calvarial osteoblasts cultured in the calcification medium expressed ALP, OPN, OCN, and podoplanin (Fig. 4). The production of these was not changed by CLEC-2 administrations in cultured calvarial osteoblasts from the Col1a1-Cre;Pdpn fl/fl mice (Fig. 4), or in cells from the Col1a1-Cre;Pdpn Δ/Δ mice (not shown). In the quantitative analysis for the formic acid solution of alizarin red-stained products and for ALP activity on the cultured calvarial osteoblasts, the amounts of calcified products and ALP activity of calvarial osteoblasts from both Pdpn fl/fl and Col1a11-Cre;Pdpn Δ/Δ mice were significantly higher in the calcification medium than in the α-mem, suggesting that the calcification was promoted by the calcification medium in the matured osteoblasts (Fig. 5). Both the amounts of calcified products and ALP activity of the calvarial osteoblasts from the Pdpn fl/fl mice were significantly smaller in cells with the calcification medium with CLEC-2 than in cells without CLEC-2. The amounts of calcified product and ALP activity of osteoblasts from the Pdpn fl/fl mice were also significantly larger in the calcification medium with 0.1μg/ml of CLEC-2 than in that with 1μg/ml of CLEC-2. There were no significant differences in the amounts of calcified product and ALP activity of calvarial osteoblasts from the Col1a11-Cre;Pdpn Δ/Δ mice in the calcification medium with CLEC-2 (Fig. 5). Our previous study showed that the calcification products in osteoblasts cultured with anti-podoplanin was significantly less than in the cells without anti-podoplanin (41). This was ascribed to the CLEC-2 being only a receptor of podoplanin on platelets, the platelet CLEC-2 may suppress the calcification via binding to podoplanin on the mature osteoblasts expressing podoplanin.  There is a weak reaction in the salivary gland (asterisk) and cross reaction to the alveolar bone. Bars:

Figure legends
100μm.   and Col1a11-Cre;Pdpn Δ/Δ mice were significantly higher in the calcification medium than in the αmem (asterisk). Both the amounts of calcified products and alkaline phosphatase activity of the calvarial osteoblasts from Pdpn fl/fl mice were significantly smaller in the calcification medium with 1μg/ml (double-asterisk) and 0.1μg/ml (three asterisks) of CLEC-2 than in the calcification medium without CLEC-2. Both the amounts of calcified products and alkaline phosphatase activity of calvarial osteoblasts from Pdpn fl/fl mice were significantly larger in the calcification medium with 0.1μg/ml of CLEC-2 (circle) than in that with 1μg/ml of CLEC-2. There were no significant differences in the amounts of calcified products and alkaline phosphatase activities of calvarial osteoblasts from the Col1a11-Cre;Pdpn Δ/Δ mice in the calcification medium with CLEC-2 (nil, 0.1, and 1μg/ml). Asterisks and circles: significantly different in ANOVA (P<0.01); *(vs α-mem), **(vs calcification medium with 0μg/ml of CLEC-2), ○ (vs calcification medium with 1μg/ml of CLEC-2).