Identification of an Alternatively Spliced Variant of Ca2+-promoted Ras Inactivator as a Possible Regulator of RANKL Shedding*

The receptor activator of NF-κB ligand (RANKL), a critical regulator of osteoclastogenesis, is synthesized as a membrane-anchored protein and cleaved into a soluble form by ectodomain shedding. We developed an assay system to identify molecules regulating the RANKL shedding. Using this system, we found that a splice variant of Ca2+-promoted Ras inactivator (CAPRI), ΔCAPRI, which is expressed in primary osteoblasts, promoted the RANKL shedding. The wild type CAPRI is a member of the Ras GTPase-activating protein (GAP) family and suppresses Ca2+-dependent Ras activation, whereas ΔCAPRI, which lacks one exon in the GAP-related domain, activated the Ras pathway. Overexpression of ΔCAPRI or a constitutive active form of Ras up-regulated the expression level of matrix-metalloproteinase 14 (MMP14), which directly cleaves the ectodomain of RANKL, whereas Erk activation by expressing the constitutive active Mek1 did not affect the MMP14 expression or RANKL shedding. These results suggest that ΔCAPRI is a possible regulator of RANKL shedding by modulating MMP14 expression through Ras signaling cascades other than the Erk pathway.

The receptor activator of NF-B ligand, RANKL (also known as TNFrelated activation-induced cytokine, TRANCE, osteoprotegrin ligand, OPGL, and osteoclast differentiation factor, ODF) is a type II transmembrane glycoprotein with a molecular mass of ϳ45 kDa, which belongs to the tumor necrosis factor (TNF) 2 ligand family (1)(2)(3)(4)(5). RANKL is expressed on the membrane of osteoblasts and bone marrow stromal cells and binds to and activates TNF family receptor RANK expressed on monocyte-macrophage lineage osteoclast precursors (3,4). Upon binding to RANKL, RANK activates the intracellular signaling pathways including NF-B, Erk, JNK, and NFATc1, and leads to osteoclast differentiation, activation, and survival (6,7). The essential role of RANKL in normal bone turnover was further established by a series of knock-out mice, i.e. both RANKL-and RANK-deficient animals exhibited severe osteopetrosis because of the lack of osteoclast differentiation (8,9), whereas the targeted disruption of osteoprotegerin, a natural inhibitor of RANKL, developed severe osteoporosis because of enhanced osteoclastogenesis (10).
Some transmembrane proteins are extracellularly cleaved and released into the surrounding environment. RANKL is also made as a membrane-bound protein, cleaved by some proteinases, and converted to the soluble RANKL (11,12). This process, known as ectodomain shedding, has a diverse effect on a wide variety of membrane-bound proteins. For example, when TNF-␣ is cleaved by the TNF-␣ converting enzyme and released into the circulatory system, it exhibits strong systemic effects (13,14). In contrast, the Fas ligand is a strong apoptosis inducer in its membrane-bound form, but the soluble Fas ligand has fewer effects on apoptosis induction (15).
Although some proteinases have been demonstrated to have the RANKL shedding activity, no definite RANKL sheddase(s) have yet been identified. TNF-␣ converting enzyme was reported to be a candidate of RANKL sheddase (11); however, a more recent study has shown that TNF-␣ converting enzyme had no apparent effect on the RANKL shedding and that the RANKL shedding in TNF-␣ converting enzymedeficient cells was indistinguishable (16). A disintegrin and metalloproteinase domain family (ADAM)19 has also been reported to exhibit RANKL shedding activity (17), but embryonic fibroblasts from ADAM19 knock-out mice showed almost the same RANKL shedding activity as the cells from the wild type animals (16). Matrix metalloproteinase 14 (MMP14, also called the membrane-type 1 matrix metalloproteinase, MT1-MMP) can also cleave RANKL, although its cleavage site differs from that previously reported (16). These results suggest that there are other molecules implicated in RANKL shedding.
To identify molecules involved in the regulation of RANKL shedding, we developed a novel screening system, in which expression plasmids encoding secreted placental alkaline phosphatase (SEAP) fused with mouse C-terminally truncated RANKL (tRANKL-SEAP) were co-transfected with cDNA library pools of ST2 cells. Utilizing this screening system, we found that an alternatively spliced variant of Ca 2ϩpromoted Ras inactivator (CAPRI), ⌬CAPRI led to an increase in the RANKL shedding.

MATERIALS AND METHODS
Reagents-p-Nitrophenyl phosphate was purchased from Sigma-Aldrich. DNA polymerase, Pyrobest (for subcloning), was purchased from Takara biochemicals (Shiga, Japan), KOD plus (for reverse transcription-PCR) was from Toyobo (Osaka, Japan), and ionomycin was from Merck. Antibodies were purchased as follows: for the His tag, Erk, phosphorylated Erk, integrin-␤ and HSP-90 from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), for the V5 tag from Invitrogen, for actin from Sigma-Aldrich, and for RANKL from Active Motif (Carlsbad, CA).
Division of ST2 cDNA Library and Plasmid Purification-Procedures for the construction of the ST2 cDNA library were previously described (3). Competent high DH5␣ (Toyobo) Escherichia coli cells were transformed by heat shock with DNA solution of the ST2 cDNA library, and the library was divided into 1000 subpools, each of which contained ϳ100 clones. After the first screening was completed, every positive pool was divided into 48 subpools each of which contained ϳ10 clones and was subjected to a second screening, and the resulting positive pools were divided into 48 single clones. The plasmids were purified from E. coli using the QIAprep Spin Miniprep Kit (Qiagen, Hilden, Germany).
Constructs-The expression vectors for the mutants of Ras, pcDNA3-RasV12 (a constitutively active mutant) and pcDNA3-RasN17 (a dominant negative mutant), were generous gifts from Dr. C. Kitanaka (Yamagata University). An expression vector for a constitutively active form of Mek1, pcDNA3.1(ϩ)-Mek CA , was constructed by insertion of the cDNA fragments digested from pGEM-Mek CA (generous gift from K. Arai, Research Institute, National Rehabilitation Center for Persons with Disabilities) to NotI restriction site of pcDNA3.1(ϩ) (Invitrogen). The tRANKL-SEAP expression vector was constructed as follows. The cDNA encoding for SEAP was subcloned from pSEAP2-Control (Clontech) by PCR using a set of primers, 5Ј-CGCTCGAGAATCATCCCAGTTGAGGAGGAGA-ACC-3Ј and 5Ј-GCGTCTAGAGTAACCCGGGTGCGCGG-3Ј, and the cDNA fragment of the cytoplasmic region, transmembrane domain, and stalk region of the mouse RANKL (corresponding to amino acids 4 -157) were subcloned from ST2 cDNA library by PCR using a set of the following primers: 5Ј-AAGCTTGCCACCATGGCCAGCCGAGACTACGGCA-A-3Ј and 5Ј-GCGGCCGCCGCCTCGCTGGGCCACATCCA-3Ј. The PCR fragments were ligated into pCR-blunt II TOPO (Invitrogen) using protocols recommended by the manufacturer. The SEAP fragment and the RANKL fragment were digested from TOPO vectors with XhoI/XbaI and HindIII/NotI, respectively, and inserted into the corresponding restriction sites of pcDNA3.1-V5HisA (Invitrogen). The full-length RANKL was cloned from the cDNA of primary osteoblasts by PCR and inserted into the HindIII/NotI sites of pcDNA3.1-V5HisB (Invitrogen). The wild type CAPRI was cloned from the cDNA of primary osteoblasts, and ⌬CAPRI was subcloned from pcDL-⌬CAPRI by PCR (for pcDL-SR␣ vector, see Ref. 18) and inserted into the EcoRI/NotI sites of pEF1-HisC (Invitrogen) and the EcoRI/XhoI sites of pcDNA3.1-V5HisA (Invitrogen). The cDNA for full-length MMP14 was digested from pcDL-MMP14 at SalI sites and inserted into XhoI sites of pcDNA3.1(ϩ). The cDNA for MMP13 was cloned from the cDNA of primary osteoblasts by PCR, and inserted into the HindIII/XbaI sites of pcDNA3.1-V5HisA. Small interfering RNA plasmids for MMP14 were constructed using piGENE U6 vector (iGENE Therapeutics Inc., Ibaraki, Japan) according to manufacturer's protocol. The target sites were 5Ј-GGGCTGAGATCAAGGCCAA-TG-3Ј and 5Ј-GCGGGTGAGGAATAACCAAGT-3Ј.
Cell Culture and Transfection-The human kidney cell line 293T, human osteosarcoma cell line SaOS2, and mouse fibroblast cell line NIH3T3 were cultured with Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal bovine serum and 1% penicillinstreptomycin solution (Sigma) at 37°C in a humidified atmosphere containing 5% CO 2 . The plasmids were transfected into the cells using the FuGENE 6 transfection reagent (Roche Diagnostics) according to the manufacturer's instructions.
Alkaline Phosphatase Assay of Culture Medium-293T cells were seeded in 96-well cell culture plates at a concentration of 5 ϫ 10 5 cells/ ml. 24 h later, 625 pg of pcDNA3.1-tRANKL-SEAP and 49.375 ng of a subpool of the ST2 cDNA library or other constructs were co-transfected to 293T using FuGENE 6. 72 h after the transfection, 50 l of the culture medium was collected from each well and subjected to an alkaline phosphatase assay. In brief, the culture medium was incubated with 125 l of dH 2 O, 100 mM NaHCO 3 , 100 mM Na 2 CO 3 , 1 mM MgCl 2 , and 2 mM p-nitrophenyl phosphate at 65°C for 90 min, and then the absorbance of the solution at a wavelength of 405 nm was measured using the MTP-300 microplate reader (CORONA Electric, Ibaraki, Japan).
DNA Sequencing of Positive Clones-DNA sequences of the positive clones were determined by the cycle sequencing method. In brief, samples were prepared using the BigDye terminator Ver 1.0 (Applied Biosystems, Foster City, CA) and DyeEx 2.0 spin kit (Qiagen) and then analyzed using the ABI PRISM 310 Genetic Analyzer (Applied Biosystems).
Reverse Transcription-PCR-The procedures for preparation of the primary osteoblasts from the calvaria of newborn C57/BL6 mice and for formation of osteoclasts were previously described (19,20). The total RNA was purified from primary osteoblasts and osteoclasts using ISO-GEN (Nippon Gene Co. Ltd., Toyama, Japan) according to the method recommended by the manufacturer. The cDNA was synthesized from purified RNA using SuperScript II reverse transcriptase (Invitrogen). The primers used for the PCR analysis of ⌬CAPRI were 5Ј-GGCTGGC-CAAGGACTTTCTG-3Ј and 5Ј-CTATGTTCTGGACCGCCTGG-3Ј.
Western Blotting-The procedure for Western blotting was described previously (20). To detect the soluble RANKL released into the supernatants, 3 ml of the culture medium were incubated with 6 l of recombinant protein G-agarose (Invitrogen) and 1 g of recombinant osteoprotegerin-Fc chimeric protein (R&D Biosystems) for 16 h at 4°C, then recovered by brief centrifugation. The pellets were suspended in TNE buffer (10 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40) and subjected to SDS-PAGE. Membrane fraction was gained using Mem-PER mammalian protein extraction reagent (Pierce Chemical Co.).
Statistical Analysis-A statistical analysis was performed using the Student's t test for alkaline phosphatase assay and the unpaired Student's t test for real time PCR.

Development of a Screening System for Molecules with RANKL Shed-
ding Activity-To identify molecules potentially involved in the regulation of RANKL shedding, we developed a new library screening system. We constructed an expression vector encoding the fusion protein of SEAP with the C-terminally truncated form of RANKL, which contains the stalk region, transmembrane domain, and intracellular domain of RANKL (tRANKL-SEAP) (Fig. 1A). Clear up-regulation of the alkaline phosphatase activity was detected in the culture medium when this plasmid was co-transfected with the MMP14 expression vector into 293T cells (data not shown), indicating that this assay system is suitable for screening molecules with the RANKL shedding activity. Using this assay system, we screened the cDNA library of the mouse bone marrowderived stromal cell line, ST2 cells for molecules with RANKL shedding activity. From 1 ϫ 10 6 clones, 12 positive clones were isolated. Nucleotide sequences of these cDNA fragments revealed that one of the positive clones encoded a full-length cDNA for a novel splice variant of CAPRI, which lacks one exon (138 bp) in the RasGAP domain of the wild type CAPRI (⌬CAPRI) (Fig. 1, B and C). The other clones included full-length MMP14 (Fig. 1B), which was already reported to increase the RANKL shedding (16). The reverse transcription-PCR analysis showed that ⌬CAPRI is expressed in mouse primary osteoblasts and osteoclasts (Fig. 1D).
⌬CAPRI Expression Increases RANKL Shedding-We next examined the effect of ⌬CAPRI and wtCAPRI on the RANKL shedding. Either the ⌬CAPRI or the wtCAPRI expression plasmid was transfected to the 293T cells with the tRANKL-SEAP construct, and the alkaline phospha-tase activity in culture media was measured using p-nitrophenyl phosphate as substrate. The 293T cells co-transfected with the tRANKL-SEAP expression plasmid and empty vector exhibited a very low level of alkaline phosphatase activity in the supernatant. Similarly, a low alkaline phosphatase activity was detected in the conditioned medium of the cells expressing wtCAPRI. In contrast, the co-expression of ⌬CAPRI significantly increased the amount of tRANKL-SEAP released into the supernatant, suggesting that the ⌬CAPRI expression induces the cleavage of membrane-anchored RANKL in transfected 293T cells ( Fig. 2A). In addition, the co-transfection of ⌬CAPRI and the full-length RANKL expression plasmids markedly promoted the release of RANKL as detected by Western blotting with the anti-RANKL antibody (Fig. 2B). We next examined the effect of ⌬CAPRI on the RANKL shedding in osteoblastic cells. Overexpression of ⌬CAPRI also promoted the RANKL shedding in the osteoblastic cell line SaOS2 cells (Fig. 2C).  ⌬CAPRI Activates Ras Signaling Pathways-Lockyer et al. (21) reported that the elevation of the intracellular Ca 2ϩ causes a rapid C2 domain-dependent translocation of wtCAPRI to the plasma membrane, resulting in the activation of the RasGAP activity and the inactivation of the Ras/Mek/ Erk pathways. Therefore, we investigated the translocation of wt and ⌬CAPRI after the Ca 2ϩ influx by the Ca 2ϩ ionophore ionomycin treatment and found that both the wt and ⌬CAPRI moved to the plasma membrane of NIH3T3 and 293T cells 2 min after the ionomycin stimulation as shown in Fig. 3A by immunocytochemistry and Fig. 3B by cell fractionation analysis. To investigate whether ⌬CAPRI modulates the Ras/Erk signaling pathway, we examined the Erk activity in the ⌬CAPRIor wtCAPRI-overexpressing 293T cells. As shown in Fig. 3C, the overexpression of ⌬CAPRI enhanced the ionomycin-induced Erk phosphorylation, whereas the wtCAPRI expression reduced the Erk activation, indicating that ⌬CAPRI works in a dominant negative fashion. In fact, wtCAPRI suppressed RANKL shedding promoted by ⌬CAPRI (Fig. 3D).
Ras Activation, but Not Mek/Erk Pathway, Is Involved in RANKL Shedding-Because CAPRI is a Ca 2ϩ -sensitive RasGAP, we next tested whether the Ras activity is involved in the RANKL shedding. As shown in Fig. 3E, the expression of the constitutively active Ras strongly promoted the RANKL shedding in the 293T cells. RANKL shedding induced by the ⌬CAPRI overexpression was significantly suppressed by the co-expression of a dominant negative mutant of Ras (RasN17). These results indicate that the Ras activity plays an important role in the regulation of the RANKL shedding induced by ⌬CAPRI. Interestingly, however, the up-regulation of Erk activity in the 293T cells with the constitutively active Mek1 expression (Mek CA ) did not increase the alkaline phosphatase activity released into the supernatant (Fig. 3E).  A and B, translocation of wtCAPRI and ⌬CAPRI to plasma membrane in response to ionomycin stimulation. A, NIH3T3 cells transfected with either pcDNA3.1-⌬CAPRI-V5HisA or -wtCAPRI were unstimulated (a and c) or stimulated with 5 g/ml ionomycin for 2 min (b and d) and immunostained with anti-V5 antibody. Note the translocation of both wtCAPRI and ⌬CAPRI from cytoplasm to plasma membrane upon stimulation with ionomycin. B, 293T cells were transfected with pcDNA3.1-⌬CAPRI-V5HisA, and 48 h later, membrane fraction and total cell lysates of cells unstimulated (left lane) or stimulated (right lane) with 5 g/ml ionomycin for 2 min were subjected to SDS-PAGE. Integrin-␤ was used as a marker for the membrane fraction, and HSP90 as an internal control of total cell lysates. Ionomycin treatment up-regulated the membrane localization of ⌬CAPRI as shown by the anti-V5 blotting, whereas the amount of ⌬CAPRI in the total cell lysates did not appear to differ. TCL, total cell lysates. C, ⌬CAPRI promotes ionomycin-induced Erk activation. 293T transfected with pcDNA3.1-V5HisA, -⌬CAPRI, or -wtCAPRI were incubated for 36 h and serum-starved for 12 h. Cells were stimulated by 5 g/ml of ionomycin for 2 min and subjected to Western blot analysis. ⌬CAPRI expression increased ionomycin-induced Erk activation, whereas wtCAPRI expression suppressed it. P, pcDNA3.1-V5HisA; ⌬, pcDNA3.1-⌬CAPRI-V5HisA; wt, pcDNA3.1-wtCAPRI-V5HisA. D and E, alkaline phosphatase activity of the supernatants of the 293T cell cultures transfected with pcDNA3.1-tRANKL-SEAP and pcDNA3.1-V5HisA, -⌬CAPRI, wtCAPRI, pcDNA3-RasV12, -RasN17, or pcDNA3.1(ϩ)-Mek CA . *, significantly different, p Ͻ 0.01.
Ras Activity Regulates the Expression Level of MMP14-ST2 cDNA library screening for the RANKL shedding confirmed that MMP14 could be a candidate for the RANKL sheddase, and in fact, the cotransfection of MMP14 expression vector with pcDNA3.1-tRANKL-SEAP significantly increased the amount of tRANKL-SEAP released into the supernatant (Fig. 1B). We next examined whether or not the Ras activation downstream of ⌬CAPRI is involved in the expression level of MMP14. In the real time PCR analysis from mRNA of the 293T cells, ⌬CAPRI up-regulated MMP14, which was significantly suppressed by the co-expression of Ras DN . Ras CA induced the up-regulation of MMP14, whereas Mek CA did not affect the expression level of MMP14 (Fig. 4A). This result suggested that ⌬CAPRI would promote the RANKL shedding through up-regulation of MMP14. To confirm this, we constructed siRNA vectors for MMP14. As shown in Fig. 4B, siMMP14-1  To examine the relevance of this proteinase in RANKL shedding, we transfected the MMP13 expression vector and tRANKL-SEAP to 293T cells. Overexpression of MMP13 exhibited little effect on the up-regulation of the alkaline phosphatase activity in the supernatant, indicating that MMP13 had much weaker RANKL shedding activity compared with MMP14 (Fig. 5).

DISCUSSION
Ectodomain shedding is a highly regulated process that affects a number of transmembrane proteins, and is considered to play an important role in regulating various pathophysiological events. The role of ectodomain shedding varies between substrate proteins. For example, shedding allows some local growth factors, such as TNF-␣ and epidermal growth factor, to be released from the local environments and participate in the paracrine and endocrine signalings (13,14,23). Interestingly, proteolytic processing via shedding is important even for the local effects of the growth factors such as epidermal growth factor (24).
RANKL is a key molecule for the osteoclastogenesis and bone-resorbing activity of mature osteoclasts. RANKL is produced as a membranebound cytokine and released into the paracrine and the endocrine milieu via ectodomain shedding, although the biological significance of RANKL shedding is still unknown. The soluble form of RANKL induces in vitro osteoclastogenesis from bone marrow cells, and several studies have revealed an increase in the intraarticular level of soluble RANKL under pathological conditions such as rheumatoid arthritis (25). Recently, Mizuno et al. (26) generated transgenic animals overexpressing soluble RANKL in the liver after birth that exhibited a marked decrease in bone mineral density with aging, indicating that the excessive production of soluble RANKL can promote in vivo osteoclastogenesis. On the other hand, it was reported that the membrane-bound RANKL is more potent in stimulating osteoclast differentiation than its soluble form (12). To reveal the biological and pathological relevance of RANKL shedding, elucidating the molecular mechanism underlying RANKL shedding is indispensable.
It is difficult to detect soluble RANKL in the culture medium by the usual Western blot analysis because the expression level of RANKL is relatively low. Nakashima et al. (12) developed a ligand-receptor precipitation Western blot analysis using osteoprotegerin, which specifically binds to RANKL, but this system is not suitable for systematic screening of a large number of molecules. Blobel and coworkers (27) reported a simple and quantitative assay for TNF-␣ shedding using alkaline phosphatase-tagged TNF-␣. This method allows the rapid and reproducible quantitation of the TNF-␣ shedding. We first constructed an expression vector that encodes a fusion protein of mouse full-length RANKL and  SEAP (full RANKL-SEAP) and transfected it into 293T cells. Although we could find the expression of the full RANKL-SEAP by Western blotting, we failed to detect alkaline phosphatase activity in the supernatants even when the vector was co-transfected with a putative RANKL sheddase, MMP14, the reason of which remains unclear (data not shown). For several type II transmembrane proteins including TNF-␣, the juxtamembrane sequence surrounding the cleavage site has been shown to be sufficient to target the protein for regulated shedding (16). Therefore, we next constructed an expression vector encoding a fusion protein of SEAP with the C-terminally truncated RANKL, which contained the stalk region but lacked the TNF-like domain of RANKL (pcDNA3.1-tRANKL-SEAP), and transfected it together with the ST2 cell-derived cDNA library. Using this assay system, we could isolate 12 independent positive clones, which showed alkaline phosphatase activity in the supernatant when transfected with pcDNA3.1-tRANKL-SEAP. They could also increase the amount of soluble RANKL when transfected into the 293T cells with the full-length RANKL, confirming the relevance of this screening system.
One of the positive clones was a spliced variant of CAPRI. The protease activity inducing ectodomain shedding of the transmembrane proteins has been reported to be modulated by protein kinase signaling such as the Ras/Mek/Erk pathway (28). CAPRI was originally identified as a member of the RasGAP, negative regulators of Ras signaling pathways (21,29), and the R473S mutation decreased the RasGAP activity of CAPRI and enhanced ATP-or ionomycin-induced Erk phosphorylation (21). ⌬CAPRI lacks the Arg-473-containing exon in the RasGAP domain in which the FLR motif stabilizes the catalytic arginine-finger loop (30), suggesting that it can work in a dominant negative fashion like the R473S mutant, and in fact, the ⌬CAPRI expression increased the ionomycin-induced Erk activation, and co-expression of wtCAPRI diminished the RANKL shedding promoted by ⌬CAPRI. The expression of a dominant negative mutant of Ras (RasN17) suppressed the RANKL shedding induced by ⌬CAPRI, and a constitutively active mutant of Ras (RasV12) expression stimulated the RANKL shedding, indicating that ⌬CAPRI induces RANKL shedding via activating Ras pathways.
The only protease we could isolate in this screening system was MMP14 (MT1-MMP), which was previously reported as a candidate of the RANKL sheddases (16). It should be noticed that MMP14 knock out mice showed an increase in osteoclast number and developed severe osteopenia (31). This suggests that MMP14 negatively regulates the local osteoclastogenesis by reducing membrane-bound RANKL through ectodomain shedding, although it may increase the amount of soluble RANKL. The serum concentration of soluble RANKL in the physiological condition is reportedly less than 1 ng/ml, which is not high enough to induce general osteopenia (32).
Our data do not exclude the possibility that other proteases than MMP14 are involved in this process. Recent studies have demonstrated the role of MMP13 in regulating bone integrity; however, RANKL shedding activity of MMP13 was much weaker than that of MMP14. The overexpression of ⌬CAPRI or Ras CA increased the expression of MMP14, and MMP14 knockdown suppressed RANKL shedding promoted by ⌬CAPRI, indicating that the ⌬CAPRI/Ras pathways stimulate the RANKL shedding by regulating the expression of MMP14.
Interestingly, the constitutively active Mek expression failed to upregulate the amount of the cleaved RANKL or MMP14 expression, although the Erk activity was strongly up-regulated (data not shown). These results suggest that the downstream cascades of Ras other than the Mek/Erk pathway are implicated in the RANKL shedding. A further investigation is required to clarify the signal transduction pathway(s) downstream of Ras that regulates the RANKL shedding.
In conclusion, we established a novel library screening system for identifying the molecules involved in the RANKL shedding and identified a splice variant of CAPRI, ⌬CAPRI, as a possible candidate. ⌬CAPRI activates the Ras pathways, which increase the expression of MMP14 in a Mek/Erk-independent manner and lead to the RANKL shedding. These data suggest that the ⌬CAPRI-Ras-MMP14 axis plays an important role in the RANKL shedding.