Inhibition Effect of Enteropeptidase on Ranklâ€“RANK Signalling By

FEBS Letters 587 (2013) 2958–2964

journal homepage: www.FEBSLetters.org

Inhibition effect of enteropeptidase on RANKL–RANK signalling by cleavage of RANK

Yunfeng Zhao a,b, Mengmeng Jin c, Juan Ma a, Shiqian Zhang d, Wei Li a, Yuan Chen e, Yingsheng Zhou f, ⇑ Hong Tao f, Yu Liu c, Lei Wang a,b, Huamin Han g, Ge Niu a,b, Hua Tao a, Changzhen Liu a,1, Bin Gao a, a CAS Key Laboratory of Pathogenic Microbiology and Immunology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China b University of Chinese Academy of Sciences, Beijing, China c The General Hospital of Chinese People’s Liberation Army, Beijing, China d Department of Orthopaedics, The First Clinical College of Harbin Medical University, Harbin, China e School of Life Sciences, Anhui University, Hefei, China f Department of Endocrinology and Metabolism, Beijing Anzhen Hospital, Capital Medical University, Beijing, China g Institute of Biophysics, Chinese Academy of Sciences, Beijing, China article info abstract

Article history: Enteropeptidase can cleave trypsinogen on the sequence of Asp-Asp-Asp-Asp-Lys and plays an Received 29 April 2013 important role in food digestion. The RANKL–RANK signalling pathway plays a pivotal role in bone Revised 28 June 2013 remodelling. In this study, we reported that enteropeptidase can inhibit the RANKL–RANK signalling Accepted 1 August 2013 pathway through the cleavage of RANK. A surrogate peptide blocking assay indicated that entero- Available online 13 August 2013 peptidase could specifically cleave RANK on the sequence NEEDK. Osteoclast differentiation assay and NF-jB activity assay confirmed that enteropeptidase could inhibit osteoclastogenesis in vitro Edited by Zhijie Chang through the cleavage of RANK. This is the first study to prove that the RANKL–RANK signalling pathway can be inhibited by cleavage of RANK instead of targeting RANKL. Keywords: Enteropeptidase/EP Structured summary of protein interactions: RANKL–RANK signalling pathway EP cleaves hRANK by cleavage assay (View interaction) Osteoclastogenesis EP cleaves mRANK by cleavage assay (View interaction)

1. Introduction activation by duodenase [1] or trypsin. Trypsinogen is specifically cleaved after the sequence of Asp-Asp-Asp-Asp-Lys, which is Enteropeptidase (EP), a type II transmembrane serine protease highly conserved in vertebrates. This cleavage results in trypsin- localised to the brush border of the duodenal and jejunal mucosa, dependent activation of other pancreatic zymogens, such as chy- is synthesised as a zymogen (proenteropeptidase), which requires motrypsinogen, proelastase, procarboxypeptidase and prolipase in the lumen of the gut [2,3]. Diarrhoea, oedema, vomiting, hypo- proteinaemia, anaemia, and failure to gain weight are common Abbreviations: EP, enteropeptidase; BMM, bone marrow-derived macrophage; symptoms of patients with congenital EP deficiency during early RANKL, receptor of activator of NF-jB ligand; RANK, receptor of activator of NF-jB; infancy [4,5]. Conversely, duodeno-pancreatic reflux of proteolyti- TNFRSF18, tumour necrosis factor receptor superfamily 18; M-CSF, macrophage- cally active EP can cause acute pancreatitis [6]. colony stimulating factor; TRAP, tartrate-resistant acid phosphatase ⇑ Corresponding author. Address: The Centre for Molecular Immunology, Insti- EP is a two-chain polypeptide consisting of an N-terminal tute of Microbiology, Chinese Academy of Sciences, 1 Beichen West Road, Beijing 120 kDa heavy chain disulphide linked to a C-terminal 100101, China. Fax: +86 10 6480 7338. 47 kDa light chain including a chymotrypsin-like serine prote- E-mail addresses: [email protected] (Y. Zhao), jinmengmeng1980@ ase domain [7]. The membrane association of the enzyme is not yahoo.com.cn (M. Jin), [email protected] (J. Ma), [email protected] required for substrate recognition since trypsinogen activation (S. Zhang), [email protected] (W. Li), [email protected] (Y. Chen), [email protected] (Y. Zhou), [email protected] (H. Tao), [email protected] (Y. Liu), was not impaired when the transmembrane domain of EP was [email protected] (L. Wang), [email protected] (H. Han), deleted. The catalytic capacity of the light chain in the absence [email protected] (G. Niu), [email protected] (H. Tao), [email protected] of the heavy chain decreased by approximately 500-fold (C. Liu), [email protected] (B. Gao). compared to the complete EP, suggesting that the heavy chain 1 Co-corresponding author. Address: The Centre for Molecular Immunology, is necessary for optimal cleavage. However, the light chain of Institute of Microbiology, Chinese Academy of Sciences, 1 Beichen West Road, Beijing 100101, China. Fax: +86 10 6480 7338. EP was sufficient for the normal recognition of substrate target

0014-5793/$36.00 Ó 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.febslet.2013.08.005 Y. Zhao et al. / FEBS Letters 587 (2013) 2958–2964 2959 sequences [8]. In 1998, Lu and Sadler summarised the features of 2.2. The synthesis of a surrogate peptide substrates required by EP [7]: the substrates of it should contain Lys/Arg residue at position P1 and four Glu/Asp residues at posi- A surrogate peptide of RANK, DTWNEEDKCLL was synthesised tions P2–P5, but more variation is shown at position P5. Receptor on 4-MeBHA-resin using manual ‘in situ neutralisation’ Boc chem- of activator of NF-jB (RANK) is expressed on mammary gland istry protocols for stepwise SPPS[16]. Boc amino acids were used cells, osteoclast precursors, intestinal epithelial cells and some with the following side chain protection: Asp(OcHxl), Asn(Xan), cancer cells. There is peptide motif with the sequence Asn-Glu- Glu(OcHxl), Cys(4-CH3Bzl), Lys(2C1-Z), Thr(Bzl), Glu(OcHex), Glu-Asp-Lys (NEEDK) (Fig. 1) in the sequence of RANK, which Trp(CHO). Other amino acids were used side chain unprotected. indicates that RANK may be recognised and cleaved by EP. Since The synthesis was carried out on a 0.25 mmol scale. Prior to HF RANK plays a pivotal role in the differentiation of osteoclasts, we deprotection and cleavage, the CHO group of the residue Trp was predicted that EP could inhibit the differentiation of osteoclasts removed from the resin-bound peptide segment by treatment with by cleaving RANK expressed at the surface of osteoclast precur- 10% v/v piperidine plus 5% v/v water in DMF for 2 h at 0 °C. After sors. There are several inhibitors which could block RANKL–RANK removal of the N-aBocgroup, the peptide was cleaved from the re- signalling, such as denosumab [9], an antibody drug approved by sin and simultaneously deprotected by treatment at 0 °C for 1 h the FDA, OPG-Fc [10], RANK-Fc [11], and peptides that simulate with anhydrous HF containing 5%v/v p-cresol as scavenger. After the interaction between RANKL and RANK [12,13]. However, all removal of HF by evaporation under reduced pressure, the crude these strategies were based on the blockade of RANKL. By cleav- peptide was precipitated and washed with diethyl ether, then dis- age of RANK EP would be the first inhibitor for RANK–RANKL sig- solved in 50% aqueous acetonitrile containing 0.1% TFA and lyoph- nalling by targeting RANK instead of RANKL. ilised. Analytical RP-HPLC was performed on a Shimadzu LC-15C The aim of this study was to confirm the potential ability of EP system with a C-18 silica column (4.6 Â 250 mm, Venusil to inhibit osteoclast differentiation by the cleavage of RANK. In this ASB150 Å, 5 lm, Bonna-Agela Technologies, China) at flow rate of study, RANK could be efficiently cleaved by EP, this was demon- 1 ml/min. The peptide was eluted from the column using a gradi- strated by bands shift on SDS–PAGE. The inhibition effect of osteo- ent of acetonitrile/0.08% TFA (solvent A) versus water/0.1% TFA clast development was demonstrated in a model of osteoclast (solvent B). Preparative HPLC of the crude peptide after SPPS was differentiation with bone marrow-derived macrophages (BMMs) performed on Shimadzu LC-15C system on aVenusil ASB C-18 in the presence of EP. (250 Â 10 mm, 150 Å, 5 lm, Bonna-Agela Technologies, China) column. 2. Materials and methods 2.3. RANK cleavage assay 2.1. Cell lines, reagents and protein purification Recombinant mouse, human RANK, and the mixture of recom- The Raw264.7 cell line was purchased from American Type Cul- binant mouse RANK and RANKL at different ratios were incubated ture Collection and propagated in culture according to the manu- with different concentrations of EP at 37 °C for six hours. Recombi- facturer’s protocol. BMMs were isolated from the bone marrow, nant mouse RANK and EP were incubated with PMSF at 37 °C for which was obtained by flushing femurs and tibiae from 7-wk-old six hours. TNFRSF18 was treated at the same conditions. Different BALB/c mice as previously reported [14]. Human EP was purchased concentrations of the surrogate peptide were added into the mix- from Sigma-Aldrich (St. Louis, MO, USA). The TRAP kit was pur- ture of mouse RANK and EP then incubated at 37 °C for six hours. chased from Sigma-Aldrich (St. Louis, MO, USA). Anti-RANK All the samples were detected by SDS–PAGE. The recombinant (Ab10498) antibody was purchased from Abcam (Hong Kong, Chi- mouse RANK was incubated with or without EP at 37 °C for six na). The rat anti-mouse RANKL mAb (IK22.5) was the gift from Dr. hours, and the samples were detected by Western blotting. All pro- Hisaya Akiba (Juntendo University, JP). Phenylmethanesulfonyl teins, proteases and peptide above were dissolved in PBS. fluoride (PMSF) was purchased from MP Biomedical (Santa Ana, California, USA). Macrophage-colony stimulating factor (M-CSF) 2.4. Differentiation of BMMs was purchased from eBioscience (San Diego, CA, USA). The Cell Counting Kit was purchased from Dojindo Laboratories (Kuma- BMMs were cultured in -MEM media containing 10% foetal bo- moto, Japan).The NF-jB luciferase plasmid was the gift from pro- a fessor Xin Ye (Institute of Microbiology, CAS, China). vine serum (FBS) and 25 ng/ml M-CSF with different concentrations of EP at a density of 1 105 cells per well in a 24-well Lipofectamine 2000 was purchased from Invitrogen (Carlsbad, Â CA, USA).The SuperSignal West Pico Trial Kit was purchased from culture plate. After adding EP for 1 day, 20 ng/ml mouse RANKL Thermo (Rockford, USA). The purifications of recombinant human was added to each well except for the control and the EP alone RANK, mouse RANK and mouse RANKL were performed as previ- group. The media with fresh growth factors was changed every ously reported [14,15]. TNFRSF18 was prepared by our laboratory two days. Ten days later, the cells were fixed and stained for tar- using standard E coli expression system. trate-resistant acid phosphatase (TRAP). Surrogate peptide

Fig. 1. Sequence alignment of the extracellular domain of RANKs from different species. 2960 Y. Zhao et al. / FEBS Letters 587 (2013) 2958–2964

(0.2 mg/ml) was added into each well of BMMs treated as men- present (Fig. 2A). Higher concentrations of EP yielded more pro- tioned above. The cells with the same treatment as above were found cleavage. Based on the features of substrates required by used to quantify TRAP activity as previously described [12]. Briefly, EP, only one potential cleavage site, NEEDK, exists in the sequence Naphthol AS phosphate substrate in 100 mM sodium acetate (pH of RANK (Fig. 1) and thus the expected cleavage fragment is about 5.2) reacted with substrates at 37 °C for 30 min, and the reaction 19 kD. The analysis of SDS–PAGE showed that a digested band ap- product was then quantified by spectroscopic measurement at peared around 19 kD (shown by arrow, Fig. 2A), which matched 405 nm. Each experiment was repeated four times for statistical with an expected size. However more bands could also be seen be- analysis. low the expected band, which might be due to that the protein be- comes more vulnerable to enzyme degradation after being 2.5. Cell survival assay digested and the conformation destroyed. Recombinant human RANK could also be specifically cleaved by EP under the same con- The toxicity of EP to BMMs was assessed by Cell Counting Kit-8 ditions (Fig. 2B). The 19 kDa of fragment observed by Western blot- assay as previously reported [17]. Briefly, BMMs were cultured at a ting further proved that the cleavage of RANK by EP would happen density of 5 Â 104 cells per well in 100 ll mediain a 96-well plate. at the site of NEEDK sequence (Fig. 2C). One day later, different concentrations of EP (0, 0.2, 0.5, 1 lg/ml) Since RANKL and RANK bind to each other, it is worthy of con- were added to each well except for the blank control wells, which sideration whether RANKL plays a protective role in the process of contained the media. After incubation for 2 days, the media was re- cleavage. In this study, different mass ratios (1:0, 1:1, 1:2, 1:4) of moved, and the cells were washed twice with fresh media. Then, RANK and RANKL were used to test the potential of protective ef- 100 ll of fresh serum-free a-MEM containing 1/10 (v/v) Cell fect of RANKL. In the group of 1:0, RANK was clearly cleaved. With Counting Kit-8 reagent was added to each well and incubated for increased concentrations of RANKL, the cleavage effects were an additional 4 hours. After incubation, the viability of BMMs weakened but still present (Fig. 2D). These results implied that was measured with Cell Counting Kit-8 using a 96-well plate read- RANKL had a limited protective effect on the EP-mediated cleavage er (DG5032, Huadong, Nanjing, China) at 450 nm. The group with- of RANK. out EP served as a negative control. Toxicity was assessed by cell Since EP is a serine protease, we attempted to block its activity survival rate (optical density of the treated cells minus the OD of with PMSF, a serine protease inhibitor, to confirm that cleavage the blank control)/(OD of negative control minus the OD of the was EP-dependent. Mouse RANK was incubated with EP, a mixture blank control) Â 100%). of EP and PMSF, PMSF alone or nothing. The results of SDS–PAGE analysis demonstrated that RANK could be cleaved only in the EP 2.6. Transfection of NF-jB reporter plasmid and the measurement of group (Fig. 2E). These results suggested that the digestion was spe- luminescence cifically mediated by EP. To confirm the specificity of the EP-dependent cleavage, we RAW264.7 cells were transfected with NF-jB luciferase plasmid used TNFRSF18, another member of tumour necrosis factor recep- with lipofectamine 2000 according to the manufacturer’s protocol. tor superfamily (TNFRSF) as a control. Since there is no substrate After 48 hours, transfected cells were divided into 24-well culture motif presented in TNFRSF18 for EP we expected that there would plates at the density of 1x105. 24 hours later, the cells were trans- be no cleavage on TNFRSF18 by EP. After incubation of TNFRSF18 ferred in the medium containing mouse RANKL (50 ng/ml) and with EP for 2 h at 37 °C, SDS–PAGE was carried out and subjected incubated for 8 h. The luminescence was measured as previously to Coomassie brilliant blue staining to demonstrate the cleavage. described[18]. Briefly, 0.5% NP40 dissolved in PBS was added into As expected, TNFRSF18 was not cleaved while RANK was clearly the well for the preparation of cell lysates, and the cells were broke down by EP (Fig. 2F). scraped from dishes. The supernatant was obtained through the To further prove that the digestion site was on the sequence of centrifugation at 13000 rpm for 2 min. 50 ll of supernatant was NEEDK, we designed a surrogate peptide, DTWNEEDKCLL corre- placed into wells of an opaque black 96-well plate and an equal sponding to residues 78-88 of mouse RANK and containing the pre- volume of luciferin was added into each well. Then luminescence dicted cleavage amino acids. Recombinant mouse RANK was was measured by Synergy TM H4 Multi-mode Microplate Reader incubated with EP along with different concentrations of peptides (Biotek). at 37 °C for six hours. SDS–PAGE analysis showed that the surrogate peptide could compete with mouse RANK and protected RANK 2.7. Western blotting from being digested by EP in a dose-dependent manner (Fig. 2G). The data indicates that the sequence NEEDK was the specifically Prior to Western blotting, 5 Â 104 RAW264.7 cells were cul- cleavage site on RANK by EP. tured in each well of a 24-well culture plate and stimulated by 25 ng/ml M-CSF and 20 ng/ml mouse RANKL for two days. Differ- 3.2. EP can block the differentiation of BMMs into osteoclasts ent concentrations of EP were added to the wells and incubated for ten hours. The cell lysates were subjected to 10% SDS–PAGE BMMs were isolated from bone marrow that was obtained by and Western blotting. The signal was detected by SuperSignal West flushing femurs and tibiae from 7-wk-old BALB/c mice, and the dif- Pico Trial Kit. ferentiation assay was carried out to assess the ability of EP to inhibit the differentiation of BMMs. All the groups of BMMs were 3. Results cultured in a-MEM media with 25 ng/ml M-CSF. The control group was cultured without EP, while the EP group was cultured with 3.1. RANK can be specifically cleaved by EP 200 ng/ml EP. The osteoclasts were observed after the staining of TRAP. There were no TRAP staining-positive cells or mature osteo- Recombinant mouse RANK was incubated with different con- clasts (TRAP staining-positive, multinuclear cells with diame- centrations of EP at 37 °C for six hours. SDS-PAGE analysis showed ter > 100 lm) formed in either group. Extra two groups were that mouse RANK was specifically cleaved in the presence of cultured in a-MEM media containing 50 ng/ml mouse RANKL with different concentrations of EP, by contrast there was no cleavage or without 200 ng/ml EP. The cells in the group without EP clearly on mouse RANK when thrombin, a different serine protease was differentiated, and the number of mature osteoclasts was Y. Zhao et al. / FEBS Letters 587 (2013) 2958–2964 2961

Fig. 2. Cleavage of hRANK and mRANK by EP. Recombinant mouse (A) or human (B) RANK (0.8 mg/ml) was incubated with different concentrations of EP or 1 lg/ml thrombin at 37 °C for 6 h. The samples were loaded onto a 15% SDS–PAGE gel, and the gel was stained with Coomassie brilliant blue. RANK (0.2 mg/ml) was incubated with EP (1 lg/ml) at 37 °C for 6 hours and then subjected to Western blotting (C). (D) RANKL had a limited protective effect from the cleavage of RANK by EP. mRANK (0.2 mg/ml) was mixed with mRANKL in different mass ratio (1:0, 1:1, 1:2 and 1:4, room temperature for 15 min) and then the mixtures were incubated with 1 lg/ml EP at 37 °C for 6 h. (E) PMSF could block the cleavage of RANK mediated by EP. mRANK (0.8 mg/ml) was incubated with or without 3 lg/ml EP or 1 mM PMSF. (F) RANK (0.8 mg/ml) and TNFRSF18 (0.5 mg/ml) were incubated with or without 1 lg/ml EP at 37 °C for 6 h. The samples were loaded onto a 15% SDS–PAGE gel, and the gel was stained with Coomassie brilliant blue. (G) Recombinant mouse RANK (0.15 mg/ml) was incubated with 0.1 lg/ml EP and different concentrations of surrogate peptide at 37 °C for 6 h. The samples were loaded onto a 15% SDS–PAGE gel, and the gel was stained with Coomassie brilliant blue.

112.5 ± 43.8. Conversely, the group with EP had few TRAP staining- To confirm that the differentiation inhibition was not due to positive cells and no mature osteoclasts (Fig. 3A and B). toxicity of EP, the Cell Counting Kit-8 assay was performed. EP We further measured the TRAP activity of the BMMs under con- exhibited little toxicity to the BMMs in high concentration groups ditions of different EP concentrations by spectroscopic measure- (1, 0.5 lg/ml), in which the survival rates were 96 and 98%, and no ment at 405 nm. The control group was cultured without EP or toxicity at the low concentration (0.2 lg/ml), which was the work- mouse RANKL, while the other groups were cultured with 20 ng/ ing concentration for the differentiation assay (Fig. 3D). These re- ml mouse RANKL and EP at different concentrations. With de- sults indicated that there was not a link between the inhibition creased concentrations of EP, the TRAP activity increased. The TRAP effect of EP and its toxicity to BMMs. activity of the group cultured with 400 ng/ml EP was similar to the To demonstrate downstream NF-jB pathway was affected control group. Meanwhile, after the removal of the background va- when RANK was digested by EP enzyme, we transfected NF-jB lue, the TRAP activity of the group cultured without EP was luciferase report plasmid into RAW264.7 cells and monitor the approximately fifty-fold higher compared to the group cultured activation of NF-jB signalling to confirm that the inhibition of with 200 ng/ml EP (Fig. 3C). osteoclastogenesis by EP was achieved by the cleavage of RANK. 2962 Y. Zhao et al. / FEBS Letters 587 (2013) 2958–2964

Fig. 3. Inhibition of osteoclast differentiation mediated by EP. (A) The control group was cultured in media without EP or RANKL. The group cultured in the presence of 50 ng/ ml RANKL contained many multinucleated, large diameter (>100 lm) TRAP staining-positive cells. The group cultured in media both EP (200 ng/ml) and RANKL (50 ng/ml) contained few incompletely differentiated, TRAP staining-positive and small diameter cells. The group cultured in the presence of 200 ng/ml EP did not contain differentiated cells. All the groups were cultured with 25 ng/ml M-CSF. Scale bar, 300 lm. (B) The multinucleated, TRAP staining-positive and large diameter (>100 lm) cells were counted. The bar represents the average of four independent experiments, and the data were shown as the mean ± S.D. ⁄⁄P < 0.01. (C) TRAP activity of the BMMs cultured with different concentrations of EP. All the groups were cultured in a-MEM media containing 25 ng/ml M-CSF. The control group was cultured without RANKL and EP, while the other groups were cultured with 20 ng/ml RANKL and different concentrations of EP. The curve represents the average of four independent experiments with S.D. bars. (D) Cytotoxicity assay for BMMs cultured with different concentrations of EP. The negative control was untreated cells, and the blank control contained the Cell Counting Kit-8 reagent without cells. The curve represents the average of three independent experiments, and the data were shown as the mean ± S.D. (E) Luciferase activity were presented as the folds of basal value (control). RAW264.7 cells were transfected with NF-jB luciferase plasmid for 48 hours and then the cells were incubated with 50 ng/ml mouse RANKL for 8 hours. Then the luminescence was measured. Each group was repeated for four times. The data were shown as the mean ± S.D. ⁄P < 0.05. (F) The control group was cultured with or without RANKL, EP or peptide. The concentrations of EP, mouse RANKL and peptide were 200 ng/ml, 50 ng/ml and 0.2 mg/ml respectively. All the groups were cultured with 25 ng/ml M-CSF. Scale bar, 300 lm. (G) The multinucleated, TRAP staining-positive and large diameter (>100 lm) cells were counted. The bar represents the average of four independent experiments, and the data were shown as the mean ± S.D. ⁄⁄⁄P < 0.001. (H) TRAP activity of the BMMs cultured in the same condition as that shown on Fig. 3F. The colomn represents the average of four independent experiments with S.D. bars. ⁄⁄P < 0.01.

The activity of NF-jB was proportional to the activity of luciferase. We further tested whether the surrogate peptide could inhibit As shown in Fig. 3E, the activity of NF-jB in the RANKL group, was the effect of EP on RANK by BMM differentiation assay. There about 3-fold of the group with EP added and the group with anti- was no differentiated cell in the EP group when the cells were cul- RANKL antibody. The activity of NF-jB of the RANKL group was tured with mouse RANKL in the presence of EP. While in the group close to the EP and surrogate group (Fig. 3E).These results indicate cultured with mouse RANKL in the presence of both EP and the that EP can inhibit RANKL-induced NF-jB activation. peptide, there was a similar number of differentiated cells Y. Zhao et al. / FEBS Letters 587 (2013) 2958–2964 2963

Fig. 4. Western blotting of RANK expression on the surface of RAW264.7 cells. After cleavage by EP, Western blotting was performed to estimate the cleavage effect. In the 400 and 200 ng/ml groups, two bands could be clearly seen. compared with either the group cultured with RANKL or the group demonstrate that the RANK can be used as a target to block cultured with RANKL and peptide (Figs. 3F and G). We have also RANKL–RANK signalling. Furthermore, EP has been proved to be measured the TRAP activity of the BMMs treated as mentioned effective for inhibiting RANK functions in vitro and would be a use- above. The TRAP activities of control group and the peptide alone ful tool for osteoclastogenesis-related research. group were almost the same. The group cultured with RANKL, EP The classical substrate sequence for EP recognition is (E/D)4-K. and the peptide had a little lower TRAP activity compared with RANKs from different species have the NEEDK sequence, which is a that in the group cultured with RANKL and RANKL plus peptide little different from the classical recognition sequence. However, group (Fig. 3H). These results together suggested that the inhibi- based on the features of the substrates that EP can recognise, we tion of osteoclastogenesis by EP was achieved through the cleavage predicted that RANK could be recognised and cleaved. In deed as of RANK. predicted, RANK could be cleaved by EP. A weaker cleavage was expected, due to the non-classical sequence in RANK. Although 3.3. Cleavage of RANK expressed on the cell membrane by EP NEEDK sequence is not an ideal substrate sequence, the inhibition of BMM differentiation was not affected. After the cleavage of RANK by different concentrations of EP, As we reported previously, residues Glu84 and Asp85 of RANK Western blotting was performed to estimate the cleavage effect. in the sequence NEEDK (mouse residue 81-86), play an important We used RAW264.7 cells, a murinemacrophage-likecell line with role in the interaction of RANKL and RANK [14]. We examined the high levels of RANK expression, to examine the cleavage of mem- cleavage of RANK in the presence of RANKL to determine whether brane bound RANK by Western blotting. As shown in Fig. 4,it RANKL had a protective effect. Our results show that, although was showed that anti-RANK antibody could specifically detect in- RANKL and RANK could form a heterotrimer complex, EP could rec- tact RANK and digested RANK expressed on RAW264.7 cells, while ognise the sequence and cleave RANK in a weaker manner. The could not detect this protein on K41, a wild-type mouse embryonic three-dimensional steric hindrance may have given rise to the fibroblasts cell line used as a RANK-expression negative control. weaker cleavage. With decreased concentrations of EP, the cleavage effect became The RANK-cleavage activity of EP presented in our study sug- weaker. EP at the concentrations of 200 and 400 ng/ml had a signif- gests a possible application of this enzyme in the treatment of dis- icant cleavage effect for the RANK expressed on cells was cleaved eases. The RANKL–RANK signalling pathway plays a pivotal role in into two visible segments. regulating osteoclast development and function. Therefore, this pathway is demonstrated to be a good target for treatments of oste- 4. Discussion oporosis, tumour bone metastases and other diseases associated with bone loss. We speculate that EP would be a potential treat- In this study, we showed that EP had the ability to cleave RANK ment for bone loss related diseases since it could block osteoclast on the sequence NEEDK. By contrast, thrombin, a different serine differentiation. The supposition needs to be tested in future trials. protease without ability to cleave NEEDK sequence failed to cleave Moreover, the cleavage activity of EP may provide a clue to its RANK, and EP could not cleave TNFRSF18, another member of potential roles in the development of diseases. EP is reported to TNFRSF without NEEDK sequence. Moreover, PMSF could block be expressed at the surface of prostate cancer cells [19].We the ability of EP to cleave RANK. A surrogate peptide of RANK could hypothesize it may contribute to the formation of osteoblastic le- protect RANK from being digested by EP. In the NF-jB activation sions at the early stage of prostate cancer bone metastases through assay, NF-jB signalling could be blocked by EP by the cleavage of the inhibition of RANKL–RANK signalling pathway. Besides, duode- RANK. Similarly, in the BMM differentiation assay, the surrogate no-pancreatic reflux of proteolytically active EP can cause acute peptide could rescue RANKL-induced osteoclastogenesis. These re- pancreatitis [6], which gives rise to the impairment of mucosal sults showed that the cleavage of RANK was effective and specific. immunity and a decreased concentration of secretary IgA (SIgA). As shown in Fig. 1, RANKs from different mammals (human, This is the primary reason for bacterial and endotoxin transloca- mouse, rat, rabbit, and gorilla) contain the NEEDK sequence. There- tion [20]. In 2009, Kathryn A. Knoop et al. reported that the fore, we predicted that RANKs of various species could be cleaved RANKL-RANK pathway controls the differentiation of micro-fold by EP. In deed, both mouse and human RANK could be cleaved by cells (M cells), the antigen transporters to T cells that support B cell EP (Fig. 2A and 2B). activation, from RANK-expressing intestinal epithelial precursor There are several inhibitors available for RANKL-RANK signal- cells [21]. We hypothesize that during acute pancreatitis, EP ling pathway for instance, OPG-Fc, RANK-Fc, the human monoclo- cleaves not only trypsinogen but also RANK expressed on the sur- nal antibody denosumab. These inhibitors all interfere with face of epithelial precursor cells. This may cause the inhibition of M RANKL-RANK signalling through the blockade of RANKL. Here we cell differentiation. Furthermore, the lack of M cells may inhibit B have described EP, a novel inhibitor of the RANKL–RANK signalling cell activation and cause a decrease in SIgA. These hypotheses re- by the cleavage of RANK. We believe that this is the first time to quire further investigation. 2964 Y. Zhao et al. / FEBS Letters 587 (2013) 2958–2964

Acknowledgements [10] Body, J.J., Greipp, P., Coleman, R.E., Facon, T., Geurs, F., Fermand, J.P., Harousseau, J.L., Lipton, A., Mariette, X. and Williams, C.D. (2003) A Phase I study of AMGN-0007, a recombinant osteoprotegerin construct, in patients This work was supported by grants from the Ministry of Science with multiple myeloma or breast carcinoma related bone metastases. Cancer and Technology of China (S & T Major Program: No. 97, 887–892. 2012ZX10004701-001-002), 973 program (2013CB531502) and [11] Hsu, H., Lacey, D.L., Dunstan, C.R., Solovyev, I., Colombero, A., Timms, E., Tan, H.-L., Elliott, G., Kelley, M.J. and Sarosi, I. (1999) Tumor necrosis factor receptor the National Natural Science Foundation of China (No.31070783, family member RANK mediates osteoclast differentiation and activation 31170829, 81021003, 81270424, 81100550, 81100600, induced by osteoprotegerin ligand. Proc. Nat. Acad. Sci. 96, 3540–3545. 81041024), Beijing Natural Science Foundation (5112022), China [12] Ta, H.M., Nguyen, G.T.T., Jin, H.M., Choi, J., Park, H., Kim, N., Hwang, H.-Y. and Kim, K.K. (2010) Structure-based development of a receptor activator of International Medical Foundation (CIMF) - Novo Nordisk China nuclear factor-jB ligand (RANKL) inhibitor peptide and molecular basis for Diabetes Young Scientific Talent Research Funding (2012-25). We osteopetrosis. Proc. Nat. Acad. Sci. 107, 20281–20286. thank Prof. Fremont (Washington University, USA) for supplying [13] Aoki, K., Saito, H., Itzstein, C., Ishiguro, M., Shibata, T., Blanque, R., Mian, A.H., Takahashi, M., Suzuki, Y. and Yoshimatsu, M. (2006) A TNF receptor loop the the GST–RANKL construct, Dr. Hisaya Akiba (Juntendo Univer- peptide mimic blocks RANK ligand-induced signaling, bone resorption, and sity, JP) for supplying the rat anti-mouse RANKL mAb (IK22.5) and bone loss. J. Clin. Invest. 116, 1525. Prof. Xin Ye (Institute of Microbiology, CAS, China) for supplying [14] Liu, C., Walter, T.S., Huang, P., Zhang, S., Zhu, X., Wu, Y., Wedderburn, L.R., Tang, P., Owens, R.J. and Stuart, D.I. (2010) Structural and functional insights of the NF-jB luciferase plasmid. RANKL–RANK interaction and signaling. J. Immunol. 184, 6910–6919. [15] Zhang, S., Liu, C., Huang, P., Zhou, S., Ren, J., Kitamura, Y., Tang, P., Bi, Z. and References Gao, B. (2009) The affinity of human RANK binding to its ligand RANKL. Arch. Biochem. Biophys. 487, 49–53. [16] Schnölzer, M., Alewood, P., Jones, A., Alewood, D. and Kent, S.B. (2007) In situ [1] Zamolodchikova, T.S., Sokolova, E.A., Lu, D. and Sadler, J.E. (2000) Activation of neutralization in Boc-chemistry solid phase peptide synthesis. Int. J. Pept. Res. recombinant proenteropeptidase by duodenase. FEBS Lett. 466, 295–299. Ther. 13, 31–44. [2] Kunitz, M. and Northrop, J.H. (1936) Isolation from beef pancreas of crystalline [17] Han, H., He, W., Wang, W. and Gao, B. (2011) Inhibitory effect of aqueous trypsinogen, trypsin, a trypsin inhibitor, and an inhibitor-trypsin compound. J. dandelion extract on HIV-1 replication and reverse transcriptase activity. BMC Gen. Physiol. 19, 991–1007. Complement. Altern. Med. 11, 112. [3] Neuruth, H. (1957) The activation of zymogens. Adv. Protein Chem. 12, 319. [18] Tsai, H.-Y., Lin, H.-Y., Fong, Y.-C., Wu, J.-B., Chen, Y.-F., Tsuzuki, M. and Tang, C.- [4] Furie, B. and Furie, B.C. (1988) The molecular basis of blood coagulation. Cell H. (2008) Paeonol inhibits RANKL-induced osteoclastogenesis by inhibiting 53, 505–518. ERK, p38 and NF-jB pathway. Eur. J. Pharmacol. 588, 124–133. [5] Hadorn, B., Haworth, J., Gourley, B., Prasad, A. and Troesch, V. (1975) Intestinal [19] Cottrell, G.S., Amadesi, S., Grady, E.F. and Bunnett, N.W. (2004) Trypsin IV A enterokinase deficiency. Occurrence in two sibs and age dependency of clinical novel agonist of protease-activated receptors 2 and 4. J. Biol. Chem. 279, expression. Arch. Dis. Child. 50, 277–282. 13532–13539. [6] Grant, D. (1986) Acute necrotising pancreatitis—a role for enterokinase. Int. J. [20] Qiao, S.-F., Lu, T., Sun, J.-B. and Li, F. (2005) Alterations of intestinal immune Gastrointestinal Cancer 1, 167–183. function and regulatory effects of L-arginine in experimental severe acute [7] Barrett AJ, Rawlings ND, Woessner JF. Handbook of proteolytic enzymes: pancreatitis rats. World J. Gastroenterol. 11, 6216. Academic Press San Diego, CA; 1998. [21] Knoop, K.A., Kumar, N., Butler, B.R., Sakthivel, S.K., Taylor, R.T., Nochi, T., Akiba, [8] Lu, D., Yuan, X., Zheng, X. and Sadler, J.E. (1997) Bovine proenteropeptidase is H., Yagita, H., Kiyono, H. and Williams, I.R. (2009) RANKL is necessary and activated by trypsin, and the specificity of enteropeptidase depends on the sufficient to initiate development of antigen-sampling M cells in the intestinal heavy chain. J. Biol. Chem. 272, 31293–31300. epithelium. J. Immunol. 183, 5738–5747. [9] Rizzoli, R., Yasothan, U. and Kirkpatrick, P. (2010) Denosumab. Nat. Rev. Drug Discovery 9, 591–592.