|
Advertisement | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
J. Biol. Chem., Vol. 279, Issue 10, 9167-9175, March 5, 2004
Insulin Promotes Rat Retinal Neuronal Cell Survival in a p70S6K-dependent Manner*![]() ![]() ![]() ![]() ![]() ![]() ![]() ¶
From the
Received for publication, November 12, 2003 , and in revised form, December 4, 2003.
The purpose of this study was to examine the role of the ribosomal protein S6 protein kinase (p70S6K), a protein synthesis regulator, in promoting retinal neuronal cell survival. Differentiated R28 rat retinal neuronal cells were used as an experimental model. Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% newborn calf serum, and during the period of experimentation were exposed either to the absence or presence of 10 nM insulin. Insulin treatment induced p70S6K, mTOR, and Akt phosphorylation, effects that were completely prevented by the PI3K inhibitor, LY294002. Insulin-induced phosphorylation of p70S6K and mTOR was prevented by the mTOR inhibitor, rapamycin. Apoptosis, induced by serum deprivation and evaluated by Hoechst staining, was inhibited by insulin treatment in R28 cells, but not in L6 muscle cells. This effect of insulin was also largely prevented by rapamycin. Inhibition of p70S6K activity by exogenous expression of a dominant negative mutant of p70S6K prevented insulin-induced cell survival, whereas, overexpression of wild type p70S6K or expression of a rapamycin resistant form of the kinase enhanced the effect of insulin on survival. Enhanced cell survival under the latter condition was accompanied by increased p70S6K activity and phosphorylation. Rapamycin did not inhibit insulin induced p70S6K phosphorylation and activity in cells transfected with the rapamycin-resistant mutant. Together, these results suggest that p70S6K plays a key role in insulin stimulated retinal neuronal cell survival.
Apoptotic cell death of retinal neuronal and vascular cells contributes to the pathogenesis of diabetic retinopathy (1, 2) but the mechanism of this process is uncertain. Several studies have identified activation of phosphoinositide 3-kinase (PI3K)1 as a necessary step in the cell survival pathway that is stimulated by a number of growth factors and insulin (35). Akt/PKB, a protein kinase that functions downstream of PI3K in the insulin signal transduction pathway, is also involved in the regulation of cell survival (6). Evidence for a role of Akt in IGF-1-mediated cell survival was provided by Dudek et al. (7) showing that overexpression of Akt prevents apoptosis in primary cultures of cerebellar neurons induced by survival factor withdrawal or chemical inhibition of PI3K. The expression of dominant-negative forms of Akt interferes with growth factor-mediated survival in these cells, indicating that Akt is necessary and sufficient for neuronal survival. We have also shown that both IGF-I and insulin can rescue retinal neuronal cells from apoptosis through a PI3-kinase/Akt mediated mechanism (3), and that systemically administered insulin activates the retinal insulin receptor, PI3K, and Akt1 in normal rats (8).
The 70 kDa ribosomal protein S6 kinase, p70S6K, is one of the downstream effectors of PI3K (912), (13). It is now believed that there are five domains in the primary structure of S6K1. Beginning with the N terminus, they are: acidic, catalytic, linker, autoinhibitory, and the C-terminal domain (1418). The mechanism through which S6K1 is activated is complicated (1922), and involves interactions among four of the five domains and phosphorylation of at least seven specific regulatory sites (17). Of those 7 sites, phosphorylation of Thr389 and Thr229 are critical for p70S6K activation. To date, PI3K (23, 24), PDK1, Akt/PKB (2528), protein kinase C (PKC) (29, 30), the Rho family of small G proteins and the mammalian target of rapamycin protein kinase (mTOR) (3136) are thought to be the upstream effectors of S6K1 phosphorylation. Many growth factors, including insulin, activate p70S6K in a PI3K-dependent manner. In addition, amino acids can activate S6K1 in a PI3K-independent manner. Compared with the upstream effectors, the only known downstream effector of S6K1 is the 40 S ribosomal S6 protein. Phosphorylation of S6 allows translational up-regulation of mRNAs containing 5'-tracts of pyrimidines (TOP), which encode for components of the translational apparatus (37), cell cycle-related (G1 to S transition) transcription factor E2F and insulin transcription (3840). Most studies of S6K1 including ours have to date focused on its role in protein synthesis and cell cycle regulation. In those studies, we reported that insulin controls protein synthesis in skeletal muscle through activation of p70S6K (41, 42). However, a number of recent studies have suggested that S6K1 may be intimately involved in mediating cell survival. Wan and Helman (43) found that inhibition of the p70S6K pathway may enhance chemotherapy-induced apoptosis in the treatment of IGF-II-overexpressing tumors. Agents that cause apoptosis inactivate mTOR signaling as a common early response prior to caspase activation (44). Rapamycin, a macrolide immunosuppressant that is a specific inhibitor of mTOR, represses p70S6K, prevents phosphorylation of Ser136 on BAD, and blocks cell survival induced by IGF-I. IGF-I-induced phosphorylation of BAD Ser136 is abolished in p70S6K-deficient cells (45). In addition, rapamycin induces apoptosis in different cell types (4649). Overall, the results of these studies strongly support a role for p70S6K in promoting cell survival. Based on the results of the studies described above, we hypothesize that insulin promotes retinal neuronal cell survival in a p70S6K-dependent manner. In the present study, the rat retinal neuronal cell line R28 was used as an in vitro model to examine this hypothesis. We found that insulin promotes phosphorylation and activation of p70S6K and exogenous expression of wild type S6K1 decreased pyknotic cell numbers in serum-starved cells and insulin further decreased apoptosis in such cells. Rapamycin did not block the insulin cell survival effect in R28 cells expressing a rapamycin-resistant p70S6K, indicating that insulin promotes R28 cell survival in a p70S6K-dependent manner. These data suggest that p70S6K phosphorylation and activity may contribute to retinal cell survival in vivo.
PlasmidsS6K1 constructs were a kind gift from Dr. John Blenis (Department of Cell Biology, Harvard Medical School, Boston) (50, 51). The constructs include control vector pRK7, wild type construct HA-p70S6K1/pRK7, kinase inactive construct HA-p70S6K1-F5A, and C-terminal truncated construct HA-p70S6K1-E389- CT. Detailed primary sequence alignment has been previously reported (52). Plasmids were prepared using a Qiagen EndoFree Plasmid Kit. Cell CultureR28 cells were a generous gift from Dr. Gail M. Seigel, State University of New York, Buffalo (53). They were grown in Dulbecco's modified Eagle's medium (DMEM) containing 5 mM glucose supplemented with 10% newborn calf serum (Hyclone). The cells were differentiated to neurons on laminin-coated plates or coverslips with addition of 25 mM cell-permeable cAMP (Sigma) 24 h prior to experiments, as described previously (3, 4). Cells were seeded at 5 x 105/60-mm dish or 3 x 105/well in 6-well plates with coverslips 24 h prior to transfection or treatments. Cells were then placed in serum-free DMEM for 24 h prior to stimulation and lysis. L6 muscle cells were purchased from ATCC and were also cultured in DMEM supplemented with 10% newborn bovine serum. TransfectionR28 cells were seeded 24 h before transfection at 3 x 105 (6-well plates) or 5 x 105 (60-mm plates). PLUSTM reagent (Invitrogen Life Technologies, Inc.) was mixed with serum-free DMEM and incubated with 3 µg/well (6-well plates) or 6 µg/dish (60-mm dishes) of DNA for 15 min to form the PLUSTM-DNA complex. LipofectAMINE was then added to the above mixture and incubation was continued for another 15 min. During the incubation, fresh serum-free medium was applied to R28 cells. LipofectAMINE-PLUS-DNA complex was added, and cells were incubated at 37 °C, 5% CO2 for 3 h. After 3 h of transfection, fresh medium with 10% NCS was added. This method routinely yields 3040% transfection efficiency in R28 cells. Cells were incubated for another 24 h (for apoptosis studies) or 36 h (for kinase assay and Western blot). Immunocytochemistry and Hoechst StainingTransfected R28 cells were deprived of serum, pretreated with or without rapamycin (10 nM), and treated for 24 h with or without 10 nM insulin. The cells were fixed in 1% paraformaldehyde and blocked at room temperature for 1 h in phosphate-buffered saline containing 0.1% Triton (PBST) and 10% donkey serum. They were then incubated at room temperature for 1 h in PBST with rabbit anti-HA polyclonal antibody (1:200, Clontech, Palo Alto, CA). The cells were washed and incubated at room temperature for 1 h with rhodamine red X-conjugated donkey anti-rabbit IgG (1: 2000, Jackson Immunologicals, West Grove, PA). Cells were simultaneously stained with the nuclear dye bisbenzimide (Hoechst dye 33258, 0.5 µg/ml, Sigma). For all other experiments in which R28 cells were stained with Hoechst, they were incubated in Hoechst (1:2000) for 1 h at room temperature, and mounted to slides with Aqua mount. Apoptosis Quantification(A) For cells stained only with bisbenzimide Hoechst 33258 (0.5 µg/ml; Sigma), five fields were randomly sampled from each coverslip by fluorescence microscopy, and all the cells stained with Hoechst dye in each field were counted. The number of pyknotic cells with condensed or fragmented nuclei was summated in the five sampled regions. The percentage of pyknotic cells per coverslip was then calculated as described previously (3, 4). (B) For R28 cells transfected with S6K1 constructs, at least 100 HA-positive cells were counted per coverslip. Pyknotic cell numbers were also measured for HA-positive cells. Protein Extraction and Protein AssayR28 cells were harvested as described previously (54) in Triton buffer (10 mM HEPES, 42 mM KCl, 5 mM MgCl2, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 mM sodium pyrophosphate, 10 mM benzamidine, 1% Triton X-100, and protease inhibitor tablet; Roche Applied Science, Mannheim, Germany). Protein was extracted by incubating samples at 4 °C for 15 min followed by 10min centrifugation at 14,000 x g. Protein concentration was determined by using Bio-Rad DC reagent compared with bovine serum albumin standard. S6 Kinase AssayS6 kinase assay kit was from Upstate Biotechnology. Manufacturer's instructions were followed in performing the assay. Immunoprecipitation of p70 S6 kinase: 2 µg of anti-p70 S6 kinase or anti-HA antibody were prebound to 100 µl of a 50% protein A-agarose slurry that had been washed in PBS for 1h at 4 °C. The supernatant was removed and the protein A-agarose was washed twice with Buffer A. The washed beads were resuspended in 200 µl of Buffer A, and cell lysates containing 200 µg of protein were then immunoprecipitated using the prepared beads for 2 h at 4 °C. Protein A-agarose/enzyme immunocomplex was sequentially washed in 500 µl of Buffer A containing 0.5 M NaCl, Buffer A alone, and ADBI.
Kinase Assay with Protein A Enzyme Immunocomplex10 µl of ADBI, inhibitor mixture, substrate mixture, and [ Antibodies and ImmunoblottingAnti-p70S6 kinase SC-18 was from Santa Cruz Biotechnology. Anti-phospho-p70S6K (Thr389), anti-Akt/phospho-Akt (Ser473), and anti-mTOR/phospho-mTOR(Ser2448) were from Cell Signaling Technologies. Anti-phospho-p70S6 kinase (Thr229) was from R&D System and anti-HA-Tag was from Clontech (BD Biosciences). Proteins were separated on 10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were blocked using 5% nonfat milk (TBST) for 1 h at room temperature with rocking. Primary antibody (1:1000) incubations were carried out in 5% nonfat milk-TBST (total antibodies) or 5% bovine serum albumin-TBST (phosphospecific antibodies) overnight at 4 °C. Membranes were washed in TBST three times for 10 min. Membranes were then incubated in donkey-anti-rabbit IgG-HRP (1:5000) for 1 h at room temperature followed by another three washes in TBST. Proteins were visualized using LumiGlo chemiluminescence reagent (Cell Signaling Technologies), using a Genome bioimaging system and analyzed using GeneTool software (Syngene).
Statistic AnalysisThe minimum level of statistical significance was set at
Insulin Induces Phosphorylation/Activation of p70S6 Kinase in a Time-dependent Manner in R28 Rat Retinal Neuronal CellsInsulin induction of P70S6K phosphorylation has been demonstrated in several types of cells and tissues, but the response has not been previously studied in retinal cells. Previous studies from this laboratory (3) showed that insulin induces phosphorylation of Akt in R28 cells through PI3K-mediated signaling, and that 10 nM insulin activates the insulin receptor but not the IGF-I receptor. Because p70S6K acts downstream of PI3K, we asked if p70S6K is phosphorylated in response to insulin in R28 cells. To this end, we first sought to demonstrate an effect of insulin on p70S6K. Serum-deprived cells were incubated in medium containing 10 nM insulin for time periods ranging from 15 min to 24 h. The results show that insulin induced phosphorylation of p70S6K within 15 min, and the effect was maintained for at least 4 h (Fig. 1A). A similar time course for changes in phosphorylation was also observed for Akt (Fig. 1A). After 24 h, both Akt and p70S6K phosphorylation returned to basal values. Insulin also stimulated S6 kinase activity by 3-fold (p < 0.01) within 15 min as measured in extracts of insulin-treated compare with control cells (Fig. 1D). These data indicate that insulin can induce phosphorylation and activation of p70S6K in rat retinal neuronal cells, as has been observed for other types of cells.
Insulin Induces Phosphorylation of p70S6K through the PI3K/Akt/mTOR Pathway in R28 CellsWe next asked if the activation of p70S6K by insulin occurs through the PI3K/Akt pathway, so the PI3K inhibitor LY294002 and the mTOR inhibitor rapamycin were employed. As shown in Fig. 2, both LY294002 and rapamycin completely inhibited insulin-induced p70S6K phosphorylation. In contrast, LY294002 inhibited insulin-induced phosphorylation of Akt, but rapamycin did not (Fig. 2). To determine if insulin induces mTOR phosphorylation in R28 cells, cells were treated with either the PI3K inhibitor LY294002 or the mTOR inhibitor rapamycin prior to analysis of mTOR phosphorylation on Ser2448. As shown in Fig. 3, insulin increased mTOR phosphorylation by 2.5-fold (p < 0.01) in R28 cells, and both LY294002 and rapamycin completely inhibited insulin-induced mTOR phosphorylation. Together, these data suggest that insulin induces phosphorylation of mTOR through the PI3K pathway, and that insulin activates p70S6K via PI3K/Akt/mTOR in retinal neurons.
Rapamycin Blocks the Survival Effect of Insulin in R28 CellsPrevious studies have suggested that insulin protects R28 cells from apoptosis through a PI3K/Akt mediated mechanism (3). As a downstream effector of Akt, p70S6K may be involved in the cell survival effect of insulin. This idea is supported by the results of several studies showing that rapamycin is a proapoptotic agent (39, 47, 49), but its role in retinal cells is unknown. Here we investigated whether or not rapamycin affects the ability of insulin to promote cell survival in R28 cells. Differentiated R28 cells were incubated in serum free medium in the presence or absence of insulin (10 nM) or rapamycin for 24 h. R28 cells were maintained in serum-containing medium supplemented with or without rapamycin served as controls. Cells were then fixed in 1% paraformaldehyde and stained with Hoechst dye. Consistent with the previous studies, insulin significantly (p < 0.01) reduced the number of pyknotic cells in serum-starved R28 cells. Rapamycin alone had no effect on the number of pyknotic cells in the presence of serum but completely inhibited the cell survival effect of insulin on serum-starved cells (Fig. 4, A and B). This observation implies that components of serum other than insulin act through a rapamycin-independent mechanism to promote cell survival in R28 cells, and insulin acts via rapamycin dependent pathways. However, in cultures of L6 myoblasts, insulin rescued cells to the same extent as for R28 cells, but rapamycin had no effect on either insulin- or serum-induced cell survival (Fig. 4, C and D). Thus, the rapamycin effect is cell type-specific, and neurons may be more susceptible than muscle cells to effects of p70S6K inhibition.
Insulin Promotes R28 Cell Survival in a p70S6K-dependent MannerRecent evidence suggests p70S6K is involved in suppressing apoptosis in various cell types (39, 45, 55, 56). We then asked if the cell survival effect of insulin is p70S6K-dependent. To address this question, variants of p70S6K were introduced into R28 cells by cDNA transfection. The cDNA used encoded the wild type enzyme HA-p70S6K1, a kinase-inactive variant HA-p70S6K1-F5/A, a C-terminally truncated variant HA-p70S6K1-E389- CT, which is rapamycin-resistant, and the empty vector, pRK7. For each construct three different conditions were examined: control, insulin, and rapamycin+insulin. 24 h after transfection, cell culture medium was replaced with medium lacking serum with or without insulin or rapamycin for additional 24 h. As expected, the percentage of pyknotic cell numbers of control vector transfection in different conditions is similar to those in Fig. 4B. Counting of pyknotic cells that were also HA-positive showed that wild type S6K1 transfection reduced the percentage of pyknotic cell number as compare with control vector pRK7 transfection (Fig. 5, A and D). Insulin treatment further significantly reduced the pyknotic cell number to 25% of control (p < 0.001), and the effect of the hormone was completely blocked by rapamycin. In contrast, in cells transfected with the kinase inactive HA-p70S6K-F5/A variant, insulin did not affect HA-positive pyknotic cell number (Fig. 5, B and D). In cells transfected with the rapamycin-resistant HA-p70S6K1-E389- CT variant, significantly (p < 0.01) reduced pyknotic cell number was observed in all three conditions: control, insulin, and rapamycin plus insulin (Fig. 5, C and D). Rapamycin failed to block the cell survival effect of insulin in HA-p70S6K1-E389- CT-transfected cells, strongly indicating that insulin promotes R28 cell survival in a rapamycin-sensitive and p70S6K-dependent manner. The percentage of pyknotic cells was equivalent among different transfections (Fig. 5D). We conclude therefore that insulin specifically promotes retinal cell survival in a rapamycin-sensitive and p70S6K-dependent manner.
Insulin Induces Phosphorylation/Activation of both Endogenous and Transfected p70S6K in R28 CellsAmong the 8 phosphorylation sites in p70S6K, Thr389 and Thr229 are vital for p70S6K activation (16). The phosphorylation of both these residues on endogenous p70S6K was consistently increased in response to insulin, and rapamycin completely abolished the response (Fig. 6A). In transfected cells, no phosphorylation of exogenous p70S6K could be observed in control vector and kinase inactive vector transfections, but increased phosphorylation was observed in wild type p70S6K transfection for both sites. Insulin increased p70S6K phosphorylation at both Thr389 and Thr229. As predicted, rapamycin did not block phosphorylation of Thr229 in the HA-p70S6K1-E389- CT variant because of the loss of rapamycin-sensitive domain. Moreover, while endogenous p70S6K responded to insulin, no HA-p70S6K1-E389- CT phosphorylation on Thr389 was observed (Fig. 6A), because Thr389 was mutated to Glu in the variant. As shown, the HA-p70S6K1-E389- CT variant migrated farther in the gel because of the C-terminal truncation. These results are consistent with the kinase assay results (Fig. 6B). To assess the effect of phosphorylation on kinase activity, exogenously expressed p70S6K was immunoprecipitated with HA antibody and S6 kinase assays were performed. The results revealed a high basal S6Kinase activity in HA-S6K1 and HA-p70S6K1-E389- CT-expressing cells, and the activity of both was significantly increased (p < 0.01) by insulin. Rapamycin reversed the increase of p70S6K activity caused by insulin in cells expressing the HA-S6K1 variant but not in cells expressing the HA-p70S6K1-E389- CT variant. Only basal S6 kinase activity was detected in cells transfected with the control vector and cells expressing the kinase-inactive variant. These data provide further support for the hypothesis that p70S6K plays a key role in insulin-mediated retinal neuronal cell survival.
Mammalian retinas exhibit high level of expression of cell surface receptors for a variety of growth-promoting hormones including insulin (5760) and IGF-I (6163). The purpose of this study was to determine the effects of insulin on p70S6K activity and its role in retinal cell survival. We show here for the first time that p70S6K in rat retinal neuronal cells is phosphorylated and activated in response to insulin via the PI3K/Akt/mTOR signal transduction pathway, and that p70S6K mediates nearly all of the insulin-mediated survival effects, similar to reports on skeletal muscle (64), liver (65), kidney (66), and myocardium (67). These studies are important because they provide further evidence that the retina is sensitive to the anabolic effects of insulin (8).
P70S6K is a protein synthesis regulator that mediates the effects of hormones on global and selective patterns of mRNA translation (37). The results of the present study demonstrate that p70S6K is also involved in rat retinal neuronal cell survival. When R28 cells were pretreated with rapamycin, insulin did not rescue serum-starved R28 cells from apoptosis. In contrast, rapamycin had no effect on serum-mediated cell survival, suggesting a specific role of p70S6K in the insulin-mediated cell survival effect. p70S6K also appears to exert a cell-specific survival effect in response to insulin because rapamycin did not block the survival effect of insulin in L6 myocytes. The cause for this difference is not known currently. Wild type S6K1 transfection resulted in a high basal level of phosphorylation and activity in the absence of insulin. With insulin stimulation, wild type p70S6K phosphorylation and activity were further increased. These results are consistent with those of other studies in 293 cells (68). Counting of pyknotic cells in HA-positive cells showed a reduced number of pyknotic cells in both control and insulin stimulated conditions, but rapamycin blocked the anti-apoptotic effect of insulin. The other evidence that also strongly supports the hypothesis comes from cell transfected with a rapamycin-resistant construct, HA-F5A-E389-
Phosphorylation of the C-terminal autoinhibitory domain that is lacking the HA-F5A-E389- The specific mechanism by which p70S6K supports insulin-stimulated retinal cell survival remains uncertain. Recent work by Holcik et al. (69) suggests that internal ribosomal initiation of mRNA translation, a step that is affect by rapamycin and p70S6K, is critical for survival of cells under transient apoptotic stress. Postmitotic retinal neurons require protein synthesis for survival (70), so it is possible that insulin stimulates protein synthesis via a rapamycin-dependent mechanism. Indeed, the same concentration of rapamycin used in this study completely blocks insulin stimulated protein synthesis in R28 cells.2 In addition, rapamycin partially blocks insulin-induced Foxo1 phosphorylation and translocation from nucleus to the cytosol, and co-transfection of wild type S6K1 and Foxo1 causes Foxo1 translocation to the cytosol3 suggesting a potential mechanism for S6K1-mediated cell survival in retinal neuronal cells.
The physiologic significance of these in vitro studies remains to be determined but p70S6K in whole retinas also responds to insulin stimulation,2 so it is likely that p70S6K activity also supports retina cell survival and protein synthesis in vivo. Studies are in progress to examine this question and the role of p70S6K activity in diabetic retinopathy.
* This study was supported by grants from the American Diabetes Association, the JDRF Diabetic Retinopathy Center at Penn State University, Fight for Sight (to X. W.), Pennsylvania Sight Conservation and Eye Research Foundation (to X. W.), and National Institutes of Health Grants EY12021, DK13499, and DK15658. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ¶ To whom correspondence should be addressed: Dept. of Ophthalmology, The Pennsylvania State University College of Medicine, 500 University Dr., Hershey, PA 17033. Tel.: 717-531-6711; Fax: 717-531-0631; E-mail: tgardner{at}psu.edu.
1 The abbreviations used are: PI3K, phosphoinositide 3-kinase; DMEM, Dulbecco's modified Eagle's medium; HA, hemagglutinin.
2 X. Wu, T. W. Gardner, manuscript submitted.
3 X. Wu, P. Quinn, and T. W. Gardner, unpublished data.
We thank Dr. John Blenis (Department of Cell Biology, Harvard Medical School, Boston) for providing S6K1 constructs.
This article has been cited by other articles:
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Advertisement | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||