Inorganic polyphosphate elicits proinflammatory responses through activation of mTOR complexes 1 and 2 in vascular endothelial cells

Summary

Background

Inorganic polyphosphate (polyP) elicits proinflammatory signaling responses in endothelial cells through interaction with two receptors, RAGE and P2Y₁. It is known that polyP activates mTOR signaling in breast cancer cells.

Objectives

The objective of this study is to understand the mechanism of polyP-mediated signaling pathway in endothelial cells and to determine whether polyP exerts its proinflammatory effect through activation of mTOR.

Methods

mTOR activation by polyP or platelet releasates in cellular and animal models were monitored in the absence and presence of pharmacological inhibitors and/or siRNA knockdown of specific signaling molecules.

Results

PolyP effectively induced phosphorylation of mTOR complex 1 (mTORC1) substrate, p70S6K, in endothelial cells by an AKT-dependent but ERK-independent mechanism. The siRNA knockdown of both RAGE and P2Y₁ or specific inhibitors of PI3K/PLC/PKC/Ca²⁺ signaling axis inhibited polyP-mediated p70S6K phosphorylation. Moreover, either rapamycin or siRNA knockdown of raptor (mTORC1-specific component) abrogated polyP-mediated phosphorylation of p70S6K. By contrast, the siRNA knockdown of rictor (mTOR complex 2-specific component) but not raptor eliminated barrier-disruptive effect of polyP. Specific NF-κB inhibitors abrogated polyP-mediated phosphorylation of p70S6K and rapamycin suppressed polyP-induced activation of NF-κB. Finally, specific inhibitors of mTOR signaling network eliminated polyP-mediated vascular leakage and leukocyte recruitment in animal models.

Conclusions

PolyP, through interaction with RAGE and P2Y₁, activates both mTORC1 and mTORC2 signaling network. Both proinflammatory and mTOR signaling functions of polyP are linked.

Introduction

Inorganic polyphosphates (polyP) are linear polymers of 3 to over 1,000 inorganic phosphate residues, which are linked together by ATP-like phosphoanhydride bonds [1]. PolyP containing ~60–100 phosphate units at high concentrations is stored in dense granules of human platelets and can be released to circulation upon activation [2,3]. Longer chain polyP polymers containing over 1000 phosphate units may be synthesized by microorganisms under different environmental conditions [1,3]. The role of polyP in regulating coagulation, inflammation and metabolic pathways is under investigation by several laboratories [3–7].

It is known that polyP can activate mammalian target of rapamycin (mTOR) in breast cancer cells, thereby stimulating proliferative signaling pathways in these cells [8]. mTOR is a serine-threonine kinase that is composed of at least two distinct multi-protein complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2) [9–11]. Although both mTOR complexes have several common protein components, the regulatory-associated protein of mTOR (raptor) and rapamycin-insensitive companion of mTOR (rictor) are specific subunits of mTORC1 and mTORC2 respectively [9,10]. It is established that mTORC1, through phosphorylation of its downstream effectors, regulates cell growth, proliferation and metabolism in mammalian cells [9–11]. Thus, phosphorylation and stimulation of p70S6 kinase (p70S6K) activity by mTORC1 leads to an increase in mRNA synthesis and translation of ribosomal proteins which promote protein synthesis during cell growth [9–11]. In contrast to mTORC1, relatively little is known about mTORC2 biology, though it has been demonstrated that mTORC2 can play a key role in cytoskeleton organization and cell morphology through regulation of Rho family of small GTPases [9,11]. The mechanism through which polyP activates mTOR signaling has not been investigated and neither has the type of mTOR complex that is involved in polyP-mediated signaling.

We recently demonstrated that polyP containing 70 phosphate units (polyP-70), similar to the size released by activated platelets, and polyP-700 (similar to the size synthesized by bacteria) can elicit potent proinflammatory responses in vascular endothelial cells [12]. Thus, we discovered that polyP interacts with two receptors, receptor for advanced glycation end products (RAGE) and the purinergic receptor (P2Y₁) [13], thereby activating NF-κB, promoting expression of cell adhesion molecules and inducing barrier-disruptive effects in endothelial cells [12]. In this study, we investigated the role of polyP in modulating mTOR signaling in both cellular and animal models. Results suggest that polyP, through interaction with the same two receptors, activates both mTOR complexes 1 and 2 through AKT- and NF-κB-dependent phosphorylation of the upstream regulatory tuberous sclerosis complex 1/2 (TSC1/2), thereby inducing phosphorylation of the downstream mTORC1 substrate, p70S6K, in endothelial cells. Results further suggest that barrier-disruptive effect of polyP in endothelial cells is mediated through activation of mTORC2 signaling pathway. Pharmacological inhibitors of PI3K, AKT, PLC and rapamycin all abrogate polyP-induced mTOR signaling in both in vitro and in vivo models. Moreover, boiled platelet releasates exerted a similar signaling effect which was abrogated if releasates were pretreated with the specific polyP inhibitor, EcPPXc, or alkaline phosphatase.

Materials and Methods

Preparation of human platelet releasates

Platelet units were centrifuged (1000g) at 25°C for 10min to pellet. Supernatants were removed, and platelets were then washed twice with calcium-free Tyrode’s buffer (137mM NaCl, 11.9mM NaHCO₃, 0.36mM NaH₂PO₄, 5.55mM D-glucose, 1.05mM MgCl₂, 2.7mM KCl) with inhibitors (2μM PGE₁, and 1mM theophylline) at 25°C followed by centrifugation and re-washing with buffer lacking inhibitors. Resulting platelet pellets were resuspended at 10¹¹/mL in Tyrode’s buffer containing 1.8 mM CaCl₂. Releasates were stimulated with either 5 μM TRAP or 10 μM A23187 at 37°C for 20min as described [14]. The sources of all reagents are presented in Supplementary Materials. The platelet suspension was centrifuged at 2500g, the supernatant was removed, and centrifuged again at 20,000g for 30min to deplete microparticles. Platelet releasates (protein contents 300μg/mL) were aliquotted and stored at −80°C.

Permeability assays and transfection with siRNA

Endothelial cell permeability in response to polyP and boiled (30min) platelet releasates with or without pretreatment with either recombinant polyP-binding domain of Escherichia coli exopolyphosphatase (EcPPXc) [15] or alkaline phosphatase (ALP) was monitored in a dual chamber model by measuring flux of Evans blue-labeled BSA across cell monolayer as described [6,12]. In the presence of siRNA, 3.5×10⁵ cells were seeded in 6-well plates and transfected with 30nM siRNA as described [16].

In vivo permeability and leukocyte migration assays

Six weeks old male C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME) were used for in vivo studies after a 5-day acclimatization period on a 12h light/dark cycle in a controlled environment. All animals were treated in accordance with Guidelines for Care and Use of Animals at Saint Louis University. Vascular permeability was evaluated according to described methods [12]. In the presence of inhibitors, mice (six/group) were intraperitoneally injected with following inhibitors: rapamycin (mTOR inhibitor, 5mg/kg), API-2 (AKT inhibitor, 2mg/kg), U-73122 (PLC inhibitor, 20mg/kg), wortmannin (PI3K inhibitor, 1mg/kg). Following 1h of inhibitor administration, 1% Evans blue dye solution in normal saline was injected intravenously to each mouse, immediately followed by intraperitoneal injection (μg/g body weight) of either polyP-70 (200, 400), polyP-700 (50, 100) or 0.7% acetic acid as a positive control as described [6,12]. For assessing leukocyte migration, animals (six/group) were intraperitoneally injected with above concentrations of inhibitors and stimuli for 4h before sacrificing and counting the number of leukocytes as described [6,12]. The details of data analysis and experimental approaches have been described previously [6,12].

In vivo permeability in response to boiled platelet releasates was monitored by same procedures described above. In this case, mice were injected with 300μL boiled platelet releasates (estimated to contain ~7.8μg polyP) plus either HMGB1 (1μg/g body weight) or histone H4 (2.5μg/g body weight). After 30min, mice were sacrificed, peritoneal exudates were collected and vascular permeability was monitored as described above.

Supplementary Material

For a detailed description of methods relative to cell culture, Western blot analysis, NF-κB activation, and cell viability (MTT) assay (17) and statistical analysis see Methods’ section of the online-only Supplementary Material.

Results

PolyP activates mTORC1

PolyP increased p70S6K phosphorylation in a concentration- and time-dependent manner in endothelial cells (Fig. 1A–F). Both primary and transformed HUVECs (EA.hy926) yielded identical results (only data for EA.hy926 is presented). A significant increase in p70S6K phosphorylation by polyP-70 and polyP-700 (polyP concentration is expressed in terms of phosphate monomer) was observed at concentrations of 5μM and 1μM, respectively (Fig. 1A–C). Maximal mTORC1 activation occurred after 15min of polyP treatment (Fig. 1D–F). PolyP-700 activated mTORC1 much more effectively than polyP-70 in all assays similar to those presented in Figs. 1C and F, however, only data for polyP-70 is presented in the text described below. Data with polyP-700 is presented as Supplementary Figs. S1,S2). In agreement with a role for polyP in mTORC1 activation, rapamycin inhibited polyP-mediated phosphorylation of p70S6K (Fig. 1G). In light of findings that the mTORC1 function, under certain conditions, may be resistant to rapamycin [18], and that prolonged rapamycin treatment, can inhibit mTORC2 assembly [19], the role of polyP in mTORC1 activation was also analyzed by siRNA knockdown of raptor and rictor, which are specific components of mTORC1 and mTORC2, respectively. The siRNA knockdown of both raptor and rictor effectively inhibited their expression (Fig. 1H), however, only siRNA for raptor but not for rictor inhibited polyP-mediated phosphorylation of p70S6K (Fig. 1I), suggesting that polyP elicits intracellular signaling responses through activation of mTORC1. Our recent results have indicated that polyP exerts its cellular effect through interaction with two receptors; RAGE and P2Y₁ [12]. Thus, we first studied the efficacy of siRNA knockdown of each receptor (Fig. 1J,K). The siRNA knockdown of each receptor individually had a partial inhibitory effect, but knockdown of both receptors effectively inhibited polyP-mediated phosphorylation of p70S6K (Fig. 1L), suggesting polyP activates mTORC1 through interaction with the same two receptors. To rule out off-target effects of siRNA, we used two different siRNAs for each individual gene, however, since the results were identical we only presented one data set in the manuscript (Fig. 1L).

Figure 1.

PolyP-70 and polyP-700 activate mTOR in a concentration- and time-dependent manner. (A–C) Endothelial cells were incubated with increasing concentrations of polyP-70 (A) or polyP-700 (B) for 15 min followed by measuring and quantitation of phosphorylation of p70S6K (70, 85 kDa) for each polyP derivative (C). (D–F) The same as A–C except that the time-dependent effect of polyP on p70S6K activation was measured and quantitated. (G) The same as A except that polyP-mediated p70S6K activation was monitored in the presence of different concentrations of rapamycin. (H) Endothelial cells were transiently transfected with control siRNA or siRNA specific for either raptor (150 kDa) or rictor (200 kDa) and the efficiency of gene knockdown was determined 48h post transfection by western blotting using specific antibodies. (I) The same as H except that polyP-mediated p70S6K activation was monitored after siRNA knockdown of either raptor or rictor. (J–K) The same as panel H except that the efficiency of gene knockdown of RAGE (46 kDa) and P2Y₁ (45 kDa) was determined 48h post transfection by western blotting using specific antibodies. (L) The same as panel I except that polyP-mediated p70S6K activation was monitored after siRNA knockdown of RAGE and P2Y₁ individually or in combinations. All results are means ± s.e.m. of three different experiments. *p<0.05; **p<0.01; ***p<0.001.

PolyP activates mTORC1 by AKT-dependent but ERK1/2-independent mechanisms

Next, we investigated the role of PI3K/AKT and ERK signaling in polyP-mediated mTORC1 activation by monitoring phosphorylation of p70S6K. PolyP enhanced phosphorylation of AKT at both Thr-308 and Ser-473 sites (Fig. 2A) which was abrogated by co-incubation of polyP with ALP (Supplementary Fig. S3A). AKT phosphorylation at Ser-473 is required for its maximal activity [20]. To further investigate the role of AKT in polyP-induced mTOR activation, cells were pretreated with either AKT inhibitor VIII or 3-phosphoinositide-dependent protein kinase-1 (PDK-1) inhibitor (OSU-03012). PDK1 can activate AKT through phosphorylation of AKT at Thr308 [21]. PDK-1 inhibitor (Fig. 2B) and AKT inhibitor VIII (Fig. 2C) both effectively suppressed phosphorylation of p70S6K in polyP-stimulated cells. AKT can activate mTORC1 through phosphorylation of both PRAS-40 (a component of mTORC1) and tuberous sclerosis complex 1/2 (TSC1/2) [9,10]. PRAS-40 and TSC1/2 are both negative regulators of mTOR and their phosphorylating inactivation by AKT activates mTORC1 [9,22,23]. PolyP induced phosphorylation of both PRAS-40 and TSC-2 in endothelial cells by a time-dependent manner (Fig. 2D). In agreement with these results, wortmannin (PI3K inhibitor) inhibited polyP-induced PRAS-40 and TSC-2 phosphorylation (data not shown). Wortmannin also inhibited polyP-mediated phosphorylation of p70S6K (Fig. 2E). Similar to activation of AKT, polyP also phosphorylated ERK1/2 in a time-dependent manner (Fig. 2F) which was inhibited by co-incubation of polyP with ALP (Supplementary Fig. S3B). However, ERK1/2 inhibitor, PD-98059, had no effect on polyP-mediated phosphorylation of either p70S6K (Fig. 2G) or TSC-2 (Fig. 2H). To further support the hypothesis that effect of polyP on mTORC1 is independent of ERK, we investigated phosphorylation of p70S6K in polyP-stimulated cells transfected with ERK1/2 siRNA. Two different siRNAs were used for ERK1/2 to rule out off-target effects of siRNA. The efficiency of gene knockdown was determined 48h post transfection by Western-blotting using specific antibody against ERK1/2 (Fig. 2I). Consistent with results obtained with specific PD-98059 ERK1/2 inhibitor, neither one of ERK1/2 siRNAs abrogated polyP-induced mTOR activation (Fig. 2J).

Figure 2.

PolyP-mediated activation of mTOR upstream signaling pathways in the absence and presence of specific inhibitors. (A) Time course of polyP-mediated phosphorylation of AKT (60 kDa) at Thr-308 and Ser-473 in endothelial cells. (B) PolyP-mediated phosphorylation of p70S6K (70, 85 kDa) in the absence and presence of increasing concentrations of PDK-1 inhibitor (OSU-03012). (C) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of AKT inhibitor VIII. (D) Time course of polyP-mediated phosphorylation of PRAS40 (40 kDa) and TSC2 (200 kDa). (E) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of PI3K inhibitor (wortmannin). (F) Time course of polyP-mediated phosphorylation of ERK1/2 (42, 44 kDa). (G) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of ERK inhibitor (PD98059). (H) PolyP-mediated phosphorylation of TSC2 at Thr-1462 in the presence of ERK inhibitor (PD98059). (I) Endothelial cells were transiently transfected with control siRNA or two different siRNA specific for ERK1/2 and the efficiency of gene knockdown was determined 48h post transfection by western blotting using specific antibody. (J) The same as I except that polyP-mediated p70S6K activation was monitored after siRNA knockdown of ERK1/2. (K) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of intracellular calcium chelator (BAPTA-AM). (L) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of intracellular calcium releaser (Thapsigargin). (M,N) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of PKC inhibitors (BIS and Gö 6983). (O) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of PLC inhibitor (U-73122). (P) PolyP-mediated phosphorylation of p70S6K in the absence and presence of increasing concentrations of inactive analog of PLC inhibitor (U-73343). (Q) The viability of HUVECs following incubation with different inhibitors. With all inhibitors, cells were treated with each inhibitor for 30 min prior to incubation with or without polyP-70 (25 μM) for 15 min. All results are means ± s.e.m. of three different experiments. *p<0.05; **p<0.01; ***p<0.001. NS, not significant.

It was recently demonstrated that polyP mediates calcium release from intracellular stores in astrocytes through interaction with purinergic receptors [13]. To determine whether calcium signaling is required for polyP-mediated mTORC1 activation, cells were pretreated with increasing concentrations of the Ca²⁺ chelator, BAPTA-AM, or the intracellular calcium releaser, Thapsigargin, before treating cells with polyP. Both BAPTA-AM and Thapsigargin inhibited effect of polyP on p70S6K activation by a concentration-dependent manner (Fig. 2K,L). Inhibitors of PK, either bisindolylmaleimide I hydrochloride (BIS) or Gö 6983 and PLC (U-73122) also inhibited polyP-induced p70S6K phosphorylation (Fig. 2M–O) whereas the PLC-inactive analog, U-73343, had no effect on polyP signaling (Fig. 2P). These results clearly demonstrate that polyP exerts its modulatory effect through PI3K/AKT- and PLC/PKC-dependent activation of mTOR signaling network. Data with polyP-700 for this section is presented in Supplementary Fig. S2A–C. To investigate the toxicity of inhibitors, the viability of endothelial cells following incubation with the highest concentration of inhibitors used in the experiments, was measured by the MTT assay as described [17]. The results demonstrate that none of the inhibitors used in this study was toxic to cells (Fig. 2Q).

Interplay between mTORC1 and NF-κB pathways

We previously demonstrated that polyP activates NF-κB in endothelial cells [6,12]. Noting that a role for NF-κB in regulation of mTOR has been reported [24,25], we investigated polyP-mediated interplay between the two pathways by determining if inhibition of one pathway by specific inhibitors is mirrored in the other. Inhibition of NF-κB signaling pathway by pharmacological inhibitors, IKKα (BAY11-7082), IKKβ (BMS-345541) and IKKα/β (Wedelolactone), dramatically decreased polyP-mediated activation of p70S6K (Fig. 3A–C). Similarly, pretreatment with rapamycin suppressed stimulatory effect of polyP on NF-κB activation (Fig. 3D). siRNA knockdown of raptor but not rictor also effectively attenuated NF-κB activation by polyP (Fig. 3E). PolyP up-regulated expression of NF-κB-regulated adhesion molecules, VCAM-1 and E-selectin, and rapamycin effectively inhibited this function (Fig. 3F). These results suggest that polyP regulates both NF-κB and mTORC1 signaling, thereby linking inflammation to metabolic regulatory pathways in endothelial cells. The common regulatory step in the pathway appears to be TSC1/2 since inhibitors of AKT and NF-κB both inhibited polyP-mediated phosphorylation of TSC1/2 (Supplementary Fig. S1A–C).

Figure 3.

PolyP-70-mediated regulation of mTOR and NF-κB pathways in endothelial cells. (A) PolyP-mediated phosphorylation of p70S6K (70, 85 kDa) was monitored in the absence and presence of increasing concentrations of IKK-α inhibitor. Cells were treated with IKK-α inhibitor for 30 min prior to incubation with polyP-70 (25 μM) for 15 min. (B) PolyP-mediated phosphorylation of p70S6K was monitored in the absence and presence of increasing concentrations of IKK-β inhibitor. Cells were treated with IKK-β inhibitor for 30 min prior to incubation with polyP-70 (25 μM) for 15 min. (C) PolyP-mediated phosphorylation of p70S6K was monitored in the absence and presence of increasing concentrations of IKKα/β inhibitor. (D) Endothelial cells were pretreated with rapamycin (50 nM) for 30 min and then stimulated with two different concentrations of polyP. (E) PolyP-mediated activation of NF-κB with or without transient transfection of endothelial cells with control siRNA or siRNA specific for either raptor or rictor. (F) PolyP-mediated cell surface expression of VCAM-1 and E-selectin was measured by a cell-based ELISA in the absence and presence of rapamycin (50 nM). Cells were treated with rapamycin (50 nM) for 30 min prior to incubation with polyP-70 for 4h. All results are means ± s.e.m. of three different experiments. *p<0.05; **p<0.01.

PolyP disrupts barrier-permeability through mTORC2

PolyP has been shown to disrupt barrier-permeability of endothelial cells in both cellular and in vivo systems [4,6,12]. Interestingly, the siRNA knockdown of rictor, but not raptor, effectively attenuated the barrier-disruptive function of polyP in endothelial cells (Fig. 4A), suggesting that polyP exerts its effect on cell permeability through activation of mTORC2. Specific inhibitors of PI3K (wortmannin), AKT (OSU-03012), PLC (U-73122), PKC (BIS), Ca²⁺ (BAPTA-AM) and rapamycin all inhibited the barrier-disruptive effect of polyP (Fig. 4B). These results suggest that upstream signaling molecules participating in mTOR activation by polyP are shared by both mTORC1 and mTORC2 pathways. Consistent with results obtained in cellular model, polyP effectively induced vascular leakage in vivo as evidenced by mediating passage of BSA-bound Evans blue dye from plasma into peritoneal cavity of mice treated with polyP (Fig. 4C). Moreover, polyP markedly enhanced binding of leukocytes to peritoneal tissues and their subsequent extravasation to peritoneal cavity (Fig. 4D). The intravenous pre-administration of rapamycin and specific inhibitors of AKT (API-2), PLC (U-73122) and PI3K (wortmannin) all abrogated proinflammatory effects of polyP in these in vivo assay systems (Fig. 4C,D). Histology of peritoneal tissues for untreated negative control (A), CMC-Na-treated positive control (B), polyP-treated (C) and polyP+rapamycin-treated (D) groups are presented as Supplementary Fig. S4.

Figure 4.

Analysis of proinflammatory functions of polyP in in vitro and in vivo models. (A) PolyP-mediated endothelial cell permeability was measured following transiently transfecting cells with control siRNA or siRNA specific for either raptor or rictor as described under Methods. (B) The same as A except that PolyP-mediated cell permeability was measured in the absence and presence pharmacological inhibitors of different signaling molecules. (C) Analysis of polyP-mediated vascular leakage in mice with or without pretreatment with pharmacological inhibitors. Mice (n=6 for each group) were intravenously injected with 1% BSA-bound Evans blue dye followed by an immediate intraperitoneal injection of two concentrations of polyP-70 (200 and 400 μg/g body weight) with 0.7% acetic acid as a positive control. Vascular permeability was determined from the extent of extravasation of Evans blue to the peritoneal cavity as described in Methods. In the presence of inhibitors, mice were first injected (i.p.) with the following inhibitors: rapamycin (mTOR inhibitor, 5 mg/kg), API-2 (AKT inhibitor, 2 mg/kg), U-73122 (PLC inhibitor, 20 mg/kg) and wortmannin (PI3K inhibitor, 1 mg/kg) for 1h before administration of BSA-bound Evans blue dye. (D) The same as C except that the effect of polyP was monitored on the migration of leukocyte to peritoneal cavity in the absence and presence of inhibitors. All results are presented as means ± s.e.m. *p<0.05 and **p<0.01.

Platelet releasates activate mTOR signaling pathways

It has been demonstrated that activated platelets secrete polyP with polymer lengths of ~60–100 phosphate units [3–5,14], the same polymer size (polyP-70) that we used in this study. To determine whether polyP secreted by activated platelets have the same signaling function in endothelial cells, phosphorylation of p70S6K by activated platelet releasates was monitored employing the same assays described above. It was found that platelet releasates phosphorylate the mTORC1 substrate by a dose-dependent manner (Fig. 5A). Thus, incubation of endothelial cells with a platelet releasates ratio of 0.04 (platelet releasates/total volume, estimated to be ~10μM if its concentration was expressed in terms of phosphate monomer) resulted in a significant activation of p70S6K (Fig. 5A). Since platelets were activated with TRAP, the possible stimulatory effect of a small amount of TRAP present in platelet releasates was analyzed. A TRAP concentration of 0.5μM, which exceeds the concentration of TRAP present in the highest platelet releasates ratio (0.1), did not have any mTOR activation property in this assay (Fig. 5A), excluding any contribution from trace amounts of TRAP present in the platelet releasates (Fig. 5A). To further confirm the hypothesis that polyP present in platelets is responsible for mTOR activation, incubation of platelet releasates with specific polyP inhibitor EcPPXc or ALP abrogated the signaling activity (Fig. 5B). Furthermore, boiling platelet releasates for 30 min did not impact phosphorylation of p70S6K, however, boiling releasates followed by treatment with either recombinant EcPPXc (polyP-binding domain of Escherichia coli exopolyphosphatase) [15] or ALP abrogated the signaling effect (Fig. 5B). It has been previously demonstrated that boiling platelet releasates for 30 min denatures all proteins without negatively affecting the cofactor function of purified polyP [14]. These results were reproduced with three different batches of releasates derived from platelets activated by TRAP +Ca²⁺ (Fig. 5B, lanes 5 and 6) calcium ionophore A23187 (lane 9), and TRAP+EDTA (lane 11). These results suggest that platelet polyP is responsible for activating mTORC1 signaling pathway in endothelial cells.

Figure 5.

Platelet releasates activate mTOR and elicits proinflammatory responses in in vitro and in vivo models. (A) Dose response for platelet releasates (30 min) in inducing the phosphorylation of p70S6K (70, 85 kDa) in endothelial cells. PolyP-70 is shown as a control in the last lane. (B) Phosphorylation of p70S6K by boiled (30 min) platelet releasates (plt-rel, 0.05 ratio) from three different batches is inhibited by coincubation of platelet releasates with EcPPXc (250 μg/ml) and alkaline phosphatase (ALP 2 units/ml). (C) Analysis of phosphorylation of p70S6K by boiled (30 min) platelet releasates by the inhibitors of mTOR upstream signaling pathway. With all inhibitors, cells were treated with each inhibitor for 30 min prior to incubation with boiled platelet releasates for 30 min. (D) Dose response for boiled platelet releasates (4h) in inducing barrier-permeability in endothelial cells. (E) Platelet releasates (0.2 ratio)-mediated cell permeability (4h) before and after treating platelet releasates with two different concentrations of EcPPXc (125 and 250 μg/ml) and ALP (2 and 4 units/ml) was measured as described in Methods. (F) Platelet releasates (0.05 ratio)-mediated cell permeability before and after treating platelet releasates with EcPPXc (250 μg/ml) and ALP (4 units/ml) was measured in the presence of either HMGB1 (10 nM) or histone 4 (0.9 μM) (4h) as described in Methods. (G) Analysis of platelet releasates (300 μl)-mediated vascular leakage in mice with or without pretreatment with ALP (4 units/ml) and addition of either HMGB1 (1 μg/g body weight) or histone 4 (2.5 μg/g body weight). Mice (n=6 for each group) were intravenously injected with 1% BSA-bound Evans blue dye followed by an immediate intraperitoneal injection of 300 μl boiled platelet releasates with 0.7% acetic acid as a positive control. Vascular permeability was determined from the extent of extravasation of Evans blue to the peritoneal cavity as described in Methods. All results are presented as means ± s.e.m. *p<0.05.

In agreement with results presented above, wortmannin, BIS, U-73122, BAPTA-AM, AKT VIII, rapamycin, BAY11-7082, and BMS-345541 all inhibited platelet releasates-mediated phosphorylation of p70S6K (Fig. 5C). However, ERK1/2 inhibitor, PD-98059, had no effect (Fig. 5C), suggesting that polyP-70 and platelet releasates activate mTORC1 through same mechanisms.

Similar to polyP, platelet releasates (0.1 ratio, estimated to be ~25μM) also enhanced barrier-permeability of peritoneal tissues and that the effect was not inhibited by boiling, but eliminated by either EcPPXc or ALP (Fig. 5D and E). We recently demonstrated that polyP dramatically amplifies barrier-disruptive effects of nuclear proteins H4 and HMGB1 by binding and presenting nuclear proteins to specific proinflammatory receptors in endothelial cells [12]. To establish the hypothesis that polyP derived from platelet releasates is responsible for the observed barrier-disruptive effect, the assay was conducted with a mixture of platelet releasates and nuclear proteins which were individually inactive in this assay. Thus, neither HMGB1 (10nM) and H4 (0.9μM) nor boiled platelet releasates (0.05 ratio) individually influenced cellular barrier-permeability function, however, combinations of platelet releasates with either HMGB1 or H4 disrupted barrier-permeability function of endothelial cells (Fig. 5F), suggesting that platelet releasates-derived polyP is responsible for mediating this effect. Consistent with the effect being due to polyP, pretreatment of platelet releasates with either EcPPXc or ALP abrogated proinflammatory signaling effects of nuclear proteins in endothelial cells (Fig. 5F). In agreement with results in the cellular model, platelet releasates amplified nuclear protein-mediated vascular leakage in vivo as measured by the passage of BSA-bound Evans blue dye from plasma into the peritoneal cavity of mice (Fig. 5G). This effect was mediated through platelet releasates’ polyP since both EcPPXc and ALP inhibited proinflammatory effects of nuclear proteins in this in vivo model (Fig. 5G). Taken together, these results strongly suggest that polyP in platelet releasates is responsible for eliciting these proinflammatory signaling responses in both cellular and animal models.

Discussion

We have demonstrated in this study that polyP, through interaction with RAGE and P2Y₁ receptors activates mTORC1 and mTORC2, thereby eliciting diverse intracellular signaling responses in endothelial cells. PolyP induced phosphorylation and activation of the mTORC1-specific substrate, p70S6K, in endothelial cells. When phosphorylated by polyP-activated mTORC1, p70S6K can bind to different mRNA translation initiation machinery proteins to increase the rate of protein synthesis. In support of polyP-mediated activation of mTORC1, rapamycin effectively inhibited polyP-mediated phosphorylation of p70S6K. Since rapamycin may also inhibit mTORC2 under certain conditions [19], this question was further investigated by the siRNA approach. The siRNA knockdown of raptor (a specific component of mTORC1), but not rictor (a specific component of mTORC2), inhibited phosphorylation/activation of p70S6K in endothelial cells, supporting the hypothesis that polyP activates this pathway through activation of mTORC1 in endothelial cells. PolyP activates mTORC1 by an AKT-dependent but ERK-independent mechanism. This was evidenced by findings that two specific inhibitors of AKT abrogated polyP-mediated phosphorylation of p70S6K. However, a specific inhibitor of ERK1/2 had no effect on this signaling function of polyP. Interestingly, the barrier-permeability regulating function of polyP was found to be mediated through mTORC2 as evidenced by the siRNA knockdown of rictor, but not raptor, inhibiting the barrier-disruptive effect of polyP. These results suggest that polyP can activate both mTORC1 and mTORC2 pathways in endothelial cells. The mTORC2-depedent barrier-disruptive effect of polyP is consistent with the proposed role of mTORC2 in cytoskeleton organization and actin polymerization through regulation of Rho family of GTPases [9,10]. It was interesting to note that platelet releasates recapitulated all proinflammatory signaling functions of purified polyP-70 in both cellular and in vivo models. The observations that boiling did not inhibit signaling effect of platelet releasates, but incubation with specific polyP inhibitor, EcPPXc and/or digestion by alkaline phosphatase effectively inhibited signaling effects of platelet releasates strongly suggest that platelet polyP is a physiological activator of mTOR pathways in vascular endothelial cells.

We demonstrated that the siRNA knockdown of raptor inhibited both mTORC1 and NF-κB signaling functions of polyP, suggesting that the two pathways may be linked and that polyP modulates both signaling pathways through a common mechanism in endothelial cells. In further support of a polyP-mediated linkage between the two pathways, rapamycin not only inhibited mTORC1 signaling, but also inhibited polyP-mediated NF-κB activation. Toward understanding the mechanism of polyP-mediated activation of NF-κB and mTOR signaling pathways, we discovered that polyP activates mTORC1 by AKT-dependent inactivation of the tumor suppressor complex, tuberous sclerosis 1/2 (TSC1/2), which is known to function as a key negative regulator of mTOR signaling [9,10,23–25]. It is known that IKKβ, a kinase responsible for IκB-dependent activation of NF-κB, can activate mTORC1 through phosphorylation and suppression of the TSC1/2 activity [25]. Thus, polyP-mediated phosphorylation of TSC1/2 appears to be the common denominator in linking inflammation to activation of mTOR signaling in endothelial cells. Cytokine-mediated integration of NF-κB and mTOR signaling pathways has also been reported by other groups [24–26]. Thus, polyP, through phosphorylation-dependent inactivation of the TSC1/2 complex is able to link the inflammatory NF-κB pathway to metabolic regulatory mTOR signaling pathway in endothelial cells. It is worth noting that TSC1/2 activity is also known to be inhibited by ERK1/2 signaling which is also activated by polyP. However, a specific inhibitor of ERK1/2 signaling (PD-98059) did not have any effect on polyP-mediated phosphorylation of either TSC1/2 or p70S6K, suggesting that polyP activates mTORC1 by an ERK-independent mechanism. In addition to phosphorylation of TSC2, polyP also induced phosphorylation of the mTORC1 subunit, PRAS-40, which can also function as a negative regulator of mTORC1 signaling [9]. It is known that PRAS-40 can be directly phosphorylated by AKT, thereby leading to its dissociation from the signaling complex and activation of the mTORC1 pathway [9]. Thus, in addition to TSC1/2-dependent activation of mTORC1, polyP may also activate mTORC1 by AKT-dependent phosphorylation of PRAS-40 by a TSC1/2-independent mechanism.

The observations that polyP mediated its barrier-disruptive effect through activation of mTORC2 and that specific inhibitors of PI3K/AKT and PLC/PKC pathways all abrogated the barrier-disruptive effect of polyP suggest that polyP can activate the mTORC1/2 signaling network through activation of same upstream signaling molecules in endothelial cells as presented in Fig. 6. The pathophysiological relevance of this signaling mechanism of polyP was validated by our findings that specific inhibitors of mTORC1, PI3K, PLC and AKT, all attenuated polyP-mediated vascular leakage (edema) and recruitment of leukocytes to inflammatory sites in an in vivo model. The hypothesis, based on the raptor and rictor siRNA knockdown data, that polyP-mediated activation of mTORC2 is responsible for the barrier-disruptive effect of polyP may appear to be at odds with the finding that rapamycin inhibits vascular permeability effect of polyP in both in vivo and cellular systems. However, as indicated above, rapamycin is not specific to mTORC1 and can also inhibit mTORC2 by interfering with assembly of the signaling complex in endothelial and other cell types [19], explaining why the barrier-permeability regulating function of polyP is susceptible to inhibition by rapamycin.

Figure 6.

Schematic representation of the mechanism of polyP-mediated mTOR activation in vascular endothelial cells. PolyP, through interaction with RAGE and P2Y₁ receptors, triggers PI3K/AKT and PLC/PKC/Ca⁺² signaling pathways. Both pathways activate mTOR complexes 1 and 2 through the inhibition of the mTOR suppressor TSC1/2 complex. Although ERK is also activated by polyP but the effect of polyP on mTOR activation is independent of ERK signaling in endothelial cells (as evidenced by the ERK inhibitor not inhibiting polyP-mediated phosphorylation of either TSC1/2 or p70S6K). The TSC1/2 complex is a negative regulator of mTOR, thus phosphorylation-dependent inactivation of the TSC1/2 complex by AKT and/or PKC activates both mTORC1 and mTORC2. Phosphorylation of p70S6K by polyP is mediated through activation of mTORC1 whereas the effect of polyP on cytoskeleton reorganization and endothelial barrier-permeability is specifically mediated through activation of mTORC2.

We recently demonstrated that both polyP-70 and polyP-700 can form high affinity complexes with nuclear proteins, high mobility group box 1 (HMGB1) and histone H4, to dramatically amplify proinflammatory signaling responses in endothelial cells through interaction and activation of two cell surface receptors, RAGE and P2Y₁ [12]. In this study, we further demonstrated that polyP in platelet releasates can similarly amplify proinflammatory effects of nuclear proteins in both cellular and animal models. Noting the importance of nuclear proteins in inflammatory disorders and the abundance of polyP in biology (polyP is synthesized by all eukaryotic cells), we hypothesize that polyP through activation of RAGE and mTORC1/2 signaling pathways can make a critical contribution to pathogenesis and complications of metabolic- and inflammation-based disorders including severe sepsis, diabetes, cardiovascular disease, and cancer. A fine-tuned mTOR signaling is essential for many developmental and physiological processes and its deregulation is associated with pathogenesis of acute and chronic inflammatory diseases [10]. mTOR signaling is physiologically initiated when sufficient nutrients and growth factors are available for cellular growth and proliferation, but is inhibited during starvation and/or when cells are under stressed environmental conditions [9,10]. It has been established that during cellular stress, inhibition of protein synthesis, which requires significant energy expenditure, has a survival advantage as it prevents cells from depleting the energy needed to cope with cellular stress [10]. The polyP-mediated proinflammatory RAGE signaling not only activates mTOR and NF-κB pathways but also induces apoptotic oxidative stress by dramatically amplifying the mitochondrial generation of reactive oxygen species [12,27]. Thus, the capacity of polyP to effectively activate RAGE and mTORC1/2 signaling network in vascular endothelial cells, as demonstrated in this study, can dramatically disrupt normal cellular function. In support of a pathological role for RAGE signaling, RAGE null mice, or wild-type mice treated with soluble RAGE lacking the transmembrane spanning and cytoplasmic domains of the receptor, exhibit increased survival rates in various inflammatory and injury animal models [28–30]. High levels of polyP and nuclear proteins can be released to circulation by platelets and other cells under various cellular stress, injury and proinflammatory conditions [3,12,31]. Our results suggest that polyP can dramatically exacerbate these pathophysiological processes by activating mTOR and RAGE pathways in the vascular system. Thus, polyP may constitute a novel therapeutic target for treating/preventing metabolic and proinflammatory disorders.