Chaperone Insufficiency Links TLR4 Protein Signaling to Endoplasmic Reticulum Stress*

Background: Activation of TLR4 leads to endoplasmic reticulum stress. However, the mechanisms involved in this phenomenon are unknown. Results: In TLR4 signaling, insufficient GRP94 and GRP78 mediate the activation of endoplasmic reticulum stress. Conclusion: The insufficiency of chaperone expression links TLR4 signaling to endoplasmic reticulum stress. Significance: This study may improve our understanding about the inflammatory response in metabolic and infectious disease. Inflammation plays an important pathogenic role in a number of metabolic diseases such as obesity, type 2 diabetes, and atherosclerosis. The activation of inflammation in these diseases depends at least in part on the combined actions of TLR4 signaling and endoplasmic reticulum stress, which by acting in concert can boost the inflammatory response. Defining the mechanisms involved in this phenomenon may unveil potential targets for the treatment of metabolic/inflammatory diseases. Here we used LPS to induce endoplasmic reticulum stress in the human monocyte cell-line, THP-1. The unfolded protein response, produced after LPS, was dependent on CD14 activity but not on RNA-dependent protein kinase and could be inhibited by an exogenous chemical chaperone. The induction of the endoplasmic reticulum resident chaperones, GRP94 and GRP78, by LPS was of a much lower magnitude than the effect of LPS on TLR4 and MD-2 expression. In face of this apparent insufficiency of chaperone expression, we induced the expression of GRP94 and GRP78 by glucose deprivation. This approach completely reverted endoplasmic reticulum stress. The inhibition of either GRP94 or GRP78 with siRNA was sufficient to rescue the protective effect of glucose deprivation on LPS-induced endoplasmic reticulum stress. Thus, insufficient LPS-induced chaperone expression links TLR4 signaling to endoplasmic reticulum stress.

The anomalous activation of macrophages plays an important role in metabolic diseases such as atherosclerosis, obesity, and type 2 diabetes mellitus (1)(2)(3). Pattern recognition receptors (PPRs), 2 particularly TLR4 (4), and endoplasmic reticulum stress (ERS) (5) have been independently identified as important inducers of dysfunctional macrophage activation in these diseases, and both genetic and pharmacological targeting of the TLRs and ERS can attenuate or revert the metabolic phenotypes (4, 6 -8). Recent data have revealed that signals generated by the activation of TLRs or the induction of ERS can integrate with each other and, thus, boost the pathological inflammatory response, eventually leading to macrophage apoptosis, which in the case of atherosclerosis is a well known complicating event (8 -10). However, the mechanisms involved in the integration of TLR signaling with ERS are virtually unknown.
The expression of TLR2 and TLR4 on the cell surface is necessary for pattern recognition and the activation of signal transduction (11). To reach the cell membrane, the nascent TLRs must be correctly folded, which depends on their binding to the ER resident chaperone GRP94 (also known as gp96) (12). Whole body deficiency of GRP94 is lethal at embryonic day 5 (12,13). However, when only macrophages are depleted of GRP94, rodents are viable, presenting a defective response to Listeria monocytogenes, which results from an almost complete absence of cell surface TLRs expression (14).
In macrophages, lipopolysaccharide (LPS) stimulation produces ERS, a phenomenon believed to play an important role in eliciting the cytokine response required for optimal bacterial destruction (15). However, upon LPS stimulation, cells undergo a rapid reduction in the surface expression of TLR2 and TLR4, which is promptly accompanied by an increase in TLRs mRNA expression (16). Because replacement of surface TLR depends upon adequate chaperoning in the ER, we suspected that under LPS stimulation insufficient chaperone expression could mediate TLR4-induced ERS. Here, we show that, upon LPS stimulation, both GRP94 and GRP78 undergo only a discrete increase in protein expression level. After glucose deprivation, the substantial increase in the expression of these chaperones completely inhibits LPS-induced ERS, an effect that can be rescued by targeting either GRP94 or GRP78 with interference RNA.
Cell Line-The monocyte cell line, THP-1, was plated at an initial density of 10 6 cells/ml in Roswell Park Memorial Institute (RPMI) 1640 medium containing 10% fetal bovine serum (FBS) in 1.5-cm culture dishes. The cells were incubated with LPS (0.5 g/ml or other concentrations as stated in the legends to the figures), and the expression/activation of proteins involved in ERS was determined by real-time PCR, immunoblot, or flow cytometry. In some experiments, cells were deprived of glucose for 6 h or treated with siRNA (GRP94 siRNA, sc35523 and GRP78 siRNA, sc29338 from Santa Cruz Biotechnology) to increase or decrease, respectively, the expression of the chaperones. Pharmacological inhibition of endoplasmic reticulum stress was achieved by preincubation for 1 h at 37°C with the chemical chaperone PBA at a final concentration of 3 mM; after this period the cells were chased by LPS (0.5 g/ml) for 24 h. For experiments aimed at evaluating TLR4 signaling inhibition, the cells were incubated with 2 g/ml anti-CD14 for 1 h; after this period the cells were chased by LPS (0.5 g/ml) for 24 h. Inhibition of AMPK was achieved by incubation of cells at 37°C with compound-C at a final concentration of 20 M (17). Incubation with compound-C started 2 h before the beginning of glucose deprivation, which lasted for 6 h. Cells were then returned to medium containing glucose, let to rest for 24 h, and finally chased with LPS (0.5 g/ml) for 24 h. Compound-C was maintained in medium throughout the experimental procedure.
Glucose Deprivation Protocol-THP-1 cells were deprived of glucose for 6 h in RPMI 1640 without glucose containing 10% of FBS. After this period glucose was reintroduced at a final concentration of 11 mM. The evaluation of chaperone expression and the incubation of LPS were performed after 24 h of glucose reintroduction.
GRP78 and GRP94 siRNA-To inhibit chaperone expression, cells were transfected with 50 nM GRP78 siRNA or with 80 nM GRP94 siRNA conjugated with Lipofectamine TM ; 2000 for 7 h. The experiment was performed 35-60 h after the transfection. Optimal concentrations of siRNAs were obtained after dose-response and time-course experiments.
Flow Cytometry-At the end of the incubation period with LPS, THP-1 cells were harvested and resuspended in PBS. For F4/80, TLR4, and MD2 labeling, the cells were not submitted to any fixation or permeabilization protocol. For the remainder of the proteins (p-JNK, p-PERK, p-elF2␣, GRP78, and GRP94), cells were permeabilized with 0.1% saponin. Primary antibodies were used at a final concentration of 2.5 g/ml, and the secondary, FITC-conjugated antibody was used at a final dilution of 1:100. Signal detection was performed in a FACSCalibur flow cytometer equipped with an argon laser, and analyses of data were performed with the CellQuest software (BD Biosciences). Differences are presented as percent variation of control.
Reverse Transcription PCR and XBP-1 Splicing-Standard reverse transcription PCR was performed using total RNA from THP-1 cell line, as described previously (19).
Immunoprecipitation and Immunoblotting-For evaluation of TLR activation and ER stress induction, the THP-1 cells were homogenized in solubilization buffer at 4°C. Aliquots of the resulting protein extract containing 0.35 mg of total protein were used for immunoprecipitation with antibodies against TLR4, GRP78, GRP94, and MyD88 at 4°C overnight followed by SDS-PAGE transfer to nitrocellulose membranes and blotting with anti-TLR4, anti-GRP78, anti-GRP94, or anti-MyD88 antibodies. In direct immunoblot experiments, 0.06 mg of protein extracts were separated by SDS-PAGE, transferred to nitrocellulose membranes, and blotted with anti-pPERK, anti-PERK, anti-pEIF2␣, anti-EIF2␣, anti-ATF6, anti-sXBP-1, anti-GRP78, or anti-GRP94. In immunodepletion experiments, 0.35 mg of total protein extracts were submitted to a first round of immunoprecipitation with 10 l of TLR4 antibody. Immunocomplexes were precipitated with protein-A-Sepharose and saved. The supernatant was incubated with 10 l of TLR4 antibody in the second round of immunoprecipitation. This approach continued for a total of four rounds. All immunoprecipitates collected were separated by SDS-PAGE, transferred to nitrocellulose membranes, and blotted with TLR4 and GRP94 antibodies. In animal experiments, 0.06 mg of total protein extracts obtained from the spleen were separated by SDS-PAGE, transferred to nitrocellulose membranes, and blotted with anti-pPERK, anti-PERK, anti-pEIF2␣, and anti-EIF2␣. Specific bands were detected by chemiluminescence, and visualization was performed by exposure of the membranes to X-ray film.
Statistical Analysis-Specific protein bands present in the blots were quantified by digital densitometry (UN-SCAN-IT). Mean values Ϯ S.E. obtained from densitometry scans and from real-time PCR were compared using Tukey-Kramer test (analysis of variance); p Ͻ 0.05 was accepted as statistically significant.

LPS Activates Signal Transduction through TLR4 and Induces Expression of F4/80 in Human
Monocytes-THP-1 is a human monocyte cell line that under non-stimulated conditions expresses virtually no markers of either ERS or inflamma-FIGURE 1. LPS-induced activation of TLR4 signaling in THP-1 cells. THP-1 cells were exposed for 2-24 h to 0.5 g/ml LPS, and protein extracts were obtained for immunoprecipitation (IP) using anti-MyD88 antibody and immunoblotting (IB) using anti-TLR4 antibody (A). The homogeneity of the cell preparation was evaluated by flow cytometry, as determined by the forward scatter (FSC-H) and the side scatter (SSC-H) (B), and the homogeneous cell preparation is shown in R1. The basal (C) as well as the LPS (D-F)-stimulated expressions of F4/80 were evaluated by flow cytometry; cells were exposed to 0.5-5.0 g/ml LPS for 24 h (D and E). au, absorbance units. In F, cells were exposed to 0.5 g/ml LPS and harvested after 0, 12, or 24 h. In all experiments, n ϭ 6. In A: *, p Ͻ 0.05 versus CTR; in E and F: *, p Ͻ 0.05 versus LPS 0 g/ml and time 0 h, respectively. tion. A time-course experiment showed that the continuous exposure of THP-1 cells to LPS results in the activation of signal transduction through TLR4/Myd88 as long as the LPS is present for at least 24 h (Fig. 1A). The homogeneity of the cells was determined by flow cytometry employing the forward scatter (FSC-H) and the side scatter (SSC-H) (Fig. 1B). To evaluate the capacity of LPS to induce an inflammatory response by the cells, we determined the expression of the membrane protein F4/80, which is typically expressed in activated macrophages. Before LPS stimulus, the expression of F4/80 was virtually absent in the THP-1 cell line (Fig. 1C). However, a dose-response experiment showed that a LPS concentration of as low as 0.5 g/ml was capable of inducing F4/80 expression (Fig. 1, D-E). The induction of F4/80 by 0.5 g/ml LPS was detectable at 24 h but not at 12 h (Fig. 1F). In all subsequent experiments LPS was employed for 24 h at the concentration of 0.5 g/ml.

LPS Induces ERS in Human Monocytes-ERS can be induced by a number of different inflammatory stimuli, including LPS.
To define the best approach to induce ERS in the THP-1 cell line, we performed time-course experiments and evaluated a number of markers of ERS. The three canonical pathways involved in the unfolded protein response (UPR) were activated by LPS. The IRE-1 pathway was evaluated by the determination of spliced XBP-1, which was increased as early on as 2 h after the beginning of LPS exposure, lasting for at least 48 h ( Fig. 2A). The PERK pathway was also activated as demonstrated by the pPERK/PERK (Fig. 2B) and peIF2␣/eIF2␣ (Fig. 2C) ratios using immunoblot and also by the determination of pPERK (Fig. 2E) and peIF2␣ (Fig. 2F) by flow cytometry. The activation of the ATF-6 pathway was determined by immunoblot using an antibody that reacts against the cleaved/active form of the ATF-6 protein (Fig. 2D). Additional markers of ERS were also induced by LPS, such as the serine kinase, pJNK (supplemental Fig. 1A), ATF-3 (supplemental Fig. 1B), the calcium-regulatory proteins stanniocalcin-2(supplementalFig.1C)andcalreticulin(supplemental Fig. 1D), and the co-chaperone DNAJC3 (supplemental Fig. 1E). The timings for activation of each of the evaluated proteins varied; however, all proteins except calreticulin were induced 24 h after the beginning of LPS exposure. Thus, 24 h of LPS exposure was employed in most of the subsequent experiments.
CD14/TLR4, but Not PKR, Are Required to Transduce LPS Signal toward ERS-As double-stranded PKR is implicated in the connection of the nutrient and inflammatory cues leading to ERS through eIF2␣ (20), we evaluated the capacity of LPS to activate ERS through PERK and eIF2␣ in PKR-KO mice. As our working hypothesis is that LPS-induced ERS depends on the inappropriate handling of TLR4 by the ER, we had to rule out the alternative pathway linking inflammation to ERS through PKR. As depicted in Fig. 3, A and B, LPS was equally effective at inducing the phosphorylation of PERK and eIF2␣ in the spleen of either C57 or PKR-KO mice. Nevertheless, blocking CD14 with an inhibitory antibody resulted in the complete inhibition of LPS-induced ERS through XBP-1 (Fig. 3C), pPERK/peIF2␣ (Fig. 3, D and E). and ATF-6 ( Fig. 3F).
LPS Induces Only Discrete Increase of Chaperone, GRP94-LPS induced the expression of GRP94 in THP-1 cells, as determined by immunoblot (Fig. 4A), flow cytometry (Fig. 4E), and realtime PCR (Fig. 4I). The magnitude of the increase in GRP94 was comparable to the increase in GRP78 (Fig. 4, B and F). However, when compared with the impact of LPS on TLR4 expression ( Fig.  4. C, G, and J), there was always an imbalance favoring TLR4 expression independently of the method employed for the measurement of the respective proteins/mRNAs (Fig. 4, L-N). Conversely, MD2, another protein that binds to TLR4 and is required for its cell-membrane expression, is induced by LPS at levels much higher than TLR4 (Fig. 4, D, H, K-N).
LPS-induced ERS Is Inhibited by Chemical Chaperone-LPS induced the increased physical association of TLR4 with GRP94 ( Fig. 5A) but not with GRP78 (Fig. 5B). However, in immunedepletion experiments, we observed that in non-LPS-stimulated cells, the complex TLR4/GRP94 could be recovered even after four rounds of depletion, whereas in LPS-stimulated cells, this In all experiments, cells were exposed to 0.5 g/ml LPS and harvested at times depicted in the panels (A-D) or after 24 h (E and F). In A-D, protein extracts were separated by SDS-PAGE and immunoblotted with anti-spliced XBP1 (A), pPERK (B), peIF2␣ (C), or ATF6 (D) antibodies. All membranes were stripped and reblotted with anti-␤-actin antibody. In E and F, the expressions of pPERK (E) and peIF2␣ (F) were evaluated by flow cytometry. Controls were labeled or not with the specific antibodies, CTRϩ(pPERK or peIF2␣), or CTR, respectively. For the flow cytometry evaluation some groups of cells were treated with thapsigargin (THAP). In all experiments, n ϭ 6. *, p Ͻ 0.05 versus CTR.
complex was almost non-detectable already after three rounds of depletion (Fig. 5C). To test the hypothesis that insufficient chaperoning of TLR4 would provide the link between TLR4 signal transduction activation and ERS, we treated the cells with the chemical chaperone PBA, which acts by binding and providing an appropriate folding to nascent proteins in the ER (6). The treatment with PBA resulted in the complete inhibition of LPS-induced ERS as determined by the expression of spliced XBP-1 (Fig. 5D) and pPERK/PERK, peIF2␣/eIF2␣ ratios (Fig. 5, E-F).
Glucose Deprivation Increases Steady-state and LPS-induced GRP94 Expression-The protein expressions of both GRP94 and GRP78 are known to be induced by glucose deprivation (21). Therefore, we decided to employ this strategy as an attempt to increase chaperone availability in the ER. THP-1 cells were glucose-deprived for 6 h and used in experiments 24 h later (Fig. 5G). The protein expressions of GRP94 (Fig. 5H) and GRP78 (Fig. 5I) were significantly increased by glucose deprivation. However, when glucose deprivation was followed by LPS exposure, only GRP94, but not GRP78, demonstrated an additional significant increase as compared with LPS alone (Fig.  5H). As glucose deprivation is known to induce the activity of the enzyme AMPK, which can per se modulate ERS through mTOR (22), we pretreated THP-1 cells with the AMPK inhib-itor, compound-C, and determined the impact of glucose deprivation on ERS. As depicted in Fig. 5J, although glucose deprivation was indeed capable of activating AMPK, the inhibition of AMPK produced no impact on either the expression of chaperones or the inhibition of ERS induced by glucose deprivation.
Glucose Deprivation Protects Human Monocytes from LPSinduced ERS-The capacity of LPS to induce ERS though all the three canonical pathways was completely inhibited by the preincubation of THP-1 cells in glucose-deprived medium (Fig. 6, A-C).
Small-interfering RNA Targeting of GRP94 or GRP78 Rescues Protective Effect of Glucose Deprivation on LPS-induced ERS-To obtain further evidence of the positive impact of glucose deprivation-induced chaperone expression on the protection against LPS-induced ERS, we inhibited the expression of the chaperones employing siRNA. The use of siRNA to target GRP94 or GRP78 resulted in an almost complete disappearance of steady-state or glucose-induced protein levels of either protein (Fig. 6, D-E). This approach rescued the cells from the protective effect of glucose deprivation on LPS-induced ERS, as determined by the evaluation of spliced XBP-1 ( Fig. 6F and 6I, for GRP94 and GRP78, respectively), pPERK/PERK ( Fig. 6G and 6J, for GRP94 and GRP78, respectively) and ATF-6 ( Fig. 6H and 6K, for GRP94 and GRP78, respectively). Finally, the inhibition of GRP94 expression restored the capacity of LPS to induce the expression of the inflammatory cytokine IL-8, even when cells were glucose-deprived (Fig. 6L).

DISCUSSION
LPS is a potent activator of the innate immune response, which plays an important role in the defense against Gramnegative bacteria (23). It is currently accepted that a LPS receptor complex is formed by the association of TLR4 with MD-2, and also depends on two accessory proteins, CD14 and LPSbinding protein (LBP) (24). Upon ligand binding, TLR4 signal transduction proceeds though at least two distinct intracellular pathways: Myd88, which activates gene transcription in response to NFB; and Trif, which can activate either NFB or IRF3 (25). The excessive activation of this pathway in living organisms can lead to a life-threatening condition known as toxic shock, which can result in multiple organ failure and, eventually, death (26). Thus, the termination of TLR4 signaling is tightly regulated, depending on the endocytosis and lysosomal degradation of the protein (27). If the LPS signal is brief, the TLR4 synthetic machinery in the ER will replace the recycled proteins, and the response to LPS will be rapidly normalized (28). However, upon prolonged exposure, cells become desensitized to LPS at least in part due to insufficient TLR4 replacement (29,30).
Recent studies have reported the induction of ERS as a consequence of TLR4 activation (8 -10). This integration of signals has a pathogenic role in diseases such as obesity, type 2 diabetes, and atherosclerosis, and the complete characterization of this phenomenon may offer new therapeutic possibilities for a number of metabolic and inflammatory diseases. In the hypothalamus of obese rodents, the inhibition of TLR4 signaling produces a complete inhibition of ERS (8). Conversely, the inhibition of ERS with a chemical chaperone blocks ERS alone but has no effect on TLR4 signaling (8). With this information in mind, we hypothesized that TLR4/ERS integration would occur as a consequence of a primary and prolonged stimulus of the TLR4 pathway, which would activate the synthesis and folding of new TLR4 units in the ER, therefore, generating a stress condition for the organelle. As a type I transmembrane protein, the nascent TLR4 molecules are folded in the ER, requiring the activity of the resident chaperone GRP94 in association with PRAT4A (11,30). In addition to the chaperoning activity of GRP94, the association of TLR4 with MD-2 is a required condition for the receptor to functionally reach the cell membrane (31). The importance of this association is illustrated by the fact that in MD-2 knock-out mice, the presence of TLR4 in the cell membrane is drastically reduced and the response to LPS is severely compromised (31). Thus, it is expected that, during an increased demand for novel TLR4 units in the cell membrane, the increased   TLR4 (B) antibodies. In C, THP-1 cells were maintained in steady state (CTR) or treated with LPS; protein extracts were submitted to four consecutive rounds (R1-R4) of immunoprecipitation with anti-TLR4 antibody. Immunoprecipitates collected in each round were separated by SDS-PAGE and blotted with TLR4 antibody. Membranes were stripped and reblotted with anti-GRP94 antibody. In D-F, THP-1 cells were maintained in steady state (CTR) or treated with PBA, LPS, PBAϩLPS, or thapsigargin. Protein extracts were separated by SDS-PAGE and immunoblotted with anti-spliced XBP1 (D), pPERK (E), or peIF2␣ (F) antibodies. All membranes were stripped and reblotted with anti-␤-actin. In G, the protocol for glucose deprivation followed by 24 h LPS treatment is depicted. In H-I, THP-1 cells were maintained in steady state (CTR) or treated with LPS or submitted to glucose deprivation without LPS treatment (GLU-R) or submitted to glucose deprivation followed by LPS treatment (GLU-RϩLPS). Protein extracts were separated by SDS-PAGE and immunoblotted with anti-GRP94 (H) or GRP78 (I) antibodies. All membranes were stripped and reblotted with anti-␤-actin. In J, THP-1 cells were maintained in steady state (CTR) or treated with LPS, compound-C (CC), CCϩLPS, submitted to glucose deprivation without LPS treatment (GLU-R), submitted to glucose deprivation followed by LPS treatment (GLU-RϩLPS), treated with CC followed by glucose deprivation (CCϩGLU-R), or treated with CC followed by glucose deprivation and then followed by LPS (CCϩGLU-RϩLPS). Protein extracts were separated by SDS-PAGE and immunoblotted with antibodies against pAMPK, pACC, GRP94, GRP78, pPERK, peIF2␣, spliced XBP1, or ATF6. In all experiments, n ϭ 6. To investigate the mechanisms involved in TLR4-induced ERS, we treated a human monocyte cell line with LPS. The induction of ERS in monocytes/macrophages can be considered to be a physiological phenomenon to which the cells are perfectly adapted (32). In these cells ERS plays an important role in the phenotypic changes and cytokine expression that are necessary for the most effective handling of pathogens (33). In THP-1 cells, LPS produced in a dose-and time-dependent fashion the activation of ERS with the induction of UPR through all the three canonical pathways. There were some expected variations in the time-course of activation of markers of ERS (34), but as rule 24 h of continuous exposure to LPS was sufficient to activate all markers tested. In addition, this activation was dependent on the adequate function of the LPS receptor complex, as immunoinhibition of CD14 was capable of completely restraining the phenomenon.
As expected, LPS produced some increase in the main TLR4 processing chaperone, GRP94, and in the chaperone that acts as the main sensor for stress, GRP78. However, the magnitude of chaperone increase after LPS exposure was somewhat lower than that of either TLR4 or MD-2. In fact, the increase in MD-2, which is equally required for TLR4 to reach the cell membrane, is up to 10-fold higher than the increase in GRP94.
As a proof of the importance of chaperone insufficiency for the induction of ERS by LPS, we were capable of completely inhibiting UPR by treating the cells with the chemical chaperone, PBA, an approach with proven efficiency at inhibiting ERSdependent insulin resistance and diabetes (6).
Recently, PKR has been identified as yet another component of the machinery that connects metabolic stress to inflammation (20). It acts in part by inducing the phosphorylation of the ER protein eIF2␣. As PKR can also be activated by inflammatory signals, we asked if the connection between TLR4 and ERS could occur indirectly through PKR. The experiments with PKR-KO FIGURE 6. GRP94 and GRP78 control ERS in THP-1 cells. In A-C, THP-1 cells were maintained in steady state (CTR) or treated with LPS (24 h) or submitted to glucose deprivation without LPS treatment (GLU-R) or submitted to glucose deprivation followed by LPS treatment (GLU-Rϩ24 h LPS). Protein extracts were separated by SDS-PAGE and immunoblotted with anti-spliced XBP1 (A), peIF2␣ (B), or ATF6 (C) antibodies. In D and E, THP-1 cells were treated with Lipofectamine alone (Lipo) or in the presence of glucose deprivation (LipoGlu-R) or with a control siRNA (siCTR), a GRP94 (D), or a GRP78 (E) siRNA (siGRP94 or siGRP78, respectively), or a GRP94 (D) or a GRP78 (E) siRNA in the presence of glucose deprivation (siGRP94Glu-R or siGRP78Glu-R, respectively). Protein extracts were separated by SDS-PAGE and immunoblotted with anti-GRP94 (D) or GRP78 (E) antibodies. In F-L, THP-1 cells were treated with Lipofectamine alone (Lipo) or in the presence of LPS, glucose deprivation (LipoGlu-R), glucose deprivation plus LPS (GLU-RϩLPS) in the absence of presence of a GRP94 (F-H and L), or a GRP78 (I-K) siRNA (siGRP94 or siGRP78, respectively). Protein extracts were separated by SDS-PAGE and immunoblotted with anti-spliced XBP1 (F and I), pPERK (G and J), or ATF6 (H and K) antibodies or real-time PCR was used to determine the expression of IL-8 (L). In A-K, all membranes were stripped and reblotted with anti-␤-actin. In all experiments, n ϭ 6. *, p Ͻ 0.05 versus CTR in A-C or versus Lipofectamine alone in D-L. In A and B, #, p Ͻ 0.05 versus 24-h LPS. asu, arbitrary scanning units. mice clearly show that even in the absence of this kinase, LPS, acting through TLR4, is capable of inducing ERS.
Next, we used an approach to increase endogenous chaperone expression. Both GRP94 and GRP78 are inducible by glucose deprivation. In the THP-1 cell line we obtained 2.4-and 10-fold increases in the expressions of GRP94 and GRP78, respectively, at 24 h after glucose deprivation. However, when compared with the effect of LPS exposure, glucose deprivation produced an additional increase only in GRP94 but not in GRP78. This was accompanied by a complete inhibition of LPSinduced ERS.
The nutrient gauge, AMPK, which is also activated by glucose deprivation (35), can inhibit ERS through mTOR (22). To investigate if the effect of glucose deprivation on ERS could occur at least in part as a consequence of the activation of the AMPK pathway, we used the chemical inhibitor of AMPK, compound-C, that was capable of inhibiting the glucose-induced activation of AMPK but produced no effect on LPS-induced ERS.
In the final part of the study we targeted GRP94 and GRP78 with siRNA. Using this approach, we obtained a complete inhibition of glucose deprivation-induced expression of both chaperones. Under chaperone-interference conditions, glucose deprivation was no longer capable of restraining LPS-induced ERS through all three canonical pathways.
Thus, we conclude that in human monocytes prolonged TLR4 activation by LPS produces ERS, leading to UPR due to the insufficient expression of chaperones, particularly GRP94, to match the folding demands of the newly synthesized TLR4 molecules. This can be completely reversed by exogenous chemical chaperones or by the induction of endogenous chaperones by glucose restriction. Fig. 7 summarizes the mechanism involved in LPS/TLR4-induced ERS/UPR. In all experiments, cells were exposed to 0.5 g/ml LPS and harvested after 24h (A) or at times depicted in the panels (B-E). In A, the expression of pJNK, was evaluated by flow-cytometry. Controls were labeled or not with the specific antibodies, CTR+pJNK, or CTR, respectively. For the flow-cytometry evaluation some groups of cells were treated with thapsigargin (THAP). In B-E, real-time PCR was used to determine the expressions of ATF3 (B), stanniocalcin-2 (C), DNAJC3 (D) and calreticulin (E). In all experiments, n = 6. *p<0.05 vs. CTR. E.