Pdcd4 Is Involved in the Formation of Stress Granule in Response to Oxidized Low-Density Lipoprotein or High-Fat Diet

Stress granules (SGs) in response to various stresses have been reported in many diseases. We previously reported the implication of programmed cell death 4 (Pdcd4) in obesity-induced stress responses, but the possible link between Pdcd4 and SGs remains lacking. In this study we showed that oxidized low-density lipoprotein (ox-LDL) or high-fat diet (HFD) induced SG formation in mouse macrophages and liver tissues, and Pdcd4 deficiency in mice remarkably reduced its formation. In response to ox-LDL, either endogenous or ectopic Pdcd4 displayed granule-like expression and co-localized with SG markers including T-cell-restricted intracellular antigen-1, fragile X mental retardation-related protein 1, and eukaryotic initiation factor 4A. Ectopic expression of truncated Pdcd4 that depleted specific RNA-binding motif significantly disrupted the SG formation, suggesting the direct involvement of Pdcd4 in ox-LDL-induced SGs through its RNA-binding activity. Additionally, Pdcd4 deficiency drove AKT activation and suppression of eIF2α phosphorylation, thereby contributing to the resistance to ox-LDL or HFD-induced SG formation. Collectively, our data suggest that Pdcd4 as a crucial regulator in SGs induced by ox-LDL or HFD maybe a potential target for mitigating SG-associated stress responses in obesity and related diseases.


Introduction
Stress granules (SGs) are translationally silent cytoplasmic ribonucleoprotein complexes that assemble during various types of cellular stresses such as heat shock, viral infection, oxidative and/or endoplasmic reticulum (ER) stresses [1][2][3]. In response to environmental emergency, cells can trigger a sudden translational arrest, leading to rapid polysome disassembly and subsequent assembly of heterogeneous mRNAs and translation factors [4]. Although the composition can vary depending on various stimuli, SGs typically contain mRNA, 40S ribosomal subunits, eukaryotic translation initiation factor (eIF) 4E, eIF4G, eIF4A, eIF4B, poly(A)-binding protein, the institutional guidelines for animal care and utilization. The protocol was approved by Ethical Review Committee of Animal Experiments of Shandong University School of Medicine (Permit Number: NO.ECAESDUSM 2012033). Mice were sacrificed under anesthesia, and all efforts were made to minimize mice suffering.

Cell culture
Primary peritoneal macrophages were harvested from peritoneal exudates of mice, which were injected intraperitoneally with 6% hydroxyethyl starch solution 3 days before the harvest. After 6 h of incubation, non-adherent cells were removed and the adherent cells were used as peritoneal macrophages. HeLa (human cervical carcinoma) and HepG2 (human hepatoblastoma) cell lines were obtained from Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, China), and maintained at 37°C under 5% CO 2 in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Invitrogen, Carlsbad, CA), 100 U/ml penicillin, and 100 mg/ ml streptomycin.

Cell treatment
Primary macrophages from WT or Pdcd4 -/mice were treated with ox-LDL (50μg/ml, Yiyuan, Beijing, China) for 24 h, and then were used to perform western-blot or immunofluorescence.
In some experiments, AKT inhibitor MK2206 (Selleck, Houston, TX) was added 6 h before the addition of ox-LDL. For HeLa or HepG2 cells, ox-LDL (50μg/ml) treatment for 24 h or 16 h was performed after overnight transfection with different plasmids.

Statistical analysis
Statistical analysis was performed using one-way ANOVA or student-t test (GraphPad Prism 5). Data are presented as the mean ± s.e.m. Statistical significance was established at P values< 0.05.

Pdcd4 deficiency reduces the SG formation in macrophages and liver tissues from HFD-fed mice
Pdcd4 has been demonstrated to be involved in HFD-induced obesity and associated stress responses [13]. To determine the possible link between Pdcd4 and SGs upon HFD challenge, we detected the formation of SGs in macrophages and liver tissues in established HFD-fed WT and Pdcd4 -/mice models. Based on our previous study, although WT and Pdcd4 -/mice were both fed on HFD, WT mice developed obesity, while Pdcd4 -/mice displayed lean phenotype [13]. Using TIA-1 as a SG marker, we found that about 16.5% of WT macrophages isolated from HFD-fed mice displayed TIA-1 positive SGs, a significant increase when compared to macrophage isolated from ND-fed mice (about 6.8%). This result suggests that indeed, HFD induces a stress response in WT macrophages. However, macrophages isolated from Pdcd4 -/mice on HFD did not display this increase, showing that Pdcd4 is required for SG formation under HFD (Fig 1A and 1B). Similar effects were observed in liver sections from WT and Pdcd4 -/mice. HFD-fed WT mice showed conspicuous hepatic steatosis and much more cells displayed SGs when compared to mice fed on ND ( Fig 1C). These data indicate that HFD induces the formation of SGs, and importantly, reveal that Pdcd4 serves as key contributor to SG formation induced by HFD stimulus.
Pdcd4 deficiency reduces the formation of TIA-1 + SGs in macrophages in response to ox-LDL Since ox-LDL is a typical biomarker for oxidative stress in HFD-induced obesity and related metabolic syndromes, we next used ox-LDL to treat primary peritoneal macrophages from WT and Pdcd4 -/mice, and examined TIA-1 + SGs in these macrophages. In response to ox-LDL, the TIA-1 + SGs in WT macrophages significantly increased as suggested by elevated percentage of cells containing TIA-1 + SGs compared with non-treated macrophages (Fig 2A and 2C). However, Pdcd4 -/macrophages treated with ox-LDL did not show this increase (Fig 2B and  2C), showing that Pdcd4 is required for SG formation under ox-LDL. The SGs were only observed in a few of Pdcd4 -/macrophages, which is similar to non-treated macrophages ( Fig  2B and 2C). These results suggest that Pdcd4 deficiency causes reduced SG formation in response to ox-LDL, thus revealing a key role of Pdcd4 in ox-LDL-induced SGs.  Pdcd4 is co-localized with specific markers of SGs in ox-LDL-stimulated macrophages To verify the role of Pdcd4 in ox-LDL-induced SGs, the expression of Pdcd4 and other related SG marker proteins were examined in macrophages by immunofluorescence. Fig 3 showed that in response to ox-LDL, Pdcd4 clearly displayed granule-like expression in the perinuclear area of the macrophages and all the detected SG markers including TIA-1, FXR1 and eIF4A were positively expressed. Notably, these markers apparently co-localized with Pdcd4 in the fluorescence granules. While in non-treated macrophages, the expression pattern of TIA-1, FXR1 and eIF4A exhibited rare fluorescence granules (S1 Fig). These data suggest that ox-LDL can induce SG formation in the primary macrophages, and Pdcd4 may be an important component of ox-LDL-induced SGs. Ectopic Pdcd4 is co-localized with specific markers of SGs in ox-LDLstimulated HeLa cells To confirm the involvement of Pdcd4 in ox-LDL-induced SGs, HeLa cells were transfected with pEGFP-C1-Pdcd4 plasmid, and then were treated with ox-LDL, the formation of SGs were determined. In non-treated HeLa cells, exogenous GFP-fused Pdcd4 was mostly expressed in nuclei and only a few SGs were detected. While in ox-LDL-treated HeLa cells, exogenous Pdcd4 was expressed in both nuclear and cytoplasmic location, and obvious SGs were observed. Of note, the cytoplasmic Pdcd4 displaying granule-like expression was also co-localized with SG markers including TIA-1, FXR1 and eIF4A (Fig 4A and 4B). These data are partially in consistent with the observed finding in macrophages, thereby suggesting that ox-LDL treatment also drives exogenous Pdcd4 into SG assembly and further confirming the involvement of Pdcd4 in ox-LDL-induced SGs.
AKT-eIF2α axis is involved in the regulation of Pdcd4 on ox-LDLstimulated SGs As eIF2α phosphorylation plays a pivotal role in the formation of SGs [2], we further examined the expression of p-eIF2α in HeLa cells. Fig 6A and 6B showed that ox-LDL treatment significantly enhanced the expression of p-eIF2α in Pdcd4-overexpressed HeLa cells, indicating that ox-LDL triggers the activation of eIF2α which contributes to the SG formation. Similarly, upon ox-LDL stimulation, a marked elevation in p-eIF2α levels was detected in WT macrophages but not in Pdcd4 -/macrophages compared with non-treated ones (Fig 6C and 6D). Recently, Mounir et al reported the negative effect of AKT signaling on eIF2α phosphorylation which might contribute to relieved ER or oxidative stress [10]. To investigate the possible roles of AKT pathway in eIF2α phosphorylation and SG formation, we further detected the activation status of AKT in ox-LDL-treated WT and Pdcd4 -/macrophages. In response to ox-LDL, Pdcd4 -/macrophages displayed a moderate but significant elevation in AKT phosphorylation as compared to WT macrophages, which was consistent with the decline in eIF2α phosphorylation (Fig 6C and 6E). These results hint the reduction of p-eIF2α in Pdcd4 -/macrophages in response to ox-LDL might be caused by the activation of AKT. To confirm this hypothesis, we further blocked AKT activation using an inhibitor MK2206 in ox-LDL-stimulated Pdcd4 -/-  (C) HepG2 cells were transfected with various truncated plasmids and then were stimulated with ox-LDL (50 μg/ml) for 16 h, the SG formation was examined by immunofluorescence. Exogenous Pdcd4 is visualized with GFP (green), TIA-1 was stained with rhodamine (red). Cells were macrophages. As expected, after blocking AKT phosphorylation, the expression of p-eIF2α substantially reversed to a relatively higher level (Fig 6F and 6G), accompanied by an obvious increase of SG formation (S3 Fig). On the other hand, knockdown of AKT in Pdcd4 -/macrophages using AKT siRNA, also led to increased level of p-eIF2α and SG formation in response to ox-LDL (S4 Fig). In addition, results from WT and Pdcd4 -/mice fed on HFD also supported the changing trend of AKT-eIF2α axis in Pdcd4-associated SGs. Obvious reduction of eIF2α phosphorylation accompanied by enhancement of AKT activation were detected in macrophages from HFD-fed Pdcd4 -/mice compared with those from HFD-fed obese WT mice ( Fig  6H-6J). These findings indicate the potential regulatory role of AKT pathway in Pdcd4-involved SG formation through p-eIF2α.

Discussion
SG assembly is a highly complex process during the cells response to adverse environmental conditions. Transient SGs appear to have cytoprotective effects that influence cell metabolism and survival [1,2,30], whereas prolonged SG activity is predicted to lead to a persistent stressed state and sustained translational repression that is deleterious [31,32]. Sustained SG assembly has been recognized as a conspicuous feature of some degenerative diseases such as amyotrophic lateral sclerosis and Alzheimer's disease [31,33,34]. In the present study, we showed that ox-LDL, a biomarker of persistent oxidative stress in obesity and related metabolic syndromes [20,21], induced obvious SG formation in WT macrophages. In HFD-fed obese mice, numerous SGs were also detected in macrophages and fatty liver tissues. However, compared with ox-LDL, HFD induces relatively few SGs in macrophages (Figs 1B and 2C), possibly because of its indirect effects on macrophages or the influence by in vitro manipulation in the absence of HFD. Since ox-LDL is not only the product of oxidative stress, but also enhances oxidative stress by promoting the generation of reactive oxygen species, it is very likely that HFD-induced SGs are caused by ox-LDL-related oxidative stress. These observations reveal the involvement of SGs in obesity and related diseases, in which either HFD or ox-LDL brings long and chronic oxidative stress, thus contributing to sustained SG assembly. Therefore, SG assembly in obesity is not only a result of stress stimuli, but also indicates persistent unsolved oxidative or ER stress. On the other hand, sustained SGs in turn lead to persistent blockade of translation that may produce detrimental effects on cell survival and function. Although the precise roles of SGs need to be further investigated, it still raises the possibility that mitigating SG-associated stress responses in these pathological conditions.
Pdcd4 has been recognized as a translation repressor which inhibits translation in a capdependent way. Using its MA-3 domain, Pdcd4 binds to eIF4A and blocks the interaction of eIF4A and eIF4G, thus interrupting the formation of translation initiation complex [29]. These findings provide the possibility of Pdcd4 to be involved in SG assembly. The co-localization of Pdcd4 with SG markers like eIF4A, TIA-1 and FXR1 confirmed that Pdcd4 is a key component of ox-LDL-induced SGs. Of note, we have demonstrated that the RNA-binding activity of Pdcd4 plays a critical role during SG assembly. Depletion of RNA-binding motifs RBM2 or both RBM1 and RBM2 led to almost vanished SG formation in response to ox-LDL. Previous reports have verified the RNA-binding activity of Pdcd4 and the association of Pdcd4 with small ribosomal subunits [29]. These data provide some evidence and further strengthen the importance of RNA-binding by Pdcd4 in SG formation, suggesting that the involvement of Pdcd4 in SG assembly is dependent on RNA-binding activity.
As the central trigger of the integrated stress response, eIF2α phosphorylation on Ser51 reduces levels of the eIF2-GTP-tRNA i Met ternary complex that is required for cap-dependent translation initiation, thereby leading to translational arrest, polysome disassembly, and SG assembly [1,2,30]. Our data showed markedly increased SG formation in WT macrophages stimulated with ox-LDL, accompanied by remarkable elevation in p-eIF2α levels, thus supporting that ox-LDL-induced SG formation is dependent on eIF2α phosphorylation. More interesting, Pdcd4 deficiency conferred macrophages resistance to SG formation in response to ox-LDL or HFD, meanwhile the levels of p-eIF2α dramatically decreased. Consistently, increased levels of p-eIF2α were detected in Pdcd4-overexpressed HeLa cells in response to ox-LDL, which simultaneously led to obvious SG assembly. These findings suggest that Pdcd4 could regulate SG formation induced by ox-LDL, and eIF2α may act as a regulatory hub during this process.
Since AKT activation has been demonstrated to inhibit eIF2α phosphorylation [10], to dissect the possible mechanisms for the suppression of p-eIF2α in Pdcd4 -/macrophages, we further compared the expression levels of p-AKT in Pdcd4 -/and WT macrophages. As expected, in response to ox-LDL or HFD, a significant increase of p-AKT levels together with a decrease of p-eIF2α levels were detected in Pdcd4 -/macrophages compared with WT macrophages. Blockade of AKT activation in ox-LDL-stimulated Pdcd4 -/macrophages led to significant increase in p-eIF2α level and SG formation. Therefore, as a sensor to stress signals ox-LDL or HFD, Pdcd4 promotes eIF2α phosphorylation and SG formation, at least partially by repressing AKT activation. There are also other possibilities that p-AKT regulates SG formation independent of eIF2α phosphorylation. One recent report has demonstrated that TORC2, which is involved in SG formation in Drosophila, is required for AKT phosphorylation on S505 and stability upon heat shock [9]. Although the phosphorylation of eIF2α is not required for SG formation in Drosophila during heat stress [4], whether and how AKT plays a role in SG formation in Drosophila remains to be investigated [9]. Notably, the possible regulatory roles of TORC2-AKT signaling in SG formation in Drosophila may not be applied to mammalian cells used in our study, because of the differences in species, stress stimulus, and phosphorylation at AKT residues. In addition, it has been reported that AKT can activate TORC1, which is involved in SGs [11,35]. So, it is important to consider the role of TORC1 pathway when studying the regulation of p-AKT in SG formation. As for the regulation of Pdcd4 on AKT activity, it has been reported in several types of cells, but detailed mechanisms remain to be elucidated. One possible explanation is that low level of Pdcd4 de-represses the activity of eIF4A RNA helicase, which contributes to the activation of PI3K-AKT pathway by increasing the translation of eIF4A-sensitive mRNA [36][37][38].
Collectively, stress signals like ox-LDL or HFD have the ability to induce SG accompanied by eIF2α phosphorylation. This process is dependent on the involvement of Pdcd4, which not only functions as a key component of SG assembly via its RNA-binding region, but also promotes the phosphorylation of eIF2α by repressing AKT activation (Fig 7). Furthermore, it is also interesting to clarify the role of Pdcd4 in SG formation induced by other stimuli like heat shock or arsenite, which need to be further investigated. In conclusion, our data suggest that Pdcd4 plays an indispensible role in the pathogenesis of obesity and related diseases, thus may serve as a potential therapeutic target for these diseases. Pdcd4 -/macrophages were transfected with control or AKT siRNA (Sigma) for 36 h using GenePORTER 1 2 Transfection Reagent (Genlantis), and then were treated with ox-LDL (50 μg/ml) for additional 24 h. The effect of AKT siRNA on AKT expression and the influence of AKT pathway on the level of p-eIF2α were determined by western-blot. Representative (A, C) and statistic (B, D) data was shown. ÃÃ P <0.01, ÃÃÃ P <0.001. (E) The influence of AKT pathway on the formation of TIA-1 + SGs was assayed by immunofluorescence. TIA-1 (red); nuclei (blue). The original magnification is 1000. Scale bar = 10μm. (TIF)