PDCD4–MAPK–NF-κB Positive Loop Simultaneously Promotes Microglia Activation and Neuron Apoptosis During Neuroinammation

Neuroinammation and neuron injury are common features of the central nervous system (CNS) diseases. It is of great signicance to identify their shared regulatory mechanisms and explore the potential therapeutic targets. Programmed cell death factor 4 (PDCD4), an apoptosis-related molecule, extensively participates in tumorigenesis and inammatory diseases, but its expression and biological function during CNS neuroinammation remain unclear. In the present study, utilizing the lipopolysaccharide (LPS)-induced neuroinammation model in mice, we reported an elevated expression of PDCD4 both in injured neurons and activated microglia of the inamed brain. A similar change in PDCD4 expression was observed in vitro in the microglial activation model. Silencing PDCD4 by shRNA signicantly inhibited the phosphorylation of MAPKs (p38, ERK, and JNK), prevented the phosphorylation and nuclear translocation of NF-κB p65, and thus attenuated the LPS-induced microglial inammatory activation. Interestingly, LPS also required the MAPK/NF-κB signaling activation to boost PDCD4 expression in microglia, indicating the presence of a positive loop. Moreover, a persistent elevation of PDCD4 expression was detected in the H 2 O 2 -induced neuronal oxidative damage model. Knocking down PDCD4 signicantly inhibited the expression of proapoptotic protein BAX, suggesting the proapoptotic activity of PDCD4 in neurons. Taken together, our data indicated that PDCD4 may serve as a hub regulatory molecule that simultaneously promotes the microglial inammatory activation and the oxidative stress-induced neuronal apoptosis within CNS. The microglial PDCD4–MAPK–NF-κB positive feedback loop may exaggerate the vicious cycle of neuroinammation and neuronal injury and thus may become a potential therapeutic target for neuroinammatory diseases. on in the subsequent experiments to establish the experimental model. demonstrated both


Introduction
Neuroin ammation, an in ammatory reaction occurring within the central nervous system (CNS), extensively participates in pathological processes in the CNS [1][2][3]. Microglia, the principal resident immune cells in CNS that account for approximately 10% of brain parenchymal cells, act as the main executor of neuroin ammation [4][5][6]. Under physiological conditions, microglia secrete multiple neurotrophic factors, dynamically monitor synaptic functions, and clear dead cell debris to maintain brain homeostasis. Stimulated by pathogen-associated molecular patterns (PAMPs) following pathogen infection, or by damage-associated molecular patterns (DAMPs) during ischemia, trauma, or neurodegeneration, microglia proliferate, migrate to the damaged area, and act as the rst line of defense in the brain [7,8]. Apart from the neuroprotective effects, microglia have proin ammatory functions, considering that they are the main source of interleukin (IL)-1β, IL-6, tumor necrosis factor-α (TNF-α), chemokines, reactive oxygen species (ROS), and nitric oxide (NO), which further exacerbate the neuroin ammation and result in the subsequent neuronal injury [9][10][11][12].
Accumulating evidence suggests that neuroin ammation and neuron injury can form a vicious cycle to promote neuropathological disorders [13]. For example, we previously reported that soluble HSP60, which is released by the injured neurons, can act as extracellular DAMPs, combining with the microglial receptor LOX-1 and driving neuroin ammation [14] [15]. The mitogen-activated protein kinases (MAPKs) family, including intracellular signal-regulated kinase 1/2 (ERK1/2), the c-Jun N-terminal (JNK), and p38 MAPK, is the backbone of the major proin ammatory signaling pathway within microglia, which can be activated by various PAMPs (such as LPS) or DAMPs (such as sHSP60 or the conditioned medium collected from the injured neurons) [15] [16]. After the activation, MAPKs subsequently trigger the phosphorylation and nuclear translocatin of transcription factors, such as nuclear factor-κB (NF-κB) or activating protein-1 (AP-1), thereby enhancing the transcription of the downstream proin ammatory genes [15] [16]. Although a large number of studies have focused on the in ammatory activation of microglia, the exact molecular mechanism has not been fully elucidated. Considering the close interaction of neuroin ammation and neuronal injury, it is of great signi cance to identify their shared regulatory molecule, analyze their internal relationship, and explore the potential therapeutic targets.
Human programmed cell death factor 4 (PDCD4) gene, discovered in 1999, is located on chromosome 3q21.3 with a total length of 28212bp and encodes a protein containing 496 amino acid residues [17] [18].
Based on proapoptotic and cell-cycle inhibiting activity of PDCD4, intensive efforts have been invested to clarify its role as a tumor suppressor [19] [20]. PDCD4 exerts its antineoplastic effects by promoting cellular apoptosis and inhibiting malignant transformation, cellular proliferation, invasion, and metastasis [19][20] [21] Mechanistically, PDCD4 binds to eIF4A/E to inhibit the RNA helicase activity or directly interacts with the target mRNAs to block protein translation [18] [19]. PDCD4 can also combine with certain cytoplasmic proteins or interact with some transcription factors to modulate the target gene transcription [18] [19].
Besides the tumor-suppressive activity, PDCD4 is also widely expressed in immune cells. It closely participates in in ammatory responses, but its role during in ammation is still uncertain. The available con icting evidence showed both the proin ammatory and anti-in ammatory effects of PDCD4 under different experimental conditions [18] [19]. For example, PDCD4 de ciency protected mice against dietinduced obesity, in ammation in white adipose tissue, and insulin resistance via restoring LXR-α expression [22]. LPS required PDCD4 to induce NF-κB activation and IL-6 expression, and mice de cient in PDCD4 were protected from LPS-induced death [23]. In contrast, other studies reported that PDCD4 knockout mice displayed upregulation of proin ammatory cytokines (such as IL-6) and enhanced activation of the proin ammatory STAT3 activation in the experimental colitis model [24]. The expression and biological function of PDCD4 in CNS have been rarely reported, especially in relation to microglial activation and neuroin ammation [25].
In this study, we analyzed the temporal-spatial pattern of PDCD4 expression in the classical LPS-induced neuroin ammation model in mice and detected its association with microglial activation and neuronal apoptosis for the rst time. Utilizing an in vitro microglial activation model stimulated by LPS, we demonstrated the proin ammatory function of PDCD4 in microglia; we also observed its proapoptotic activity in the oxidative damaged neuronal model. More importantly, our data suggested the existence of a PDCD4-MAPK-NF-κB positive feedback loop to promote the microglial activation and neuronal injury during neuroin ammation.

Animals and Treatments
Six-to eight-week-old male C57BL/6 (n = 63) mice were purchased from the Experimental Animal Center of Nantong University (China) and then randomly assigned into seven groups, including one control group and six experimental groups. The experimental mice had a free access to food and water for 7 days in a 12-hour light-dark cycle temperature-controlled environment (21°C). To establish the CNS in ammation model [26], the mice were intraperitoneally injected with 9 mg/kg of LPS (Escherichia coli O111-B4, Sigma, St. Louis, MO, USA), whereas the same amount of normal saline was injected in the control group. The mice were anesthetized at a speci c time point after the injection to harvest the brain tissue [27]. All of the animal experiments were performed in accordance with the protocol approved by the Institutional Animal Ethics Committee. All efforts were made to minimize the number of animals and their suffering in this experiment.

Cell Cultures and Stimulation
The mouse microglia cell line BV2 and the mouse hippocampal neuron cell line HT22 were cultured with 10% (v/v) fetal bovine serum and 0.1% penicillin-streptomycin in DMEM medium (C119955OOBT, Gibco, China) under an atmosphere of a humidi ed air and 5% CO2 at 37°C. To establish the microglial activation model, 100 ng/mL LPS (L4391, Sigma, USA) was used to stimulate BV2 cells at different time points (0 h, 1 h, 3 h, 6 h, 12 h, and 24 h). To mimic the oxidative damage of neurons in vitro, HT22 cells were exposed to various concentrations of H 2 O 2 (50, 100, 200, 400, 500, and 100 μM) for 24 h. Then, we selected the appropriate concentration of H 2 O 2 (400 μM) to stimulate HT22 cells at different time points.

Quantitative Real-Time Polymerase Chain Reaction Analysis
Total RNA from mouse brain tissue and BV2 was extracted using TRIzol (15596018, Ambion, USA) and quanti ed using NanoDrop spectrophotometer (NanoDrop ONE C, Thermo Fisher Scienti c, USA). Next, the cDNA synthesis was conducted using 5× PrimeScript RT Master Mix (TaKaRa) in line with the manufacturer's instructions. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using 2×QuantiNova SYBR Green (QIAGEN) to analyze gene expression on the Bio-Rad CFX Maestro 1.0 system. The relative mRNA expression was analyzed by using the 2-ΔΔCT method, normalized to GAPDH. Primer pair sequences are shown in Table 1. Table 1 DNA sequences of primers used in PCRs and expected product sizes. The immunohistochemistry of the brain sections was performed according to the DAB Detection Kit's instructions (GK600511, Gene Tech, Shanghai, China). After returning to room temperature, the sections were washed with phosphate-buffered saline (PBS). After blocking endogenous peroxidase for 20 min in the dark, the sections were incubated with anti-PDCD4 (1:100, CST, USA) overnight at 4°C. Next, they were incubated with horseradish peroxidase (HRP) at room temperature for 30 min and stained with diaminobenzidine (DAB), followed by gradient dehydration, drying, and resin sealing. The stained sections were observed using a microscope (ECLIPSE Ni-U, Nikon, Japan) on NIS-Elements F 4.6 system.

Immunofluorescence Staining
The brain tissue sections were stained by standard immunohistochemistry procedures. Brie y, the frozen sections were permeabilized in 0.3% Triton-100 for 20 min and blocked with 5% bovine serum albumin (BSA) for 1 h. Next, the sections were incubated at 4°C overnight with the following antibodies, as Scienti c, USA) at room temperature in the dark for 2 h, and the nuclei were stained with DAPI (Sigma, USA). All the sections were observed using a uorescence microscope (ECLIPSE Ni-E, Nikon, Japan) on the NIS-Elements D 5.11 system.m

Western Blot
The lysis buffer (RIPA effective lysis solution: PMSF=100:1) was used to lyse and release the proteins from the cells and mouse cerebral cortex tissues at 4°C, and the supernatant was collected after centrifugation. The protein concentration was determined using the BCA assay (23225, Thermo Fisher, USA). The proper amount of protein samples was subjected to 10% SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene uoride (PVDF) membrane (Millipore). The membrane was blocked with 5% skim milk for 2 h, followed by an overnight incubation at 4°C with the speci c primary antibodies, including anti-PDCD4 (

Plasmids and Transfection
PDCD4-shRNA and Neo-shNC (negative control) plasmids were purchased from China Gene Pharma. We used Lipofectamine3000 reagent to transiently transfect the plasmids into BV2 cells or HT22 cells in accordance with the manufacturer's instructions. Table 2 The target sequences of PDCD4-shRNA. with serum-free cell culture medium for three times and observed using a uorescence microscope (ECLIPSE Ti-E, Nikon, Japan

Statistical Analysis
Statistical analysis was conducted using GraphPad Prism 8.4.3 software. The data were calculated as mean ± standard deviation (SD). The differences between the groups were detected by t tests and followed by Tukey's post hoc multiple comparison tests. The P values below 0.05 were considered statistically signi cant.

PDCD4 expression was upregulated in the in amed mice cerebral cortex following LPS intraperitoneal injection
To explore the expression and function of PDCD4 in neuroin ammation, we established the classical neuroin ammation model induced by LPS intraperitoneal injection in mice [6]. The mRNA expression levels of proin ammatory genes (TNF-α, IL-1β, IL-6, and inducible nitric oxide synthase (iNOS)) in the cerebral cortex of mice were signi cantly upregulated after LPS stimulation, reaching their peak values around the third day (Fig. 1A). Western blot con rmed the increased levels of proin ammatory protein iNOS and proapoptotic protein cleaved-parp in the in amed cerebral cortex (Fig. 1B). All of these results indicated that LPS successfully induced CNS in ammation and neuronal injury in mice. Next, we used western blot to examine the changes in PDCD4 expression in the brain tissue (Fig. 1B). The PDCD4 level was lower in the cerebral cortex of the saline-injected control group, gradually increased 12 h after the LPS injection, reached the peak on day 1, and then decreased back to the basal level by day 7.
Immunohistochemical staining further clari ed the expression and distribution of PDCD4. Compared with the control group, the number and intensity of PDCD4-positive cells in the cerebral cortex increased signi cantly one day after the LPS administration, which was consistent with the results of the western blot analysis (Fig. 1C) Elevated PDCD4 expression was associated with neuronal apoptosis and microglial activation in the in amed mice brainn To further determine the cell types expressing PDCD4 protein, we employed double immuno uorescence staining and investigated the possible co-localization of PDCD4 with different cell markers, including NeuN (neuronal marker), GFAP (astrocytic marker), and Iba1 (microglial marker) in the mouse brain. As shown in Fig. 2, an elevated PDCD4 expression level was mainly observed in NeuN-positive neurons and Iba1-positive microglia in the mouse cerebral cortex one day after the LPS administration ( Fig. 2A, B). No apparent co-localization was detected between PDCD4 and GFAP (the astrocyte marker) (Fig. 2C). Double immuno uorescence staining demonstrated a clear co-localization of PDCD4 with the in ammatory marker iNOS and cellular apoptotic marker cleaved-caspase3 in the LPS-injected group ( Fig. 2D and E).
These results indicated that PDCD4 may be a hub molecule that simultaneously participates in microglial in ammatory activation and neuronal apoptosis in the LPS-triggered neuroin ammation.
Increased PDCD4 expression in the LPS-activated BV2 microglia cells To further analyze the contribution of PDCD4 to neuroin ammation, we established an in vitro microglial activation model. LPS (100 ng/mL) was used to stimulate BV2 microglia cells. Real-time PCR assay demonstrated a time-dependent increase in proin ammatory marker genes (iNOS, TNF-α, and IL-1β) in BV2 cells following the LPS administration (Fig. 3A). Western blot analysis con rmed the elevation of the in ammatory proteins (iNOS and IL-1β), with the maximum levels achieved at 12 h (Fig. 3B). In line with the in vivo results, microglial PDCD4 expression was signi cantly induced by LPS stimulation, reached the peak at 12 h, and declined thereafter (Fig. 3B). These data further indicated that PDCD4 was closely associated with microglial in ammatory activation.
PDCD4 knockdown suppressed the expression of proin ammatory factors and ROS production in microglia To explore the potential function of PDCD4 in the in ammatory activation of microglia, we silenced PDCD4 expression in BV2 cells by RNA interference (RNAi) and con rmed the interference e ciency by western blot (Fig. 4A). Since sh-PDCD4#4 exerted the best silencing effect, it was employed in the subsequent experiments. Western blot revealed that LPS triggered the proin ammatory iNOS expression in BV2 cells (Fig. 4B). RT-PCR also showed the elevated transcription level of proin ammatory genes, such as TNF-α, IL-1β, and iNOS (Fig. 4C). Compared with the scrambled control shRNA group (sh-Ctrl), knocking down PDCD4 signi cantly attenuated the expression of the LPS-induced proin ammatory factors in BV2 microglia ( Fig. 4B and C). Additionally, silencing PDCD4 largely suppressed the LPSinduced ROS production in BV2 cells (Fig. 4D). Thus, despite con icting data in the previous reports regarding the function of PDCD4 in in ammation, our data demonstrated that PDCD4 facilitates the LPSinduced microglial in ammatory activation.
Silencing PDCD4 inhibited the MAPK-NF-κB proin ammatory signaling pathway in the LPS-treated microglia Next, we investigated the possible downstream signaling pathway through which PDCD4 regulates microglial activation. Western blot analysis showed that LPS sharply induced the phosphorylation of JNK (p-JNK), p38 (p-p38), and ERK (p-ERK) in BV2 cells, indicating the activation of the classical proin ammatory MAPK signaling pathway (Fig. 5A). Moreover, apparent phosphorylation of NF-κB p65 (p-p65) and its nuclear transportation represented the activation of the pivotal proin ammatory transcription factor ( Fig. 5A and B). Compared with the control group, silencing PDCD4 signi cantly inhibited the LPS-induced phosphorylation of all three MAPK branches (JNK, p38, and ERK) in BV2 microglia (Fig. 5A). Moreover, knocking down PDCD4 suppressed the NF-κB p65 phosphorylation and its nuclear accumulation in the LPS-treated cells ( Fig. 5A and B). Our data suggested that PDCD4 promoted the LPS-induced microglial activation by upregulating the classical MAPK-NF-κB proin ammatory signaling pathway.

LPS required the MAPK-NF-κB signaling pathway to elevate PDCD4 expression in microglia
To examine the possible mechanism regulating PDCD4 expression in microglia, we utilized the speci c inhibitors of MAPKs or NF-κB to pre-treat BV2 cells followed before LPS administration (Fig. 6A). Western blot showed that all of the MAPK inhibitors (SB20358 for p38, PD98059 for ERK, and SP600125 for JNK) prevented the LPS-induced iNOS expression (Fig. 6B), proving their anti-in ammatory function.
Interestingly, inhibition of the MAPK signaling pathways signi cantly reversed the LPS-induced elevated PDCD4 protein expression in BV2 cells (Fig. 6B). Moreover, speci c inhibition of NF-κB activity by Bay11-7082 signi cantly attenuated the LPS-triggered expression of PDCD4 and iNOS (Fig. 6B). Taken (Fig. 7B). A timedependent continuous elevation in PDCD4 protein level was observed in this neuronal oxidative damage model (Fig. 7B). More importantly, knocking down PDCD4 by shRNA signi cantly inhibited the expression of the proapoptotic factor BAX in HT22 cells (Fig. 7C). These data indicated that PDCD4 played a pivotal role in promoting reactive oxygen-related neuronal damage and acted as a shared regulatory molecule for both microglial activation and neuronal injury during neuroin ammation.

Discussion
Neuroin ammation, the immune response in the nervous system, normally plays a protective role to promote neuronal repair and regeneration. However, persistent neuroin ammation disturbs the brain homeostasis and causes progressive neuron loss and dysfunction, which is a common pathological mechanism for CNS diseases[1] [7]. Identifying a shared hub molecule which is simultaneously involved in both neuroin ammation and neuronal apoptosis would be of great signi cance for the in-depth understanding of the pathogenesis of neurological diseases. In this study, we demonstrated that the tumor suppressor PDCD4 acts as a common molecule that simultaneously promotes in ammatory activation of microglia and oxidative stress-mediated apoptosis of neurons. Moreover, we proved the existence of a positive feedback loop of PDCD4-MAPK-NF-κB that accelerates the LPS-induced neuroin ammation; indeed, this loop might become a novel therapeutic target for neuroin ammatory diseases.
PDCD4 (programmed cell death protein 4) was originally cloned as a novel apoptosis-inducible gene in the experimental apoptosis models [29]. It is highly conserved during evolution and ubiquitously expressed in normal tissues and organs [8]. Decreased PDCD4 expression was observed in various tumors, such as lung cancer, colorectal cancer, breast cancer, hepatocellular carcinoma, and glioblastoma [19][30] [31]. Some proapoptotic drugs (such as ionomycin), antineoplastic drugs (such as retinoic acid receptor agonist), or cytokines (including IL-12 and TGF-β1) can stimulate PDCD4 expression [18][19] Considering its pivotal functions in the regulation of cellular apoptosis, the potential roles of PDCD4 during neuronal injury have gradually attracted research attention. Chronic restraint stress increased PDCD4 expression in mice hippocampus by decreasing the mTORC1-mediated proteasomes degradation [32]. Elevated PDCD4 expression was observed in the rat spinal cord injury model [33], the chronic sciatic nerve injury (CCI)-induced rat neuropathic pain model [34], the oxygen-glucose deprivation/reoxygenation (OGDR) injury model of rat hippocampal neurons [35], and the ischemia and reperfusion (I/R)-induced neuronal lesion model in mouse retina [25]. The available studies have indicated that PDCD4 may be extensively involved in neuronal damage. However, to the best of our knowledge, the expression and biological function of PDCD4 during LPS-induced neuroin ammation, especially its role in the interaction between neuroin ammation and neuronal injury, have not been systematically studied. In this study, we detected an inducible expression of PDCD4 in the mouse neuroin ammation model by LPS intraperitoneal injection; the expression peaked after one day and regressed to the baseline on day 7. Immunohistochemistry staining revealed the presence of PDCD4 both in injured neurons and activated microglia in the in amed mouse brain ( Fig. 1 and 2). A similar change in PDCD4 expression was observed in the LPS-induced BV2 microglial in ammatory activation model (Fig.  3). These data indicate a potential involvement of PDCD4 during the LPS-induced neuroin ammation and neuronal injury in mouse brain.
Although previous studies implied a close association between PDCD4 and in ammatory diseases, the con icting experimental data made it di cult to determine its exact biological function and mechanisms [8]. There is some evidence that PDCD4 is a proin ammatory molecule. For example, PDCD4 de ciency signi cantly inhibited the JNK phosphorylation, NF-κB activation, and IL-6 production in mice macrophages [23], and PDCD4-de cient mice were protected from LPS-induced death [20]. PDCD4 overexpression exaggerated the apoptosis and proin ammatory cytokine production in toxin-treated intestinal porcine epithelial cells, thereby aggravating cellular damage [36]. However, other results suggested an anti-in ammatory role for PDCD4. For instance, PDCD4 de ciency aggravated the experimental colitis and colitis-associated colorectal carcinoma (CRC) in mice by accelerating the typical proin ammatory IL-6/STAT3 pathway [24]. In the LPS or D-galactosamine (D-GalN)-induced acute liver injury model, compared with the wild-type mice, PDCD4-de cient mice presented more necrotic and apoptotic hepatocytes, in ammatory cells in ltration, in ammatory cytokine (IL-6 and TNF-α) release, and liver internal hemorrhage [37]. In the current study, to identify the biological function of PDCD4 in neuroin ammation, we silenced its expression by shRNA; our results showed that knocking down PDCD4 signi cantly attenuated the expression of proin ammatory cytokines (IL-1β and TNFα), iNOS, and ROS production in microglia (Fig. 4). More importantly, silencing PDCD4 inhibited the phosphorylation and activation of all three substreams of the MAPK signaling (p38, ERK, and JNK) and prevented the phosphorylation and nuclear transportation of NF-κB p65 (Fig. 5). Taken together, our data demonstrated that PDCD4 facilitates the LPS-induced microglial in ammatory activation by promoting the pivotal MAPK/NF-κB signaling pathway.
The mechanisms regulating PDCD4 expression during in ammation remain unclear, but PDCD4 involvement in oncology may provide some clues. Namely, PDCD4 expression is reportedly regulated at different levels [18] [19]. The epigenetic silencing by promoter methylation and some transcription factors such as FOXO and v-myb directly regulate the PDCD4 expression at the transcriptional level [18] [19].
SRSF3 modulates PDCD4 expression by inhibiting the alternative splicing and nuclear export of Pdcd4 mRNA at the post-transcriptional level [8,9]. Some non-coding RNAs, typically miR-21 and lncRNA MALAT1, bind and form sponging to target PDCD4 expression directly at the translational level [18] [19]. Moreover, protein-protein interactions (such as BCL6-PDCD4), phosphorylation, and ubiquitination modi cations reportedly regulate PDCD4 activity, nuclear transfer, and protein degradation at the posttranslational level [18] [19]. In this study, utilizing the speci c chemical inhibitors to pre-treat BV2 cells, we found that speci cally blocking the MAPKs (p38, ERK, or JNK) or NF-κB activation signi cantly suppressed the LPS-induced PDCD4 protein expression in BV2 cells (Fig. 6). Our results indicated that LPS requires the MAPK-NF-κB signaling pathway to stimulate PDCD4 expression, which suggests that a PDCD4-MAPK-NF-κB positive feedback loop is formed to drive the microglial in ammatory activation.
Further investigations are needed to explore the exact underlying mechanisms.
PDCD4 was rst discovered as an apoptosis-related gene [29], and its involvement in cellular apoptosis has been extensively documented. In Huh7 hepatoma cells, PDCD4 overexpression elevated the expression of BAX, the proapoptotic member of Bcl2 protein family; moreover, it promoted the release of cytochrome C from mitochondria and activation of caspases 8, 9, and 3, thereby accelerating cellular apoptosis [38]. PDCD4 speci cally binds to the internal ribosome entry site (IRES) elements of both the XIAP and Bcl-x(L) messenger RNAs, represses the translation of these anti-apoptotic proteins by inhibiting the formation of the 48S translation initiation complex, and thus promotes apoptosis [39]. A recent study implied that miR-183-5p may directly target PDCD4 and RIPK3 to protect neurons from apoptosis and necroptosis during amyotrophic lateral sclerosis (ALS) [40]. MiR-21 derived from the exosomes of mesenchymal stem cells (MSCs) suppressed neuronal death via inhibiting PDCD4 and PTEN signaling pathways [17]. Moreover, miR-340-5p performed neuroprotective function by targeting PDCD4 and then activating the PI3K/Akt pathway in rat hippocampal neurons exposed to OGDR injury [35]. LncRNA-H19 facilitated PDCD4 expression via sponging miR-21 and regulated the ischemia and reperfusion (I/R)induced sterile in ammation and neuronal lesion in mice retinas [25]. Oxidative stress, de ned as excess production of reactive oxygen species (ROS), is a typical feature of neuroin ammation, and directly leads to subsequent neuronal damage [4]. In the present study, a persistent elevation of PDCD4 expression was detected in the H 2 O 2 -induced neuron oxidative damage model. Knocking down PDCD4 signi cantly inhibited the expression of the proapoptotic protein BAX, suggesting the proapoptotic activity of PDCD4 in the oxidative stress-induced neuronal injury (Fig. 7).
In summary, we found that PDCD4 can serve as a hub regulatory molecule that promotes both microglial in ammatory activation and oxidative stress-induced neuronal apoptosis within CNS. Moreover, we proved that PDCD4/MAPKs/NF-κB can form a positive feedback loop to exaggerate the vicious cycle of neuroin ammation and neuronal injury, which might become the potential therapeutic target for the treatment of neuroin ammatory diseases.

Con icts of interest
The authors declare that they have no con ict of interest.

Availability of Data and Material
All data generated and analyzed during this study are included in this published article.    DCFH-DA method revealed that knockdown PDCD4 largely inhibited the production of reactive oxygen species (ROS) in the LPS-stimulated BV2 cells. (*p<0.05, #p<0.05).
(A) BV2 microglial cells were stimulated by LPS for 3 h. Western blot showed that inhibiting PDCD4 expression by shRNA signi cantly reduced the phosphorylation of NF-κB p65 (p-p65) and the phosphorylation of MAPKs (p-JNK, p-ERK, p-P38) in BV2 cells either with or without LPS treatment. (B) The protein levels of NF-κBp65 and p-p65 in the cytoplasm and nucleus of BV2 were also detected by western blot. (*p<0.05, #p<0.05) Figure 6 Page 24/25 The expression of PDCD4 in microglia was positively regulated by MAPK/NF-κB signaling.. (A) BV2 cells were pre-treated with p65 inhibitor Bay11-7082 (20 μM), p38 inhibitor SB203580 (40 μM), ERK inhibitor PD98059, or JNK inhibitor SP600125 for 1 h, followed by LPS-stimulation for 3 h. Western blot analysis con rmed the inhibitory effect on the phosphorylation of p65, p38, ERK, or JNK. (B) The protein expression of iNOS and PDCD4 was analyzed by western blot. The LPS-stimulated PDCD4 expression in BV2 cells was largely reduced by the above inhibitors. (*p<0.05, #p<0.05) Figure 7