Differences in Proinflammatory Property of Six Subtypes of Peroxiredoxins and Anti-Inflammatory Effect of Ligustilide in Macrophages

Background Peroxiredoxins (Prxs) are proposed to function as damage-associated molecular patterns (DAMPs) and contribute to post-ischemic neuroinflammation and brain injury by activating Toll-like receptor (TLR) 4 at the acute and subacute phases after ischemic stroke. However, there are few studies concerning the inflammatory profiles of six distinct subtypes of Prxs (Prx1–Prx6). Our previous study demonstrated that the protective effect of ligustilide (LIG) against cerebral ischemia was associated with inhibition of neuroinflammatory response and Prx/TLR4 signaling in rats. Herein, the present study explored the inflammatory members of Prxs and the effect of LIG on their inflammatory responses in macrophages. Methodology/Principal Findings The murine RAW264.7 macrophages were treated with each of exogenous recombinant Prxs at a range of 1 to 50 nM for 24 h. The WST-1 test showed that Prx3 exhibited a significant cytotoxicity, whereas the rest five Prxs did not affect cellular viability. The quantitative measurements with spectrometry or ELISA indicated that three subtypes, Prx1, Prx2 and Prx4, increased production of proinflammatory mediators, including nitric oxide (NO) metabolites, tumor necrosis factor-α (TNF-α), and interleukin-6 (IL-6) in a concentration-dependent manner. Immunostaining demonstrated that 20 nM Prx1, Prx2 or Prx4 significantly increased expression of TLR4 and iNOS and nuclear translocation of NF-κB p65. However, Prx5 and Prx6 showed no poinflammatory effect in macrophages. Remarkably, LIG treatment effectively inhibited the inflammatory response induced by Prx1, Prx2 and Prx4. Conclusion Three members of Prxs, Prx1, Prx2 and Prx4, are inflammatory DAMPs that induce TLR4 activation and inflammatory response in macrophages, which is effectively inhibited by LIG. These results suggest that inflammatory Prxs-activated macrophages may provide a novel cellular model for screening the potential inhibitors of DAMPs-associated inflammatory diseases such as stroke. Moreover, selective blocking strategies targeting the inflammatory subtypes of Prxs probably provide promising therapeutic approaches with a prolonged time window for stroke.


Introduction
Neuroinflammation has come to denote central nervous system (CNS)-specific inflammatory responses that are implicated in the progression of cerebral ischemic injury [1]. It is well known that the major inflammatory cells in ischemic brain are the activated macrophages, consisting of resident microglia in brain and infiltrating monocytes from peripheral blood. A number of studies have demonstrated that overactivated macrophages in brain play a pivotal role in post-ischemic neuroinflammatory response by induction of various proinflammatory mediators production [1,2]. Consistent with this notion, various anti-inflammatory neuroprotectants, such as anti-intracellular cell adhesion molecule 1 (ICAM-1) antibody (Enlimomab) and cyclooxygenase inhibitor (Paracetemol), have been evaluated for their therapeutic potential in ischemic stroke [3]. To date, however, these attempts have not yielded the desired outcomes in clinical trials because of either disappointing efficacy or serious side-effects, and the optimal therapy for post-ischemic neuroinflammation and brain injury has not yet been developed. Therefore, there is a huge unmet need for the development of innovative therapeutic strategies targeting post-ischemic neuroinflammatory cascades in ischemic stroke.
Toll-like receptor 4 (TLR4) constitutively expressed in various brain cells and blood-borne immune cells has been demonstrated to be intimately associated with experimental and clinical neuroninflammatory responses and outcomes in ischemic brain [4,5]. Cerebral ischemia stress robustly induces the extracellular release of damage-associated molecular patterns (DAMPs) from necrotic and dying neural cells within ischemic territory, which may function as the endogenous ligands for TLR4 and induce the activation of TLR4/nuclear factor κB (NF-κB) signaling pathway, thereby enhancing both innate and adaptive immunity and promoting post-ischemic neuroinflammation and brain infarction [6,7]. Recent studies have revealed two major protein DAMPs with a differential temporal profile, high mobility group box-1 protein (HMGB-1) and peroxiredoxins (Prxs) [8,9]. In contrast to the early release of HMGB-1 within 6 h after brain ischemia, Prxs are extracellularly released from damaged brain cells predominantly 12-72 h after stroke onset in mice [9]. Remarkably, the extracellular Prxs without antioxidant capacity may act as the specific DAMPs responsible for TLR4 activation and promote post-ischemic neuroinflammation and neuronal damage via release of neurotoxic cytokines during the acute phase, and production of IL-23 and IL-12 during the delayed phase that thereby causes release of neurotoxic IL-17 and IFN-γ and late brain injury [9]. Moreover, systemic administration of a mixture of antibodies specific for Prxs even at 12 h after cerebral ischemia significantly ameliorated both cerebral infarct and motor deficits in mice [9]. It is reported that the mammalian Prxs family has six subtypes (Prx1-Prx6) with brain cell typespecific expression patterns: Prx1 and Prx6 are expressed in astrocytes and oligodendrocytes, while Prx2-Prx5 are mainly expressed in neurons [10][11][12]. To the best of our knowledge, the inflammatory profile of each subtype of Prxs has not been investigated in macrophages in vitro, which has become an obstacle to translational medicine research. Therefore, identifying the specific Prx(s) that activate TLR4 may promote the development of novel aseptic neuroinflammation models in vitro and innovative approaches with a prolonged time window for the treatment of neuroinflammation and brain injury after ischemic stroke.
In addition, Ligustilide (3-butylidene-4,5-dihydrophthalide, LIG, S1 Fig) is a main effective component of various traditional Chinese medicines, such as Radix Angelicae sinensis and Ligusticum chuanxiong [13]. We have reported the potent anti-neuroinflammatory and neuroprotective effects of LIG against various cerebral ischemic insults in vivo, and the strong inhibitory effect of LIG on inflammatory response induced by lipopolysaccharide (LPS), a natural ligand for TLR4, in primary rat cerebral microglia in vitro [14,15]. Moreover, our recent study showed that LIG protected against ischemic brain injury, accompanied by inhibition of extracellular release of Prx5 and Prx6, TLR4/NF-κB signaling and neuroinflammatory response [16]. Therefore, the present study explored the inflammatory members of Prxs and the effect of LIG on their inflammatory response in cultured murine RAW246.7 macrophages. Prx1-Prx6 were prepared as a 10 μM stock solution in sterile water following the manufacturer's recommended procedure, while LIG (purity > 98.5%, S1 Fig), prepared by a well-established procedure in our laboratory (Kuang et al., 2008), was dissolved in DMSO as a 50 mM stock solution. Both Prx and LIG stock solutions were kept at -80°C, and diluted with cell culture medium before use. A final concentration of 0.1% DMSO was used as vehicle for the treatment of the control cultures.
On the day of the experiment, macrophage cells were seeded in 96-well plates and left to acclimate for 24 h. The culture medium was then replaced with fresh medium and the cells were pretreated with or without different concentrations of LIG for 1 h, followed by co-treatment with each Prx at the indicated concentration(s) for another 24 h. Cellular supernatants were harvested and stored at -80°C until further measurement of the levels of proinflammatory mediators, and the remaining cells were used for immunostaining.

Cellular viability assay
The effects of distinct Prxs with or without LIG on the viability of RAW264.7 cells were assessed using the WST-1 assay as described previously [17]. In brief, after treatment with different Prx in the absence or presence of LIG at the indicated concentrations for 24 h, the cells in 96-well plate (1×10 5 cells/well) were washed with serum-free medium and incubated with the WST-1 reagent (10 μL/well) in fresh serum-free medium at 37°C for an additional 2 h. At the end of the incubation period, the intensity of the color formation was detected using a microplate reader at 450 nm (Bio-Rad, Hercules, California, USA). Data are expressed as a percentage of control treatment alone [17].

Nitrite assays
In order to detect the nitric oxide (NO) production in macrophages, 5 × 10 5 cells per well were treated with Prx in the absence or presence of LIG at the indicated concentrations for 24 h, and then 100 μL supernatant of cells was harvested for determination of NO contents by measuring nitrite levels using NO assay kit (JianCheng Bioengineering Institute; NanJing, Jiangsu, China) according to the manufacturer's instruction [15].

Measurement of cytokine levels
For the measurement of tumor necrosis factor α (TNF-α) and interleukin-6 (IL-6) production in macrophages, cells (5 × 10 5 cells per well) were treated with Prx in the absence or presence of LIG at the indicated concentrations for 24 h. And then the culture supernatants were collected for cytokines measurement using ELISA kits (Dakewe Biological Technology; Shenzhen, Gungdong, China) according to the manufacturer's instruction [15].

Immunocytochemistry and Immunofluorescence analysis
Immunocytochemistry analysis was used to detect the expression of iNOS and TLR4 in RAW264.7 cells according to the SABC kit procedure as reported previously [15]. In brief, the cells at a density of 1 × 10 4 cells/mL were treated with Prx in the absence or presence of LIG at the indicated concentrations for 24 h. The cells were then fixed with 4% paraformaldehyde for 30 min at 37°C and washed with PBS for 3 times. The fixed cells were then permeabilized with 0.3% Triton X-100 in PBS for 30 min at 37°C and washed with PBS for 3 times. The permeabilized cells were then rinsed with 3% H 2 O 2 for 30 min to block endogenous peroxide activity and incubated with 5% bovine serum albumin from the SABC kit for 1 h to block nonspecific binding. The cells were then incubated with rabbit polyclonal antibody against iNOS or TLR4 (1:100), respectively, overnight at 4°C, followed by detection with a biotinylated secondary antibody and the SABC kit at 37°C for 1 h, respectively. Immunoreactions were visualized using 3, 3'-diaminobenzidine tetrahydrochloride (DAB; Boster Biological Technology) as a chromagen. The expression of iNOS and TLR4 in cells was observed under microscope (Nikon, Japan), and photomicrographs were acquired using a 40 × objective and processed using SPOT advanced software 4.6 (SPOT Imaging Solutions, Sterling Heights, MI, USA) [15].

Statistical analysis
Data are expressed as the mean ± SEM. of at least three independent experiments. To compare three or more groups, one-way analysis of variance was used, followed by a Newman-Keuls post-hoc test. Statistical analyses were performed with GraphPad Prism software, version 5.01 (GraphPad Software Inc., San Diego, CA, USA). Values of P < 0.05 were considered statistically significant.

Effects of Prxs on macrophage cellular viability
The WST-1 assay was applied to determine the effect of each Prx on the cellular viability of murine RAW264.7 macrophages. As shown in Fig 1, the present results showed that 24 h treatment with Prx 1, 2, 4, 5 or 6 at concentrations ranging from 1 to 50 nM had no significant effect on RAW264.7 cell viability (P > 0.05), whereas Prx 3 showed a concentration-dependent cytotoxicity, leading to a significant decrease in cell viability at doses of 10 nM and greater levels (P < 0.01).

NO production induced by Prxs in macrophages
To evaluate the potential inflammatory effect of each of Prxs in macrophages in vitro, we first examined the effects of Prx1~Prx6 at the indicated non-cytotoxic concentrations on NO  Fig 2A). Moreover, Prx5 and 6 at concentrations of ranging from 5 to 50 nM also had no significant effect on NO production (P > 0.05; Fig 2B). However, Prx1, 2 and 4 concentration-dependently induced NO production, resulting in a significant increase in NO levels at a dose of 20 or 50 nM (P < 0.01; Fig 2B).

Proinflammatory cytokine production induced by Prxs in macrophages
To further explore the proinflammatory activity of each of Prxs in macrophage cells, ELISA was applied to detect the effects of distinct subtype of Prxs (5-50 nM) on proinflammatory cytokine production in vitro. Consistent with the induction of NO production by Prx1, 2 and 4, these three subtypes of Prxs also concentration-dependently enhanced the levels of TNF-α and IL-6 in cell culture medium, whereas Prx5 and 6 did not affect cytokine secretion (Fig 3). Treatment with Prx1 (20 or 50 nM), Prx2 (10-50 nM), or Prx4 (10-50 nM) could significantly induce increases in TNF-α and/or IL-6 levels (P < 0.05 or P < 0.01). Effect of LIG on cellular viability in Prx-stimulated macrophages To avoid the potential cytotoxic effect of LIG, the effect of LIG on cellular viability was determined in the presence of Prx in RAW264.7 cells. The results showed that LIG at concentrations ranging from 5 to 20 μM had no significant effect on the cellular viability of macrophages cotreated with 20 nM Prx1, Prx2, or Prx4, respectively (P > 0.05; Fig 4). These data showed that the results from the corresponding quantitative analysis were not influenced by the cell counts.

Effect of LIG on proinflammatory mediator production and iNOS expression induced by Prxs in macrophages
To study whether LIG could directly inhibit inflammatory response induced by Prx1, 2 and 4 in macrophages in vitro, we first investigated the concentration-effect relationship of LIG on NO production in RAW264.7 cells. As shown in Fig 5A, pretreatment of macrophages with LIG at 5, 10, and 20 μM for 1 h prior to Prx (20 nM) stimulation significantly decreased NO production in a concentration-dependent manner compared with the corresponding Prx group, respectively, with maximal inhibitory efficacy at 20 μM LIG (P < 0.01). To explore whether proinflammatory cytokines are amenable to modulation by LIG in Prx-stimulated macrophages, the levels of TNF-α and IL-6 in cell culture medium were measured with ELISA. Consistent with the inhibitory effect of LIG on NO production in Prx-stimulated RAW264.7 cells, LIG at a concentration of 20 μM tended to exhibit significant inhibition on the release of proinflammatory cytokines (P < 0.05, Fig 5B and 5C).
Moreover, to further confirm the inhibitory effect of LIG on NO production in Prx-treated macrophages, the immunoreactivity of inflammatory enzyme iNOS was detected with immunocytochemistry 24 h after treatment with 20 nM Prx1, 2, or 4, respectively. Fig 5D showed that each of three subtypes of Prxs potently increased iNOS expression, which was significantly inhibited by pretreatment with 20 μM LIG.

Inhibitory effect of LIG on Prxs-induced TLR4/NF-κB activation in macrophages
To investigate the potential mechanism that underlies the anti-inflammatory effect of LIG in Prx-treated macrophages, we further explored the effect of LIG on the TLR4/NF-κB signaling pathway. The TLR4 expression and transcription factor NF-κB activation were examined with immunostaining 24 h after treatment with 20 nM Prx1, 2, or 4, respectively. The present results showed that each of three subtypes of Prxs strongly enhanced the immunoreactivity of TLR4 and nuclear translocation of NF-κB p65 subunit that was predominantly located in the cellular cytoplasm in the control group (Figs 6 and 7). Notably, LIG treatment at a concentration of 20 μM greatly inhibited TLR4 expression and NF-κB activation compared with the corresponding Prxtreated group, suggesting that the inhibitory effect of LIG against Prx-induced inflammation is probably associated with downregulation of TLR4/NF-κB signaling activation in macrophages.

Discussion
Accumulated evidence suggests that immunomodulation associated with DAMPs/TLRs signaling plays a pivotal role in the destructive neuroinflammatory response accompanying ischemic  Identification of Inflammatory Peroxiredoxins and Anti-Inflammatory Effect of Ligustilide stroke [6,7,19]. TLR4 is the most widely investigated pattern recognition receptor in ischemic stroke. Studies of TLR4 knockout mice showed smaller infarct sizes and improved neurological deficits after cerebral ischemia [4]. Clinical studies found that TLR4 polymorphisms were associated with ischemic stroke outcome, and TLR4 expression was independently associated with lesion volume and correlated with the intense inflammatory response after ischemic stroke [5]. As a member of highly conserved innate immunity receptors, TLR4 recognizes not only highly preserved structures in infectious pathogens, called pathogen-associated molecular patterns (PAMPs, e.g. LPS), but also DAMPs that are released from aseptic damaged cells. Recent studies have suggested the existence of various types of DAMPs, including nucleic acid, modified lipids, and proteins etc. [19]. Of the known DAMPs in ischemic brain, Prxs have the prominent translational potential at least due to three major reasons: (1) Prxs are highly expressed in brain than in other tissues, which is further upregulated by an increase in oxidative stress after ischemic brain; (2) Prxs are extracellularly released at 12-72 h after stroke onset, which may provide a wide therapeutic window available for the much-needed late phase treatment of ischemic brain injury; (3) the accumulated evidence gradually reveal the potential molecular mechanisms of intracellular and extracellular Prxs [9,[19][20][21][22][23].
The previous studies suggest that Prxs, a redox-active intracellular enzyme family, may function as neuroprotective antioxidants within cells in ischemic brain. Several Prxs (Prx2, 3 and 5) have been reported to protect against ischemic cerebral injury through scavenging peroxides and peroxynitrites that is increased by ischemic stress [24][25][26][27][28]. Moreover, intracellular Prx1 could govern the progression of neuroinflammation by suppressing microglial activation in a ROS/p38MAPK-dependent manner, and Prx1 deficiency increased NF-κB mediated-iNOS induction as proinflammatory mediators in microglia [23], suggesting that endogenous Prx1 probably exhibits neuroprotective effect through both anti-oxidative and anti-neuroinflammatory properties. On the other hand, Prxs may act as neurotoxic DAMPs outside cells in ischemic brain. Studies showed extracellular Prxs released from dead or dying brain cells become potent proinflammatory signals to TLR2/4 located predominantly on residential and  Identification of Inflammatory Peroxiredoxins and Anti-Inflammatory Effect of Ligustilide infiltrating immune cells in ischemic territory, leading to the destructive immune responses [9,14]. Notably, the recent study showed that an increase in Prx2 release could modify the redox status of cell surface receptors and enable induction of inflammatory responses [22]. Collectively, the seemingly paradoxical findings that intracellular and extracellular Prxs have beneficial or harmful effect in progression of ischemic stroke put forward the necessity of further exploring the inflammatory profiles of exogenous Prx subtypes.
Therefore, in the current study, we first measured cell viability by WST-1 assay after treatment with each recombinant mouse Prxs at a range of concentrations from 1 to 50 nM in murine RAW264.7 cells. Strikingly, we observed that Prx3 induced a significant and concentration-dependent cytotoxicity, and 10 nM or more Prx3 caused a great decrease in cellular viability. On the contrary, the rest five members of Prxs at the tested concentrations had no effect on cell viability. Furthermore, we investigated whether each of Prxs at the non-cytotoxic concentrations could induce inflammatory response in macrophages. We found the distinct inflammatory properties of Prxs subtypes in macrophages. The results show that treatment with recombinant Prx3 (1 nM), Prx5 and Prx6 (1 to 50 nM) had no significant effects on the production of proinflammatory mediators (NO, TNF-α or IL-6), iNOS and TLR4 expression, and NF-κB activation, whereas Prx1, 2 and 4 concentration-dependently and significantly induced production of proinflammatory mediators and activation of TLR4/NF-κB signaling in RAW264.7 cells (Figs 2, 3, 5-7 and S2). Other studies have indicated that Prx3, a mitochondrion-specific member of Prxs, plays an indispensable role in apoptotic regulation via direct mitochondrial pathway and interaction with other gene products [29]. The current study shows for the first time that exogenous Prx3 causes the strongest cytotoxicity of Prxs family in macrophages. Further research is needed to explore the cytotoxicity mechanism of Prx3 in future. Altogether, the present results provide the direct and visible evidence to identify the specific members of Prxs with proinflammatory profile in macrophages. Our findings suggest that only Prx1, Prx2 and Prx4 outside cells are potent inflammatory subtypes of Prxs that may directly activate TLR4/NF-κB signaling pathway and trigger inflammatory responses in macrophages.
LPS is a well-characterized PAMP and LPS-stimulated macrophages have widely used as an acute inflammatory model in vitro for the evaluation of potential anti-inflammatory agents [15,30]. Similarly, it is reasonable propose that activated macrophages induced by inflammatory DAMPs, such as Prx1, 2 and 4, probably provide a novel cellular model suitable for screening novel anti-inflammatory neuroprotants that inhibit Prxs/TLR4 signaling pathway in ischemic stroke. Natural phthalide compound LIG has been reported to exert potent anti-neuroinflammatory and neuroprotective effects against various cerebral ischemic insults. We and other group have reported that LIG showed a concentration-dependent anti-inflammatory effect in LPS-activated primary microglia and RAW264.7 macrophages in vitro [15,30]. An in vivo study showed that intravenous LIG injection (20-80 mg/kg) protected rabbits against LPS-induced endotoxic shock, with decreased levels of TNF-α, IL-1β, and nitric oxide and partially restored function of the injured heart, liver, lungs, and kidneys [31]. Moreover, our recent study showed that LIG protected against ischemic brain injury, accompanied by inhibition of the Prx6/TLR4 signaling pathway and neuroinflammatory response [16]. To further investigate whether the neuroprotection of LIG against ischemic brain injury is associated with its direct neutralization on the proinflammatory effect of Prxs, the present study examined the effect of LIG on the inflammatory responses induced by exogenous Prx1, 2, or 4 in RAW264.7 macrophages, respectively. Consistent with the inhibitory effect of LIG against LPS-induced inflammatory response in macrophages [15,30], treatment with LIG (5, 10 or 20 μM) could concentration-dependently and significantly decrease NO release from the macrophages stimulated by Prx1, 2, or 4, respectively. Moreover, 20 μM LIG could significantly inhibited the proinflammatory cytokine production, TLR4 and iNOS expression, and NF-κB activation in the Prx-stimulated cells. Our results suggest that LIG is a small molecule inhibitor targeting TLR4/Prx6 signaling and subsequent inflammatory response, which contributes to its potent protection against neuroinflammation and ischemic brain [16]. The potential effect of LIG on binding of Prx on TLR4 needs to be investigated in future.
Collectively, our study shows that three members of exogenous recombinant Prxs, Prx1, Prx2 and Prx4, are potent inflammatory DAMPs that induce TLR4/NF-κB pathway activation in macrophages, which could be effectively inhibited by LIG treatment. These results suggest that the inflammatory Prx-activated macrophages provide a novel screening model for small molecule inhibitors on aseptic neuroinflammation in ischemic brain. Moreover, the selective blocking strategies targeting the inflammatory Prx/TLR signaling pathway probably provide promising approaches with a wide therapeutic window for ischemic stroke treatment.
Supporting Information S1 Fig. Chemical structure and high-performance liquid chromatography chromatogram of ligustilide (LIG). The HPLC analysis was performed using a Waters ACQUITY UHPLC system equipped with a photodiode array detector (Milford, MA, USA). LIG was analyzed using a Waters BEH C18 column (1.7 μm, 50 mm × 2.1 mm inner diameter). The isocratic mobile phase consisted of acetonitrile-water (55:45, v/v) at a flow rate of 0.5 ml/min. The column temperature was 30°C, and the detection wavelength was set at 280 nm for acquiring the chromatograms. The purity of LIG was found to be > 98.5%, based on the percentage of total peak area.