Activation of the Omega-3 Fatty Acid Receptor GPR120 Protects against Focal Cerebral Ischemic Injury by Preventing Inflammation and Apoptosis in Mice

Ischemic stroke remains the leading cause of mortality and adult disabilities in both developed and developing countries (1, 2). Middle cerebral artery occlusion (MCAO) is the most common cause of ischemic stroke, resulting in high mortality rates of 40–80% (3). Although thrombolytic therapy has been approved to treat ischemic stroke by the Food and Drug Administration of the United States, it can only be used in acute conditions, and contraindications are present in many patients. Thus, exploring novel therapeutic methods is needed. Inflammatory response and neuronal apoptosis have been indicated to play a role in the pathogenesis of cerebral ischemia (4–7). If so, agents that exhibit anti-inflammatory and antiapoptotic properties will be potentially beneficial for patients with cerebral ischemia.

G protein–coupled receptors (GPCRs) are members of the largest and most diverse family of cell membrane protein receptors, which regulate a variety of physiological processes and are targets of 40% of drugs. Approximately 100 GPCRs remain classified as “orphan” receptors without identified ligands (8). Among the orphan GPCRs, GPR40 and GPR120 were found to be activated by long-chain fatty acids: in particular, omega-3 fatty acids (ω-3 FAs) (9, 10). ω-3 FAs, mainly including docosahexaenoic acid (DHA) and eicosapentaenoic acid, are rich in fish oil and significantly improve the health of patients suffering from chronic inflammatory diseases (11, 12). A recent study has demonstrated that eicosapentaenoic acid can prevent arterial calcification via GPR120 in klotho mutant mice (13). In addition, some small molecules have been found to activate the GPR120 receptor with an excellent activity (14).

Increasing evidence suggests that stimulation of GPR120 by ω-3 FAs exerts anti-inflammatory effects in a broad type of cells, including macrophages, mature adipocytes, Kupffer cells, and hypothalamic neurons (15–17). The coupling of ligand-stimulated GPR120 to β-arrestin2 induces receptor endocytosis with subsequent inactivation of TGF-β–activated kinase 1 (TAK1), which in turn provides an inhibitory mechanism on inflammatory signaling pathways (15). GPR120 activation also suppressed inflammation and metabolic disorder by inhibiting inflammasome activation in macrophages (18). In addition to its anti-inflammatory effect, GPR120 signaling also regulates ERK and PI3K-AKT pathways and leads to an antiapoptotic effect in a murine enteroendocrine cell line (19); another study showed that it counteracted dexamethasone-induced cell apoptosis in mesenchymal stem cells (20).

A recent study showed that fish oil supplements in a daily diet improved the prognosis of patients with ischemic stroke (21). GPR120 in the hypothalamus and anterior pituitary gland has been shown to control their function (16, 22–24). Given the anti-inflammatory and antiapoptotic functions of GPR120, we hypothesize that GPR120 activation may protect against cerebral ischemic injury by modulating the inflammatory and apoptotic response in the ischemic regions. In this study, we tested this hypothesis by using an in vivo MCAO model in mice and an in vitro model of oxygen-glucose deprivation (OGD) in cultured cells.

DHA (D2534) and LY294002 (L9908) were purchased from Sigma-Aldrich (St. Louis, MO). U0126 (S1901) was purchased from Beyotime (Haimen, China).

Eight-week-old male wild-type (WT) C57BL/6 mice were purchased from the Experimental Animal Center of Shandong University. Mice were maintained under specific pathogen-free conditions in individual cages at 21 ± 3°C with 12 h light/dark cycle and free access to food and water. All studies were approved by the Institutional Animal Care and Use Committee of Shandong University.

Four experiments were conducted in our study (Fig. 1). The first experiment examined the temporal profile and cellular localization of GPR120 expression in cortex after focal ischemic brain injury in mice (Fig. 1A). Assessment methods included quantitative real-time PCR (qPCR), Western blot, and double immunofluorescence staining. The second experiment used qPCR and Western blot to measure the expression of GPR120 in microglia and neurons subjected to OGD (Fig. 1B). The third experiment investigated the function and mechanism of GPR120 in microglia and neurons subjected to OGD (Fig. 1C). Endogenous GPR120 in BV2 and PC12 cells was knocked down using small interfering RNA (siRNA). DHA (80 μM) was used to activate GPR120 before the OGD injury. To assess the anti-inflammatory effect of GPR120 in BV2 cells, qPCR, Western blot, ELISA, and coimmunoprecipitation were performed. For evaluation of the role and mechanism of GPR120 in PC12 cells in the condition of OGD, 10 μM U0126 (inhibitor of ERK1/2) or 10 μM LY294002 (inhibitor of PI3K) was used prior to DHA pretreatment (Fig. 1D). Assessments included quantitative qPCR, Western blot, cell counting kit-8 (CCK-8) assay, and flow cytometry. The fourth experiment was performed to evaluate the role of GPR120 in MCAO mice model (Fig. 1E). Adeno-associated virus (AAV)–Gfp or AAV containing short hairpin RNA (shRNA) targeting GPR120 (AAV-shGPR120) was injected into the left cortex of mice using a stereotaxic apparatus; 3 w after the AAV virus injection, DHA (200 mg/kg, once a day) was administered orally for 2 w; mice were subsequently subjected to the model of MCAO. Evaluation was performed as follows: 1) neurologic score, 2) infract volume by TTC staining, 3) qPCR, 4) Western blot, and 5) cytometric bead array (CBA) analysis.

The model of cerebral focal ischemia was established by intraluminal occlusion of the left middle cerebral artery using a silicone rubber-coated nylon monofilament (25). A successful occlusion was confirmed by a marked reduction in the regional cerebral blood flow to <20% of the baseline using a laser Doppler blood flow monitor (Moor, England) with a probe attached to the skull in the area of cerebral cortex supplied by the middle cerebral artery. DHA (200 mg/kg) was orally administered for 2 w before MCAO.

For histological analysis, brain sections were prepared in coronal section (10 μm). Samples were double-labeled with anti-GPR120 Ab and one of the following Abs: anti–glial fibrillary acidic protein (clone GA5, 1:200 dilution; Millipore) for astrocytes, anti-NeuN (clone A60, 1:200 dilution; Millipore) for neurons, or anti-CD11b (1:200 dilution; Millipore) for microglia. After 4°C overnight incubation, secondary Abs including Alexa Fluor 488 donkey anti-rabbit IgG Ab (1:200 dilution; Invitrogen, Gaithersburg, MD) or 594 rabbit anti-Mouse IgG Ab (1:200 dilution; Invitrogen) were added for 2 h at room temperature. After washing three times, the sections were incubated with a solution containing 200 mg/ml of DAPI (Beyotime) for nuclear staining. Immunofluorescence staining was performed to determine the cell distribution of GPR120 in the cortex of the mice.

The microglial cell line BV2 cells from China Center for Type Culture Collection (Wuhan, China) and neuronal cell line PC12 cells from the Cell Center of Chinese Academy of Sciences (Shanghai, China) were cultivated to 80% confluence in DMEM (Hyclone) supplemented with 10% FBS (Gibicol). Primary microglia was isolated and cultured as described previously by Butovsky et al. (26). The cells were subjected to the model of OGD for 2 h, followed by reperfusion for different time points. To establish OGD conditions, cells were incubated with glucose-free Earle’s balanced salt solution for 2 h in a hypoxic chamber that was continuously flushed with 95% N₂ and 5% CO₂ at 37°C to obtain <0.5% O₂. Reoxygenation was obtained by exposing cells at normal culture conditions (37°C, 95% air, 5% CO₂) in regular medium. Cells were pretreated with DHA (80 μM) for 2 h prior to OGD (Fig. 1C). The proper doses of DHA were selected following a dose response experiment (Supplemental Fig. 1A–C). For experiments with ERK1/2 and PI3K inhibitors, PC12 cells were treated with 10 μM U0126 (inhibitors of ERK1/2) or 10 μM LY294002 (inhibitors of PI3K) for 1 h prior to DHA pretreatment (Fig. 1D).

Specific siRNAs against GPR120 and the siRNA control were purchased from Sigma-Aldrich. GPR120 siRNA sequences were as follows: no. 1, sense: 5′-GCGAAAUGACUUGUCUGUUdTdT-3′ and antisense: 5′-AACAGACAAGUCAUUUCGCdTdT-3′, no. 2, sense: 5′-CUUACGCUGAGCUUGGCAUdTdT-3′ and antisense 5′-AUGCCAAGCUCAGCGUAAGdTdT-3′, no. 3, sense: 5′-CUGCUCUUCUACGUGAUGAdTdT-3′ and antisense 5′-UCAUCACGUAGAAGAGCAGdTdT-3′.

β-Arrestin2–specific shRNA and scrambled shRNA were kind gifts from Dr. D. Yin. Transfection was performed using Lipofectamine 3000 (Invitrogen).

Western blot was performed as described previously (27). Briefly, total proteins were isolated from mouse penumbral cortex and cultured cells with RIPA buffer (Beyotime). Protein concentrations were determined using a BCA protein assay reagent kit (Beyotime). The primary Abs used in this study included anti-GPR120 (SAB4501490) and anti-ERK (M5670) from Sigma Aldrich; anti–β-arrestin2 (sc-13140) from Santa Cruz Biotechnology (Dallas, TX); anti–phospho-JNK (9251), anti–phospho-p38 (9216), anti–phospho-ERK (4370S), anti–phospho-AKT (4058S), anti-COX2 (4842), and anti–Bcl-2 (2870) from Cell Signaling Technology (Beverly, MA); anti-p38 (14064-1-AP), anti-JNK (24164-1-AP), anti-AKT (60203-I-Ig), and anti-Bax (50599-2-Ig) from Proteintech (Wuhan, China). The signals were quantified by scanning densitometry and data within a linear range were quantified using Image Quant software (GE Amersham, Piscataway, NJ).

Total RNA was extracted using acid-phenol reagent TRIzol (Sangon Biotech). The cDNA was synthesized using Takara reverse transcriptase (Bio-Rad, Hercules, CA). Real-time PCR analysis was performed using SYBR Green reagents (Invitrogen Applied Biosystems, Carlsbad, CA) and a Bio-Rad iCycler system (Bio-Rad). The mRNA levels were calibrated using β-actin or GAPDH as controls. The primers for target genes in this study were synthesized by Generay Biotechnology (Beijing, China) and are listed in Table I.

Supernatants from BV2 cell were assayed for mouse IL-1β (DRKEWE, DKW12-2012-096) according to manufacturer’s instruction.

IL-6, IL-12p70, MCP-1, TNF-α, and IL-10 in mice cortex were captured by CBA (552364; BD Biosciences) according to the manufacturer’s manual. Cytokine levels were then quantified by flow cytometry (Beckman Coulter).

Lysates of BV2 cells (600 μg protein content) were incubated with 1 mg of anti-GPR120 Ab at 4°C for 6 h. Forty microliters of Protein A+G was then added into immune complexes for shaking at 4°C for overnight. Immune complex elutes were separated by SDS-PAGE, and membranes were incubated with either β-arrestin2 or GPR120 Abs.

Cell viability was assessed by CCK-8 (Beyotime).

The number of apoptotic cells was determined by flow cytometry (Beckman Coulter) using annexin V–FITC/propidium iodide apoptosis detection kit (4A Biotech, Beijing, China).

AAV-Gfp and AAV-shGPR120 were designed and purchased from the Shanghai Hanbio. Eight-week-old male WT C57BL/6 mice were anesthetized and placed on a stereotaxic apparatus for left cortical injection of 4 μl of AAV-Gfp or AAV-shGPR120 suspension (2 × 10¹² μg/ml) at a rate of 0.2 μl/min. The coordinates (Supplemental Fig. 4A) were adjusted according to the previous report (28) as follows: point 1, 0.3 mm anterior to the bregma, 3 mm lateral, 2 mm deep; point 2, 1.9 mm posterior to the bregma, 3 mm lateral, 2 mm deep. Three weeks after the AAV virus injection, DHA (200 mg/kg) was administered orally. After another 2 w, mice were subjected to model of MCAO.

Twenty-four hours after MCAO, a four-tiered neurologic scoring system and infarct volume were used to determine the outcome by a blinded observer as described previously (27, 29).

TUNEL assay was performed to measure apoptotic cells. Using the In-Situ Cell Death Detection Kit, TMR red (catalog no. 18173100; Roche), brain sections were treated following the procedures illustrated by the manufacturer.

Data were expressed as means ± SEM. The significance of the differences in mean values between and within multiple groups was examined by one-way ANOVA followed by LSD post hoc test using the SPSS statistical software. When equal variances were not assumed, a Dunnett multiple comparisons test was used to compare the differences between groups. A value of p < 0.05 was considered statistically significant.

We measured the protein and mRNA expression of GPR120 in a well-established mouse model of cerebral ischemia induced by MCAO (Fig. 1). As shown in Fig. 2A, 2B, GPR120 expression was upregulated in ischemic penumbra of cortex with a peak expression at 24 h and last until 48 h after ischemia. Double immunofluorescence staining showed GPR120 was present in microglia, neurons, and astrocytes. We further quantified the numbers of microglia, neurons, and astrocytes that expressed GPR120. Quantification of the fluorescent images in the cortex showed that 31% of microglia in sham group were immunopositive for GPR120, whereas the number was increased in MCAO group (91%) (Fig. 2C, 2D). In addition to microglia, GPR120 immunoreactivity was much higher in neurons in MCAO group (34%) when compared with neurons in sham group (8%) (Fig. 2E, 2F). In contrast, similar numbers of GPR120-positive astrocytes were observed in sham and MCAO group (Fig. 2G, 2H). Because GPR120 was upregulated in microglia and neurons but not in astrocytes, we then focused on microglia and neurons in our in vitro studies to explore the role and mechanisms of GPR120 in ischemic stroke. Consistent with the findings in mouse MCAO model, GPR120 was also upregulated in microglia BV2 cells (Fig. 3A, 3B) and PC12 cells (Fig. 3C, 3D) when subjected to OGD and reperfusion. The results suggested that GPR120 is involved in the pathological process of cerebral ischemia.

To further investigate the role of GPR120 in the process of cerebral ischemia, siRNA was transfected to reduce the endogenous level of GPR120 in BV2 cells. Among three siRNAs targeting GPR120, siGPR120 no. 1 showed the highest efficiency, with ∼71% reduction in GPR120 mRNA (Fig. 4A) and 75% reduction at protein level (Fig. 4B). Therefore, it was used to knock down GPR120 in the following experiments. In the absence of DHA and GRP120 knockdown, cells subjected to OGD showed increased JNK and p38 phosphorylation (Fig. 4C), increased COX2 protein expression (Fig. 4D), increased mRNA expression of IL-6, TNF-α, and IL-1β (Fig. 4E–G), and increased secretion of IL-1β (Fig. 4H). All these changes was significantly repressed by DHA pretreatment (Fig. 4C–H, comparison between columns no. 3 and columns no. 2). However, the effects of DHA were completely abolished by GPR120 knockdown (Fig. 4C–H, comparison between columns no. 6 and columns no. 3). In addition, the knockdown of GPR120 strongly exacerbated the inflammation induced by OGD in the absence of DHA (Fig. 4C–H, comparison between columns no. 5 and columns no. 2). As Butovsky et al. (26) recently reported that monocytes recruited to the CNS as well as microglia cell lines behaved differently from in vivo microglia, we next cultured primary mouse microglia and tested the expression and the effect of GPR120 on the inflammation induced by OGD. As shown in Supplemental Fig. 2, OGD also significantly induced the expression of GPR120, and GPR120 activation by DHA negatively regulated the inflammation induced by OGD in primary microglia. These are consistent with the phenomenon observed in BV2 cells. The data strongly suggested that GPR120 played an anti-inflammatory role in microglia subjected to OGD.

It is known that β-arrestin2 interacts with activated GPR120 to inhibit the inflammatory response. However, whether this interaction is affected by cerebral ischemic injury remains unknown. As shown in Fig. 5A, the binding between GPR120 and β-arrestin2 was reduced after OGD and reperfusion and reached its lowest level at 12 h. The decreased interaction between GPR120 and β-arrestin2 might explain the persistent inflammation even though the GPR120 expression was significantly increased in microglia subjected to OGD (Fig. 3A, 3B). DHA pretreatment markedly increased the binding between GPR120 and β-arrestin2 (Fig. 5B). Moreover, the knockdown of β-arrestin2 abolished the anti-inflammation function of DHA (Fig. 5C, 5D). These results indicated that activated GPR120 inhibited OGD-induced inflammation by interacting with β-arrestin2.

As GPR120 expression was increased in neurons after cerebral ischemia, it is of interest to investigate the direct role of GPR120 in neurons. The knockdown efficacy of siGPR120 no. 1 in PC12 cells was confirmed by Western blot analysis (Supplemental Fig. 3A). Interestingly, different from microglia, neither GPR120 activation nor GPR120 knockdown affected the inflammatory response induced by OGD in PC12 cells (Supplemental Fig. 3B, 3C). Next, we tested the role of GPR120 in OGD-induced apoptosis. GPR120 knockdown impaired the antiapoptotic effect of DHA, demonstrated as the increased percentage of apoptotic cells and the decreased ratio of Bcl-2/Bax (Fig. 6A, 6B, 6D). Moreover, GPR120 knockdown also reduced the cell viability in PC12 cells subjected to OGD (Fig. 6C). It has been known that the downstream targets of GPR120 that are involved in modulating cell apoptosis include two signaling pathways: Ras-Erk1/2 (30) and PI3K-AKT (19). In PC12 cells, we found that OGD promoted the phosphorylation of ERK and inhibited AKT phosphorylation (Fig. 6E). DHA pretreatment significantly inhibited the ERK phosphorylation and enhanced AKT phosphorylation, whereas these effects were abolished by GPR120 knockdown (Fig. 6E, comparison between columns no. 6 and columns no. 3). To determine whether both or one of these two pathways are involved in apoptosis, we next used inhibitors targeting these two pathways. We found that the ratio of Bcl-2/Bax was reduced by LY294002 (the inhibitors of AKT) treatment, whereas U0126 (the inhibitors of Erk1/2) treatment did not affect DHA-induced the antiapoptotic effect (Fig. 6F). These results indicated that antiapoptotic effect of GPR120 in PC12 cells subjected to OGD was at least partially AKT dependent.

Because our in vitro studies demonstrated that GPR120 played an important role in inhibiting inflammation and apoptosis, we next aimed to confirm the role of GPR120 in MCAO mouse model. AAV harboring shGPR120 was used to inhibit GPR120 expression. Unilateral injections of the shGPR120 virus into brain resulted in ∼50% reduction of GPR120 expression as shown in Supplemental Fig. 4B–E. DHA pretreatment resulted in a significant decrease in neurologic scores and an obviously smaller infarct volume when compared with control mice at 24 h after MCAO (Fig. 7A–C). Similar to in vitro results, GPR120 knockdown also abolished DHA’s effect (Fig. 7A–C). In addition, mice with GPR120 knockdown showed worse neurologic dysfunctions (Fig. 7A) and larger infarction volumes when compared with WT mice at 24 h after MCAO (Fig. 7B, 7C, comparison between column no. 4 and column no. 1).

Next we examined the impact of DHA and GPR120 shRNA on MCAO-induced inflammation in vivo. Similar to the in vitro findings, MCAO induced phosphorylation of JNK and p38 (Fig. 8A), enhanced COX2 protein expression (Fig. 8B), and increased the secretion of various cytokines including MCP-1, IL-6, IL-12, and IL-10 (Fig. 8C–F). Pretreatment with DHA reversed these effects (Fig. 8A–H, comparison between column no. 3 to column no. 2). GPR120 knockdown completely abolished DHA’s effects (Fig. 8A–H, comparison between column no. 6 to column no. 3). Furthermore, MCAO-induced inflammation was remarkably exacerbated in GPR120 knockdown mice (Fig. 8A–H, comparison between column no. 5 to column no. 2). In this study, we also measured the level of TNF-α and IFN-γ in MCAO model. As shown in Fig. 8G, 8H, their levels were not changed significantly. In addition to their roles in inflammation, we also observed that DHA treatment significantly increased the ratio of Bcl-2 and Bax (Fig. 9A) and inhibited the apoptosis (Fig. 9B, 9C). GPR120 knockdown reversed the DHA’s effect (Fig. 9A–C). These results indicated that activation of GPR120 protected against focal cerebral ischemic injury in mice by preventing inflammation and apoptosis in MCAO mice model.

Previous studies have demonstrated that ω-3 FAs, especially DHA, exerted protective effects in ischemic stroke (31–34). GPR40, a receptor for ω-3 FAs, is expressed in a wide variety of cells in the CNS and β cells in the pancreas (35). Recently, studies have showed that GPR40 expression was significantly upregulated in the subgranular zone and promoted hippocampal neurogenesis after transient global brain ischemia (36). GPR120 is another important receptor for ω-3 FAs (15, 18, 37), and its role in cerebral ischemic injury remains unclear. A previous study has demonstrated that the expression level of GPR120 in the brain, including cortex, was much lower than that in peripheral tissues (22). In our study, we found that the expression of GPR120 was remarkably increased in the penumbra of cortex after ischemic injury. Consistent with a previous report (22), the immunofluorescence results demonstrated that GPR120 is predominantly expressed in microglia in the sham group. However, the expression of GPR120 was significantly upregulated in both microglia and neurons after MCAO. These data suggested a potential role of GPR120 in microglia and neurons after ischemic stroke. Thus, it is valuable to investigate the role and underlying mechanisms of ω-3 FA receptor GPR120 in cerebral ischemic stroke.

We further found that GPR120 agonist DHA inhibited OGD-induced inflammatory response in microglia, whereas knockdown of GPR120 exacerbated the inflammation induced by OGD and abolished the anti-inflammatory effects of DHA. These data strongly suggested that GPR120 is a critical regulator of inflammation. It is known that β-arrestins serve as important adaptor molecules to mediate the functions of various GPCRs, as well as other receptor subtypes (38). The C-terminal region of GPR120 contains several putative β-arrestin2 binding motifs, and the anti-inflammation function mediated by GPR120 are β-arrestin2 dependent (15, 38, 39). DHA inhibits both TLR and TNF-α inflammatory signaling by affecting TAK1 activation through GPR120/β-arrestin2/TAB1 interaction (15). GPR120 has also been indicated to scaffold to nucleotide-binding domain and leucine-rich repeat containing protein through β-arrestin2 to prevent the formation of the NLRP3 inflammasome (18). In our study, we demonstrated that the association between GPR120 and β-arrestin2 was reduced after OGD, whereas DHA treatment led to the anti-inflammatory effects by promoting the interaction between GPR120 and β-arrestin2 in microglia. These results suggested that GPR120 activation inhibited the inflammation in ischemic stroke via its interaction with β-arrestin2.

Because GPR120 was markedly upregulated in neurons after ischemic injury, we further explored the potential role of GPR120 in neurons subjected to OGD. Interestingly, in contrast to its anti-inflammatory effect in microglia, GPR120 did not affect the inflammation induced by OGD in neuron PC12 cells. However, pretreatment of DHA inhibited OGD-induced apoptosis and improved cell viability in neurons, which were abolished by knockdown of endogenous GPR120. The ERK pathway has been linked to cell proliferation and growth (40). PI3K translocates to the plasma membrane and subsequently phosphorylates AKT. As a key effector of PI3K, AKT also regulates various cellular responses, including cell proliferation and cell survival (41–43), and can sustain Bcl-2 expression in certain diseases (44, 45). Previous studies have shown that ω-3 FAs promoted the activation of ERK and PI3K/AKT pathways mainly via GPR120, exerting the antiapoptotic effect in certain cell lines (19, 30, 46, 47). Thus, we examined PI3K/AKT and ERK pathways to investigate the antiapoptotic mechanism of GPR120 in brain ischemia. PC12 cells were treated with inhibitors of the above two signaling pathways in the presence of DHA. Only LY294002, a PI3K inhibitor, had the ability to partially inhibit the antiapoptotic effect of DHA, which demonstrated that the PI3K/AKT pathway is involved in the antiapoptotic activity of GPR120 in brain ischemia. Of note, besides the PI3K/AKT pathway, other pathways may also be involved in antiapoptotic effects mediated by GPR120. Further studies are necessary to fully understand the mechanisms underlying the antiapoptotic activity of GPR120 and to investigate how the PI3K/AKT pathway participates in the antiapoptosis action mediated by GPR120 in neurons in brain ischemic injury.

We further confirmed the above observations in mice. Using an in vivo model of MCAO, stimulation of GPR120 by DHA ameliorated the stroke outcomes, whereas GPR120 knockdown aggravated the stroke outcomes. GPR120 activation significantly inhibited MCAO-induced phosphorylation of JNK, p38, as well as the expression of COX2 and promoted AKT phosphorylation, thereby exhibiting anti-inflammatory and antiapoptotic effects. Knockdown of GPR120 significantly suppressed the anti-inflammatory and antiapoptotic effects induced by DHA. These results demonstrated that activation of GPR120 protected against focal cerebral ischemic injury by preventing inflammation and apoptosis. To the best of our knowledge, this is the first study that illustrated the function of GPR120 in cerebral ischemia.

Of note, DHA, a GPR120 agonist, may also play its role via other mechanisms. Recent studies have indicated that metabolic products derived from ω-3 FAs, such as resolvins, protectins, and maresins, may play a role in the resolution of inflammation (48, 49). D-series resolvins, the bioactive metabolite of DHA, displayed potent anti-inflammatory actions in microglia (50), which could provide another possible mechanism for DHA-induced anti-inflammatory response.

In summary, we concluded that the ω-3 FA receptor GPR120 protected against cerebral ischemic injury through dual mechanisms: inhibiting the inflammatory responses via GPR120/β-arrestin2 signaling pathway in microglia and inhibiting the apoptosis via PI3K/AKT signaling pathway in neurons. Our results strongly suggested pharmacological targeting of GPR120 may provide a novel approach for the treatment of patients with ischemic stroke.

We thank Dr. Deling Yin for providing specific shRNA targeting ARRB2. We thank Dr. Wei Wang (Assistant Professor, Department of Hematopathology, The University of Texas MD Anderson Cancer Center) and Dr. Haitao Wen (Assistant Professor, Department of Microbial Infection and Immunity, Ohio State University College of Medicine) for polishing the language. We also thank the Animal Facility of our Institute for mouse care.

The authors have no financial conflicts of interest.