Lysophosphatidic Acid Receptor 5 (LPA5) Knockout Ameliorates the Neuroinflammatory Response In Vivo and Modifies the Inflammatory and Metabolic Landscape of Primary Microglia In Vitro

Systemic inflammation induces alterations in the finely tuned micromilieu of the brain that is continuously monitored by microglia. In the CNS, these changes include increased synthesis of the bioactive lipid lysophosphatidic acid (LPA), a ligand for the six members of the LPA receptor family (LPA1-6). In mouse and human microglia, LPA5 belongs to a set of receptors that cooperatively detect danger signals in the brain. Engagement of LPA5 by LPA polarizes microglia toward a pro-inflammatory phenotype. Therefore, we studied the consequences of global LPA5 knockout (-/-) on neuroinflammatory parameters in a mouse endotoxemia model and in primary microglia exposed to LPA in vitro. A single endotoxin injection (5 mg/kg body weight) resulted in lower circulating concentrations of TNFα and IL-1β and significantly reduced gene expression of IL-6 and CXCL2 in the brain of LPS-injected LPA5-/- mice. LPA5 deficiency improved sickness behavior and energy deficits produced by low-dose (1.4 mg LPS/kg body weight) chronic LPS treatment. LPA5-/- microglia secreted lower concentrations of pro-inflammatory cyto-/chemokines in response to LPA and showed higher maximal mitochondrial respiration under basal and LPA-activated conditions, further accompanied by lower lactate release, decreased NADPH and GSH synthesis, and inhibited NO production. Collectively, our data suggest that LPA5 promotes neuroinflammation by transmiting pro-inflammatory signals during endotoxemia through microglial activation induced by LPA.


Introduction
Normal brain function depends on a balanced set of diverse lipid species, which have structural as well as signaling roles. Within the body, the brain has the second highest lipid content and diversity after adipose tissue [1]. High-resolution mass spectrometry-based lipidomic approaches revealed a distinct lipid signature of the brain that is substantially different from non-neural tissues, such as muscle and kidney [2]. Bioactive signaling lipids transmit their function by binding to specific receptors [3][4][5] and activating the downstream pathways [5,6]. In the periphery and the central nervous system (CNS), lysophosphatidic acid (LPA) species represent an important subclass of bioactive signaling lipids [6]. Signaling is determined by LPA receptor-triggered pathways in which six LPA receptors (LPA 1-6 ) couple to distinct classes of heterotrimeric G proteins, leading to a variety of cellular activities [7,8]. In the CNS, the LPA/LPAR axis plays a pivotal role during neurodevelopment and homeostasis by modulating neurotransmission, synaptic plasticity, enzyme function, gene expression, and (neuro)inflammatory reactions [6].
Immune activation in the periphery affects the function of the CNS and is associated with an increased risk of developing neuropsychiatric and neurodegenerative diseases [9]. Microglia play an important role in this scenario since there is growing evidence that microglia activation and neuroinflammation can be induced by systemic inflammatory events [10,11]. This is ascribed to their ability to continuously monitor changes in the brain microenvironment through a specifically enriched set of surface receptors termed the "microglia sensome" [12]. Early studies had identified LPA receptor gene expression on rodent microglia [13], while recently validated mouse and human microglia RNAseq data sets define the microglia core sensome as a key set of 57 genes conserved between the mouse and human sensome [14] including LPA 5 in mouse and human microglia (along with additional expression of LPA 6 in humans) associated with the core sensome [14].
LPA 5 [15] and LPA 6 [16,17] are highly expressed in BV-2 cells and primary murine microglia [18,19], although LPA receptor expression in the brain is subject to developmental regulation [6,20] and depends on the genetic background of the animal model used [21]. LPA 5 is considered to be a driving factor for acute and chronic ischemic injuries in the mouse transient middle cerebral artery occlusion (tMCAO) stroke model, and the LPA 5 antagonist TCLPA5 has been shown to confer acute and long-term protection in this injury model [22,23]. These findings were ascribed to a pathogenic role of LPA 5 closely associated with microglia activation in injured brains [22], likely via RAGE-dependent pathways [24]. Targeted deletion of LPA 5 identified a novel role for this receptor in mechanically or chemically induced murine neuropathic pain models [25,26]. Consistently, the LPA 5 inhibitor AS2717638 ameliorated mechanical allodynia and thermal hyperalgesia in rodent models of neuropathic pain [27]. In microglia, LPA 5 induces a polarization program toward a neurotoxic phenotype, since genetic (siRNA) [22] and pharmacological (TCLPA5, Cpd3, AS2717638) approaches revealed that inhibition of this receptor attenuates the neuroinflammatory output of the cells [28,29].
LPA signaling plays an important role in microglia polarization [18,30], the neuroinflammatory response [31], and a shift in cellular metabolic pathways [32,33]. To monitor the role of LPA 5 in endotoxemia in the absence of pharmacological antagonists (and potentially associated off-target effects), we used a global LPA 5 knockout mouse model. Using this mouse model, we studied the peripheral and central inflammatory response in an endotoxemia model utilizing wildtype (wt) and LPA 5 -deficient ( -/-) mice. In vitro, we investigated the impact of LPA on cyto-/chemokine secretion and basic immunometabolic parameters (mitochondrial function, lactate, NADPH, GSH, and nitric oxide synthesis) in microglia isolated from newborn wt and LPA 5 -/mice.

Animals
Mice were housed and bred in a clean environment and a 12 h/12 h light-dark cycle with chow diet and water access ad libitum. All animal experiments were approved by the Austrian Federal Ministry of Education, Science and Research (BMWF-66.010/0067-V/3b/2018 and 2020_0.547.884). All measures were taken to minimize animal suffering and distress.
The genotyping protocols for LPA 5 -/genetically modified mice have been described previously [25]. Briefly, homozygous LPA 5 -null mutant mice were generated by targeted deletion of the LPA 5 gene to eliminate and replace most of the LPA 5 coding region in C57BL/6J mice. Heterozygous mice were mated to each other to obtain wildtype (wt; +/+) and null mutant offspring ( -/-). Genotypes were confirmed by PCR genotyping using the following primers: GFP Int Rev, 5-GTGGTGCAGATGAACTTCAGG-3; 92GTFor, 5-CAGAGTCTGTATTGCCACCAG-3; and Male wt and LPA 5 -/mice aged 12-16 weeks weighing 20-30 g were injected i.p. with PBS or LPS (5 mg/kg body weight in PBS). Twenty-four hours post injection, the animals were euthanized, perfused, and brains were collected in QIAzol Lysis Reagent (Qiagen, Hilden, Germany) for RNA isolation.
Blood (200-300 µL) was isolated by cardiac puncture. The tubes containing the blood samples were kept at room temperature for 1 h and then centrifuged at 5000× g for 10 min. The clear supernatant (serum) was collected, diluted 1:10 and used for ELISA analyses.

Indirect Calorimetry (Metabolic Cage Monitoring)
LPA 5 -/and wt mice were individually housed in PhenoMaster cages (TSE Systems, Bad Homburg, Germany). We determined energy intake and expenditure as well as ambulatory movements in the mice over 5 days before and for 4 days during chronic LPS i.p. application (1.4 mg/kg body weight in PBS every 24 h). We chose this chronic low-dose regimen to observe animal behavior over an extended period of time, which is not possible in the acute high-dose LPS model (5 mg/kg body weight) since animal suffering increases at time points > 1 d post LPS application [34]. A comparable chronic treatment regimen (four daily LPS injections of 1 mg/kg) was shown to induce global microglia activation in C57BL/6 mice [35]. Oxygen consumption (VO 2 ) and carbon dioxide production (VCO 2 ) were simultaneously measured every 15 min by indirect gas calorimetry.
Since food consumption from the built-in food containers (which are localized above the light beams that monitor locomotion) may be too challenging for some animals after the LPS injections, we decided to place food pellets directly in the cage and measure food consumption manually. During the adaption phase, cumulative food consumption was divided by the days of the adaption phase to measure average food consumption. During the LPS treatment, food consumption was measured manually every day just before the LPS injection.

RT-qPCR Analysis
We isolated total RNA from the brain with the RNeasy Lipid Tissue Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's protocol. Total RNA was quantitated using NanoDrop (Thermo Fisher Scientific, Waltham, MA, USA) and reverse-transcribed using the SuperScript ® III reverse transcription kit (Invitrogen, Waltham, MA, USA). Quantitative real-time PCR (qPCR) using the QuantifastTM SYBR ® Green PCR kit (QIAGEN, Hilden, Germany) was performed on the Applied Biosystems 7900HT fast real-time PCR system. Gene expression was normalized to the expression of hypoxanthineguanine phosphoribosyltransferase (HPRT) as housekeeping gene. Expression profiles and associated statistical parameters were calculated using the 2 −∆∆CT method. Primer sequences are listed in Table 1.

Primary Microglia Cultures
To isolate primary murine microglia from the cortices of newborn (P0-P4) wt and LPA 5 -/mice, we dissected brain cortices from the entire brain, and the meninges were removed and cut into small pieces with scissors. Tissues were trypsinized (0.1% trypsin, 25 min, 37 • C, 5% CO 2 ), and centrifuged at 1700 rpm for 7 min. The supernatant was aspirated and the pellet was suspended in DMEM. This cell suspension was cultured in poly-D-lysine (PDL; 5 µg/mL)-coated 75 cm 2 tissue culture flasks (4 brains per flask) in DMEM supplemented with 15% FCS, 1% penicillin, 1% streptomycin, and 1% L-glutamine. After cultivation of the cells for another 10 to 14 days, we removed microglia from the mixed glia cell cultures by vigorously tapping the culture flasks on the bench top. Microglia were then seeded onto PDL-coated cell culture plates for further use.

LPA Treatment
The aqueous LPA (1-Oleyl-2-hydroxy-sn-glycero-3-phosphate; Sigma-Aldrich, St. Louis, MO, USA; Cat. L7260) stock solution (5 mM) was aliquoted and stored at −70 • C. Fresh aliquots were used for the experiments. Primary microglia were plated out in 12-or 24-well plates and allowed to adhere for 2 days. Before treatments, cells were incubated in serum-free DMEM overnight. The following day, fresh, serum-free medium was added, followed by the addition of LPA.

Seahorse XF Analyzer Respiratory Assay
Cells (6 × 10 4 cells per well) were seeded in Seahorse XFe96 FluxPaks for metabolic analysis with an extracellular flux analyzer XF96 (Seahorse, Agilent, Santa Clara, CA, USA). The sensor cartridge was hydrated in a 37 • C non-CO 2 incubator one day before the experiment. Cells were serum-starved overnight, treated with LPA at the indicated concentrations for the given time periods, washed, and incubated with the appropriate assay medium for 1 h in a 37 • C non-CO 2 incubator according to the manufacturer's instructions. Cellular oxygen consumption rate (OCR) was determined using the XF Cell Mito Stress Test (Agilent). Optimized stressor concentrations were added as follows: 2 µM oligomycin (complex V inhibitor), 1.75 µM cyanide p-trifluoromethoxy-phenylhydrazone (FCCP; proton gradient disruption), and 2.5 µM antimycin A (inhibitor of complex I and III). OCR was normalized to protein concentrations, and data from 3 independent experiments are shown.

Lactate Measurement
Lactate content in the supernatant was measured using the EnzyChrom™ Glycolysis Assay Kit (ENZO Life Sciences, Lausen, Switzerland) according to the manufacturer's protocol. Briefly, primary microglia (96-well plate, 6 × 10 4 cells per well) were allowed to adhere, then incubated in serum-free medium and treated with LPA for indicated time periods. At the end of the treatment, the supernatant was collected and treated with the enzyme mix under brief shaking. Optical density was measured at 565 nm to quantify the lactate content.

NADPH/NADP Assay
Nicotinamide nucleotides were assayed using the NADP/NADPH assay kit (Abcam, Cambridge, UK) according to the manufacturer's instructions. Primary microglia were seeded onto PDL-coated 12-well plates at a density of 5 × 10 5 per well and serum-starved overnight prior to the experiments. Cells were treated with the given concentrations of LPA for the indicated time periods, after which the medium was removed, cells were washed twice with ice-cold PBS, and NADP/NADPH were extracted with the extraction buffer. The samples were deproteinized using 10 kDa Spin Columns (Abcam, Cambridge, UK) before performing the assay. An aliquot of the sample was used to measure total NADPt (NADP and NADPH). Another part of the sample was heated at 60 • C for 30 min to decompose NADP for NADPH measurement. Ten µL of the sample were mixed with NADP cycling mix to convert NADP to NADPH. Thereafter, 10 µL of NADPH developer were added into each well, mixed, and incubated at room temperature for 1-4 h. Multiple readings were taken at OD 450 nm. The ratio of NADPH/NADP was calculated as follows: NADPH/NADP ratio = NADPH/(NADPt − NADPH).

Glutathione Assay
Primary microglia (5 × 10 4 per well) were seeded overnight in clear-bottom black 96-well plates to allow cells to adhere, incubated in serum-free medium, and treated with LPA (1 or 5 µM) for the indicated time periods. To measure intracellular glutathione content, cells were incubated with the GSH-Glo™ reagent (GSH-Glo™ Glutathione Assay Kit; Promega Corporation, Madison, WI, USA) for 30 min. After addition of the luciferin detection reagent, luminescence was measured to quantify the glutathione content according to the manufacturer's protocol.

Determination of Nitric Oxide (NO)
Total nitrate content was measured in the supernatant of cells incubated with the indicated compounds in serum-free medium using the total nitric oxide assay kit (ENZO Life Sciences, Lausen, Switzerland) according to the manufacturer's protocol. This assay detects a colored azo-dye product after the enzymatic conversion of nitrate to nitrite by nitrate reductase, followed by the Griess reaction. Nitrite concentrations in the samples were calculated by a standard curve in the range of 0-100 µM using nitrate as standard.

Statistical Analysis
Data are presented as mean ± SEM of at least 3 independent experiments (performed in triplicate), unless otherwise stated. Statistical analyses were performed using GraphPad Prism6 software. Significance was determined by unpaired Student's t-test (two groups) or two-way ANOVA followed by Bonferroni correction (>two groups). Values of p < 0.05 were considered significant.

Endotoxemia Is Reduced in LPA 5 -/-Mice
In vivo, we investigated a potentially protective role of global LPA 5 deficiency in an LPS-induced endotoxemia mouse model. Adopting a previously published protocol [36], wt and LPA 5 -/mice were injected i.p. with LPS (5 mg/kg body weight) and sacrificed 24 h later. Cyto-/chemokine concentrations and gene expression were determined in serum and in brain homogenates, respectively. Serum ELISA measurements revealed that LPS administration significantly increased TNFα, IL6, and IL-1β concentrations ( Figure 1A-C). In LPA 5 -/mice, this increase was significantly lower for TNFα and IL-1β, while IL-6 concentrations remained unaffected ( Figure 1A-C).
wt and LPA5 -/-mice were injected i.p. with LPS (5 mg/kg body weight) and sacrificed 24 h later. Cyto-/chemokine concentrations and gene expression were determined in serum and in brain homogenates, respectively. Serum ELISA measurements revealed that LPS administration significantly increased TNFα, IL6, and IL-1β concentrations ( Figure 1A-C). In LPA5 -/-mice, this increase was significantly lower for TNFα and IL-1β, while IL-6 concentrations remained unaffected ( Figure 1A-C). Wt and LPA5 -/-mice were injected i.p. with PBS (n = 4) or LPS (5 mg/kg; n = 6). After 24 h, the animals were sacrificed, blood was collected and serum was isolated. The concentrations of (A) TNFα, (B) IL-6, and (C) IL-1β were quantified using ELISAs. Values are expressed as mean ± SEM, * p < 0.05, **** p < 0.0001 compared to wt PBS control; &&& p < 0.001 compared to LPA5 -/-PBS control; # p < 0.05, ## p < 0.01 compared to LPS-treated wt mice; two-way ANOVA with Bonferroni correction). Following brain RNA isolation, gene expression of TNFα, IL-6, IL-1β, CXCL10, CXCL2, CCL5, iNOS, and Arg1 was analyzed by qPCR. In both genotypes, cyto-and chemokine expression levels were significantly enhanced in response to LPS when compared to vehicle (PBS)-injected animals ( Figure 2A-F). iNOS (M1 marker) and Arg-1 (M2 marker) were upregulated in response to LPS; however, in wt animals, this effect was statistically not significant ( Figure 2G,H). In LPA5 -/-mice, the increase of cyto-and chemokine mRNA expression was consistently lower as compared to wt (Figure 2A-F). These observations reached statistical significance for IL-6 and CXCL2. Following brain RNA isolation, gene expression of TNFα, IL-6, IL-1β, CXCL10, CXCL2, CCL5, iNOS, and Arg1 was analyzed by qPCR. In both genotypes, cyto-and chemokine expression levels were significantly enhanced in response to LPS when compared to vehicle (PBS)-injected animals (Figure 2A-F). iNOS (M1 marker) and Arg-1 (M2 marker) were upregulated in response to LPS; however, in wt animals, this effect was statistically not significant ( Figure 2G,H). In LPA 5 -/mice, the increase of cyto-and chemokine mRNA expression was consistently lower as compared to wt (Figure 2A-F). These observations reached statistical significance for IL-6 and CXCL2.

Improved Metabolic Performance of LPA 5 -/-Mice after Short-Term, Low-Dose LPS Treatment
To monitor potential differences in animal behavior and energy expenditure, wt and LPA 5 -/mice were housed in metabolic cages and subjected to a low-dose, chronic LPS regimen (1.4 mg LPS/kg body weight every 24 h for 4 d). During the preceding 5 d adaptation phase, wt or LPA 5 -/mice with ad libitum access to food and water were individually kept in metabolic cages. Under these basal conditions, feeding behavior, locomotion during the night cycle, respiratory exchange ratio (RER) and energy expenditure (EE) during the night cycle of wt mice were not significantly different from LPA 5 -/animals ( Figure S1A-H). Locomotion and energy expenditure during the day cycle was slightly lower (p < 0.05) for LPA 5 -/mice ( Figure S1D,H). In comparison to basal conditions, animals of both genotypes exhibited classical signs of sickness behavior 24 h after the first LPS application ( Figure 3). This is reflected by decreased water and food intake ( Figure 3A Figure 3G,H). However, in contrast to wt mice, several parameters of sickness behavior were significantly less pronounced in LPA 5 -/animals during the early acute inflammatory phase 24 h after the first LPS dose. This was reflected by higher water (night cycle) and food consumption, locomotor activity, as well as RER (night cycle) in comparison to LPS-injected wt animals ( Figure 3A,B,D,F). The corresponding data for the entire 96 h monitoring period are shown in Figure S2 and indicate that the animals partially recovered from LPS-induced sickness behavior (despite the consecutive injection). Most of the metabolic parameters (except total food intake and RER during the night cycle; Figure S2A,F) were comparable between wt and LPA 5 -/mice.

LPA 5 Regulates LPA-Induced Secretion of Cyto-/Chemokines in Primary Microglia
In response to acute or chronic endotoxemia, LPA levels and gene expression of autotaxin (ATX) and several LPA receptors (including LPA 5 ) are upregulated in mouse brain homogenates [31]. Since LPA provides an induction signal for the transition of microglia toward a pro-inflammatory phenotype, we tested the hypothesis that LPA 5 deficiency might affect this phenotypic switch. Indeed, analysis of cyto-/chemokine secretion in wt and LPA 5 -/microglia revealed remarkable differences between the two genotypes. LPA     Significance was calculated by Student's t-test. * p < 0.05, **, p < 0.01, *** p < 0.001 compared to LPS-treated wt mice. might affect this phenotypic switch. Indeed, analysis of cyto-/chemokine secretion i and LPA5 -/-microglia revealed remarkable differences between the two genotypes. increased the concentrations of TNFα, IL-6, and IL-1β ( Figure 4A-C) and the chemok CXCL10, CXCL2, and CCL5 ( Figure 4D-F) at one or more time points in wt microgl LPA5 -/-cells, this pro-inflammatory response was attenuated, with significantly red secretion of TNFα and IL-6 ( Figure 4A,B). LPA-induced effects on IL-1β ( Figure 4C) chemokine secretion ( Figure 4D-F) were virtually absent in LPA5 -/-cells. To account for potential effects mediated by other LPA receptors (in particular L that is highly expressed by primary microglia [18]) or potential LPA loss during sam To account for potential effects mediated by other LPA receptors (in particular LPA 6 that is highly expressed by primary microglia [18]) or potential LPA loss during sample preparation [37], all in vitro experiments were performed also with 5 µM LPA. The higher LPA concentration was chosen to account for the lower LPA affinity of LPA 6 [17]. Similar results for cyto-/chemokine secretion were obtained in response to 5 µM (Supplementary Figure S3).

LPA 5 -/-Microglia Have Higher Mitochondrial Capacity as Compared to wt Cells
Microglia, like peripheral immune cells, are able to utilize different energy metabolites to immediately adapt to chemical alterations in the local microenvironment [38]. Earlier studies from our group have indicated that LPA alters the metabolic profile of the mouse BV-2 microglia cell line to a glycolytic phenotype via an AKT/mTOR/HIF1α-dependent pathway [32] and drives them toward a pro-inflammatory phenotype via LPA 5 [29,36]. To gain insight into whether the core sensome member LPA 5 contributes to metabolic plasticity in microglia, we monitored basic metabolic parameters and inflammatory output in response to LPA in wt and LPA 5 -/cells. In the first set of experiments, we examined mitochondrial function in real time using the Seahorse XF Cell Mito Stress Test. During Seahorse flux analysis, we compared mitochondrial function between wt and LPA 5 -/cells under basal (PBS) and LPA-activated conditions and assessed changes in the oxygen consumption rate (OCR) after treatment with an ATP synthase inhibitor (oligomycin), H + ionophore (FCCP), and under electrontransport chain inhibition (rotenone and antimycin A). These experiments revealed that, under basal conditions, LPA 5 -/microglia showed higher OCR after 2 and 24 h in comparison to wt cells ( Figure 5A,B). In response to LPA (1 µM, 2 h), maximal respiration ( Figure 5E) was significantly higher in LPA 5 -/microglia as compared to wt cells, whereas basal respiration ( Figure 5C), ATP production ( Figure 5D), and spare respiratory capacity ( Figure 5F) showed an upward trend. The increase in the latter two parameters was sustained up to 24 h in LPA 5 -/cells ( Figure 5E,F). Qualitatively and quantitatively comparable observations were made during Seahorse analysis of primary cells cultured in the presence of 5 µM LPA ( Figure S4).

LPA5 Deletion Attenuates LPA-Induced Lactate, NADPH, GSH, and NO Synthesis
In response to pro-inflammatory stimuli, metabolism of microglia shifts OXPHOS to aerobic glycolysis [39]. To investigate potential differences in metabo wiring in LPA-polarized wt and LPA5 -/-cells, we quantified the levels of extracellul tate, the major end product of aerobic glycolysis. These analyses revealed that lact cretion by wt microglia was significantly increased by LPA (0.07 vs. 0.23 and 0.38 v mM; control vs. LPA at 2 h and 24 h, respectively; Figure 6A). In LPA5 -/-cells, this LPA-induced lactate secretion was virtually absent, indicating an important role for

LPA 5 Deletion Attenuates LPA-Induced Lactate, NADPH, GSH, and NO Synthesis
In response to pro-inflammatory stimuli, metabolism of microglia shifts from OXPHOS to aerobic glycolysis [39]. To investigate potential differences in metabolic rewiring in LPA-polarized wt and LPA 5 -/cells, we quantified the levels of extracellular lactate, the major end product of aerobic glycolysis. These analyses revealed that lactate secretion by wt microglia was significantly increased by LPA (0.07 vs. 0.23 and 0.38 vs. 0.52 mM; control vs. LPA at 2 h and 24 h, respectively; Figure 6A). In LPA 5 -/cells, this rise in LPA-induced lactate secretion was virtually absent, indicating an important role for LPA 5 during maintenance of metabolic plasticity. The production of NO was determined by measuring the total nitrate concentration in the supernatants. Results are presented as mean values ± SEM of three independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to wt controls, # p < 0.05, ## p < 0.01, ### p < 0.001 for LPA5 -/-compared to wt cells, two-way ANOVA with Bonferroni correction.
To get an indication about the cellular redox status of LPA-treated wt and LPA5 -/microglia, we analyzed the intracellular NADPH/NADP ratio, reduced glutathione (GSH) content, and nitric oxide (NO) release. The rationale for these analyses is based on earlier observations from our group [32], in which we showed that LPA treatment of BV-2 cells induces phosphorylation of nuclear factor erythroid 2-related factor 2 (Nrf2). Activation of the Nrf2 pathway transcriptionally regulates several key enzymes involved in the antioxidant response, including glucose-6-phopshate dehydrogenase (G6PD) and glutamate cysteine ligase subunits that catalyze the first step of GSH synthesis. Activation of G6PD activates the pentose phosphate cycle, which (in the oxidative branch) generates NADPH as an indispensable co-factor for NO synthesis via inducible NO synthase (iNOS). The NADPH/NADP ratio in wt cells was increased by 5-fold in response to a 24 h LPA exposure, whereas this response was significantly attenuated in LPA5 -/-cells ( Figure 6B). In response to LPA, wt microglia time-dependently increased their intracellular GSH content. In contrast, under basal conditions, the GSH content was slightly lower (statistically not significant) in LPA5 -/-cells, whereas the LPA response of intracellular GSH levels was almost absent at all time points investigated ( Figure 6C). As previously reported [40], LPA treatment increased NO concentrations in the cellular supernatant leading to a 2.8-fold increase after 24 h. This response was not observed in LPA5 -/-cells ( Figure 6D). Comparable results were obtained with cells exposed to 5 µM LPA ( Figure S5).
To get an indication about the cellular redox status of LPA-treated wt and LPA 5 -/microglia, we analyzed the intracellular NADPH/NADP ratio, reduced glutathione (GSH) content, and nitric oxide (NO) release. The rationale for these analyses is based on earlier observations from our group [32], in which we showed that LPA treatment of BV-2 cells induces phosphorylation of nuclear factor erythroid 2-related factor 2 (Nrf2). Activation of the Nrf2 pathway transcriptionally regulates several key enzymes involved in the antioxidant response, including glucose-6-phopshate dehydrogenase (G6PD) and glutamate cysteine ligase subunits that catalyze the first step of GSH synthesis. Activation of G6PD activates the pentose phosphate cycle, which (in the oxidative branch) generates NADPH as an indispensable co-factor for NO synthesis via inducible NO synthase (iNOS). The NADPH/NADP ratio in wt cells was increased by 5-fold in response to a 24 h LPA exposure, whereas this response was significantly attenuated in LPA 5 -/cells ( Figure 6B). In response to LPA, wt microglia time-dependently increased their intracellular GSH content. In contrast, under basal conditions, the GSH content was slightly lower (statistically not significant) in LPA5 -/cells, whereas the LPA response of intracellular GSH levels was almost absent at all time points investigated ( Figure 6C). As previously reported [40], LPA treatment increased NO concentrations in the cellular supernatant leading to a 2.8-fold increase after 24 h. This response was not observed in LPA 5 -/cells ( Figure 6D). Comparable results were obtained with cells exposed to 5 µM LPA ( Figure S5).

Discussion
Although it is not entirely clear under which conditions peripheral LPS crosses the blood-brain barrier (BBB) to trigger neuronal damage [41,42], there is a consensus that peripherally induced endotoxemia has the potential to cause neuroinflammation in experimental animal models and the human organism [43]. This results in microglia and astrocyte activation, memory loss, as well as destruction of synapses and apoptosis of neurons [44]. We have previously demonstrated that pharmacological interference with the ATX/LPA/LPA 5 axis attenuates LPS-induced neuroinflammation in vivo and in a microglia cell line in vitro [36]. Here, we confirm and extend these findings to LPA 5 -/mice and primary LPA 5 -/microglia, thereby eliminating concerns regarding potential off-target effects of synthetic LPA 5 antagonists. In particular, this global knockout model enabled us to directly demonstrate the involvement of LPA 5 during the LPS-mediated peripheral and central inflammatory response, and a certain (short-term) contribution to energy expenditure and sickness behavior. In addition, the pro-inflammatory response of microglia towards LPA was less pronounced in LPA 5 -/microglia, a fact that might be partially attributed to the different immunometabolic phenotype induced in the knockout cells. A question arising from the present study is clinical translatability. As a lipid-activated receptor belonging to the GPCR family, LPA 5 (among the other members of the receptor family) clearly qualifies as a druggable target in the CNS [45]. GPCRs are of high interest as pharmacological targets, since they are involved in pathophysiology, and druggable sites within this receptor class are accessible at the extracellular leaflet. Consequently, approximately 35% of FDA-approved drugs act on GPCRs [46]. Although drug delivery to the brain is restricted by the BBB [47], the situation for the LPA5 antagonist AS2717638 is encouraging since it accumulates in rat brain and displays neuroprotective action [27]. However, considering known differences between mouse and human immunology, findings in murine models must not necessarily reproduce in the human system [48].
Although Banks and Robinson have reported minimal BBB penetration of i.v.-injected 125 I-LPS [41], Vargas-Caraveo suggested that LPS may be transported across the BBB via a lipoprotein-mediated pathway [42]. There is also evidence that LPS associates with Aβ 1-40/42 within amyloid plaques and around vessels in brains of patients suffering from Alzheimer's disease [49]. In addition, i.p. application of LPS provokes several features of sepsis and potentiates peripheral synthesis of cytokines and chemokines. Among these, TNFα has been shown to induce neuroinflammation via the TNF receptor 1 signaling pathway [50]. Additionally, this LPS transcytosis-independent pathway causes elevated gene expression of iNOS and TNFα in the brain [50], in accordance with data obtained in the present study (Figures 1 and 2). Thus, reduced peripheral TNFα synthesis ( Figure 1) would be expected to attenuate the neuroinflammatory response. Rat and human brain microvascular endothelial cells express TLR2, TLR3, TLR4, and TLR6 [51]. LPS interaction with TLR4 on these essentially non-immune cells initiates an inflammatory response by inducing IL-1β, IL-18, IL-6, and TNFα production [52]. In response to peripheral LPS, LPA concentrations in mouse brain and serum increase significantly and could amplify the inflammatory response [31]. Of note, LPA stimulates CD14 transcription and translation [53] and could, via activation of this TLR-4-associated coreceptor, enhance the LPS-mediated inflammatory response. Thus, multiple pathways initiated at the periphery may converge at the BBB, leading to an inflammatory microglia phenotype in the CNS in response to peripheral LPS.
Microglia play a central role in the initiation of neuroinflammation by surveying the chemical composition of the environment. Of relevance for the present study, LPA 5 was identified as a core sensome member in mouse and human microglia [14]. As an extracellular, ligand-activated receptor almost exclusively expressed by microglia/macrophages (https://www.brainrnaseq.org/ (accessed on 7 February 2022); [54]), it is located in a strategic position to detect pathological alterations in the extracellular milieu and convey information to induce a phenotypic transition of microglia. Consequently, LPA 5 was shown to play a disease-amplifying role in stroke [22,23,55], neuropathic pain [25,27,56], itching sensation [57], neuroinflammation [36], and during microglia polarization toward a neurotoxic phenotype [29,36]. Our group has previously shown that pharmacological antagonism of ATX (PF8380) and LPA 5 (AS2717638) downregulated gene and protein expression of several pro-inflammatory markers in brains of LPS-injected C57BL/6 mice [36]. In line with this, results of the present study demonstrated reduced cyto-/chemokine gene expression in brains of LPS-injected LPA 5 -/animals ( Figure 2), thereby replicating findings obtained with the LPA 5 antagonist AS2717638. These data validate and re-confirm an important role of this lipid-activated receptor during induction of neuroinflammatory symptoms. Whether these observations are due to downregulation of TLR4 expression and subsequent attenuation of the inflammatory response (as observed with AS2717638; [36]) was not experimentally addressed in the present study.
In vivo energy metabolism was comparable between wt and LPA 5 -/mice under basal conditions ( Figure S1). In the endotoxemia model, global LPA 5 deficiency provided protection against LPS-induced sickness behavior, lethargy, and energy deficits within the first 24 h of treatment ( Figure 3). This protective effect was, however, lost within the following 72 h of this low-dose LPS treatment regimen ( Figure S2). Our short-term observations in LPA 5 -/mice might be due to lower peripheral TNFα or IL-1β (but unchanged IL-6) synthesis (Figure 1), both of these cytokines resembling classical inducers of sickness behavior in mice and humans [58]. Whether or not the partial recovery in response to repeated LPS injections is due to the development of endotoxin tolerance [59] is currently unclear. Of note, repeated injections of LPS trigger epigenetic modifications of microglia that result in activation of the mTOR/HIF-1α axis [60] and consequently lead to increased (aerobic) glycolysis [61], comparable to LPA-stimulated microglia observed in a previous [32] and in the present study ( Figure 6).
In vitro, LPA-stimulated LPA 5 -/microglia showed reduced cyto-/chemokine synthesis, better mitochondrial fitness, and altered metabolic properties compared to wt cells, independent of whether microglia were exposed to 1 or 5 µM LPA (Figures 4-6 and Figures S3-S5, respectively). LPA 5 was shown to play a pathogenic role during focal cerebral ischemia in mice with upregulated (mRNA and protein) expression of LPA 5 in the ischemic core region, whereas LPA 5 antagonism by TCLPA5 treatment significantly attenuated ischemic brain damage [22]. In addition, the authors demonstrated that siRNAmediated LPA 5 knockdown downregulated gene expression of pro-inflammatory cytokines in LPS-activated BV-2 microglia [22], comparable to our observations in primary LPAtreated LPA 5 -/microglia ( Figure 4). In another model, ischemic brain also exhibited increased ATX activity, LPA concentrations, and LPA receptor expression, with LPA 5 being the most pronounced [62]. These findings were accompanied by disrupted redox balance, BBB dysfunction, and reduced mitochondrial activity. All of these LPA-mediated pathological changes were reversed in response to an ATX and LPA pan-receptor inhibitor (BrP-LPA; [62]). These results are consistent with significantly higher maximal mitochondrial respiration and spare respiratory capacity in LPA 5 -/microglia compared to wt cells ( Figure 5 and Figure S4). A detrimental role of LPA on mitochondrial function is further supported by the fact that heterozygous ATX knockout leads to improved mitochondrial energy homeostasis in brown adipose tissue of mice fed a high-fat, high-sucrose diet [63]. Similarly, a microarray-based approach in brown preadipocytes revealed that ATX-LPA signaling downregulates proteins involved in mitochondrial function and energy metabolism [64].
If LPA compromises mitochondrial function and oxidative phosphorylation, cells might be expected to increase glycolytic flux to meet their energy demand through aerobic glycolysis [65]. Indeed, treatment of wt microglia with LPA led to enhanced lactate secretion ( Figure 6A), whereas this response was absent in LPA 5 -/cells, at least after 2 h, or even reversed after 24 h. The increase in the NADPH/NADP ratio ( Figure 6B) is indicative for an increased metabolite flux through the oxidative branch of the pentose phosphate pathway (PPP). The antioxidant function of NADPH for thioredoxin activity and GSH recycling is well established. However, it becomes increasingly clear that this hexose monophosphate shunt-derived metabolite can also act as a pro-oxidant during O 2 •− production by NADPH oxidases or NO generation by iNOS [66]. Our data suggest that LPA in wt microglia causes an increase in the NADPH/NADP ratio that is associated with elevated cellular GSH and NO production ( Figure 6C,D). LPA signaling via LPA 5 apparently plays a central role in these pathways, since these metabolic responses are either less pronounced or absent in LPA 5 -/microglia ( Figure 6). A recent study identified LPA as a modulator of the metabolic landscape in human pluripotent stem cells. The authors reported significantly increased relative amounts of several amino acids and glycolytic, TCA, and PPP intermediates [33], supporting our observations that LPA has the potential to induce metabolic rewiring in microglia ( [32] and Figure 6). As for mechanistic pathway analysis, combined LPS/IFNγ treatment of microglia was shown to upregulate G6PDH expression, the first and rate-limiting enzyme of the PPP [67]. The same group showed that increased PPP activity feeds NADPH into NO and ROS synthesis, but also serves as a cofactor of glutathione reductase that converts GSSG back to GSH [67], which is consistent with our data in LPA-treated microglia from wt mice. In line, upregulated expression and activity of G6PDH in Parkinson's disease (PD) models was accompanied by excessive NADPH and subsequent ROS production via NOX2 [68]. Knockdown or pharmacological inhibition of G6PDH ameliorated pro-inflammatory microglia polarization, ROS production, and NFκB activation [68]. NADPH-dependent NO synthesis by LPS/IFNγ-activated microglia relies exclusively on efficient glucose flux through the PPP [69]. Microglia use glucose as exclusive energy substrate to generate the superoxide anion radical via NOX2 that transfers electrons to molecular oxygen at the outer and oxidizing NADPH to NADP + and H + at the inner plasma membrane leaflet [70]. Thus, NADPH (generated via the PPP) can perform two seemingly opposing functions: (i) as an essential antioxidant cofactor of glutathione reductase and thioredoxin, and (ii) as a prooxidant trigger of "reductive stress" that leads to the formation of O 2 -and/or NO [66,71], highlighting the close link between redox regulation and immunometabolism [72]. Our observations that LPA 5 -/microglia have a lower NADPH/NADP ratio and produce less NO than wt microglia might suggest that this lipid-activated receptor is involved in the process termed "reductive stress".
Although our study shows that LPA 5 plays a critical role in neuroinflammation, there are also limitations: in our in vivo experiments, only male mice were used and it is unclear why LPA 5 -/mice were protected only during the first 24 h during the chronic LPS treatment regimen. For metabolic studies, it is noteworthy that the PPP is not the only source of NADPH, as substantial amounts are also generated via cytosolic and mitochondrial folate-dependent pathways [73] or by cytosolic isocitrate dehydrogenase and the malic enzyme [66]. Of note, NADPH generation during oxidation of methylene tetrahydrofolate to 10-formyl-tetrahydrofolate is also coupled to the cellular GSH/GSSG status [73]. Since we performed only enzymatic assays (and not stable isotope-labeled precursor flux analysis), we are unable to comment on the actual metabolic pathway(s) that generate NADPH. However, considering the strict glucose dependence of microglia for NO production [69], a substantial contribution of the PPP is likely.
Despite these potential shortcomings, our study clearly demonstrates that LPA 5mediated signaling cascades are centrally involved in the neuroinflammatory response. In this setting, LPA 5 -/animals and primary cells represent invaluable tools to verify and extend neurological in vivo and in vitro data obtained with pharmacological LPA 5 antagonists.

Data Availability Statement:
The data presented in this study are available on reasonable request from the corresponding author. Reagents and detailed methods of all procedures are provided in the "Materials and Methods" of this manuscript or cited accordingly.

Acknowledgments:
We gratefully acknowledge expert technical assistance by Celina Klampfer and Silvia Rainer. The authors thank A. Absenger, M. Singer, and I. Hindler (Medical University of Graz, Austria) for mice care.

Conflicts of Interest:
The authors declare no conflict of interest.