A protective role for N-acylphosphatidylethanolamine phospholipase D in 6-OHDA-induced neurodegeneration

N-acylphosphatidylethanolamine phospholipase D (NAPE-PLD) catalyzes the cleavage of membrane NAPEs into bioactive fatty-acid ethanolamides (FAEs). Along with this precursor role, NAPEs might also serve autonomous signaling functions. Here, we report that injections of 6-hydroxydopamine (6-OHDA) into the mouse striatum cause a local increase in NAPE and FAE levels, which precedes neuronal cell death. NAPE, but not FAE, accumulation is enhanced in mice lacking NAPE-PLD, which display a substantial reduction in 6-OHDA-induced neurotoxicity, as shown by increased survival of substantia nigra dopamine neurons, integrity of striatal dopaminergic fibers, and striatal dopamine metabolite content. Reduced damage is accompanied by attenuation of the motor response evoked by apomorphine. Furthermore, NAPE-PLD silencing protects cathecolamine-producing SH-SY5Y cells from 6-OHDA-induced reactive oxygen species formation, caspase-3 activation and death. Mechanistic studies in mice suggest the existence of multiple molecular contributors to the neuroprotective effects of NAPE-PLD deletion, including suppression of Rac1 activity and attenuated transcription of several genes (Cadps, Casp9, Egln1, Kcnj6, Spen, and Uchl1) implicated in dopamine neuron survival and/or Parkinson’s disease. The findings point to a previously unrecognized role for NAPE-PLD in the regulation of dopamine neuron function, which may be linked to the control of NAPE homeostasis in membranes.

membrane fusion 19 , and consolidation of lipid raft structure 20 . Furthermore, similarly to the better known phosphoinositides 21 , the NAPEs might serve as tethers for the association of intracellular proteins to the internal facet of the lipid bilayer 22 .
Ischemic insults to the brain cause a rapid and profound elevation in NAPE levels [23][24][25] . Similar responses have been documented in primary cultures of brain neurons exposed to neurotoxic insults, such as high concentrations of the excitatory transmitter glutamate [26][27][28] . It is still unknown, however, whether damage-induced NAPE accrual plays a functional role in neurotoxicity and neurodegeneration. We have recently shown that intrastriatal injections of the dopaminergic neurotoxin 6-hydroxydopamine (6-OHDA) produce a local accumulation of N-acyl saturated NAPE species 29 . In the present study, we examined the impact of genetic NAPE-PLD deletion on 6-OHDA-induced neurotoxicity in mice and in the catecholamine-producing human cell line SH-SY5Y. The results suggest that complete or partial NAPE-PLD ablation enhances 6-OHDA-induced NAPE accumulation without significantly changing FAE levels, and concomitantly protects dopamine neurons from the toxic effects of 6-OHDA. We also explored potential molecular mechanisms through which NAPE-PLD ablation might protect mice from 6-OHDA-induced toxicity. A previous report has shown that exogenous NAPE application inhibits activity of the Rho family GTP-binding protein 1 (Rac1) 30 , which has been implicated in dopamine neuron survival 31,32 . Consistent with those results, we found that genetic NAPE-PLD ablation is associated with a significant reduction in bioactive GTP-bound Rac1. However, focused transcriptomic analyses revealed several additional changes in genes involved in dopamine neuron function, suggesting that a multiplicity of mechanisms might contribute to neuroprotection in NAPE-PLD-null mice.
Effects of 6-OHDA on NAPE levels in human SH-SY5Y cells. The human neuroblastoma cell line SH-SY5Y produces dopamine and other catecholamines and is commonly used as an in vitro model of dopamine neuron degeneration [37][38][39] . When incubated in the presence of 6-OHDA (100 μM), SH-SY5Y cells displayed an increase in reactive oxygen species (ROS) formation (Fig. 5A), which was followed by a substantial activation of the apoptosis marker caspase 3 (Fig. 5B). These effects were accompanied by a progressive down-regulation of Napepld gene transcription (Fig. 5C) and NAPE-PLD protein expression (Fig. 5D). Moreover, exposure to 6-OHDA caused a time-dependent increase in cellular NAPE content, which exclusively involved N-acyl saturated NAPE species and reached a maximum after 8 h incubation with the toxin (Fig. 5E).

NAPE-PLD silencing increases NAPE levels in SH-SY5Y cells.
Next, we silenced the Napepld gene in SH-SY5Y cells using a selective 27-mer siRNA duplex, which decreased Napepld transcription by ≈60% compared to control cells exposed to a scrambled oligonucleotide (Fig. 6A). The observed reduction in Napepld mRNA was accompanied by a ≥75% decrease in the levels of NAPE-PLD protein in both cytosolic and membrane fractions ( Fig. 6B) 8 , and was associated with an ≈80% increase in the levels of N-acyl saturated NAPE species (Fig. 6C-E). Incubation with 6-OHDA further enhanced NAPE levels in both control and siRNA-treated cells ( Fig. 6C-E), but did not significantly affect SEA levels (Fig. 6F).

NAPE-PLD silencing protects SH-SY5Y cells from 6-OHDA toxicity.
To probe the functional consequences of NAPE-PLD deletion in SH-SY5Y cells, we examined the effects of siRNA-mediated NAPE-PLD silencing on 6-OHDA-induced damage. Consistent with the neuroprotective phenotype observed in NAPE-PLD −/− mice, NAPE-PLD silencing significantly blunted the effects of 6-OHDA on cellular ROS production ( Fig. 7A) and caspase 3 activation (Fig. 7B). Furthermore, overall cellular viability was greater in NAPE-PLD-silenced cells than in scrambled-treated controls (Fig. 7C). We interpret the results as indicating that NAPE-PLD down-regulation elevates NAPE levels and protects SH-SY5Y cells from 6-OHDA-induced toxicity.
Mechanistic studies. The molecular targets of NAPEs in neural cells are unknown, but studies in mouse peritoneal macrophages have shown that treatment with exogenous NAPE (36:2-16:0) inhibits the activity of Rac1 30 , a small G protein that has been implicated in dopamine neuron survival 31,32,40 . We asked therefore whether NAPE-PLD deletion might alter Rac1 expression and function in TH + dopamine neurons of the SN. Confirming previous data 32 , immunofluorescence studies showed that Rac1 is detectable in these cells (Fig. 8A). Of note, immunoreactive Rac1 levels appeared to be higher in NAPE-PLD −/− mice than in wild-type controls (Fig. 8B). Confirming this finding, western blot analyses showed that total Rac1 protein content was significantly, albeit modestly, elevated in midbrain extracts of NAPE-PLD −/− mice (Fig. 9A). More importantly, treatment with 6-OHDA produced an increase in the activated GTP-bound form of Rac1 in wild-type mice, but not in animals lacking NAPE-PLD (Fig. 9B,C), suggesting that accumulation of endogenous NAPEs may inhibit Rac1 activity in midbrain neurons.

Discussion
Ischemic and toxic insults elevate NAPE levels in rodent brain tissue and neural cell cultures 23,29 , but the functional significance of this response, if any, remains unknown. In this study, we report on a possible contribution of NAPE-PLD, a zinc hydrolase that converts membrane NAPEs into FAEs 7,8 , to the neurotoxic response elicited by 6-OHDA. In mice treated with the toxin, NAPE-PLD deletion protected both dopaminergic neurons in the Individual NAPE levels (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA with Bonferroni post hoc test.
SN and dopamine fibers in the striatum, while attenuating the motor response to apomorphine. Furthermore, in SH-SY5Y cells incubated with 6-OHDA, NAPE-PLD silencing reduced ROS formation and caspase-3 activation, and enhanced cell viability. In both models, NAPE-PLD down-regulation was accompanied by significant increases in the levels of NAPE species containing saturated N-acyl substituents, but only by minor non-significant changes in FAE content. Together, the findings point to a previously unrecognized role for NAPE-PLD in the control of dopamine neuron survival, which might be mediated through regulation of membrane NAPE levels. www.nature.com/scientificreports www.nature.com/scientificreports/ We have previously shown 29 , and confirmed in the present study, that unilateral injections of 6-OHDA into the mouse striatum cause a marked local accumulation of N-acyl saturated NAPE species such as NAPE (38:4-18:0) and NAPE (40:6-16:0). This result is consistent with published reports indicating that ischemia and other neurotoxic insults stimulate NAPE accumulation both in vitro and in vivo [41][42][43] . The molecular mechanism(s) underlying damage-induced NAPE accrual is unclear, but two lines of evidence suggest that changes in NAPE-PLD activity and/or expression might be involved. First, pharmacological or genetic blockade of NAPE-PLD causes profound elevations in cellular NAPE levels ( 33 and present study). Second, a broad range of pro-inflammatory and tissue-damaging interventions 44-47 -including administration of LPS 45 or incubation of SH-SY5Y cells with  www.nature.com/scientificreports www.nature.com/scientificreports/ 6-OHDA (present study) -suppress transcription of the Napepld gene, possibly by epigenetic processes similar to those recruited in macrophages 45 . Moreover, induction of focal cerebral ischemia 48 in mouse brain is accompanied by reduced NAPE-PLD activity, suggesting that expression of this protein may be also downregulated following 6-OHDA administration. These results suggest that NAPE-PLD activity is an important contributor to NAPE homeostasis in membranes and that transcriptional suppression of Napepld mediates, at least in part, damage-induced NAPE accrual. Another potential mechanism may be accelerated NAPE biosynthesis, which occurs via enzyme-mediated transfer of an acyl group from the sn-1 position of PC to the amino group of PE 2 . This reaction is catalyzed by the cytosolic phospholipase PLA2A4E, whose activity is stimulated by calcium 3 . Since neural cell damage is almost invariably accompanied by profound alterations in intracellular calcium homeostasis, it is possible that PLA2A4E activity might also contribute to damage-induced NAPE accumulation. Future experiments will need to address this possibility.
Because NAPE-PLD catalyzes the final step in FAE biosynthesis, tissue levels of these lipid molecules are often, but not always, lower in mice lacking NAPE-PLD than they are in wild-type mice 33 . In the present study, we did not observe any significant difference in striatal FAE levels between NAPE-PLD −/− and wild-type mice or between NAPE-PLD-silenced SH-SY5Y cells and their controls. Similarly, pharmacological inhibition of NAPE-PLD activity in intact Hek293 cells was found to cause a substantial accumulation of NAPEs without changes in FAE content 49 . These results are consistent with the non-rate limiting role of NAPE-PLD in FAE biosynthesis 8 , which is most likely controlled by PLA2A4E activity 1,50 . However, the available data do not allow us to exclude that compensations may occur in mice lacking NAPE-PLD, which might offset the genetic removal of this enzyme via isofunctional substitution. The existence of such compensations has been documented, for example, in bone marrow macrophages isolated from NAPE-PLD −/− mice 45 . Despite these uncertainties, our results clearly show that NAPE-PLD deletion increases the levels of N-acyl saturated NAPE species (e.g., 38:4-18:0) without affecting those of the corresponding FAEs (e.g., SEA). As such, the findings raise the intriguing possibility that NAPE-PLD activity might regulate dopamine neuron survival through its ability to control NAPE homeostasis in membranes. Testing this hypothesis will require further experimentation, but biophysical studies with synthetic NAPEs -such as 1,2-dioleoyl-phosphatidylethanolamine N-dodecanoyl [NAPE (36:2-12:0)] and 1,2-myristoyl-phosphatidylethanolamine N-myristoyl [NAPE (28:0-14:0)] -in reconstituted systems suggest several possible ways by which NAPEs might affect cell function, which include stimulation of calcium-dependent membrane fusion 19 , consolidation of lipid raft structure 20 and stabilization of the lipid bilayer 17,18 . Akin to the phosphoinositides 21 , the NAPEs might also regulate the association of cytosolic proteins to the internal facet of the cell membrane 22 .
An initial exploration of the molecular mechanisms underlying the neuroprotective effects of NAPEs suggests that multiple factors are likely to be involved. Previous work has shown that treatment with exogenous NAPE (36:2-16:0) inhibits the activity of Rac1 in peritoneal mouse macrophages and J774A.1 cells 30 . Rac1 is a Rho family small G protein that has been implicated, among other functions, in dopamine neuron survival and Parkinson's disease 32,40 . These data prompted us to ask whether NAPE-PLD deletion and consequent membrane NAPE accrual might influence Rac1 activity in mouse brain. We confirmed that midbrain dopamine neurons express Rac1, and further found that intrastriatal 6-OHDA injections cause an increase in the amount of activated GTP-bound form of Rac1 in the SN of wild-type mice, but not of mice lacking NAPE-PLD. This result points to a possible role for Rac1 in 6-OHDA-induced neurotoxicity, and suggests that deactivation of this small G protein may contribute to the neuroprotective effects of NAPE-PLD deletion. The mechanism through which NAPE-PLD deletion regulates Rac1 activity is unknown, though one possibility is that accrual of membrane NAPE levels might influence the association of Rac1 to cell membranes, which is required for GDP to GTP exchange and Rac1 activation 51 .

Continued
It seems unlikely that Rac1 regulation is the only mechanism through which NAPEs influence neuronal viability. Indeed, the focused gene array study presented in Table 1 identified six genes whose transcription may be significantly attenuated by 6-OHDA in midbrain extracts of NAPE-PLD −/− mice compared to wild-type controls. These include Cadps (Calcium Dependent Secretion Activator), a peripheral membrane protein involved in vesicle fusion and monoamine neurotransmission 52 and Parkinson's disease pathogenesis 53 ; Casp9 (Caspase 9) a mediator of apoptosis whose activity may be elevated in peripheral blood cells of persons with sporadic Parkinson's disease 54 ; Egln1 (Egl-9 Family Hypoxia Inducible Factor 1), whose transcription may be enhanced in the SN of sporadic PD patients 55,56 ; Kcnj6 (G Protein-Activated Inward Rectifier Potassium Channel 2), whose transcription may also be enhanced in the SN of sporadic Parkinson's disease patients 57 ; Spen (Spen Family Transcriptional Repressor) a hormone inducible repressor; and, finally, Uchl1 (Ubiquitin C-Terminal Hydrolase L1), a key component of the ubiquitin-proteasome pathway that is mutated in some familial forms of Parkinson's disease 58 . The multiplicity of molecular effectors associated with NAPE-PLD deletion makes it difficult to identify a univocal mechanism through which membrane NAPEs might influence cell death-related signals (e.g., ROS, caspase 3 activation) and dopamine neuron survival. Such multiplicity might reflect a pleiotropic role for membrane NAPEs in intracellular signaling, which is further supported by the broad effects exerted by these lipid molecules on membrane structure and function (discussed above).
Another question raised by the present results pertains to the functional significance of NAPE heterogeneity, which involves substituents in the sn-1 (16:0/18:0), sn-2 (20:4/22:6) and N-position (16:0/18:0). Such heterogeneity is well established in the literature (see for example 23,28,59 ), but can be only partially attributed to species and tissue variability. Other factors are likely to play a role, including the potential role of NAPEs as either FAE precursors or autonomous membrane signals. Additional work is needed to address these intriguing possibilities.
Abnormalities in lipid composition have emerged as important pathogenic factors in neurodegenerative disorders such as Parkinson's disease 60 and Alzheimer's disease 61 . For example, mutations in the glycolipid-metabolizing enzyme glucocerebrosidase (GBA1) are common in the familiar forms of Parkinson's disease, while single-nucleotide polymorphisms of genes involved in other aspects of lipid metabolism [e.g., ASAH1 62 , PLA2G6 63 and SMPD1 64 ] have been detected in the sporadic and more frequent form of this disorder. Furthermore, genome-wide association studies have identified several lipid-processing genes (e.g., APOE4, PLD3 and ABCA7) as risk loci that may increase susceptibility for Alzheimer's disease 61 . Indeed, select alterations in   www.nature.com/scientificreports www.nature.com/scientificreports/ plasma phospholipid profile were shown to predict, with over 90% accuracy, phenoconversion of cognitively normal older adults to either Alzheimer's disease or mild cognitive impairment 65 . In this context, it is worth noting that that the activities of two enzymes -phosphoethanolamine cytidylyltransferase and phosphocholine cytidylyltransferase -that are rate-limiting for the biosynthesis for the NAPE precursors, PC and PE, are abnormally elevated in the SN of persons with Parkinson's disease 66 . By revealing a protective role for NAPE-PLD in neurodegeneration, the present results underscore the need for further and more detailed explorations in the lipidomics of neurodegenerative disorders.
In sum, we have shown that complete or partial genetic deletion of NAPE-PLD, the membrane-associated zinc hydrolase that converts NAPEs into FAEs, increases NAPE levels and concurrently protects mouse dopamine neurons and human dopamine-producing SH-SY5Y cells from the neurotoxic effects of 6-OHDA. The results suggest that NAPE-PLD activity may participate in the regulation of dopamine neuron survival, possibly by controlling membrane NAPE homeostasis.

Materials and Methods
Animals. All procedures were performed in accordance with the Ethical Guidelines of the European Union . NAPE-PLD −/− mice were generated on a C57BL6J background as previously described 33 . All mice were group-housed in ventilated cages and had free access to food and water. They were maintained under a 12 h light/dark cycle (lights on at 8:00 am) at controlled temperature (21 ± 1 °C) and relative humidity (55% ± 10%). All efforts were made to minimize animal suffering and to use the minimal number of animals required to produce reliable results. chemicals. 6-OHDA hydrochloride, chloral hydrate, ketamine, xylazine, paraformaldehyde (PAF), dopamine, serotonin (5-HT), DOPAC and ascorbic acid were purchased from Sigma Aldrich (Milan, Italy). NAPE standards and internal standards were synthetized in the laboratory as previously described 29 . cell cultures. SH-SY5Y cells were obtained from Sigma Aldrich and were cultured at 37 °C and 5% CO 2 in Dulbecco's Modified Eagle's Medium (DMEM) (Euroclone, Milan, Italy) supplemented with 10% fetal bovine serum (FBS, Thermo Fisher, Waltham, MA, USA), L-glutamine (2 mM) and antibiotics (Euroclone). Cells were treated with 6-OHDA (100 μM), or vehicle (saline containing 0.2% ascorbic acid) for the indicated times.
nApe-pLD silencing. siRNA experiments were performed using a Napepld-specific 27mer siRNA duplex (Origene, Rockville, MD, USA). A siRNA duplex carrying TYE-563 fluorescence was used to monitor transfection. A siRNA duplex carrying a 27-mer sequence targeting the hypoxanthine phosphoribosyltransferase 1 (HPRT) gene was used as positive control. Scrambled siRNA was included in each experiment as negative control. Napepld siRNA complexes (10 nM) were formed by mixing siRNA with lipofectamine (Invitrogen, Carlsbad, CA, USA) for 10 min at room temperature and then added to SH-SY5Y cells cultured with 1% FBS Optimem medium (Gibco, Waltham, MA, USA). Cells were incubated with siRNA oligonucleotide for 6 h. After an 18 h incubation with fresh full-growth medium, 6-OHDA (100 μM) was added for additional 8 hours.

Lipid extraction and LC/MS analyses. NAPE and FAE levels in cells and tissues
were measured as previously described 29 (for further details see Supplementary). tissue processing and immunohistochemistry. Mice were deeply anaesthetized with chloral hydrate (450 mg-kg −1 , i.p.) and perfused transcardially with ice-cold sterile saline (20 ml), followed by ice-cold PFA [4% in phosphate-buffered saline (PBS), 60 ml]. The brains were excised and stored in a sucrose solution (25% in PBS) at 4 °C. Three series of sections (thickness: 40 µm) were collected in the coronal plane using a cryostat, and stored at −20 °C. Single immunostaining protocols were performed by incubation with primary antibody (for details see Table S3) followed by secondary Alexa Fluor 546 or 488 antibodies (1:1000; Invitrogen Carlsbad, CA, USA). Multiple labeling were conducted sequentially. Images were collected using a Nikon A1 confocal microscope with a 10 1.4 numerical aperture objective lens. Quantification of TH optical density was performed using the ImageJ software.
Stereological measurements. Dopamine neurons of the SN were identified after TH staining followed by Alexa fluor 488 secondary antibody of midbrain regions with a 4x objective. The SN was outlined using the Paxinos and Franklin's mouse brain atlas as reference 67 . TH + neurons were counted in every 6th section. Briefly, unbiased sampling and blinded stereological counting were performed using the optical fractionator probe of the Stereo Investigator software (MBF Bioscience, Williston, VT, USA). Parameters used included a 60x oil objective, a counting frame size of 60 × 60, a sampling site of 100 × 100, a dissector height of 15 μm, 2 μm guard zones. The Gunder's coefficient of error was less than 0.1. A total of 4 animals per group were used and 5 to 8 sections per animal were counted in the red channel.
www.nature.com/scientificreports www.nature.com/scientificreports/ Western blot analyses. Cell pellets were homogenized in 150 μl of a radioimmunoprecipitation assay buffer (RIPA), consisting of 50 mM Tris-HCI (pH 7.4), 1% Tryton X 100, 0.5% sodium-deoxycholate, 0.1% sodium dodecyl sulfate, 150 mM sodium chloride, 2 mM ethylenediaminetetraacetic acid. Protein concentrations were measured using the bicinchoninic acid (BCA) method, following manufacturer's instructions (Thermo Fisher Scientific). Proteins (30 µg) were denatured in SDS (8%) and β-mercaptoethanol (5%) at 95 °C for 5 min. After separation by SDS-PAGE on a 4-15% gel, the proteins were electrotransferred to nitrocellulose membranes. The membranes were blocked with 5% non-fat dry milk in tris-buffered saline (TBS) and incubated overnight with primary antibodies (for details see Table S3) in 1% milk-TBS containing 0.1% Tween-20, followed by incubation with horseradish peroxidase-linked to the secondary antibody (1:5,000, Millipore) in TBS 0.1% Tween-20 at room temperature for 1 h. Finally, proteins were visualized using an ECL kit (Bio-Rad, USA) and the chemiluminescence image was recorded using a LAS-4000 lumino-image analyzer system (Fujifilm, Tokyo, Japan). neurotransmitter measurements. Striatal tissue was removed, snap frozen in liquid N 2 and stored at −80 °C. Samples were weighed and homogenized in 0.9 ml/methanol:water (1:1) containing 0.1% formic acid. After stirring and centrifugation, the supernatants were dried under N 2 and the samples were reconstituted in 90 µl of mobile phase A (0.1% acetic acid in water) for LC-MS/MS analyses. LC-MS/MS analyses were carried out on an Acquity UPLC system coupled with a Xevo TQ-MS triple quadrupole mass spectrometer. Chromatographic separation was achieved using a BEH C18 column (2.1 × 100 mm, 1.7 μm particle size) eluted at a flow rate of 0.35 mL/min, using the following gradient conditions: 0-1.0 min 5% solvent B (methanol) in solvent A (0.1% acetic acid in water), 1.0-2.5 min 5% to 100% B, and 2.5-3.5 min 100% B. The column was re-equilibrated to initial conditions from 3.5 to 4.5 min. Total run time for analysis was 4.5 min, and injection volume was 5 μL. The column temperature was kept at 45 °C. The MS was operated in both positive and negative ESI mode with cone voltage, collision energy and capillary voltage set at 10 V, 20 V and 3kv respectively. The source temperature was 120 °C. Desolvation gas and cone gas (nitrogen) flow were set at 800 and 50 l/h, respectively. Desolvation temperature was 450 °C. Analytes were quantified by MRM with the following transitions (m/z): dopamine, 153.8 > 136.7; serotonin 176.9 > 159.9; DOPAC, 166.5 > 122.9. Dopamine and serotonin were acquired in positive mode and DOPAC in negative mode. Data were acquired by MassLynx software and quantified by TargetLynx software; individual standard calibration curves were used for the quantification of each analyte.