NAPE-PLD in the ventral tegmental area regulates reward events, feeding and energy homeostasis

The N-acyl phosphatidylethanolamine-specific phospholipase D (NAPE-PLD) catalyzes the production of N-acylethanolamines (NAEs), a family of endogenous bioactive lipids, which are involved in various biological processes ranging from neuronal functions to energy homeostasis and feeding behaviors. Reward-dependent behaviors depend on dopamine (DA) transmission between the ventral tegmental area (VTA) and the nucleus accumbens (NAc), which conveys reward-values and scales reinforced behaviors. However, whether and how NAPE-PLD may contribute to the regulation of feeding and reward-dependent behaviors has not yet been investigated. This biological question is of paramount importance since NAEs are altered in obesity and metabolic disorders. Here, we show that transcriptomic meta-analysis highlights a potential role for NAPE-PLD within the VTA®NAc circuit. Using brain-specific invalidation approaches, we report that the integrity of NAPE-PLD is required for the proper homeostasis of NAEs within the midbrain VTA and it affects food-reward behaviors. Moreover, region-specific knock-down of NAPE-PLD in the VTA enhanced food-reward seeking and reinforced behaviors, which were associated with increased in vivo DA release dynamics in response to both food and non-food-related rewards together with heightened tropism towards food consumption. Furthermore, midbrain knock-down of NAPE-PLD, which increased energy expenditure and adapted nutrient partitioning, elicited a relative protection against high-fat diet-mediated body fat gain and obesity-associated metabolic features. In conclusion, these findings reveal a new key role of VTA NAPE-PLD in shaping DA-dependent events, feeding behaviors and energy homeostasis, thus providing new insights on the regulation of body metabolism.


Introduction
The regulation of feeding behaviors and energy homeostasis is a cardinal and evolutionarily conserved physiological feature in mammals.By mobilizing several and functionally distinct brain circuits (1,2), such regulation tightly depends on metabolic and nutritional demands as well as on the reinforcing and hedonic properties of foods.Among the different regulatory pathways which signal homeostatic states and scale feeding behaviors, lipids, either nutritional and/or endogenous species, represent key mediators in gating the functional adaptability of complex neuronal networks that synchronize food intake and energy expenditure (3,4).
Endogenous bioactive lipids are a major class of biologically active mediators which critically regulate several functions, spanning from homeostasis to cognition.Among these, the membrane phospholipidderived long-chain fatty acids N-acylethanolamines (NAEs) are potent signaling molecules in peripheral tissues as well as in the central nervous system (CNS).Given their wide distribution and multiple biological functions, the modulation of NAEs tone represents an interesting target for the development of new therapeutic approaches (5).
While NAPE-PLD expression, enzymatic activity and related byproducts have been detected in the mouse, rat and human brains (7,(22)(23)(24), the key functions of this enzyme within the brain remain elusive and not well-elucidated.In fact, most of the current literature has focused on distinct NAEs based on their transducing effectors (receptors, transcription factors) and not on the enzyme itself.Only recently, a few studies have shown that pharmacological blockade of NAPE-PLD activity (25) or selective ablation of NAPE-PLD in stress-activated neurons (21) impaired limbic functions (fear extinction, anxiety) mainly through the regulation of the hypothalamus-pituitary-adrenal (HPA) axis, thus highlighting the importance of NAPE-PLD and its NAEs in brain functions.Furthermore, NAPE-PLD silencing and consequent increased availability of its NAPE substrate are neuroprotective in response to 6-OHDA-induced loss of dopamine (DA)-neurons (20), suggesting an important role of NAPE-PLD in regulating physiological and cellular functions of DA-neurons.
Moreover, the developmental invalidation of the Napepld gene alters very-long chain fatty acids composition in the brain, suggesting a more complex role for NAPE-PLD than previously appreciated and the existence of alternative biosynthetic pathways (26,27).Furthermore, the non-CB1R-related signaling engaged by fatty acid ethanolamines (FAEs) produced by the speci c enzymatic activity encoded by the Napepld gene (ID: 242864) points toward a role for this enzyme in the generation of bioactive FAEs with a wide array of functions including the regulation of energy balance (OEA), in ammation (PEA) or pain sensitivity (SEA).
Interestingly, the presence of a single-nucleotide polymorphism (SNP) on the coding region of the Napepld gene (rs17605251) has been associated with severe obesity (BMI ≥ 35 kg/m 2 ) (28, 29), suggesting a potential role for NAPE-PLD in the regulation of energy homeostasis, food-motivated behaviors and metabolic disorders.
In the present study, we took advantage of several integrative in vivo approaches following genetic and/or viral induction of tissues-speci c deletion of NAPE-PLD in basal, food-motivated and obese conditions.Within the mouse mesolimbic reward system, notably the ventral tegmental area (VTA), we found that the enzyme NAPE-PLD functions as a ne-tuning gatekeeper of reward events and dopamine dynamics as well as an important regulator of energy homeostasis and metabolic e ciency in both physiological and pathological (obesity) contexts.

In vivo ber photometry
For in vivo dopamine imaging (GRAB-DA2m, ( 35)), a chronically implantable cannula (Doric Lenses, Québec, Canada) composed of a bare optical ber (400 µm core, 0.48 N.A.) and a ber ferrule was implanted 100 µm above the location of the viral injection site in the NAc.The ber was xed onto the skull using dental cement (Super-Bond C&B, Sun Medical).Real time uorescence was recorded using ber photometry as described in (36, 37).Fluorescence was collected in the NAc using a single optical ber for both delivery of excitation light streams and collection of emitted uorescence.The ber photometry setup used 2 light emitting LEDs: 405 nm LED sinusoidally modulated at 330 Hz and a 465 nm LED sinusoidally modulated at 533 Hz (Doric Lenses) merged in a FMC4 MiniCube (Doric Lenses) that combines the 2 wavelengths excitation light streams and separate them from the emission light.The MiniCube was connected to a ber optic rotary joint (Doric Lenses) connected to the cannula.A RZ5P lock-in digital processor controlled by the Synapse software (Tucker-Davis Technologies, TDT, USA), commanded the voltage signal sent to the emitting LEDs via the LED driver (Doric Lenses).The light power before entering the implanted cannula was measured with a power meter (PM100USB, Thorlabs) before the beginning of each recording session.The light intensity to capture uorescence emitted by 465 nm excitation was between 25-40 µW, for the 405 nm excitation this was between 10-20 µW at the tip of the ber.The uorescence emitted by the GRAB was collected by a femtowatt photoreceiver module (Doric Lenses) through the same ber patch cord.The signal was then received by the RZ5P processor (TDT).On-line real time demodulation of the uorescence due to the 405 nm and 465 nm excitations was performed by the Synapse software (TDT).Signals were exported to Python 3.0 and analyzed o ine as previously described (36).Data are presented as z-score of ΔF/F.

Metabolic e ciency analysis
Metabolic e ciency was measured as previously described (36).Brie y, mice were monitored for whole energy expenditure (EE), O 2 consumption, CO 2 production, respiratory exchange rate (RER = VCO 2 /VO 2 , V = volume), and locomotor activity using calorimetric cages (Labmaster, TSE Systems GmbH, Bad Homburg, Germany).Gases ratio was determined through an indirect open circuit calorimeter.This system monitors O 2 and CO 2 at the inlet ports of a tide cage through which a known ow of air is ventilated (0.4 L/min) and regularly compared to an empty reference cage.O 2 and CO 2 were recorded every 15 min during the entire experiment.EE was calculated using the Weir equation for respiratory gas exchange measurements.Food intake was measured with sensitive sensors for automated online measurements.Calorimetric studies to investigate voluntary exercise-induced metabolic adaptions were performed in metabolic cages equipped with running wheels (Promethion, Sable Systems, Nevada, USA).Mice were monitored for body weight and composition at the entry and exit of the experiments using an EchoMRI (Whole Body Composition Analyzers, EchoMRI, Houston, USA).Data analysis was performed on Excel XP using extracted raw values of VO 2 (ml/h), VCO 2 (ml/h), and EE (kcal/h).

Behaviors
Operant conditioning -Mice were food-restricted and maintained at 90% of their initial body weight to facilitate learning and performance during the whole operant conditioning.Computer-controlled operant conditioning was conducted in 12 identical conditioning chambers (Phenomaster, TSE Systems GmbH, Bad Homburg, Germany) during the light phase, at the same hour every day until the end of the procedure.Each operant wall had two levers (one active and one inactive) located 3 cm lateral to a central pellet dispenser.The reinforcer was a single 20-mg peanut butter avored sucrose tablet (TestDiet, Richmond, USA).Operant training was carried out daily with no interruption for 1h under a xed-ratio 1 (FR1, 1 lever press = 1 pellet).When the discrimination score between active and inactive lever press (active lever presses/inactive lever presses) exceeded chance level, mice were shifted to sessions under a FR5 (5 lever presses = 1 pellet) and/or a progressive ratio (PR) [3 lever presses more for each subsequent reinforcer (r = 3N + 3; N = reinforcer number)].Whenever of interest, PR was conducted in both foodrestricted and sated mice.
Conditioned-place preference (CPP) -The CPP paradigm was performed during the light phase either in food-restricted (maintenance at 90% of initial body weight) or normally fed mice.All the compartments were cleaned before each conditioning session.Locomotor activity was recorded with an infrared beambased activity monitoring system and analyzed with the provided software (Phenomaster, TSE Systems GmbH, Bad Homburg, Germany).The least preferred compartment during the exploration phase was designated as the reward (HFD)-baited compartment whereas the more preferred compartment as the chow-baited compartment (biased protocol).Animals with more than 65% of preference for a compartment on the pre-test day were removed.To reduce anxiety, during the rst two days, animals were carefully put in the middle of the apparatus and allowed to freely explore the two compartments for 1h.
The subsequent days included alternating conditioning sessions of 1h.After 8 days of conditioning [4 sessions in each compartment (chow and HFD)], animals freely explored the two compartments for 30 minutes.The time spent in the reward-paired compartment before vs after conditioning was the primary outcome variable (preference score).
T-Maze -Mice were food-restricted (90% of initial body weight) during the whole paradigm and tested for learning and cognitive exibility in a T-maze apparatus (arm 35-cm length, 25-cm height, 15-cm width) (38).First, they were habituated to the apparatus (15 min of exploration) for two consecutive days.Then, mice underwent a 5-days training protocol with one arm reinforced with a palatable food pellet (HFD, cat #D12492, 5.24 kcal/g).Each mouse was placed at the starting point and allowed to explore the maze by choosing one of the two arms (reinforced and non-reinforced arms).The chosen arm was then blocked for 20 seconds and the mouse replaced again in the starting arm.This process was repeated for 10 sessions per day.At the end of this training period, cognitive exibility and relearning processes were assessed in a reversed learning task which consisted in exchanging the reinforced with the non-reinforced arm.Again, mice underwent a 5-days training protocol (10 sessions/day).Time-locked wheel running -Mice had access to a running wheel connected to an automatic revolution counter (Intellibio Innovation) during a limited amount of time (30 min per session, one session per day) during 5 consecutive days.
Food preference and choice -Mice were tested for food choice and preference by using non-caloric and/or caloric solutions.Notably, they were exposed to graduated small bottles containing either water, sucralose (2 mM), sucrose (10% w/v) or intralipids 20%.During different days of exposure (1h session with free choice between 2 bottles), preference was measured by comparing the consumption of sucralose vs water, sucrose vs water and lipids vs water.
GBR-induced locomotor activity -Locomotor activity induced by GBR12909 (10 mg/kg, i.p.) was recorded in an automated online measurement system using an infrared beam-based activity monitoring system (Phenomaster, TSE Systems GmbH, Bad Homburg, Germany).

Lipidomics
Tissue extracts and HPLC/MS/MS were performed as previously described (40).In brief, samples were placed in 50 volumes of HPLC-grade methanol then spiked with 500 pmols deuterium-labeled Narachidonoyl glycine (d8NAGly; Cayman Chemical, Ann Arbor, MI) as an internal standard to determine extraction e ciency.Samples were placed on ice in darkness for 2 hours then individually homogenized.Homogenates were then centrifuged at 19,000g for 20 minutes at 20°C.Supernatants were decanted and diluted with HPLC H 2 O to make a 75:25 water to supernatant solution.Partial puri cation was achieved using C-18 solid phase extraction columns (Agilent Technologies, Lake Forest, CA).A series of 4 elutions with 1.5 mL of 60%, 75%, 85%, and 100% methanol were collected for analysis.Samples were analyzed using an Applied Biosystems API 3000 triple quadrupole mass spectrometer with electrospray ionization.20µL from each elution were chromatographed using an XDB-C18 reversed phase HPLC analytical column (Agilent) and optimized mobile phase gradients.

Statistics
All data are presented as mean ± SEM.Statistical tests were performed with Prism 7 (GraphPad Software, La Jolla, CA, USA).Detailed statistical analyses are listed in the Suppl.Table 1.Normality was assessed by the D'Agostino-Pearson test.Depending on the experimental design, data were analyzed using either Student's t-test (paired or unpaired) with equal variances, one-way ANOVA or two-way ANOVA.The signi cance threshold was automatically set at p < 0.05.ANOVA analyses were followed by Bonferroni post hoc test for speci c comparisons only when overall ANOVA revealed a signi cant difference (at least p < 0.05).

Results
NAPE-PLD is functionally expressed in the brain and mediates motivational food-responses.
In order to explore the role of the NAEs-synthetizing enzyme NAPE-PLD in food-motivated behaviors, we rst addressed the consequence of whole-body NAPE-PLD developmental knock-out in a food-reward seeking behavioral paradigm.Here, we used the Napepld f/f mouse line in which the two LoxP sites span the exon 3 (24,30,41), the gene sequence that encodes for the catalytic activity of the enzyme and that e ciently leads to a reduction of NAPE-PLD-derived bioproducts (18, 24, 30, 41).Napepld f/f mice were bred with mice expressing Cre under the pan-promoter phosphoglycerate kinase 1 (Pgk-Cre) which result in whole-body full knock-out (KO, ( 34)) of NAPE-PLD in subsequent generations.To study the reinforcing and motivational properties of food, we performed an operant conditioning paradigm where animals were trained, under different schedules, to press a lever to obtain a palatable sugar pellet.In both a xed ratio 1 schedule (FR1, 1 lever press for 1 sugar pellet during 4 daily sessions) or a progressive ratio schedule (PR), which assesses the motivational component of reinforcement behaviors, Napepld +/+ (controls) and Napepld KO mice displayed similar performances (Suppl.Figure 1A, B).These results suggest that either physiological compensations occurred during development, as reported by previous genetic invalidations (27), and/or that, despite the expression of NAPE-PLD in the brain (23), brain NAPE-PLD plays a marginal role in reward-seeking behavior.To disentangle these two hypotheses, we moved to brain-restricted ablation of NAPE-PLD.Ablation of NAPE-PLD in the central nervous system (CNS) was achieved by crossing Napepld f/f mice with mice expressing Cre under the control of the promoter Nestin (Nes Cre+/− mice) (31,32).We observed that CNS genetic deletion of NAPE-PLD was associated with an enhanced response to operant behavior.In fact, under FR1 schedule, Napepld ΔCNS mice (males and females) collected a higher number of pellets and had a higher number of active lever presses (Fig. 1A, A 1 ).However, this enhanced reward-like phenotype was not related to differences in learning (% of active lever over inactive lever) as both genotypes were characterized by very similar discrimination scores (Fig. 1A 2 ).
Once the operant conditioning established, mice were moved to the PR schedule.Again, we noticed that, despite similar learning scores, Napepld ΔCNS mice showed enhanced performances (number of rewards and active lever presses) compared to control mice (Fig. 1B, B 1 , B 2 ).To exclude that such phenotype was driven by the presence of Cre (Nes Cre+/− (32, 42)) rather than the proper deletion of NAPE-PLD, we performed the same behavioral battery in Nes Cre−/− (controls) and Nes Cre+/− mice which both displayed a very similar phenotype on this paradigm (Suppl.Figure 1C, D), thus indicating that genetic deletion of neuronal NAPE-PLD is responsible for the enhanced reward-behavior observed in Napepld ΔCNS mice.This result revealed that tissue-speci c ablation of NAPE-PLD generates different outcomes than whole-body gene deletion.This nding, which points to an effective role of brain NAPE-PLD in food-reward seeking behaviors, also raises the possibility that the contribution of NAPE-PLD in multiple organs (full KO mice) might lead to physiological adjustments eventually driving opposite consequences on a particular behavioral output with an overall mitigated consequence.
Among the main organs that might contribute to food-dependent reward processes, the gut has emerged as a critical modulator of reinforced behaviors (3,36,43,44).It has been previously shown that mice with a speci c and inducible deletion of NAPE-PLD in the intestinal epithelial cells (IEC) (Napepld ΔIEC ) exhibited a phenotype associated with speci c changes in the homeostatic regulation of food intake and altered metabolic adaptations to high-fat diet (18, 45).Therefore, we explored the potential contribution of intestinal NAPE-PLD in reward-seeking behavior.Interestingly, Napepld ΔIEC and control mice showed comparable performances in the operant conditioning paradigm (Suppl.Figure 1E, F), indicating that, while intestinal NAPE-PLD is critical for metabolic control (18) and short-term regulation of food intake (45), brain NAPE-PLD might represent a more direct target as acute regulator of food-reward behaviors.
Reinforced behaviors tightly depend on key brain regions that constitute the reward system, notably the midbrain dopamine (DA)-producing ventral tegmental area (VTA) and its dopaminoceptive structures, including the dorsal striatum (DS)/nucleus accumbens (NAc), the prefrontal cortex (PFC) and the hippocampus (Hippo) (46).Therefore, we rst investigated whether NAPE-PLD-produced NAEs were altered within these structures in Napepld ΔCNS mice.Lipidomic analyses revealed a signi cant decrease of several NAEs species [AEA, OEA, PEA, N-stearoylethanolamine (SEA), N-linoleoylethanolamine (LEA) and N-docosahexaenoylethanolamine (DEA)] in the midbrain VTA following CNS deletion of NAPE-PLD (Fig. 1C-H).Interestingly, either no major differences (DS/NAc and PFC), speci c signi cant reductions (SEA, LEA and DEA for the hippocampus) or trends of decrease were detected in the levels of NAEs in the other reward-associated brain regions (Fig. 1C-H).Of note, no alterations were detected for the endocannabinoid (eCB) 2-AG (Fig. 1I).Moreover, within the VTA we did not detect alterations in the levels of other fatty acids (linoleic, arachidonic and oleic acids) (Suppl.Figure 2A-C) or N-acylamides (Suppl.Table 2), thus indicating that lack of NAPE-PLD speci cally affects a subset of endogenous bioactive lipids.
In addition, lipidomic analyses also revealed that NAEs, but not 2-AG (Fig. 1I), levels were higher in the VTA compared to the DS/NAc, PFC and hippocampus (Fig. 1C-H).
Altogether these observations underline the potential role for NAPE-PLD in the midbrain VTA as a regulator of food-associated reward processes.

VTA NAPE-PLD scales food-motivated behaviors and dopamine releasing dynamics
To precisely interrogate the structure-speci c functions of NAPE-PLD in driving food-motivated behaviors, we knocked-down the Napepld gene in the VTA using a local and virally mediated delivery of Cre in the VTA of Napepld f/f mice (Fig. 3A).Next, we tested the reinforcing and motivational properties of palatable food using a food-dependent operant conditioning paradigm.In line with the results obtained with Napepld ΔCNS mice (Fig. 1A, B), we observed that viral deletion of NAPE-PLD in the VTA promoted foodoperant conditioning (increased number of rewards and active lever presses) during both FR1 and PR schedules (Fig. 3B, C), with no differences in learning performances as both groups showed similar active/inactive discrimination index (Fig. 3B, C).
This enhanced reward phenotype was also present following a FR1→FR5→PR training schedule (Suppl. Figure 3A-C, food restriction) and even in sated conditions (Suppl.Figure 3D), therefore excluding the potentially confounding effect of hunger onto motivational drive.Importantly, this phenotype was also con rmed in female mice (Suppl.Figure 3E-H), again in both food-restricted and sated conditions.Of note, in both males and females, no signi cant differences in initial body weight and body weight loss (food restriction) were observed between experimental groups (Suppl.Figure 4A, B).
Aside from the motivational component, the liking and learning components of feeding are an integral part of food-reward processes (46).These components can be assessed through behavioral measurements of the positive valence assigned to palatable food in the conditioned-place preference (CPP, Fig. 3D) and T-maze (Fig. 3E) paradigms, which both rely on the association between reward value and context.In food-restricted conditions, we observed an increased and similar CPP score in both Napepld VTA−GFP and Napepld ΔVTA mice (Fig. 3F).However, in sated conditions, only Napepld ΔVTA mice showed an HFD-induced increase in CPP score (Fig. 3G), indicating enhanced susceptibility to the reinforcing properties of palatable foods.Using the T-maze paradigm, we next assessed the ability and exibility of mice to actively learn in discriminating between a rewarded (HFD) and a non-rewarded arm.During the learning phase ( rst 5 days), we observed that both groups showed a progressive increase in correct responses (%) over training days, with Napepld ΔVTA mice performing signi cantly better than Napepld VTA−GFP control mice (Fig. 3H).Then, mice were tested for their exibility to relearn the task under a reversal learning schedule (in which the food reinforcer was switched to the previously unreinforced arm of the T-maze).While both groups displayed good performance in learning/ exibility, VTA-speci c deletion of NAPE-PLD resulted in a better performance with a more rapid acquisition of the correct entry into the reinforced arm as compared to Napepld VTA−GFP control mice (Fig. 3H).
Next, we investigated the role of VTA NAPE-PLD in driving palatable food preference during a time-locked window (1h of exposure).First, we tested the reinforcing properties of the non-caloric sweetener sucralose (2 mM).As shown in Fig. 3I, Napepld ΔVTA mice consumed more sucralose than Napepld VTA−GFP control mice.A very similar pattern of enhanced preference was measured with the natural caloric sugar sucrose (10%, Fig. 3J) and with emulsi ed lipids (20%, Fig. 3K).
We therefore decided to investigate whether the enhanced reward-like behavior observed in Napepld ΔVTA mice was associated to an increased activity of the nucleus accumbens (NAc), a region highly innervated by VTA projections and whose activity is correlated with food-reward processes (46).However, the enhanced neural response within the reward system might result either from the higher tropism/consumption of palatable food of Napepld ΔVTA mice or from the increased rewarding value despite a xed amount of food-reinforcer.In order to dissociate these two possibilities, we exposed our experimental groups (sated conditions) to an equal amount of HFD during a time-locked window (1h during which all mice consumed the HFD pellet) and then we analyzed the induction of cFos, a molecular proxy of neuronal activity, in the NAc (Fig. 3L).Interestingly, we detected more cFos-positive neurons in the NAc of Napepld ΔVTA mice (Fig. 3L), thereby indicating an enhanced responsiveness of the VTA→NAc mesolimbic axis to an equal amount of food-reward consumption.
These results led us to hypothesize that VTA NAPE-PLD and its local NAEs bioproducts may contribute to the regulation of DA dynamics within the VTA→NAc mesolimbic axis.To test this hypothesis, we took advantage of in vivo ber photometry coupled to virally expressed DA biosensors (GRAB-DA2m ( 35)) to measure DA dynamics in the NAc of Napepld VTA−GFP and Napepld ΔVTA mice (Fig. 4A, B).First, we observed that exposing both fasted (Fig. 4C, D) and ad libitum fed mice (Fig. 4E, F and Suppl.Figure 5A) to HFD triggered a higher DA accumulation/release in the NAc of Napepld ΔVTA mice compared to control animals.
Second, to further explore whether and how NAPE-PLD may contribute to the regulation of DA-dependent events, we tested in vivo DA dynamics also in two non-food-dependent paradigms: the administration of cocaine (Fig. 4G, H) and the tail suspension (TS, Fig. 4I, J and Suppl.Figure 5B).In both cases, we observed that Napepld ΔVTA mice were characterized by an enhanced accumulation/release of DA in the NAc than Napepld VTA−GFP mice.Lastly, we administered the selective DAT blocker GBR12909 and noticed an enhanced locomotor response in Napepld ΔVTA mice compared to controls (Suppl.Figure 5C), further con rming an ampli ed DA release/tone as a consequence of VTA NAPE-PLD knock-down.
Overall, these results indicate that VTA NAPE-PLD tightly contributes in orchestrating the responses of midbrain DA-neurons to both food-and non-food-related reinforcers by promoting and boosting the release of DA at VTA→NAc synapses.

VTA NAPE-PLD contributes to the regulation of food intake and energy homeostasis
Although the regulation of energy homeostasis has been classically ascribed to the hypothalamus and brainstem (1), new evidence indicates that the reward system also strongly contributes in scaling wholebody metabolic functions (38,52).We therefore explored the metabolic consequences of VTA NAPE-PLD knock-down in the regulation of whole-body metabolic e ciency and peripheral substrates utilization by using longitudinal measurements of indirect calorimetry.As previously observed (Suppl.Figure 4), no major differences were observed in body weight and body composition between the two experimental groups (Fig. 5A).However, Napepld ΔVTA mice displayed a spontaneous increase in locomotor activity and in cumulative food intake compared to Napepld VTA−GFP mice during both the light and dark circadian phases (Fig. 5B, C).These phenotypes were associated with an overall enhanced energy expenditure (Fig. 5D) and to a change in peripheral substrates utilization favoring carbohydrates over lipids-based substrates as indicated by the increase in respiratory exchange ratio (RER, 1 = glucose substrate, 0.7 = lipids substrate) during the light phase (Fig. 5E) and the consequent decrease in fatty acid oxidation (FAO) in Napepld ΔVTA mice during both the light and dark phases (Fig. 5F).This feature was also associated with enhanced glucose tolerance during an oral glucose tolerance test at the expense of lower insulin release, suggesting enhanced whole-body glucose dynamics and insulin sensitivity (Suppl.Figure 6).
We then decided to investigate how Napepld ΔVTA mice adapted during manipulation of nutrients availability.We noticed that during a food deprivation period (overnight fasting) Napepld ΔVTA mice were still characterized by increased locomotor activity (Fig. 5G) and energy expenditure (Fig. 5I), but with no differences in RER (Fig. 5J).While the capability to mobilize lipids-based substrates during the fastinginduced lipolysis was similar between the two groups, as indicated by the RER (Fig. 5J), the proportion of lipids used a primary source of fuel was enhanced in the fasting period as indicated by the FAO (Fig. 5K), thus indicating a metabolic shift toward lipids-based substrates utilization.Interestingly, upon refeeding, mice displayed similar food intake (Fig. 5H), locomotor activity (Fig. 5G) or substrates utilization (Fig. 5J,   K), while a slight increase in energy expenditure was still detected in Napepld ΔVTA mice (Fig. 5I).These results con rmed the hypothesis that the integrity of NAPE-PLD within the VTA was required for the proper metabolic adaptation to changes in nutrients availability.
Since fasting increases the motivational drive and responsiveness to food, we wondered whether the expression of NAPE-PLD was required to promote DA releasing dynamics in fasted mice exposed to a chow pellet.In contrast to the acute response to palatable HFD (Fig. 4C-F), consumption of a chow pellet resulted in similar DA releasing dynamics in both Napepld VTA−GFP and Napepld ΔVTA mice (Fig. 5L, M).This may suggest that (i) the action of VTA NAPE-PLD in the modulation of adaptive metabolic responses to nutritional manipulations can be dissociated from DA release in the fast-refeeding transition and/or (ii) VTA NAPE-PLD plays an active role in discriminating between palatable (HFD, Fig. 4C-F) and regular (chow, Fig. 5L, M) foods through the control of reward-dependent DA dynamics.

VTA NAPE-PLD does not contribute to exercise-motivated behaviors but still regulates energy homeostasis
In mammals, exercise can function as a rewarding/motivational stimulus (53) and the eCBs system, especially within the VTA, has been identi ed as a key regulator of exercise-induced reinforced behaviors (54)(55)(56).We thus decided to extend our investigations to exercise-motivated behaviors (i) to investigate whether VTA NAPE-PLD was also important in mediating the reinforcing properties of exercise and (ii) to study whether metabolic adaptations observed in Napepld ΔVTA mice (Fig. 5) were solely dependent on enhanced locomotor activity.First, we performed a time-locked access (30 min session/day) to a running wheel.Despite both Napepld VTA−GFP and Napepld ΔVTA mice progressively spent more time wheel-running, we surprisingly noticed a reduced performance in Napepld ΔVTA mice compared to control animals (Fig. 6A, A 1 ).This led us to investigate whole-body metabolism and metabolic e ciency in calorimetric chambers equipped with running wheels.Again, we observed that Napepld ΔVTA mice were characterized by an enhanced spontaneous locomotor activity (Fig. 6B, B 1 ) and reduced wheel-running activity (Fig. 6C, C 1 ).When combining both forms of activity (spontaneous + wheel running activities), we detected no differences in the light phase and a lower global activity in Napepld ΔVTA mice during the dark phase (Fig. 6D).Of interest, the peculiar metabolic signature associated with VTA NAPE-PLD deletion also remained in this exercise-based paradigm and was characterized by enhanced energy expenditure (Fig. 6E, E 1 ), food intake (Fig. 6F) and RER (Fig. 6G), and lower FAO (Fig. 6H) in Napepld ΔVTA mice.
Altogether, these results indicate that the role of VTA NAPE-PLD in regulating reward-like processes cannot be generalized to all natural rewards (food vs exercise) and that the metabolic adaptations observed in VTA NAPE-PLD-deleted mice are not solely dependent on locomotor activity.
VTA NAPE-PLD controls metabolic adaptation to an obesogenic environment.
Food-reward drive, together with changes in metabolic outputs in response to food environment, are important contributors to the obesity pandemics.Given the above-mentioned results showing a key role of VTA NAPE-PLD in controlling reward and metabolic processes, we hypothesized that NAPE-PLD may in uence the (mal)adaptive responses to an obesogenic environment.Thus, Napepld VTA−GFP and Napepld ΔVTA mice were chronically exposed to an obesogenic diet (3 months of HFD) and then metabolically characterized.
First, we noticed no signi cant differences in the body weight and lean mass composition of both HFDexposed experimental groups (Fig. 7A).However, fat body mass was signi cantly lower in Napepld ΔVTA mice (Fig. 7A).The analysis of metabolic e ciency revealed that obese Napepld ΔVTA mice displayed increased nocturnal locomotor activity (Fig. 7B) and enhanced nocturnal food intake (Fig. 7C, C 1 ).Surprisingly, we detected a higher energy expenditure (Fig. 7D, D 1 ) and FAO (Fig. 7E, E 1 ) in obese Napepld ΔVTA mice, whereas the RER resulted unchanged (Fig. 7F).This metabolic blueprint suggests that, depending on diets (chow vs HFD) and metabolic pro les (lean vs obese), VTA NAPE-PLD readily allows the plastic adaptation of nutrient partitioning (Fig. 5E-F vs Fig. 7E-F) in order to maintain a higher energy expenditure.
These results indicate that, within the VTA, the deletion of NAPE-PLD partially protects against dietinduced obesity.

Discussion
NAEs represent an important family of endogenous bioactive lipids involved in several biological processes including adaptive stress responses and emotional states (21,57), pain (58), in ammation (59), feeding and metabolism (9,14,17,18,45).In an effort to characterize the phospholipase D (PLD)mediated enzymatic activity that converts NAPE into NAEs, in vitro studies identi ed NAPE-PLD as able to produce NAE derivatives from NAPE (7).Genetic invalidation of the Napepld gene (27,30) revealed that, in the brain, several alternative enzymatic pathways exist for the synthesis of polyunsaturated fatty acid NAEs.NAPE-PLD activity seems more critical for saturated/monounsaturated NAEs, with a drastic decrease of these compounds escalating with carbons chain length upon loss of NAPE-PLD activity.Later, in depth lipidomic analysis revealed a broader role for NAPE-PLD with a large spectrum and regionspeci c consequences in lipidome alteration in the brain (24,30).While constantly evolving, the sensitivity and limitations of quantitative methods to measure NAEs may underlie the di culty in formally assigning a de nitive set of substrates and bioproducts to NAPE-PLD.It is now clear that NAPE-PLD bioproducts include important lipid mediators which, by acting on a variety of transcriptional and signaling cascades, control cellular and physiological responses (5).
In the present study we explored the role of NAPE-PLD in food-and reward-driven behaviors.Whole-body deletion of NAPE-PLD did not affect food-reward operant conditioning, suggesting that compensatory mechanisms, as previously reported with a model of developmental NAPE-PLD KO mice (27), may be at play.However, using neural-speci c genetic deletion (Nestin-Cre) or midbrain-speci c viral invalidation approaches we revealed that the integrity of NAPE-PLD within the midbrain VTA is required for NAEs synthesis (AEA, OEA, PEA, SEA, LEA and DEA) and that NAPE-PLD acts as a gatekeeper for ne-tuning food-reward behaviors and for regulating, at least as a contributor, energy balance and whole-body metabolism.In fact, viral down-regulation of NAPE-PLD in the VTA was associated with an enhanced tropism towards palatable foods and a stronger conditioning for food-related rewards (operant conditioning, conditioned-place preference and T-maze paradigm).These phenomena were associated with an enhanced activity of midbrain VTA DA-neurons and their DA release dynamics in response to food rewards and also to non-food-related stimuli.Consistent with the notion of a region-speci c biosynthesis and action of NAEs, RNA-seq meta-analyses revealed low levels of Napepld in postsynaptic striatal dopaminoceptive neurons (Drd1-and Drd2-MSNs), but a higher expression of Napepld in VTAneurons, notably DA-and glutamate-neurons.
Previous studies focusing on nicotine reinforcement and tobacco use disorder (TUD) have shown that pharmacological inhibition of the fatty acid amide hydrolase (FAAH), one of the main enzymes responsible for the degradation of NAEs, reduces nicotine-enhanced DA transmission and nicotine reinforcement (19,60) through the activation of PPARα by OEA/PEA and the activation of intracellular cascades, leading to the reduction of nicotinic receptors onto midbrain DA-neurons (19,60).In line with our results showing increased dopaminergic VTA→DA transmission following downregulation of VTA NAPE-PLD, these electrophysiological studies have clearly demonstrated that OEA/PEA inhibit DAneurons, whereas inhibition of PPARα promotes their spontaneous activity (19,61,62).However, these aforementioned reports, although seminal, did not formally test nor establish the role of midbrain NAPE-PLD in these processes.In our hands, speci c knock-down of NAPE-PLD in the VTA enhanced DA transmission in response to reward stimuli.In that view, our results are perfectly in line with a putative role for NAPE-PLD-derived substrates as negative modulators of VTA DA-neurons.While our study provides an additional mechanism by identifying NAPE-PLD as a potential candidate, it also extends the role of this enzyme to the control of DA-dependent behaviors and DA releasing dynamics in response to reward stimuli well-beyond nicotine.Although our results do not rule out whether the NAPE-PLD/NAPE→NAEs machinery controls both tonic and phasic DA release (63-65), they clearly indicate that VTA NAPE-PLD activity, through the synthesis of NAEs, is an integral component of the control of DAneurons activity.Whether the cellular accumulation of NAPE and/or the decreased levels of NAEs are the primary responsible for the changes in DA-dependent behaviors and DA releasing dynamics is still unknown.Indeed, NAPE-PLD silencing seems to confer a protective action for increased NAPE species in 6-OHDA-induced neural damage (20), suggesting that the regulation of DA-neurons might be linked to NAPE/NAE membrane homeostasis.Given the multiple roles of endogenous bioactive lipids, it is possible that the invalidation of NAPE-PLD in the VTA may lead to an imbalance in the NAPE/NAEs ratio with ultimate consequences on a variety of processes including heightened DA responses to reward stimuli.In addition, NAEs, either cannabinoid-like (AEA) and non-cannabinoid-like (OEA, PEA), may be released and act through several modus operandi.In fact, while the eCB AEA is retrogradely released on demand (66), the non-eCB NAEs may act anterogradely as suggested by the presence of NAPE-PLD also in pre-synaptic terminals (67), thus potentially modulating both intra-VTA microcircuits and/or VTA-projecting circuits (i.e.VTA→NAc).However, the VTA does not only harbor DA-neurons (49,68).Importantly, the relatively high expression of NAPE-PLD in VTA Glut-neurons may also indicate a possible regulatory control of this cell type by the NAPE/NAEs ratio in the observed behavioral and metabolic features.This regulation may be exerted either through the local communication among all VTA-neurons or even through VTA Glut →NAc projections.In fact, these glutamatergic projections were recently shown to promote reinforcement independently of DA release (69) and we cannot exclude that NAPE-PLD-dependent mechanisms onto these complex neural networks may result from this additional form of VTA→NAc communication in encoding changes in reward-driven behaviors and metabolic e ciency.Despite this limitation, our in vivo imaging results clearly reveal that the VTA→NAc dopaminergic transmission is regulated by the NAPE-PLD.Indeed, further studies are warranted to fully ush out how and to which extent midbrain NAPE-PLD regulates DA events by selectively focusing on the local interconnectivity and interdependency of VTA DA-, Glut-and GABA-neurons.
While our study establishes a direct role for NAPE-PLD in the mesolimbic reward circuit in regulating DA release and DA-dependent behaviors, it also unveils the functional connection between VTA NAPE-PLD activity and the control of whole-body metabolism.NAPE-PLD ΔVTA mice displayed increased spontaneous locomotor activity in both the fed and fasting, but not refed, conditions, which is consistent with an enhanced activity of VTA DA-neurons (70).These features were associated with increased cumulative food intake and whole-body energy expenditure and, together with other recent studies (52,71,72), they point to the VTA as an important regulator of energy balance and metabolic e ciency.In fact, on chow diet the overall body weight was only marginally affected in NAPE-PLD ΔVTA compared to control mice, suggesting that increased energy expenditure (EE) was compensated by increased energy intake in a closed and well-balanced homeostatic regulation (Fig. 5).During food deprivation, NAPE-PLD ΔVTA mice showed a drastic increase in fasting-induced foraging, pointing towards a change in adaptive strategy in response to decreased nutrient availability.This increased activity was associated with increased EE and is most likely fueled by enhanced lipid-based metabolism (Fig. 5).In mice exposed to HFD, a similar increase in spontaneous activity, EE and FAO was observed together with increased food intake.Since NAPE-PLD ΔVTA mice show increased tropism and responsiveness to HFD and palatable food it is possible that this increase in HFD intake may be the result of enhanced palatability for energy-dense foods.An alternative, and not mutually exclusive, hypothesis is that increased food intake is a homeostatic mechanism to sustain enhanced EE.Although we do not provide clear evidence to disentangle these two hypotheses, the fact remains that NAPE-PLD knock-down in the VTA confers a protective phenotype against HFD-induced body fat gain and metabolic alterations (Fig. 7).This result nicely echoes a study in humans that has identi ed a common haplotype of the Napepld gene in severe obesity (28).In view of our results, one should consider that VTA NAPE-PLD deletion led to a protective effect against HFDmediated fat mass gain and metabolic (mal)adaptations but was accompanied by enhanced rewarddriven behaviors.This is important from a translational point of view as obesity is characterized by an alteration of peripheral and central eCBs and NAEs in humans (73)(74)(75).Our current results, together with those depicting a role of NAPE-PLD in peripheral tissues, culminate with the elaboration of a complex picture in which organ-and region-speci c homeostasis of NAPE/NAEs underline the complexity by which NAPE-PLD exerts its control onto reward-dependent behaviors, energy balance and body weight control.Indeed, it has been previously shown that mice lacking NAPE-PLD speci cally in adipocytes displayed spontaneous obesity, higher fat mass, glucose intolerance, and lower adipocyte browning (41).
Moreover, mice lacking NAPE-PLD in intestinal epithelial cells were more sensitive to HFD-induced body weight gain, fat mass gain and hepatic steatosis (18), a phenomenon partially explained by an alteration in food intake behavior (45).
In conclusion, our study provides direct evidence for a key role of NAPE-PLD in the control of rewarddependent behaviors, DA dynamics and energy metabolism.The main limitation of this study primarily lies in the lack of a clear cell type-speci c identi cation of cellular and molecular mechanisms occurring within the heterogenous VTA.Given the complexity of NAEs action and the variety of bioactive lipids and signaling cascades associated with NAPE-PLD activity, further research and new investigatory tools will be required to fully comprehend the role of NAPE-PLD and its bioproducts in promoting anti-obesity strategies.***p<0.001for Napepld ΔCNS vs Napepld f/f mice; ### p<0.001 for VTA vs other brain structures (Napepld f/f mice).For number of mice/group and statistical details see Suppl.Table 1.For number of mice/group and statistical details see Suppl.Table 1.

Declarations Figures
Figure 7