Zagmutt et al. Biology of Sex Differences (2023) 14:14 Biology of Sex Differences https://doi.org/10.1186/s13293-023-00498-8 RESEARCH Open Access CPT1A in AgRP neurons is required for sex‑dependent regulation of feeding and thirst Sebastián Zagmutt1,2, Paula Mera1,2,3, Ismael González‑García4, Kevin Ibeas1,2, María del Mar Romero1,2,3, Arnaud Obri5, Beatriz Martin6, Anna Esteve‑Codina6,7, M. Carmen Soler‑Vázquez1,2, Marianela Bastias‑Pérez1,2, Laia Cañes8,9,10, Elisabeth Augé1,11, Carme Pelegri1,11,12, Jordi Vilaplana1,11,12, Xavier Ariza3,13, Jordi García3,13, José Martinez‑González8,9,10, Núria Casals3,14, Miguel López3,4, Richard Palmiter15, Elisenda Sanz16,17, Albert Quintana16,17, Laura Herrero1,2,3 and Dolors Serra1,2,3*    Abstract Background Fatty acid metabolism in the hypothalamus has an important role in food intake, but its specific role in AgRP neurons is poorly understood. Here, we examined whether carnitinea palmitoyltransferase 1A (CPT1A), a key enzyme in mitochondrial fatty acid oxidation, affects energy balance. Methods To obtain Cpt1aKO mice and their control littermates, Cpt1a(flox/flox) mice were crossed with tamoxifen- inducible ­AgRPCreERT2 mice. Food intake and body weight were analyzed weekly in both males and females. At 12 weeks of age, metabolic flexibility was determined by ghrelin-induced food intake and fasting–refeeding satiety tests. Energy expenditure was analyzed by calorimetric system and thermogenic activity of brown adipose tissue. To study fluid balance the analysis of urine and water intake volumes; osmolality of urine and plasma; as well as serum levels of angiotensin and components of RAAS (renin–angiotensin–aldosterone system) were measured. At the central level, changes in AgRP neurons were determined by: (1) analyzing specific AgRP gene expression in RiboTag– Cpt1aKO mice obtained by crossing Cpt1aKO mice with RiboTag mice; (2) measuring presynaptic terminal formation in the AgRP neurons with the injection of the AAV1-EF1a-DIO-synaptophysin-GFP in the arcuate nucleus of the hypo‑ thalamus; (3) analyzing AgRP neuronal viability and spine formations by the injection AAV9-EF1a-DIO-mCherry in the arcuate nucleus of the hypothalamus; (4) analyzing in situ the specific AgRP mitochondria in the ZsGreen-Cpt1aKO obtained by breeding ZsGreen mice with Cpt1aKO mice. Two-way ANOVA analyses were performed to determine the contributions of the effect of lack of CPT1A in AgRP neurons in the sex. Results Changes in food intake were just seen in male Cpt1aKO mice while only female Cpt1aKO mice increased energy expenditure. The lack of Cpt1a in the AgRP neurons enhanced brown adipose tissue activity, mainly in females, and induced a substantial reduction in fat deposits and body weight. Strikingly, both male and female Cpt1aKO mice showed polydipsia and polyuria, with more reduced serum vasopressin levels in females and without osmolality alterations, indicating a direct involvement of Cpt1a in AgRP neurons in fluid balance. AgRP neurons from Cpt1aKO *Correspondence: Dolors Serra dserra@ub.edu Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 2 of 26 mice showed a sex-dependent gene expression pattern, reduced mitochondria and decreased presynaptic innerva‑ tion to the paraventricular nucleus, without neuronal viability alterations. Conclusions Our results highlight that fatty acid metabolism and CPT1A in AgRP neurons show marked sex dif‑ ferences and play a relevant role in the neuronal processes necessary for the maintenance of whole-body fluid and energy balance. Highlights – Fatty acid metabolism and CPT1A in AgRP neurons show marked sex-dependent differences in the control of feeding. – Cpt1a gene deletion in AgRP neurons increases energy expenditure in females but not in males. – CPT1A in AgRP neurons is involved in the control of thirst and fluid homeostasis. – Cpt1a gene deletion in AgRP neurons induces morphological, mitochondrial, and gene expression alterations in a sex-dependent manner. Keywords CPT1A, Fatty acid metabolism, AgRP neurons, Energy balance, Food intake, Thirst Background terminalis, comprising the subfornical organ (SFO), orga- Food intake and whole-body energy balance are regu- num vasculosum of the lamina terminalis (OVLT), and lated by the brain through a sophisticated neuronal net- the median preoptic nucleus (MnPO). Neurons from work located mostly in the hypothalamus. In particular, these nuclei project to paraventricular nucleus of hypo- the hypothalamic arcuate nucleus (ARC) is a fundamen- thalamus (PVH) where the neurons that produce vaso- tal sensor of the hormones and nutrients that indicate pressin (VP, also known anti-diuretic hormone, ADH) the energy status of the organism. The ARC contains two are located [20]. The involvement of AgRP neurons in populations of neurons with opposite functions: anorexi- thirst behavior is poorly understood. It is known that genic proopiomelanocortin (POMC)-expressing neurons AgRP activation induces water intake in the presence of and orexigenic agouti-related protein (AgRP)-expressing food, but this activation is not sufficient to promote water neurons hypothalamus [1–4]. AgRP neurons promote a intake independent of food consumption [21], suggesting potent orexigenic response by opposing the actions of that food and drink behavior are tightly coordinated. POMC neurons, in part through the release of the AgRP Lipid metabolism has emerged as a key regulator neuropeptide, a competitive inhibitor of melanocortin of energy balance [22–24]. Although the oxidation of receptors [5, 6]. It also has an important effect on energy fatty acids (FAs) has been explored little in neurons, expenditure (EE) via modifying the activity of the sym- FAs undoubtedly have an important role through their pathetic nervous system (SNS) [3]. The outflow onto structural role in biological membranes, their participa- peripheral tissues leads to coordination of overall body tion as a messenger in neuronal signaling pathways and metabolism, including adipose tissues, liver and other their involvement in the energy supply [25, 26]. Here, we tissues that respond to AgRP neuronal activity. For that explore some key aspects of FA utilization in the control reason, the action of AgRP neurons on energy balance of energy balance and raise new questions on whether should be studied beyond the acute control of food intake FAs could be also a signal for hypothalamic neurons. It [7]. Although our understanding of the AgRP neuronal is well documented that most enzymes in FA metabolic mechanism controlling feeding behavior has expanded pathways are expressed in the hypothalamus, includ- immeasurably, few studies have explored sex-dependent ing AMP-activated protein kinase (AMPK), acetyl-CoA regulatory mechanisms, especially considering the sex carboxylase (ACC) carnitine palmitoyltransferase 1 differences in calorie intake and body weight [8–10] and (CPT1), fatty acid synthase (FAS), and malonyl-CoA new studies are necessary to reveal sex differences in decarboxylase (MCD) [27, 28]. Of note, fasting stimu- neuronal mechanisms that control energy balance. lates hypothalamic AMPK activity and inhibits ACC and In addition to food consumption, another physiologi- FAS activities, whereas re-feeding induces the opposite cal need essential for survival is thirst. Several brain areas [26, 29–31]. Moreover, the pharmacological and genetic involved in the control of fluid balance and thirst have manipulation of some of these genes/proteins has a been already elucidated [11–19], including the lamina Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 3 of 26 profound impact on food intake and whole-body energy to generate (Cpt1a(flox/flox); AgrpCreERT2) mice and their homeostasis [32–34]. control littermates (Cpt1a(+/+); ­AgRPCreERT2), which Here, we focused on CPT1A, a key regulatory enzyme were called Cpt1aKO and control mice, respectively. All localized in the outer mitochondrial membrane that animals were on the C57BL/6J background. To evalu- is involved in the uptake of long-chain fatty acids ate AgRP neuron-specific RNA expression, we took (LCFAs) into the mitochondrial matrix for entry into advantage of the RiboTag method [42]. We bred the the β-oxidation spiral [35]. Although this isoform is RiboTag mouse line with Cpt1aKO mice to generate expressed mainly in the liver, there is evidence that its RiboTag;Cpt1a(flox/flox); ­AgRPCreERT2 (RiboTag-Cpt1aKO) expression in the brain could play an important role in mice and their RiboTag;Cpt1a(+/+); ­AgRPCre ERT2 con- peripheral metabolism. Previous results from our labora- trol littermates. These transgenic mice expressed, in tory showed that the long-term expression of a perma- an inducible manner, epitope-tagged ribosomal pro- nently activated CPT1A isoform in the hypothalamus tein L22-hemagglutinin (RPL22-HA) in only the AgRP triggers hyperphagia, leading to an overweight state [36]. neurons. Consistent with these results, the intracerebroventricular To study mitochondrial morphology specifically in the administration of 3-carboxy-4-alkyl-2-methylenebutyrol- AgRP neurons, we generated a Rosa26-flox-stop-MLS- actone (C75), a synthetic inhibitor of fatty acid synthase ZsGreen mice (ZsGreen mice). A Cre-dependent mito- (FAS) and CPT1A activity, has been reported to reduce chondrial-ZsGreen-expressing mice that allows targeting body weight and food intake by altering the melanocortin of ZsGreen to the mitochondrial matrix. The mitochon- system [37–39]. drial translocation sequence (MLS) from the Ndufs4 Most of the studies in this field lack cell type-specific- gene was fused in frame at the N-terminus of ZsGreen, ity and it is still unclear whether lipid metabolism spe- which was then inserted 3′ of a floxed Pgk-neomycin cifically regulates the physiological features modulated resistance (Neo) gene (for positive selection) in a transfer by AgRP neurons. To answer this question, we devel- plasmid. The MLS-ZsGreen-floxed Pgk-Neo region was oped an adult transgenic mouse model lacking CPT1A excised and inserted into a targeting vector that has the in their AgRP neurons. Here, we report that AgRP neu- CMV-chicken β-Actin promoter (CBAp) inserted at the rons require CPT1A to regulate some aspects of energy transcription start site of the Gt(Rosa26)Sor gene with balance, since female and male Cpt1aKO mice showed 7.7 kb of 5′ flanking and 4.1 kb of 3′ flanking Gt(Rosa26) reduced body weight gain compared with their control Sor sequence and a Pgk-DTa gene for negative selec- littermates. Notably, this regulation varied depending tion. This construct was linearized and electroporated on the sex. We also provide evidence that the metabolic into G4 embryonic stem (ES) cells. Correct gene target- changes induced by CPT1A deficiency in AgRP neurons ing was determined by Southern blot of DNA digested modify the fluid balance. Our results suggest that CPT1A with Nde1 using a probe that lies outside of the targeting in AgRP neurons affect food intake and the peripheral vector; 24 of 84 clones analyzed were correctly targeted. energy balance in a sex-dependent manner, highlighting ES cells from one correctly targeted clone were injected this enzyme as a possible target for therapeutic strategies into blastocysts of C57BL/6 recipients and chimeric pups that aim to decrease body weight and fight obesity. were bred with C57BL/6 mice. The action of Cre recom- binase deletes Pgk-Neo and allows expression of ZsGreen that is targeted to the mitochondria. We bred ZsGreen Materials and methods mice with Cpt1aKO mice to generate ZsGreen; Cpt1a(flox/ Mice flox) ; ­AgRPCre ERT2 (ZsGreen-Cpt1aKO) mice and their Mice were kept in the Unitat d’Experimentació Animal ZsGreen-Cpt1a(+/+); ­AgRPCreERT2 control littermates. facilities of the Universitat de Barcelona under stand- To induce Cre-ERT2 expression and avoid the possible ard laboratory conditions with free access to stand- toxic effect of tamoxifen, adult mice (8 weeks old) were ard chow diet (Harlan Ibérca, ref. 2014) and water at intraperitoneally (i.p.) injected with five doses of tamox- 22 ± 2 °C and 60% humidity in a 12-h light/dark cycle. ifen (150 mg/kg of body weight; Sigma, MO, USA) solved Mice were housed in groups to prevent isolation stress in corn oil. The first two injections were combined with unless otherwise stated. Mice animal welfare was regu- 24 h of food deprivation (separated by 48 h) to enhance larly monitored. Experiments were performed with 8- to Agrp promoter activity. The last three injections were 25-week-old male and female mice. administered daily with ad libitum access to food and To evaluate the role of FA metabolism in AgRP neu- water. Control mice were also injected with tamoxifen rons, we generated the time-dependent conditional solved in corn oil. All studies used age-matched litter- Cpt1aKO mouse model. Cpt1a(flox/flox) mice [40] and mates, randomly assigned to experimental groups. tamoxifen-inducible ­AgRPCreERT2 mice [41] were bred Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 4 of 26 Mice were genotyped by polymerase chain reac- again at 30 min. We monitored the eating time and food tion (PCR) using the primer listed in Additional file 7: intake for 1 h after the first injection [44]. using the QuickExtract™ DNA Extraction solution, fol- Table S1. Genomic DNA was extracted from ear samples Glucose and insulin tolerance tests lowing the manufacturer’s instructions. Each polymer- For the glucose tolerance test (GTT), mice at 14 weeks ase chain reaction (PCR) was conducted in a 20-µl final old were fasted ON. The baseline blood glucose con- volume containing 10 µl of REDExtracti-N-Amp PCR centration from a tail snip was measured using a hand ReadyMix (Sigma-Aldrich, ref. R4775), which includes all glucometer (Bayer, Contour XT, Cat# 83415194) and the reagents needed for PCR amplification. For the Cpt1a test strips (Bayer, Contour next, Cat# 84191389). A flox gene, the Cpt1a HomArm forward and reverse prim- blood sample was collected in a capillary tube (Deltalab, ers were used (Additional file 7: Table S1). The DNA frag- Cat#7301) and kept in serum separator tubes (Sarstedt, ments obtained were 990 bp for the WT Cpt1a amplified Cat# 201.280) at 4 °C to measure the insulin level. Sam- allele, 1030 bp for the Cpt1a flox allele and 219 bp for the ples were allowed to clot on ice for 15 min and then cen- recombined genomic DNA. For the ­AgRPCreERT2 gene, trifuged for 15 min at 5700 rpm. The supernatant was the AgRP CRE-ERT2 forward, reverse and control prim- collected and frozen at − 20 °C. Mice were injected i.p. ers were used (Additional file 7: Table S1). The obtained with 20% glucose (Baxter, Cat# 2B0124P) at 1.5 mg/g of amplicon was 514 bp from the WT mice and 323 bp for body weight. Blood glucose concentrations were meas- the ­AgRPCreERT2 allele. For the RiboTag gene, the RiboTag ured at 0, 15, 30, 60, 90 and 120 min after the glucose loxP forward and reverse primers were used (Additional injection. For the insulin tolerance test (ITT), mice were file 7: Table S1). The obtained amplicon was 260 bp from fasted for 6 h. Mice were i.p. injected with 1.5 U of insu- the WT mice and 290 bp for the RiboTag allele. For the lin/kg of body weight. Blood glucose concentrations were ZsGreen gene, the ZsGreen forward and reverse prim- then measured at 0, 15, 30, 60, 90 and 120 min after the ers were used (Additional file 7: Table S1). The amplicon insulin injection. obtained was 600 bp from the WT mice and 500 bp for the ZsGreen allele. Indirect calorimetry Mice at 12 weeks old were analyzed for EE, respiratory Food intake and body weight analysis quotient (RQ) ­( VCO2/VO2) and locomotor activity (LA) All mice had ad libitum food supply and after the recom- using a calorimetric system (LabMaster; TSE Systems; bination the food was monitored weekly during the first Bad Homburg, Germany). Briefly, mice were placed in a month. The initial amount of compound pelleted fodder temperature-controlled (24 °C) cage with flowing air for was weighed at the beginning of every week, using the 48 h for acclimation before starting the measurements. same precision scale. At the following week, remaining After calibrating the system with the reference gases pellets were measured again with the same scale, and (20.9% ­O2, 0.05% ­CO2 and 79.05% ­N2), the metabolic rate food was added to give the same amount as in the previ- was measured for 3 days. After that, we extended the ous week. The body weight was monitored weekly until experiment for 12 h under fasting conditions and then killing using every week the same precision scale. 4 h under refeeding conditions. ­ O2 consumption and ­CO2 production were recorded every 30 min to indirectly Fasting–refeeding satiety test determine the RQ. EE, RQ, food intake and LA were This test was assessed in 12-week-old mice. Mice were measured during the dark and light phases. The LA was housed in individual cages 2 days before the beginning assessed using a multidimensional infrared light beam of the experiment. Mice were fasted overnight (ON) for system with the parameters defined by the LabMaster 12 h and then re-fed a pre-weighed meal. Food intake system. was measured at 30 min, 1 h, 2 h, 3 h and 4 h after refeed- ing. All measurements were weighed using the same pre- Non‑invasive measurement of blood pressure cision scale [43]. Systolic, diastolic and mean blood pressures were meas- mography method (CODA® tail-cuff blood pressure ured in conscious mice using the tail-cuff plethys- Ghrelin‑induced food intake test Mice at 12 weeks old were housed in individual cages system; Kent Scientific Corporation; Torrington, CT, two days before the beginning of the experiment. Mice USA). 12-week-old mice were trained for tail-cuff meas- were deprived of food for 2 h after the dark period and urements over a period of 4 days. Blood pressure meas- i.p. injected with ghrelin (0.4 μg/g of body weight; Merck urements were performed at the same time (between 9 Millipore, Cat# 494127-100G) in physiological saline a.m. and 12 a.m.) to avoid the influence of the circadian solution (B. Braun; Cat# 12260029_1019) at 0 min and Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 5 of 26 cycle. Blood pressure values were taken from at least ten posterior, ± 0.3 mm lateral and − 5.8 mm ventral to the consecutive measurements [45]. adhesive (3 M Vetbond™, Cat# 1469Sb) was used to close bregma [47]. Once the procedure was completed, a tissue Heat production was visualized using a high-resolution infrared camera (FLIR T420; FLIR Systems, AB, Swe- the incision. Mice were caged with ad libitum access to den). Infrared thermography images were taken from the food and post-surgical drug-supplemented water con- upper half of the body to specifically analyze BAT activ- taining 10% enrofloxacin (Bayer, Cat# 572126.2) and ity. On the day before the experiment, mice were fasted 0.3 g/400 ml of buprenorphine (Indivior, Cat# 679588) ON and shaved in the interscapular area to minimize as the antibiotic and analgesic, respectively. Daily moni- interference. Interscapular BAT temperature was ana- toring was used to follow up on the general state of lyzed within a fixed area (region of interest; ROI) using the operated mice for one week. Three weeks after the the Flir Tools software (version 4.1). For each image, the adeno-associated viruses (AAVs) injection, mice were average temperature of the skin area was calculated as treated with tamoxifen to induce the recombination. the average of 3 images/animal. Tissue collection Blood and urine collection To analyze gene expression, blood metabolites, protein Mice were individually placed in metabolic cages that had levels and the histological morphology of different tis- a floor area of 370 c­ m2 and a length of 207 cm, a width of sues, mice were fasted ON and anesthetized with 4% iso- 267 cm and a height of 140 cm. The cages were adapted flurane (Piramal Healthcare, Cat# 60307-120-25) before with a grid at the bottom to collect urine. 13-week-old being maintained at a surgical plane of anesthesia by the mice were housed one day before the start of the experi- anesthetic delivery machine (Combi-Vet® Rothacher continuous inhalation of 2% isoflurane using a calibrated ment, with ad libitum access to food and water. On the day of the study, drinking water and urine volumes were Medical, Switzerland). Blood was rapidly collected in measured for 24 h and urine specimens were taken for ing aorta using a 25-gauge needle (BD Microlance™ 3, heparinized tubes (Fibrilin, Cat# 0318) from the descend- osmolality analysis. At the end of the experiment, blood samples were collected in heparinized tubes through the Cat# 300600). It was allowed to clot on ice for 15 min facial vein method [46]. Plasma was collected and frozen and then centrifuged for 15 min at 5700 rpm at 4 °C. The at − 20 °C until analysis of the plasma osmolality. At the supernatant was collected and frozen at − 20 °C. Samples end of the study, all mice were rehoused in their origi- of liver, inguinal white adipose tissue (iWAT), gonadal nal cages. One week later, the mice were placed back in white adipose tissue (gWAT), BAT, adrenal gland (AG), the metabolic cages to perform the same protocol under pancreas, testis, ovary, kidney, hypothalamus, cortex and water restriction for 24 h. Urine volume was measured, hippocampus were collected and stored immediately at and blood and urine osmolality was also analyzed. Water − 80 °C until processing. A piece of the tissues was fixed intake and urine volumes are expressed as values per in formalin solution, neutral buffered, 10% (Sigma, Cat# gram of body weight. HT501128-4L) for 24 h and then transferred to 1X phos- phate buffered saline (PBS) (Sigma, Cat# D1408-500ML) Stereotaxic procedure for histological analysis. Mice at 8 week old were anesthetized using 0.1 mg/g To obtain ARC samples, the brain was gently placed in of ketamine (Richter Pharma Ag., Cat# 580393.7) and a mouse brain matrix (Agnthos, Cat# 69-2175-1), which 0.01 mg/g of xylazine (Bayer, Cat#580393.7) before being can be used to obtain coronal sections with a width start- placed in a stereotaxic apparatus (Kopf instrument, ing from 1 mm. A 3-mm coronal section encompassing Model 900 Small Animal Stereotactic Instrument). Once most of the ARC was obtained, taking as a reference the the head was shaved with an electric trimmer, a sagittal optical chiasm to establish them for dissection. Once incision was made through the skin along the midline of the section was cut, the remaining brain was carefully the head and a hole was drilled into the skull 1.5 mm pos- removed from the matrix and the section was extracted terior and 0.3 mm lateral (left and right) to the bregma and horizontally positioned. The area was dissected with (bregma point was found in the perpendicular intersec- a crosswise cut starting from the 3rd ventricle up to the tion between the sagittal and coronal synarthroses). A base of the hypothalamus. The tissue was immediately 400 nl of solution containing either AAV9-EF1a-DIO- stored at − 80 °C until processing. mCherry or AAV1-EF1a-DIO-synaptophysin-GFP were To perform immunofluorescence assays, whole ani- injected 50 nl/min using a Hamilton Neuros syringe mal perfusion fixation was applied. Briefly, animals (5 µl, Neuros Model 75 RN, point style 3, SYR, Cat# were anesthetized with an i.p. injection of ketamine 65460-02) and a microinjection pump over 8 min into (Richter Pharma Ag., Cat# 580393.7) and xylazine the ARC. The coordinates for the ARC were − 1.5 mm (Bayer, Cat#580393.7). The delivery dose was 0.1 mg/g Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 6 of 26 of ketamine and 0.01 mg/g of xylazine. Once the ani- Tissue Mini Kit (QIAGEN, #74804), following the man- mal reached the surgical plane of anesthesia, a 25-gauge ufacturer’s instructions. For the other tissues, the Tri- blunt perfusion needle connected to a perfusion pump zol reagent was used according to the manufacturer’s (Gilson, miniplus 3) was inserted into the left ventricle. protocol. RNA samples were heated to 55–60 °C for An incision into the right atrium was performed to cre- 10–15 min using a thermoblock (JP Selecta EN, Tem- ate as large an outlet as possible. 75 ml of cold 1X PBS block, Cat# 7462200), quantified using a NanoDrop were perfused to remove red blood cells. The PBS was ND-1000 spectrophotometer and stored at − 80 °C complementary DNA (cDNA) using TaqMan® Reverse then replaced with 50 ml of 4% paraformaldehyde (PFA), until processing. RNA was reverse transcribed into pH 7.4 (Sigma-Aldrich, Cat# 158127). Once the perfu- sion was complete, the brain was extracted from the skull Transcription reagents (Thermo Fisher Scientific, Cat# and fixed in 4% PFA for 4 h at 4 °C. Brains were kept in N808-0234), following the manufacturer’s instructions. 30% sucrose (Panreach Applichem, Cat#131621) until The cDNA obtained was diluted in RNase-free water up they sank. Then, they were frozen in dry ice using pre- to a concentration of 10 ng/µl. cooled 2-methylbutane (Merck, Cat# 277258) and stored Quantitative real-time polymerase chain reaction at − 80 °C. (qRT-PCR) was performed using the Power SYBR To isolate the mRNA from AgRP neurons in vivo, Green PCR Master Mix adapted for the LightCycler RiboTag-Cpt1aKO and control mice were fasted ON and 480 system (Roche, Cat# 4887352001), according to killed by cervical dislocation. The brain was removed the manufacturer’s indication for the LightCycler 480 from the skull and gently placed in the mouse brain instrument II (Roche, Cat# 05015243001). mRNA levels matrix. A 3-mm coronal section containing the ARC was from liver tissues were normalized against the β-actin obtained. The section was extracted and horizontally level, while those from BAT and WAT were normal- positioned, and the ARC was extracted with punches ized against the hypoxanthine–guanine phosphoribo- from the base of the hypothalamus. The tissue was imme- syl-transferase (Hprt) or β-actin level. Finally, for the diately stored at − 80 °C until processing. cDNA extracted from AgRP neurons, the housekeeping gene used was glyceraldehyde-3-phosphate dehydroge- Purification of genomic DNA nase (Gapdh). All the forward and reverse primers used Genomic DNA from the different tissues was extracted are described in Additional file 7: Table S1. using the proteinase K method. Briefly, 2.5 µl of pro- teinase K (0.0001 ng/µl in the reaction) (Thermo Fisher Scientific, Cat# AM2546) were used and the tissue was Protein extracts and western blot analysis incubated in 500 µl of lysis buffer (12.2 g of Tris, 1.9 g of Tissues were disrupted by adding 1 ml or 500 µl of a EDTA, 2 g of SDS and 11.7 g of NaCl made up to a 1-L protein extraction buffer to 30–70 mg of non-fat tis- solution, pH 8.5) at 55 °C for 4 h. Once digested, 10 µl of sues or 50–100 mg of fat tissues, respectively. The RNase A (10 mg/ml) (Sigma-Aldrich, Cat# 10109142001) protein extraction buffer contained 30 mM HEPES, were added to each tube and incubated at 37 °C for 1 h pH 7.4, 150 mM NaCl, 10% glycerol, 1% Triton X-100, for RNA degradation. DNA was extracted with 700 µl 0.5% sodium deoxycholate (DOC), a Mini Protease of phenol:chloroform:isoamyl alcohol (25:24:1) (Merck, Inhibitor Tablet (Roche, Cat# 11836153001) and a Cat# P3803-100ML) and precipitated by 700 µl of 2-pro- PhosSTOP Phosphatase Inhibitor Tablet (Roche, Cat# panol (Sigma-Aldrich, Cat# 190764-1L) and 15 µl of 04906837001). To disrupt the tissue, the TissueLyser tion of the samples were quantified using the Pierce™ 5 M NaCl (Sigma-Aldrich, Cat# S7653-5 KG). The DNA LT was used for 3 min at 50 Hz. Protein concentra- pellet was washed by ethanol (70%) (Panreac, Cat# 361086.1611). The genomic DNA pellet was resuspended BCA Protein Assay Kit (Thermo Fisher Scientific, in 100 µl of a 10-mM Tris–EDTA solution and stored Cat# 23227), following the manufacturer’s instruc- at 4 °C at least ON before any further processing. The tions. Protein lysates were separated in SDS-PAGE and genomic DNA yield was quantified using a NanoDrop transferred on to 0.45-µm nitrocellulose membranes 1000 spectrophotometer (Thermo Fisher Scientific, ref. (Bio-Rad Laboratories, Cat# 1620115), which were ND-1000). incubated ON with primary antibodies (Additional file 7: Table S2). Images were acquired with an Image- Total RNA extraction, cDNA synthesis and qRT‑PCR Quant LAS 4000 mini developer system (GE Healthcare Depending on the sample, RNA was extracted using the Life Sciences). The images obtained were processed Trizol reagent (Sigma-Aldrich, Cat# T9424) or with a with Fiji ImageJ1.33 software (NIH; Bethesda, MD, specific kit. For fatty tissues or tissues with a high lipid USA) to quantify the average optical density of all the content, RNA was extracted using the RNeasy Lipid immunoreactive bands. Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 7 of 26 Histopathology 2% (v/v) goat serum (Sigma-Aldrich, Cat# G9023). The Fixed tissues were dehydrated and paraffin-embedded. slices were incubated with the primary antibody in The resulting blocks were cut into 4-µm sections and blocking solution for 1 h at room temperature and ON stained with hematoxylin and eosin (H&E) to assess at 4 °C. They were then washed 3 times for 10 min with histology. Photomicrographs were obtained using a KPBS containing 0.1% Triton X-100 before being incu- microscope camera (Leica MC 190 HD Camera) and bated with the secondary antibody in blocking solution microscope (Leica DM IL LED Tissue Culture Micro- for 2 h at room temperature and protected from the light. scope). The images shown are representative of 10 bio- Finally, the samples were washed 3 times for 10 min with logical replicates per condition. KPBS containing 0.1% Triton X-100 and mounted onto SuperFrost Plus slides (Thermo Fisher Scientific, Cat# Assessment of urine and serum osmolality J7800AMNT) with fluoromount-G containing DAPI Urine osmolarity (­Uosm) and serum osmolality ­ (Posm) (Thermo Fisher Scientific, Cat# 00-4959-52) and cover were measured with 20-µl samples using the freezing- slipped. The primary and secondary antibodies are listed point depression technique involving an osmometer in Additional file 7: Table S2 showing the dilutions used (3320 Micro-Osmometer, Advanced Instruments). A in each incubation. Fos-positive cells and synaptophysin control (Clinitrol 290) and a set of calibration standards fluorescence in PVN sections were quantified using the (50, 850 and 2000 mosm/kg H ­ 2O) were used before run- Fiji ImageJ 1.33 (NIH; Bethesda, MD, USA), consistent ning each batch. with previous studies [48]. Plasma and urine analyses Dendritic spine analysis Plasma insulin (Insulin ELISA Kit, Alpco, Cat# For detailed morphological analyses of dendritic spines, 80-INSHU-E01.1), aldosterone (Aldosterone ELISA samples were imaged with a Zeiss LSM 800 confocal Kit, LSBio, Cat# LS-F28206-1), VP/ADH (Vasopressin microscope using a 1003 Plan Apo TIRF DIC-oil immer- ADH ELISA Kit, LSBio, Cat# LS-F7592-1), renin (Renin sion objective (total magnification of 63×). To visualize ELISA kit, Elabscience, Cat# E-EL-M0061), angioten- the mCherry protein in AgRP neurons, samples were sin (Angiotensin ELISA Kit, LSBio, Cat# LS-F67331- excited with a 587-nm laser and the fluorophore emission 1), leptin (Leptin ELISA kit, Sigma-Aldrich, Cat# was captured by a 610 band-pass filter. A Z-stack was RAB0334), 17-β-estradiol (Estradiol ELISA kit, Cayman, obtained for each dendrite extending from the apex of Cat# 501890), testosterone (Testosterone ELISA kit, the cell soma. For each animal, 20 dendrites (segmented Arbor Assays, Cat# K080-H1), glucose (Monlab, Cat# by 50 µm) were analyzed. Z-stacks were used for three- MO-165086), TGs and non-esterified fatty acids (NEFAs) dimensional reconstruction. The dendritic spine density (NEFA-HR kit, Wako, Cat#434-91795, 436-91995, 270- in each dendritic segment was quantified using the Fiji 77000) levels were all measured according to the manu- ImageJ 1.33 (NIH; Bethesda, MD, USA) and is expressed facturer’s instructions. as the dendrite number/50 µm. Immunohistochemistry Adipocyte area measurement To determine the tissue distribution of the protein of To quantify the adipocyte area, three representative interest, immunofluorescence for free-floating brain images from each adipose tissue section were taken with sections was performed. Frozen brains were embed- a 20× objective using a high-sensitivity camera (Leica ded in OCT (Tissue-Tek, Cat#4583) and 30-μm slices MC 190 HD Camera). Images were analyzed with the were obtained with a microtome (Leica SM2000R). Tis- Adiposoft software, a fully automated open-source pro- sue slices were stored at 4 °C in a cryoprotectant solution gram for the analysis of WAT cellularity in histological [30% (w/v) sucrose, 30% (v/v) ethylene glycol and 250 ml sections [49]. Each image was calibrated to 4.65 pixels of PBS] until the immunostaining was performed. Tissue per micron using a 20× objective and a Leica microscope. slices were washed 3 times for 5 min in 1X PBS to remove Any adipocyte with visible alterations in the membranes the cryoprotectant solution. All steps were performed was “closed” digitally prior to continuing with the auto- under gentle agitation in a shaker (Thermo-Shaker, PST- mated quantification. To quantify the adipocyte area, 100HL). Slices were permeabilized in potassium phos- images were analyzed and adipocytes were highlighted phate buffered saline (KPBS) [0.9% (p/v) NaCl, 52 mM if they met the following criteria: (1) the boundaries for potassium phosphate dibasic and 9.6 mM potassium sizing of the cell were 40–40,000; (2) the adipocyte had dihydrogen phosphate] containing 0.1% Triton X-100, for a shape factor of 0.35–1 (a shape factor of 0 indicated a 10 min and blocked for 1 h with blocking solution (KPBS straight line, while a shape factor of 1 indicated a perfect containing 0.1% (v/v) Triton X-100, 3% (w/v) BSA and circle); and (3) the adipocyte did not border the image Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 8 of 26 frame. The results are represented as the frequency dis- the Taqman probes for Agouti-related peptide (Agrp): tribution and the average of total area counted. Mm00475829_g1; aldehyde dehydrogenase 1 family mem- ber L1 (Aldh1/1): Mm03048957_m1; Discs large MAGUK Mitochondrial content scaffold protein 4(Dlg4): Mm00492193_m1; Glutamic Sections containing the ARC from ZsGreen mice at acid decarboxylase 1 (Gad1): Mm04207432_g1; Neu- 12 weeks old fasted ON were scanned with (Multiphoton ropeptide y (Npy): 00445771m1; Proopiomelanocortin Microscope Leica TCS SP8 MP). The LAXZ software was (Pomc): Mm00435874_m1; Solute carrier family 32 mem- used to obtain high-quality images. Images were acquired ber 1 (Slc32a1): Mm00494138_m1; Solute carrier family sequentially, using 405- and 488-nm laser lines, with a 17 member 6 also called Vesicular glutamate transporter 63× oil immersion objective. The confocal pinhole was 2 (Slc17a6): Mm 00499876m1 (Thermo Fisher Scientific). set at 1 Airy unit. The format was 1024 × 1024 pixel. Spe- cific settings for the frame averaging (6 frames per image) Quantification and statistical analysis and laser gain (900 for the 488-nm laser and 800 for the Statistical significance was determined using the Prism 405-nm laser) were selected for all the images to improve 8.0 software from GraphPad (GraphPad Software, La quality. For analysis, Z-stacks over the diameter of AgRP Jolla, CA, USA). Two-way ANOVA followed by post neurons were performed with a zoom of 4×. The analysis hoc analysis with the Sidak and Bonferroni test in the was performed using the Fiji ImageJ1.33 software (NIH; multiple comparison analysis were applied when more Bethesda, MD, USA). than two groups were compared. Student’s t-test was performed when two groups were compared. Data are Transcriptomic analysis with the RiboTag technique expressed as the mean ± SEM. A p value lower than Punches containing the ARC from mice at 12 weeks old 0.05 was considered significant. Fiji ImageJ 1.33 (NIH; fasted ON were homogenized in 350 µl of homogeniza- Bethesda, MD, USA) was used to determine the levels of tion buffer (50 mM Tris, 100 mM KCl, 12 mM MgCl2, Fos-positive cells, GFP-syn, mitochondria and dendritic 1% Nonidet P-40, 1 mM DTT, 200 U/ml Promega RNa- spines as well as for western blot analyses. The numbers sin, 1 mg/ml heparin, 100 μg/ml cycloheximide, Sigma of animals used in each experimental setting and analysis protease inhibitor mixture at pH 7.5) with a 30G needle. are specified in each figure legend. After clearing, 30 μl was separated as input and stored at − 80 °C until further processing. 3 µl of the anti-HA Results antibody (BioLegend, Cat# MMS-101R) were added to CPT1A in AgRP neurons is involved in energy balance the remaining supernatant and samples were incubated sexual‑dimorphic phenotype in a cold-room spinner for 2 h. After incubation, 200 μl We generated mice with a selective deletion of Cpt1a of Dynabeads protein G magnetic beads (Thermo Fisher in their AgRP neurons at adult stage (Cpt1aKO mice). Scientific, Cat# 10004D) were added and incubated for Mice expressing tamoxifen-inducible Cre recombinase 2 h at 4 °C with rotation. Next, beads were washed 3 times (CreERT2) in their AgRP neurons ­ (AgRPCreERT2) [41] for 10 min with gentle rotation at 4 °C in high-salt buffer were crossed with mice harboring conditional alleles of (50 mM Tris, 300 mM KCl, 12 mM MgCl2, 1% Nonidet Cpt1a (Cpt1aflox/flox mice) obtained in our laboratory P-40, 1 mM DTT, 100 μg/ml cycloheximide at pH 7.5). [40]. Tamoxifen injections were done at 8-week-old mice After the final wash, beads were incubated with 350 µl of and they were all studied 4 weeks after the injection. To the RLT buffer (QIAGEN, Cat# 74034). Total RNA was confirm Cre expression in the AgRP neurons, we injected prepared according to the manufacturer’s instructions AAV9-EF1a-DIO-mCherry into the ARC in Cpt1aKO using the RNeasy-plus Mini kit (QIAGEN, Cat# 74034). mice and we observed that AgRP neurons showed red RNA was quantified with the Quant-iT RiboGreen RNA fluorescence (Additional file 1: Fig. S1a–c). To corrobo- (Thermo Fisher Scientific, Cat# R11490). cDNA was gen- rate the deletion of Cpt1a exon 4 in the AgRP neurons erated using the SuperScript IV reverse transcriptase we isolated genomic DNA from the ARC and several (Thermo Fisher Scientific, Cat# 11750150) following other tissues from both Cpt1aKO and control mice. In the manufacturer’s instructions. Briefly, 0.27 ng of RNA order to distinguish the recombined Cpt1a gene lack- from the samples and the input was incubated with 5X ing exon 4 from WT Cpt1a gene, which is also expressed SuperScript IV RT buffer, 100 mM DTT, the RNaseOUT in other cell types, we firstly digested the genomic DNA Recombinant RNase inhibitor and the SuperScript IV from all studied tissues with the restriction enzymes PstI reverse transcriptase (200 U/µl). The samples were incu- and AatII. The restriction sites in those two enzymes are bated at 23 °C for 10 min, 50 °C for 10 min and 80 °C for only present in the WT Cpt1a gene but not in the recom- 10 min. They were then stored at − 20 °C until the qRT- bined Cpt1a gene lacking exon 4 (Additional file 1: Fig. PCR was performed. qRT-PCR was performed with S1d, e). Next, we confirmed the Cpt1a recombination by Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 9 of 26 using PCR amplification of the loxP-flanked region con- Cpt1a deletion in AgRP neurons increases energy taining Cpt1a exon 4. The resulting recombined Cpt1a expenditure (EE) amplification band was 219 bp and it was only detected Analysis of metabolic parameters revealed differences in the genomic DNA from ARC. This amplicon was also in EE between the sexes. Cpt1aKO female mice showed sequenced and our results confirmed that Cpt1a exon increased EE compared to the control female mice. These 4 was only removed in the genomic DNA from ARC of changes in EE were only observed in Cpt1aKO female Cpt1aKO mice (Additional file 1: Fig. S1g). The dele- mice but not in male mice (Fig. 2a and b and Additional tion of exon 4 has been shown previously by our group file 3: Fig. S3d–f ). No changes were observed in the RQ to result in a reading frame change and an appearance (Additional file 3: Fig. S3g, h) or in locomotor activ- of a stop codon in exon 5 of the recombined Cpt1a gene ity (Additional file 3: Fig. S3i, j) in both sexes. However, [40]. Since AgRP is also expressed by chromaffin cells only Cpt1aKO female mice increased their RQ to 0.95 in of the adrenal medulla [50], we analyzed adrenal gland refeeding conditions after an ON fast (Additional file 3: histology, weight, Tyrosine hydroxylase (Th) mRNA lev- Fig. S3g). els (indicative of adrenal gland integrity). No changes in To determine whether the Cpt1a ablation from AgRP histology, adrenal gland weight and Th mRNA levels were neurons enhanced EE in female mice, we measured observed in both female and male Cpt1aKO mice with the interscapular BAT temperature as an indicator of respect to their control littermates (Additional file 2: Fig. activated BAT thermogenesis. Cpt1aKO female mice S2a–e). showed a substantial increase in the BAT temperature When we assessed the study of the phenotype we compared to the control female mice (Fig. 2c and d). observed that female and male Cpt1aKO mice gained Consistent with this, Cpt1aKO female mice showed a sig- less weight than their control littermates when fed a nor- nificant reduction in BAT weight and lipid droplet (LD) mal chow diet ad libitum (Fig. 1a, b and Additional file 3: area compared to their control littermates and the male Fig. S3a). However, only the males showed a reduction in mice (Fig. 2e–i). At the molecular level, the UCP1 pro- food intake (Fig. 1c, d and Additional file 1: Fig. S3b, c). tein concentration and mRNA levels of thermogenesis- This was confirmed by the food intake analysis during the related genes (Cidea and Mmp2), lipolytic genes (Pnpla2 dark phase (Fig. 1e and f ), indicating sex differences in the and Lipe) and FAO markers (Cpt1b) were also elevated, function of AgRP neurons. To gain more insight into this confirming an enhanced activation of thermogenesis in difference in food intake behavior, we analyzed feeding Cpt1aKO female mice compared to their control litter- patterns after an ON fast or after the i.p. ghrelin adminis- mates (Fig. 2j). All this thermogenic activity seemed to tration. Both female and male Cpt1aKO mice showed an respond to an activated sympathetic tone. The thermo- impaired feeding response after an ON fast compared to genic response and LD reduction were much lower in the the control mice (Fig. 1g and h). However, the food intake Cpt1aKO male mice than in the Cpt1aKO female mice reduction was greater in male Cpt1aKO mice. Interest- (Fig. 2d, j–m), in accordance with the higher EE observed ingly, ghrelin administration induced greater food inges- in the Cpt1aKO female mice. In addition, we analyzed tion in control than in Cpt1aKO female mice. This was the mRNA levels of leptin, adiponectin and resistin. We also observed in male mice. However, the food intake only observed an increase in the resistin mRNA levels in observed in the females was 3 times greater than that Cpt1aKO male mice. observed in the males, indicating that the females were We also analyzed the effects of Cpt1a ablation in AgRP more sensitive to ghrelin than the males (Fig. 1i). Con- neurons on different tissues. An important reduction sistent with these results, male Cpt1aKO mice spent less was observed in the selected WAT deposits in Cpt1aKO time eating food than the female mice (Fig. 1j). Alto- mice compared to the control mice (Fig. 3a and b and gether, these results suggested a different role of CPT1A Additional file 4: Fig. S4a and b), which was consistent in males and females in the AgRP neurons under ad libi- with the observed reduction in body weight. Gonadal tum conditions and after ghrelin administration. and inguinal WAT (gWAT and iWAT, respectively) of (See figure on next page.) Fig. 1 Deletion of Cpt1a in AgRP neurons affects feeding behavior in a sex-dependent manner. a, b Body weight in Cpt1aKO female in red (a, n = 8) and male mice in blue (b, n = 6) vs their control littermates (female n = 10, and male n = 9). c, d Cumulative food intake in female (c, n = 8–10) and male (d, n = 6–9) mice measured during 1 month after tamoxifen induction. e, f Analysis of food intake measured by the TSE system during light and dark phases in female (e, n = 6–6) and male mice during 3 consecutive days (f, n = 6–6). g, h Response to fasting–refeeding in female (g, n = 8–10) and male mice (h, n = 10–8). (i) Cumulative food intake measured 1 h after intraperitoneal (i.p.) ghrelin administration in female (n = 7–7) and male mice (n = 6–6). j Time that female (n = 7–7) and male mice (n = 6–6) spent eating after the ghrelin injection. Data are expressed as the mean ± SEM. In a–d, *p < 0.05, **p < 0.01, ****p < 0.0001, using two-way repeated-measures ANOVA followed by Šidák’s post hoc test. In e and f, * p < 0.05, using Student’s t-test. In g and h, *p < 0.05 using two-way repeated-measures ANOVA followed by Šidák’s post hoc test. In i–j, *p < 0.05, **p < 0.01, using Student’s t-test. In i–j *p < 0.05, ****p < 0.0001, using two-way ANOVA followed by Turkey post hoc test Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 10 of 26 Fig. 1 (See legend on previous page.) Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 11 of 26 Cpt1aKO mice of both sexes showed a reduced adipo- AgRP neurons. Although we did not observe morphologi- cyte size with respect to their control littermates (Fig. 3c, cal changes in ovaries and testis, we measured the plasma g, j and n). Interestingly, while Cpt1aKO female mice levels of 17-β-estradiol in females and testosterone in showed a major reduction in iWAT (76% iWAT reduc- males (Additional file 5: Fig. S5g and h). No changes were tion vs 45.5% gWAT reduction), gWAT was the most observed in circulating levels of 17β-estradiol in females. affected tissue in Cpt1aKO male mice (70% gWAT reduc- However, we observed a significative reduction of plasma tion vs 56.7% iWAT reduction) (Fig. 3c, g, j and n). These testosterone levels in Cpt1aKO male mice. results were consistent with the analysis of the frequency distribution of the adipocyte area, since a higher fre- CPT1A in AgRP neurons is involved in thirst control quency of smaller adipocytes (< 500 μm2) in the iWAT of One week after tamoxifen administration and Cre induc- Cpt1aKO female mice corresponded to 80% of the tissue tion, we observed that female and male Cpt1aKO mice (Fig. 3g). In addition, we observed an enhanced brown- resulted in polyuria. Thus, mice were housed in metabolic ing in the iWAT of both sexes (Fig. 3h and o), consistent cages for 24 h to study their water balance. Ad libitum with the increased Ucp1 mRNA levels (Fig. 3i and p). We water intake in both female and male Cpt1aKO mice led also measured serum leptin levels (Additional file 4: Fig. to the excretion of an abnormally large volume of urine S4c and S4d) but no significant changes were observed each day, which was more prominent in the female mice between controls and Cpt1aKO mice. Furthermore, we (Fig. 4a and e). This was accompanied by low urine osmo- measured serum TG and NEFAs levels in both sexes. A lality, indicating high urine dilution (Fig. 4b and f ). As reduction in serum TGs and NEFAs levels was observed expected, given their polyuria, female and male Cpt1aKO in Cpt1aKO female and male mice (Fig. 3e, f, l and m). mice showed water consumption that was twice that of These results are suggestive that deletion of Cpt1a from their control littermates (Fig. 4c and g). Interestingly, no AgRP neurons activated the sympathetic nervous system, significant changes were observed in serum osmolality, leading to enhanced thermogenesis mainly in Cpt1aKO suggesting that the excretion of very diluted urine was female mice, an increased browning of the iWAT in both compensated by the high intake of water (Fig. 4d and h). sexes as well as considerably reduced fat deposits consist- To distinguish the Cpt1aKO polyuria from a diabetic ent with the reduction in body weight. condition, GTT and ITT were performed. Cpt1aKO male We also analyzed other tissues such as the liver, pan- mice showed significantly increased glucose tolerance creas, testicles, and ovaries. No morphological changes compared to their control littermates (Additional file 6: were observed in these tissues (Additional file 5: Fig. S5). Fig. S6f and g). Despite females having a slight change in Moreover, no changes were observed in the liver weight the first GTT timepoint, the overall glucose sensitivity in both sexes (Additional file 5: Fig. S5b and d). However, measured as AUC was not improved (Additional file 6: when we assessed the mRNA levels of the marker genes Fig. S6a, b). No differences were observed in the ITT in involved in glucose and FA metabolism, we observed that both sexes (Additional file 6: Fig. S6c and h). Next, we Cpt1aKO male mice showed increased mRNA levels of measured serum insulin levels and urine glucose content. the gluconeogenic genes encoding phosphoenolpyruvate No alterations were observed in these two parameters carboxykinase (Pepck) and glucose 6-phosphatase (G6pc) in the Cpt1aKO mice with respect to the control mice (Additional file 5: Fig. S5c and e). These changes were (Additional file 6: Fig. S6d, e, i, j). In addition, no changes not observed in the female mice. Expression Cpt1a was were observed in the morphology of the kidneys in both increased in both sexes, but the expression of Ucp2 was sexes, excluding an intrinsic renal disease (Additional increased only in the Cpt1aKO male mice. These results file 6: Fig. S6k). These data emphasized that Cpt1aKO suggested a sex-dependent difference in the liver meta- mice did not suffer type II diabetes, but they did have dis- bolic adaptation during fasting to the deletion of Cpt1a in rupted water intake. (See figure on next page.) Fig. 2 Cpt1a ablation in AgRP neurons increases brown adipose tissue activity. a, b EE profile normalized against the lean body mass in female (a, n = 6–6) and male mice (b, n = 6–6). c Representative infrared thermal images of female (top panel, n = 11–9) and male mice (bottom panel, n = 8–7). d Quantification of the interscapular temperature adjacent to the brown adipose tissue (BAT) of female (n = 11–9) and male mice (n = 8–7). e, f Weight of BAT normalized against the body weight of female (e, n = 6–9) and male mice (f, n = 6–7). g Representative images of BAT sections dyed with the H&E stain in female (top panel) and male mice (bottom panel). Scale bar, 100 μm (×magnification 20). h, i Lipid droplet quantification using ImageJ in female (h, n = 4–4) and male mice (i, n = 4–4). j, k Analysis by qRT-PCR of the mRNA levels of Pnpla2, Lipe, Cpt1b, Slc2a4 Ucp1, Cidea, Mmp2, Leptin, Adiponectin and Resistin in female (j, n = 7–7) and male mice (k, n = 5–6). l, m Representative western blot of UCP1 protein levels in BAT (30 μg) from female (l, n = 6–8) and male mice (m, n = 6–8) normalized against β-actin levels. Data are expressed as the mean ± SEM. In d * p < 0.05, ***p < 0.001, using two-way ANOVA followed by Turkey post hoc test. In a, b, e, f, h, I, j–m, *p < 0.05, **p < 0.01, using Student’s t-test Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 12 of 26 Fig. 2 (See legend on previous page.) Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 13 of 26 To further understand the mechanisms that caused Immunofluorescence assays of Fos were used to detect polydipsia, mice were deprived of water for 24 h in the the pattern of PVN activity under ad libitum access to metabolic cages. During dehydration, male Cpt1aKO water or in response to 24 h of water restriction. No sta- mice showed a significantly increased urine osmolality, as tistical differences were observed in Fos activation in the well as their control littermates (Fig. 4f ). Unfortunately, PVN under ad libitum access to water in both female and we were not able to collect enough urine volume to per- male mice. This was interesting since male and female form this analysis in the female mice under water dep- Cpt1aKO mice showed an increase in water intake with rivation (Fig. 4b). Despite this, no differences in serum no sign of Fos activation in the PVN during free access osmolality were observed in response to water restric- to water (Fig. 5a–c). However, we observed a signifi- tion (Fig. 4d and h). These data indicated that Cpt1aKO cant reduction in Fos activation in both female and male mice partially conserved their ability to respond to water Cpt1aKO mice compared with their control littermates deprivation. We then analyzed plasma levels of vasopres- during water deprivation (Fig. 5a–c). Thus, these results sin (VP, also known as anti-diuretic hormone, ADH) an suggested that the AgRP neurons in Cpt1aKO mice important regulator of fluid balance and is synthesized in reduced their neuronal connections to the PVN. magnocellular neuronal cell bodies of the PVN and SON Next, we studied the projections from AgRP neurons to [51]. We observed reduced circulating levels of VP/ADH the PVN in female Cpt1aKO mice under fasting condi- in both sexes. However, the effect was more prominent tions. To study this, we injected AAVs that conditionally in female Cpt1aKO mice (Fig. 4i). Another critical regu- expressed synaptophysin-GFP (AAV1-EF1a-DIO-synap- lator of fluid balance and blood pressure is the renin– tophysin-GFP) under Cre recombinase activation into the angiotensin–aldosterone system (RAAS). Typically, the ARC. GFP fluorescence was analyzed in the presynaptic RAAS is activated when there is a drop in blood volume area of AgRP neurons in the PVN sections (Fig. 6a and to increase water and electrolyte reabsorption in the kid- b). We observed a reduction in the GFP fluorescence in neys. Thus, we determined the levels of renin, angioten- the PVN section from Cpt1aKO female mice compared sin II (AgII) and aldosterone as important integrators to their control littermates. This reduction was less of the RAAS and blood pressure. Both female and male observed in Cpt1aKO male mice. Altogether, this sug- Cpt1aKO mice showed increased circulating levels of gested that Cpt1a deletion in AgRP neurons reduced the renin, AgII and aldosterone (Fig. 4j–l and n–p), which projections from AgRP neurons to the PVN in female were accompanied by a mild increase in the systolic mice. blood pressure (Fig. 4m and q). Taken together, these To explore if these reduced synapses were a result of data suggest that lipid metabolism in AgRP neurons is reduced AgRP neuron viability, we analyzed the number involved in the regulation of fluid homeostasis. of AgRP neurons along the anterior–posterior axis of the ARC. To accomplish this, we injected AAVs that condi- tionally expressed mCherry (AAV9-EF1a-DIO-mCherry) CPT1A is required in AgRP neurons for spine in the presence of Cre recombinase into the ARC of and presynaptic terminal formation control mice or mice lacking Cpt1a in their AgRP neu- Considering that the PVN is an important target of AgRP rons. Female mice were analyzed 1 month after tamox- neuronal projections and since VP/ADH is released from ifen induction. Importantly, no differences in the number the axonal projections of paraventricular magnocellular of AgRP neurons were observed in the sections selected cells into different brain areas (the rostro lateral medulla, from the anterior, medial and posterior ARC (Fig. 6c and nucleus tractus solitarius and intermediolateral column d). When we analyzed the morphology of AgRP neu- of the spinal cord) that are responsible for the integra- rons by confocal microscopy, we observed an alteration tion of the peripheral sympathetic and vagal outflow [52, in the dendritic morphology of AgRP neurons lacking 53], we wanted to know whether Cpt1a deletion from Cpt1a (Fig. 6e and g), which showed a reduced number of AgRP neurons affected the activity of PVN neurons. (See figure on next page.) Fig. 3 Cpt1a ablation in AgRP neurons reduces lipid content in white adipose tissue. a, b Representative image of gonadal white adipose tissue (gWAT) (a) and inguinal white adipose tissue (iWAT) (b) of female and male mice vs their control littermates. c, j Average adipocyte area of gWAT in female (c, n = 10–10) and male mice (j, n = 10–15). d, k Morphometric analysis of the adipocyte area distribution in the gWAT of female (d, n = 5–5) and male mice (k, n = 5–5). g, n Average adipocyte area of iWAT in female (g, n = 10–10) and male mice (n, n = 8–8). h, o Morphometric analysis of adipocyte area distribution in the iWAT of female (h, n = 5–5) and male mice (o, n = 5–5). e, f TG and NEFAs measurement in female (n = 5–5) and male mice (n = 5–5) (l–m). i, p Analysis by qRT-PCR of the mRNA levels of Pnpla2, Leptin, Ucp1, Cpt1a, Il6, Adiponectin and Resistin in the iWAT of female (i, n = 8–7) and male mice (p, n = 6–6). Data are expressed as the mean ± SEM. In d, h, k, o, *p < 0.05, **p < 0.001, ****i < 0.0001, using two-way repeated-measures ANOVA followed by Šidák’s post hoc test. In c, e, f, g, i, j, l, m, n, p, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 using Student’s t-test Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 14 of 26 Fig. 3 (See legend on previous page.) Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 15 of 26 dendritic spines per 50 µm of dendrons compared to the levels only in the immunoprecipitate, indicating the spe- AgRP neurons from their control littermates. cific isolation of AgRP ribosomes (Fig. 7c). Gene expres- sion analysis of the immunoprecipitate from RiboTag Cpt1aKO mice showed altered mitochondria density Cpt1aKO mice and their control littermates showed a and gene expression in AgRP neurons reduction of mRNA levels of three genes, Agrp,Slc17a6 To further investigate the mechanisms reducing the den- and Dgl4 in the RiboTag Cpt1aKO female mice (Fig. 7d). dritic spines and presynaptic terminals, we first focused Slc17a6 codifies for the protein of vesicular glutamate our attention on mitochondria. It has been reported that transporter (VGLU2) in presynaptic terminal and, while mitochondria play a key role in the number and size of Dgl4 codifies for a postsynaptic scaffolding protein (PSD- dendritic spines [54]. We analyzed the effect of Cpt1a 95) that plays a critical role in synaptogenesis and synap- deletion in AgRP neurons on mitochondrial morphology tic plasticity by providing a platform for the postsynaptic in in vivo conditions. We first generated ZsGreen mice. clustering of crucial synaptic proteins. No changes were Cpt1aKO mice were then crossed with ZsGreen mice observed in the gene expression of Slc32a1 that codifies that expressed the green fluorescent protein ZSGreen for vesicular transporter of GABA (VGAT1) or Gad1 in the mitochondria of AgRP neurons after tamoxifen (glutamate descarboxylase), an enzyme involved in the induction. Brain sections from Cpt1aKO-ZsGreen mice synthesis of GABA (Fig. 7d). These changes were not and their control littermates were obtained and ana- observed in the males (Fig. 7e). All these results indicated lyzed by confocal microscopy. We observed an increased that FA metabolism and CPT1A are involved in the neu- number and size of the mitochondria per cell of female ronal processes related to neurotransmission and neuro- with respect to the mitochondria per cell of males. Both peptide expression in AgRP neurons in fasting conditions Cpt1aKO-ZsGreen mice sexes showed a reduced num- and play different roles depending on sex. ber of mitochondria per cell as well as a reduced mito- chondrial area (Fig. 6h–m). Although the lack of CPT1A Discussion reduced the number of mitochondria, there seemed to be In this study, we showed that CPT1A, an enzyme that enough mitochondria to support neuronal survival. regulates FAO, is relevant for orexigenic AgRP neuronal To determine if this reduced pool of mitochondria and function, energy homeostasis, and fluid balance in a sex- synapses affected the synthesis of the NPY and AgRP dependent way (Table 1). Conditional Cpt1a deletion in neuropeptides and neurotransmitters (GABA and gluta- AgRP neurons in adult mice led to a clear reduction in mate) by AgRP neurons, we performed an in vivo tran- body weight gain in Cpt1aKO male mice, which was less scriptomic analysis using the RiboTag technique [42]. pronounced in the female mice compared to their control We crossed Cpt1aKO mice with RiboTag homozygous littermates. These sex-dependent differences were clearer mice. This allowed us to isolate AgRP neuron-specific in the feeding behavior, since Cpt1aKO female mice did ribosome-associated mRNAs from the ARC of RiboTag- not show any change in food intake under ad libitum Cpt1aKO mice by immunoprecipitating the actively conditions. Sex-based differences in feeding behavior in translating polyribosomes tagged with the HA epitope mammals have been attributed exclusively to the effects (Fig. 7a). To confirm the correct labeling of the AgRP of gonadal hormones, especially estrogens and andro- ribosomes, brain sections from RiboTag Cpt1aKO mice gens, which regulate food intake and energy metabo- were obtained and we assessed the HA epitope by immu- lism by acting on the brain and diverse peripheral tissues nofluorescence (Fig. 7b). To determine the specificity [55–57]. It has been shown that in gonadal-intact mice, of the immunoprecipitation for AgRP ribosomes, we total food intake is higher in male than in female mice evaluated the transcript levels of Agrp, Pomc and Adhl1 during dark phases, with female mice eating more than as markers of AgRP neurons, POMC neurons and astro- the males during light phases. Interestingly, these differ- cytes, respectively, in the input and immunoprecipitate ences were blunted in gonadectomized animals, since no by qRT-PCR. We observed an increase in Agrp mRNA differences were observed between the sexes [58], thus (See figure on next page.) Fig. 4 Cpt1aKO mice display changes in thirst behavior. a, e 24 h of urine collection in female (a, n = 6–5) and male mice (e, n = 6–7). b, f Analysis of urine osmolality in female (b, n = 11–8) and male mice (f, n = 16–9) under 24 h of water restriction or ad libitum access to water; ND, not detected. c, g Total amount of water intake for 24 h in female (c, n = 6–7) and male mice (g, n = 6–5). d, h Analysis of serum osmolality in female (d, n = 10–5) and male mice (h, n = 9–6) under 24 h of water restriction or ad libitum access to water. i Plasma level of the vasopressin hormone in female (n = 8–8) and male mice (n = 9–6). j, n Plasma levels of renin in female (j, n = 5) and male mice (n, n = 7–9). k, o Plasma levels of angiotensin II in female (k, n = 10–8) and male mice (o, n = 9–6). l, p Plasma levels of aldosterone in female (l, n = 8) and male mice (p, n = 8–10). m, q Blood pressure in female (m, n = 5–5) and male mice (q, n = 8–8). Data are expressed as the mean ± SEM. In b, d, f, h, i, **p < 0.01, ****p < 0.0001 using two-way repeated-measures ANOVA followed by the post hoc Bonferroni test. In a, c, e, g, j–o, *p < 0.05, **p < 0.01, ***p < 0.001, using Student’s t-test Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 16 of 26 Fig. 4 (See legend on previous page.) Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 17 of 26 Fig. 5 Deletion of Cpt1a in AgRP neurons impairs PVN activation during water restriction. a Representative images of Fos activation in the paraventricular nucleus (PVN) of female (left panel) and male mice (right panel) under conditions of ad libitum access to water or 24 h of water restriction. Dashed lines indicate the PVN area analyzed. b, c Quantification of Fos-positive cells per PVN section in female (b, n = 4–4) and male (c, n = 4–4) mice. Data are expressed as the mean ± SEM. In b and a, *p < 0.05, **p < 0.001, using two-way ANOVA followed by the post hoc Bonferroni test suggesting a complex interaction between the gonadal it is not clear which molecular mechanisms and path- hormones in daily feeding rhythms and revealing sex dif- ways are involved in AgRP neuron activation. Our find- ferences depending on gonadal hormones. In our model, ing reinforces the importance of CPT1A downstream of Cpt1a deletion in AgRP neurons affected food intake the ghrelin receptor in AgRP neurons for the induction during the dark phase in male Cpt1aKO mice, which of food-seeking behavior, since both female and male dropped by half. mice lacking CPT1A in their AgRP neurons reduced food It has been reported that the fasting-mediated activa- intake after ghrelin administration. tion of AgRP neurons promotes food-seeking behav- Although the reduction in body weight gain triggered ior and energy conservation [59]. In Cpt1aKO mice, by the Cpt1a deletion in AgRP neurons was independ- we observed an important reduction in food intake in ent of sex, our findings provide evidence that males and challenging conditions, such as during fasting-induced females are hardwired differently in the regulation of refeeding and ghrelin administration, in both sexes. their energy balance. At the physiological level, Cpt1aKO These results suggested that neurons lacking CPT1A suf- male mice reduced their body weight gain through die- fer metabolic inflexibility. Thus, AgRP neurons cannot tary restriction more efficiently than the female mice, cope with the transition between fasting and refeeding. while Cpt1aKO female mice increased EE that exceeded These findings agree with those of two previous studies the energy ingested. A change in food consumption [60, 61]. In agreement with these findings, our results affects EE primarily through its effect on diet-induced suggested that CPT1A and FA metabolism are impera- EE [68]. Thus, a decrease in food intake decreases EE tive in the feeding behavior response after the activation and vice versa. Accordingly, the drop in the food intake of AgRP neurons during fasting. In addition, it has been observed in Cpt1aKO male mice probably blunted the observed that ghrelin, a hormone that stimulates appe- increase of the EE observed in Cpt1aKO female mice. tite, binds to its receptor [62–64] and promotes feeding This increased EE in Cpt1aKO female mice correlated through the activation of the AMPK–CPT1A–UCP2 with an exacerbated thermogenesis in their BAT, an axis in hypothalamic neurons [65–67]. Although the enhanced browning of their subcutaneous adipose tissue effects of ghrelin on AgRP neurons are well established, (iWAT) and increased lipolytic activity in their adipose Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 18 of 26 Fig. 6 AgRP neurons lacking Cpt1a show reduced dendritic spines and projections to the PVN. a Scheme of AAV1-EF1a-DIO-synaptophysin-GFP administration into the ARC and the projection to the PVN. b Representative fluorescence microphotograph of AgRP-syn projection to the PVN in female and male mice. Scale bar, 10 μm. Syn-GFP intensity level was quantified with ImageJ using 6 slices (n = 3–4). c Representative images of AgRP neurons along the anterior–posterior axis of the ARC. d Quantification of AgRP neurons in the anterior, medial and posterior section of the ARC in female mice (d, n = 3–4). Scale bar, 25 μm. e Representative microphotograph of the anterior ARC section in female mice. Scale bar, 100 μm. f, g Representative fluorescence microphotograph of the dendrites of AgRP neurons. Scale bar, 100 μm. Quantification of the number of dendritic spines per 50 μm of dendrites in 20/25 axons of AgRP neurons in female mice (g, n = 3). h Scheme of mitochondrial labeling in the ZsGreen Cpt1aKO mice. i Representative fluorescence microphotograph of mitochondria from ZsGreen mice. Scale bar, 10 μm. j, l Quantification of the number of mitochondria per cell. k, m The average mitochondrial area in AgRP neurons (n = 3, 20–28 neurons per genotype). Data are expressed as the mean ± SEM. In d, g, j–o, *p < 0.05, **p < 0.01, ****p < 0.001 using Student’s t-test Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 19 of 26 Fig. 7 Validation of the RiboTagCpt1aKO mouse model and gene expression analysis. a Scheme of the experimental strategy to isolate AgRP neuron-specific RNAs from Cpt1aKO–RiboTag mice. b Representative images of AgRP neurons tagged with the anti-HA antibody. Scale bar, 100 µm. c Analysis by qRT-PCR of mRNA levels of Agrp, Pomc and Adh1 was performed in the input and immunoprecipitated HA samples. d, e Analysis by qRT-PCR of the mRNA levels of Agrp, Npy, Slc17a6, Slc32a1, Gad1, Dlg4 in female mice (n = 4–6) and male mice (n = 5–6). The reference gene was Gapdh for all the samples analyzed by qRT-PCR. Data are expressed as the mean ± SEM. *p < 0.05, ****p < 0.0001, using Student’s t-test tissues, resulting in an important reduction in their fat of AgRP neurons in the regulation of peripheral nutrient mass. The connection between AgRP neurons and BAT utilization [7, 72–74]. has been well established, anatomically and function- The liver is also an essential target of AgRP neurons ally [69–71]. The results from our model are suggestive in the maintenance of energy homeostasis [61, 75]. This that Cpt1a deletion in AgRP neurons triggers an increase agrees with our findings from the gene expression analy- in the sympathetic outflow to BAT. Cpt1aKO mice pre- sis of the livers of Cpt1aKO mice. Both male and female sented enhanced BAT activity, showing an increased Cpt1aKO mice showed a significant increase in Cpt1a temperature of the suprascapular area mainly in the gene expression in the liver, suggestive of increased females. This was confirmed by measuring UCP1 protein β-oxidation of FA. We also observed a significant levels in BAT and the expression of the genes encoding increase in the expression of Pepck and G6pase in the lipolytic enzymes. In addition, an enhanced browning of male but not female mice under fasting conditions, sug- iWAT and an important reduction in the fat pads, mainly gesting a new sex-dependent difference in coping with in the females, were consistent with the physiological role challenging fasting conditions. Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 20 of 26 Table 1 Summary of metabolic phenotypes between males and females and sexual dimorphism Parameter Control Female Cpt1aKO Control Male Cpt1aKO Male Sexual Female dimorphism Body weight gain – ↓ – ↓ Cumulative food intake – – – ↓ * Food intake induced by fast – ↓ – ↓ Food intake induced by ghrelin – ↓ ↓ ↓↓ * Energy expenditure – ↑ – – * BAT temperature – ↑ ↑ – * BAT tissue – ↓ – ↓ Lipid droplet area – ↓ – ↓ Thermogenic gene expression in BAT – ↑ – ↑ UCP1 protein – ↑ – ↑ gWAT tissue – ↓ – ↓ adipocyte size of gWAT​ – ↓ – ↓ adipocyte size of iWAT​ – ↓ – ↓ TG and NEFAs – ↓ – ↓ Leptin – – – – UCP1 gene expression in iWAT​ – ↑↑ – ↑ Urine output – ↑ – ↑ Urine osmolality under water ad libitum – nd – ↓ Water intake – ↑ – ↑ Plasma osmolality – – – – Vasopressin – ↓ ↑ ↓ * Renin – ↑ – ↑ Angiotensin II – ↑ – ↑ Aldosterone – ↑ – ↑ Systolic blood pressure – ↑ – ↑ GTT (AUC) – – – ↓ * ITT – – – – Fasted insulin – – – – Urine glucose – – – – Liver weight – – – – Cpt1a liver expression – ↑ – ↑ Ucp2, Pepck and G6pc – – – ↑ * Mitochondrial content – ↓ – ↓ Agrp, Slc17ab and Dlg4 gene expresion in – ↓ – – * AgRP neurons 17β-Estradiol – – Testosterone – ↓ *Indicates sexual dimorphism considering control females as reference. nd, not detected Most studies on AgRP function have focused on the AgRP neurons project to the PVN, the observed VP/ molecular mechanisms underlying food intake, energy ADH reduction could be due to an impaired ability of metabolism and body weight changes. However, little AgRP neurons to activate the PVN. We also investigated is known about the implication of AgRP in water con- the RAAS, which is a critical regulator of fluid balance sumption. The polydipsia and polyuria observed in the and blood pressure. Typically, the RAAS is activated Cpt1aKO mice were consistent with the reduced blood when there is a drop in the blood volume to increase levels of VP/ADH in Cpt1aKO mice compared to their water and electrolyte reabsorption in the kidneys. We control littermates. Considering that the PVN is an measured the plasma levels of AgII to determine whether important nucleus for VP/ADH production and since the polyuria activated the RAAS. The increased blood Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 21 of 26 levels of AgII observed in Cpt1aKO mice indicated that only in the maintenance of the fluid balance, but also in the normal renal parenchyma was able to detect the drop the control of food intake and body weight. in water and electrolyte reabsorption, triggering RAAS Mitochondria play important roles in neurons. In addi- activation. Thus, this cascade emerges as a compensatory tion to energy production via the synthesis of ATP, this mechanism for water loss. Our finding was supported by dynamic organelle is involved in cellular metabolism, those of studies evaluating the effect of AngII on diure- innate and adaptative immune responses, and cell death sis, since the intravenous infusion of angiotensin inhib- (revised in [79]). Since CPT1A is a key enzyme in FAO, ited polyuria [76]. It is also plausible to speculate a direct we speculated that the lack of CPT1A enzyme in AgRP activation of renin release from the juxtaglomerular cells neurons may produce metabolic shifts that could affect of the kidneys. It is well stablished that the SNS triggers spine formation and synapses. Interestingly, we have renin release for the generation of angiotensin I, which is found sex differences in the size and quantity of mito- then converted to Ang II [77]. More recently, direct mul- chondria per neuron in control mice. These findings tifiber recording of sympathetic nerve activity subserving highlight the possibility that females could be more pro- the kidney was performed followed by a chemogenetic tected to cope with stressful conditions [80, 81]. We also activation of AgRP neurons. The authors showed a rapid observed a reduced size and number of mitochondria decline of sympathetic nerve activity upon AgRP activa- per neuron in mice lacking CPT1A, mainly in females, tion [78]. Based on the experiments reported here, the which could reduce the energy production necessary for lack of CPT1A in AgRP neurons is expected to upregu- many neuronal processes. In line with this, in culture of late sympathetic nerve activity to the kidney. This could embryonic primary cortical neurons we have observed also account for the activation of RAAS in our model. that the lack of CPT1A reduces GABA release [40]. AgII raises blood pressure through different mecha- Another limitation of our study is the lack of analysis nisms, the most important ones being vasoconstriction, of the functionality that reduced mitochondria have on sympathetic nervous stimulation, increased aldosterone neuronal processes in adult mice of both sexes in vivo. In biosynthesis and renal activities. Here, we observed that order to do this, the AgRP neuronal isolation and culture both systolic blood pressure and aldosterone levels were from adult mice would be needed, but currently primary increased in Cpt1aKO mice, strongly reinforcing RAAS hypothalamic cultures are only feasible in the embryonic activation. AgII has also been reported to exert effects stage. In a recent study in embryonic hypothalamic cul- on the brain. It can bind to the hypothalamus, stimulat- ture from mice with a AgRP-selective deletion of Dnm1, ing thirst and the release of VP/ADH by the posterior which codifies for a mediator of mitochondrial fission pituitary. However, the reduced level of this hormone protein (DRP1), has been shown an attenuated mito- in Cpt1aKO mice indicated that AgII requires neuronal chondrial respiration that could affect AgRP function stimuli to maintain VP/ADH levels under physiologi- [82]. Future studies assessing the mitochondrial func- cal conditions. Altogether, our results suggested that the tionality and subcellular distribution in both sexes are normal function of AgRP neurons is necessary to main- necessary to clarify the mitochondrial role in the AgRP tain fluid balance. neuronal function. The drop in the activation of PVN neurons suggested RiboTag analysis also showed important differences reduced AgRP presynaptic innervation or neuronal via- in gene expression between the females and males. In bility. We explored both and did not find differences in fasting conditions, female mice showed reduced tran- the number of AgRP neurons in the ARC sections ana- script levels of key genes involved in neurotransmission lyzed. However, the number of dendritic spines on AgRP and neuropeptide production, which was not observed neurons were reduced, as in the studies reporting a lack in the males. The decrease in AgRP neuropeptide lev- of NMDAR signaling components in spines [8, 60]. The els was consistent with the increased EE observed in analysis of AgRP presynaptic terminals in the PVN using Cpt1aKO female mice. Neurons in the PVN express- labeled synaptophysin specifically expressed in AgRP ing thyroid-releasing hormone (TRH), oxytocin (OT), neurons confirmed that female Cpt1aKO mice substan- and corticotropin-releasing hormone (CRH) all express tially reduced their presynaptic activity, which occurred MC4R [83]. The binding of α-MSH secreted by POMC to a lesser extent in the males. This suggested that the neurons to MC4R on these neurons has a positive effect female mice were very susceptible to the lack of the on the hypothalamic–pituitary–thyroid (HPT) axis and CPT1A enzyme. One limitation of this study is to under- the hypothalamic–corticotropic axis (HPA) [84]. Since stand which specific neurons of PNV or other hypotha- AgRP acts as an inverse agonist of α-MSH in the PVN lamic nucleus could be affected by the reduced AgRP in fasting conditions [85–88] its reduced expression in presynaptic activity in both sexes. This could potentially Cpt1aKO female mice could stimulate both the HPT explain the physiological sex differences observed not and HPA axes, resulting in a positive enhancement of Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 22 of 26 thermogenesis and EE. However, other brain regions and Perspectives and significance neuronal mechanisms could be involved in the enhance- Our study provided insight into understanding how fatty ment of thermogenic activity, EE and the control of body acid metabolism and, in particular, CPT1A in AgRP weight, since specific groups of AgRP neurons have been neurons play role in the control of food intake, energy reported to project to the dorsal lateral part of the dorsal expenditure and body weight. We suggest that CPT1A raphe nucleus [89] and the parabrachial nucleus [90]. We should be considered as a new target against obesity. also observed reduced Slc17a6 mRNA levels in Cpt1aKO Additionally, we provide evidence that AgRP neurons female mice, which could partly explain the reduced are involved in the control of thirst and fluid balance, activity of vasopressin neurons. Glutamate and GABA suggesting that AgRP neurons are relevant components are the main neurotransmitters in the PVN and SON of the neural circuits underlying thirst and fluid homeo- that are involved in the synaptic regulation of vasopressin stasis. Further analysis of AgRP connections with other neurons [91]. A decrease in glutamatergic inputs is partly centers of the hypothalamus and the brain that are consistent with the polydipsia and polyuria observed in involved in the control of thirst could help to understand our Cpt1aKO mouse model. Although there were no the circuitry and the mechanisms that underlie drinking changes in VGAT gene expression in fasting conditions behavior. between control and Cpt1aKO mice, we cannot discard Our study demonstrates that there are sex-specific dif- the effect of GABA in polydipsia in conditions other than ferences in the effect of CPT1A deletion, a key enzyme fasting [90]. Dgl4 mRNA levels were also decreased in in fatty acid metabolism, on specific neurons, empha- female Cpt1aKO mice. Since PSD-95, the protein codi- sizing the need to analyze both females and males when fied by Dgl4 gene, is required for the stabilization of describing the mouse phenotypes. Future studies will spines, the maturation of excitatory synapses and the be needed to specifically address the relevance of fatty synaptic function [92], we propose that this reduction metabolism in mitochondrial function and consequently on Dgl4 mRNA levels could result in a partial reduction in neuronal processes. of the postsynaptic plasticity of AgRP neurons in fast- ing conditions. The decrease of both Scl17a6 and Dgl4 Abbreviations mRNA levels suggest that Cpt1aKO female mice reduce AAVs Adeno-associated viruses AgRP neuronal synaptic processes in the fasting condi- ACC​ Acetyl-CoA carboxylase tions, which is mildly observed in Cpt1aKO male mice. AgRP Agouti-related peptide AMPK AMP-activated protein kinase This is consistent with the reduced number of mitochon- AgII Angiotensin II dria, presynaptic innervation to PVN observed in females ARC​ Arcuate nucleus in respect to the males. We propose that this reduction BAT Brown adipose tissue Cidea Cell death inducing DFFA like effector A in the neuropeptide AgRP and proteins involved synaptic CPT1A Carnitine palmitoyltransferase 1A processes could reduce the signaling to the PVN enhanc- Dlg4 Discs large MAGUK scaffold protein 4 ing HPA axis and sympathetic tone in fasting conditions. EE Energy expenditure FA Fatty acid Sexual hormones could be affected by AgRP levels. Dnm1 Dynamin-1 It has been shown that starvation-activated agouti- DRP1 Dynamin-related protein 1 related peptide (AgRP) neurons can inhibit the repro- FAO Fatty acid oxidation FAS Fatty acid synthase ductive neuroendocrine circuit, mainly in females [93]. GABA Gamma-aminobutyric acid No significant changes were observed in 17β-estradiol GAD Glutamate decarboxylase in Cpt1aKO female mice, in contrast, Cpt1aKO male Gapdh Glyceraldehyde-3-phosphate dehydrogenase G6pc Glucose 6-phosphatase mice showed a decrease of testosterone. Since levels of GTT​ Glucose tolerance test testosterone are mediated by the hypothalamic–pitui- Hmgcs2 3-Hydroxy-3-methylglutaryl-CoA synthase 2 tary–gonadal axis we cannot rule out that a drop in the iGluR Ionotropic glutamate receptor gWAT​ Gonadal white adipose tissue activation of the PVN neurons from AgRP neurons could LA Locomotor activity modulate the hypothalamic releases of gonadotropin- LCFA Long-chain fatty acid releasing hormone (GnRH) and further reducing the MCD Malonyl-CoA decarboxylase MTS Mitochondrial targeting signal secretion of follicle-stimulating hormone (FSH) and lute- Hprt Hypoxanthine–guanine phosphoribosyl-transferase inizing hormone (LH) by the anterior pituitary gland and ITT Insulin tolerance test consequently leading a reduction of testosterone produc- Lipe Lipase E iWAT​ Inguinal white adipose tissue tion. All these results highlight the different role of AgRP Mmp2 Matrix metallopeptidase 2 neurons in males and females. MCD Malonyl-CoA decarboxylase NEFA Non-esterified fatty acid NPY Neuropeptide Y PFA Paraformaldehyde Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 23 of 26 PBS Phosphate buffered saline PCR Polymerase chain reaction Additional file 5: Figure S5. Deletion of Cpt1a in AgRP neurons upregu‑ PEPCK Phosphoenolpyruvate carboxykinase lates Cpt1a expression in the liver. a Representative hematoxylin and eosin Pnpla2 Patatin-like phospholipase domain-containing 2 (H&E) staining of the livers of male and female mice. Scale bar, 50 µm POMC Proopiomelanocortin (magnification 20×). b and d Liver weight of female (b, n = 6–8) and male PVN Paraventricular nucleus mice (d, n = 6–8). (c and e) Analysis by qRT-PCR of Cpt1a, Cd36, Ucp2, RAAS Renin–angiotensin–aldosterone system Hmgcs2, Pepck and G6pc mRNA levels in female (c, n = 8) and male mice RPL22HA Ribosomal protein L22-hemagglutinin (e, n = 9–8). f Representative H&E staining of the pancreas and testes from RQ Respiratory quotient male mice (upper panel) and the pancreas and ovaries from female mice Slc2a Solute carrier family 2 member 4 (lower panel). g Plasma level of 17β-estradiol in female mice (n = 8–8) and Slc32a1 Solute carrier family 32 member 1 h testosterone level in male mice (n = 9–10). Scale bar, 50 µm (magnifica‑ Slc17a6 Solute carrier family 17 member 6 tion 20×). Data are expressed as the mean ± SEM. In b–e, * p < 0.05, using SNS Sympathetic nervous system Student’s t-test. SON Supraoptic nucleus Additional file 6: Figure S6. Cpt1a deletion in AgRP neurons does not TGs Triglycerides induce a diabetic state. a and f Glucose tolerance test (GTT) in female Th Tyrosine hydroxylase (a, n = 10–8) and male mice (f, n = 10–8). b and g area under the curve UCP Uncoupling protein (AUC) quantification in female (b, n = 10–8) and male mice (g, n = 10–8). VP/ADH Vasopressin c and h Insulin tolerance test (ITT) in female (c, n = 7–8) and male mice WT Wild type (h, n = 10–9). d and i Fasting insulin levels in female (d, n = 8–6) and male mice (i, n = 7–4). e and j Urinary glucose levels in female (e, n = 5–7) and male mice (j, n = 9–6). k Representative hematoxylin and eosin (H&E) Supplementary Information staining of the cortex (left panel) and medulla (right panel) of the kidneys The online version contains supplementary material available at https://​doi.​ from female and male mice. Scale bar, 50 µm (magnification 20×). Data org/​10.​1186/​s13293-​023-​00498-8. are expressed as the mean ± SEM. In a, f, c, h, * p < 0.05, using two-way repeated-measures ANOVA followed by Šidák’s post hoc test. In b, g, d–j, Additional file 1: Figure S1. Validation of Cre-mediated recombination in *** p < 0.001, using Student’s t-test. AgRP neurons. a Scheme of the time-course of the experiment. b Bilateral Additional file 7. Table S1: Forward and reverse primers used in the PCR injection of 400 nl of AAV9-EF1a-DIO-mCherry at a dose of 1.23 × ­1013 gc/ analysis. Table S2: Antibodies used in the study. ml into the ARC of AgRP-Cre-ERT2 mice. c Representative histological slice of Cre-dependent mCherry expression in the ARC. Scale bar, 500 μm (top image) and 200 μm (bottom image). d Schematic of the Cre-mediated Acknowledgements recombination product showing the floxed band (1030 bp) containing We thank Prof. Joel K. Elmquist from the Southwestern Medical Center, USA, the LoxP sequences surrounding exon 4 of the Cpt1a gene. After Cre for providing the ­AgRPCreERT2 mice. We thank Serveis Linguistics of Universitat recombination, the product resulted in a 219-bp DNA fragment. e Scheme de Barcelona, for the language correction of the manuscript. of the restriction sites of the restriction enzymes PstI and AatII that were used to separate Cpt1a amplicons from unrecombined genomic DNA. f Author contributions Representative PCR analysis of genomic DNA from the liver, gonadal white ZS and DS contributed to study conception and design. ZS, PM, IGG, KI, MMR, adipose tissue (gWAT), brown adipose tissue (BAT), adrenal gland (AG), AO, BM, AE, MCS, MBP, LC, EA, CP, JV, XA, JG and JMG: data acquisition, data cortex, hippocampus (Hyp) and arcuate nucleus (ARC) treated with PstI analysis and interpretation. RP, AQ and ES contributed to the generation of and AatII enzymes in Cpt1aKO mice and control mice (g) FASTA analysis of ZsGreen mice and AAVs. ZS, NC, ML, RP, ES, AQ, LH and DS contributed to the sequenced 219 bp DNA fragment extracted from the gel. manuscript writing and final review. All authors read and approved the final Additional file 2: Figure S2. Analysis of the adrenal gland after Cpt1a manuscript. deletion in AgRP neurons. a Representative H&E staining of female (left panel) and male adrenal gland (right panel). Scale bar, 500 μm (magnifica‑ Funding tion 4×) and 100 μm (magnification 20×). b and c Weight of the left and This study was supported by the Spanish Ministry of Economy and Competi‑ right adrenal glands in female (b, n = 5–5) and male mice (c, n = 5–5). (d tiveness (MINECO) (SAF2017-83813-C3-1-R to DS and LH, cofunded by the and e) Analysis by qRT-PCR of the mRNA levels of Th in the adrenal gland European Regional Development Fund [ERDF] and PID2020-114953RB-C21 of female (D, n = 5–9) and male mice (E, n = 9–7). to LH and DS, RTI2018-094727-B-100 to JMG; SAF2017-88108-R to AQ; AEI (PID2020-114977RB-I00) to AQ, ERC-2014-StG-638106 to AQ, MICINN RYC2019- Additional file 3: Figure S3. Effect of Cpt1a deletion on EE. a Representa‑ 028501-I to ES; MICIU RTI2018-101838-J-I00 to ES; a doctoral fellowship to SZ, tive image of Cpt1aKO female (left image) and male mice (right image) and a Juan de la Cierva-Incorporación Research Fellowship [IJCI-2016-28313] compared with their control littermates 3 months after tamoxifen induc‑ to PM and DS), the Centro de Investigación Biomédica en Red de Fisiopa‑ tion. b and c 24 h of food consumption in female (b, n = 8–10) and male tología de la Obesidad y la Nutrición (CIBEROBN) (CB06/03/0001 to DS and mice (c, n = 6–9). d and e EE in light and dark cycles and during fasting LH), the Government of Catalonia (2014SGR465 to DS), and La Marató de TV3 and refeeding in female (d, n = 6–6) and male mice (e, n = 6–6). f EE in (201627-30 to DS). AO is supported by a Miguel Servet contract (CP19/00083) light and dark cycles comparing control littermates and Cpt1aKO animals from Instituto de Salud Carlos III co-financed by European Regional Develop‑ (f, n = 6–6). g and h Respiratory quotient (RQ) registered by the TSE system ment Fund [ERDF]. AE is funded by ISCIII /MINECO (PT17/0009/0019) and in female (f, n = 6–6) and male mice (g, n = 6–6). h and i Locomotor activ‑ co-funded by FEDER. ity (LA) registered by the TSE system in female (h, n = 6–6) and male mice (i, n = 6–6). Data are expressed as the mean ± SEM. In b–i, * p < 0.05, using Data availability Student’s t-test. The datasets generated during and/or analyzed during the current study Additional file 4: Figure S4. Deletion of Cpt1a in AgRP neurons reduces are not publicly available due to the Universitat of Barcelona do not have a the fat mass in male and female mice. a and b Gonadal white adipose repository for research data but are available from the corresponding author tissue (gWAT) weight in female (a, n = 6) and male mice (b, n = 6). (c and on reasonable request. d) (c and d) Plasma level of leptin in female (c, n = 4–5) and male mice (d, n = 4–7). e and f Analysis by qRT-PCR of Pnpla2, Vegfa, Il6, Leptin, Mmp2, Declarations Adiponectin and Resistin mRNA levels in the gWAT of female (e, n = 8–6) and male mice (f, n = 9–6). Data are expressed as the mean ± SEM. In a–d, Ethics approval and consent to participate * p < 0.05, using Student’s t-test. All animal experiments were approved by the Comité Ètic d’Experimentació Animal of the Universitat de Barcelona (CEEA-UB). Permits 10994 and 10953 Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 24 of 26 were obtained from the Government of Catalonia, according to European 9. Dietrich MO, Liu ZW, Horvath TL. Mitochondrial dynamics controlled by Directive 2010/63/EU. mitofusins regulate Agrp neuronal activity and diet-induced obesity. Cell. 2013;155:188. https://​doi.​org/​10.​1016/J.​CELL.​2013.​09.​004. Consent for publication 10. Dietrich MO, Antunes C, Geliang G, et al. Agrp neurons mediate Sirt1’s Not applicable. action on the melanocortin system and energy balance: roles for Sirt1 in neuronal firing and synaptic plasticity. J Neurosci. 2010;30:11815–25. Competing interests https://​doi.​org/​10.​1523/​JNEUR​OSCI.​2234-​10.​2010. The material reported in this manuscript is original and has not been previ‑ 11. Abbott SBG, Machado NLS, Geerling JC, Saper CB. Reciprocal control ously published or submitted for publication elsewhere. All authors approve of drinking behavior by median preoptic neurons in mice. J Neurosci. the manuscript and declare that they have no competing interests. 2016;36:8228–37. https://​doi.​org/​10.​1523/​JNEUR​OSCI.​1244-​16.​2016. 12. Allen WE, DeNardo LA, Chen MZ, et al. Thirst-associated preoptic neurons Author details encode an aversive motivational drive. Science. 2017;357:1149–55. 1 Department of Biochemistry and Physiology, School of Pharmacy and Food https://​doi.​org/​10.​1126/​SCIEN​CE.​AAN67​47. Sciences, Universitat de Barcelona, Av. Joan XXIII, 27‑31, 08028 Barcelona, 13. Augustine V, Gokce SK, Lee S, et al. Hierarchical neural architecture under‑ Spain. 2 Institut de Biomedicina de la Universitat de Barcelona (IBUB), Univer‑ lying thirst regulation. Nature. 2018;555:204–9. https://​doi.​org/​10.​1038/​ sitat de Barcelona, Barcelona, Spain. 3 Centro de Investigación Biomédica en NATUR​E25488. Red de Fisiopatología de la Obesidad y la Nutrición (CIBEROBN), Instituto de 14. Betley JN, Xu S, Cao ZFH, et al. Neurons for hunger and thirst transmit a Salud Carlos III, Madrid, Spain. 4 NeurObesity Group, Department of Physiol‑ negative-valence teaching signal. Nature. 2015;521:180–5. https://​doi.​ ogy, CIMUS, University of Santiago de Compostela-Instituto de Investigación org/​10.​1038/​NATUR​E14416. Sanitaria, Santiago de Compostela, Spain. 5 Neuronal Control of Metabolism 15. Gizowski C, Bourque CW. The neural basis of homeostatic and anticipa‑ Laboratory, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), tory thirst. Nat Rev Nephrol. 2018;14:11–25. https://​doi.​org/​10.​1038/​ Barcelona, Spain. 6 CNAG-CRG, Centre for Genomic Regulation, Barcelona Insti‑ NRNEPH.​2017.​149. tute of Science and Technology, 08028 Barcelona, Spain. 7 Universitat Pompeu 16. Leib DE, Zimmerman CA, Poormoghaddam A, et al. The forebrain Fabra (UPF), Barcelona, Spain. 8 Instituto de Investigaciones Biomédicas de thirst circuit drives drinking through negative reinforcement. Neuron. Barcelona (IIBB-CSIC), Barcelona, Spain. 9 CIBER de Enfermedades Cardiovas‑ 2017;96:1272-1281.e4. https://​doi.​org/​10.​1016/J.​NEURON.​2017.​11.​041. culares (CIBERCV), Instituto de Salud Carlos III, Madrid, Spain. 10 Instituto de 17. Mckinley MJ, Yao ST, Uschakov A, et al. The median preoptic nucleus: Investigación Biomédica Sant Pau (IIB-Sant Pau), Barcelona, Spain. 11 Biomedi‑ front and centre for the regulation of body fluid, sodium, temperature, cal Research Networking Centre in Neurodegenerative Diseases (CIBERNED), sleep and cardiovascular homeostasis. Acta Physiol (Oxf ). 2015;214:8–32. Madrid, Spain. 12 Institute of Neurosciences of the Universitat de Barcelona, https://​doi.​org/​10.​1111/​APHA.​12487. Barcelona, Spain. 13 Department of Inorganic & Organic Chemistry, Faculty 18. Oka Y, Ye M, Zuker CS. Thirst driving and suppressing signals encoded by of Chemistry, Institut de Biomedicina de la Universitat de Barcelona (IBUB), distinct neural populations in the brain. Nature. 2015;520:349–52. https://​ Universitat de Barcelona, Barcelona, Spain. 14 Department of Basic Sciences, doi.​org/​10.​1038/​NATUR​E14108. Faculty of Medicine & Health Sciences, Universitat Internacional de Catalunya, 19. Zimmerman CA, Lin YC, Leib DE, et al. Thirst neurons anticipate Sant Cugat del Vallès, Spain. 15 Department of Biochemistry, Howard Hughes the homeostatic consequences of eating and drinking. Nature. Medical Institute, University of Washington, Seattle, WA, USA. 16 Department 2016;537:680–4. https://​doi.​org/​10.​1038/​NATUR​E18950. of Cell Biology, Physiology and Immunology, Universitat Autònoma de Barce‑ 20. Mandelblat-Cerf Y, Kim A, Burgess CR, et al. Bidirectional anticipation of lona, Bellaterra, Spain. 17 Institut de Neurociències, Universitat Autònoma de future osmotic challenges by vasopressin neurons. Neuron. 2017;93:57– Barcelona, Bellaterra, Spain. 65. https://​doi.​org/​10.​1016/J.​NEURON.​2016.​11.​021. 21. Burnett CJ, Li C, Webber E, et al. Hunger-driven motivational state compe‑ Received: 26 October 2022 Accepted: 10 March 2023 tition. Neuron. 2016;92:187–201. https://​doi.​org/​10.​1016/J.​NEURON.​2016.​ 08.​032. 22. Bruce KD, Zsombok A, Eckel RH. Lipid processing in the brain: a key regulator of systemic metabolism. Front Endocrinol. 2017;8:1–11. https://​ doi.​org/​10.​3389/​fendo.​2017.​00060. References 23. Hamilton JA, Hillard CJ, Spector AA, Watkins PA. Brain uptake and 1. Aponte Y, Atasoy D, Sternson SM. AGRP neurons are sufficient to utilization of fatty acids, lipids and lipoproteins: application to neurologi‑ orchestrate feeding behavior rapidly and without training. Nat Neurosci. cal disorders. J Mol Neurosci. 2007;33:2–11. https://​doi.​org/​10.​1007/​ 2011;14:351. https://​doi.​org/​10.​1038/​nn.​2739. s12031-​007-​0060-1. 2. Gropp E, Shanabrough M, Borok E, et al. Agouti-related peptide-express‑ 24. Cruciani-Guglielmacci C, López M, Campana M, le Stunff H. Brain ing neurons are mandatory for feeding. Nat Neurosci. 2005;8:1289–91. ceramide metabolism in the control of energy balance. Front Physiol. https://​doi.​org/​10.​1038/​nn1548. 2017;8:787. https://​doi.​org/​10.​3389/​fphys.​2017.​00787. 3. Krashes MJ, Koda S, Ye C, et al. Rapid, reversible activation of AgRP 25. Diéguez C, Frühbeck G, López M. Hypothalamic lipids and the regulation neurons drives feeding behavior in mice. J Clin Invest. 2011;121:1424–8. of energy homeostasis. Obes Facts. 2009;2:126–35. https://​doi.​org/​10.​ https://​doi.​org/​10.​1172/​JCI46​229. 1159/​00020​9251. 4. Padilla SL, Carmody JS, Zeltser LM. Pomc-expressing progenitors give rise 26. Varela L, Vázquez MJ, Cordido F, et al. Ghrelin and lipid metabolism: key to antagonistic neuronal populations in hypothalamic feeding circuits. partners in energy balance. J Mol Endocrinol. 2011;46:R43-63. https://​doi.​ Nat Med. 2010;16:403. org/​10.​1677/​JME-​10-​0068. 5. Atasoy D, Betley JN, Su HH, Sternson SM. Deconstruction of a neural 27. López M, Nogueiras R, Tena-Sempere M, Diéguez C. Hypothalamic AMPK: circuit for hunger. Nature. 2012;488:172. https://​doi.​org/​10.​1038/​natur​ a canonical regulator of whole-body energy balance. Nat Rev Endocrinol. e11270. 2016;12:421–32. https://​doi.​org/​10.​1038/​nrendo.​2016.​67. 6. Albarado DC, McClaine J, Stephens JM, et al. Impaired coordination 28. Swierczynski J, Goyke E, Korczynska J, Jankowski Z. Acetyl-CoA carboxy‑ of nutrient intake and substrate oxidation in melanocortin-4 receptor lase and fatty acid synthase activities in human hypothalamus. Neurosci knockout mice. Endocrinology. 2004;145:243–52. https://​doi.​org/​10.​1210/​ Lett. 2008;444:209–11. https://​doi.​org/​10.​1016/J.​NEULET.​2008.​08.​046. en.​2003-​0452. 29. Lane MD, Wolfgang M, Cha S-H, Dai Y. Regulation of food intake and 7. Joly-Amado A, Denis RGP, Castel J, et al. Hypothalamic AgRP-neurons energy expenditure by hypothalamic malonyl-CoA. Int J Obes (Lond). control peripheral substrate utilization and nutrient partitioning. EMBO J. 2008;32(Suppl 4):S49-54. https://​doi.​org/​10.​1038/​ijo.​2008.​123. 2012;31:4276–88. https://​doi.​org/​10.​1038/​emboj.​2012.​250. 30. López M, Lage R, Saha AK, et al. Hypothalamic fatty acid metabolism 8. Üner A, Gonçalves GHM, Li W, et al. The role of GluN2A and GluN2B mediates the orexigenic action of ghrelin. Cell Metab. 2008;7:389–99. NMDA receptor subunits in AgRP and POMC neurons on body weight https://​doi.​org/​10.​1016/j.​cmet.​2008.​03.​006. and glucose homeostasis. Mol Metab. 2015;4:678–91. https://​doi.​org/​10.​ 31. Minokoshi Y, Alquier T, Furukawa H, et al. AMP-kinase regulates 1016/J.​MOLMET.​2015.​06.​010. food intake by responding to hormonal and nutrient signals in the Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 25 of 26 hypothalamus. Nature. 2004;428:569–74. https://​doi.​org/​10.​1038/​NATUR​ 52. Bourque CW. Central mechanisms of osmosensation and systemic E02440. osmoregulation. Nat Rev Neurosci. 2008;9:519–31. https://​doi.​org/​10.​ 32. Obici S, Feng Z, Arduini A, et al. Inhibition of hypothalamic carnitine 1038/​NRN24​00. palmitoyltransferase-1 decreases food intake and glucose production. 53. Murphy D, Konopacka A, Hindmarch C, et al. The hypothalamic-neuro‑ Nat Med. 2003;9:756–61. hypophyseal system: from genome to physiology. J Neuroendocrinol. 33. Loftus TM, Jaworsky DE, Frehywot GL, et al. Reduced food intake and 2012;24:539–53. https://​doi.​org/​10.​1111/J.​1365-​2826.​2011.​02241.X. body weight in mice treated with fatty acid synthase inhibitors. Science. 54. Li Z, Okamoto KI, Hayashi Y, Sheng M. The importance of dendritic 2000;288:2379–81. https://​doi.​org/​10.​1126/​scien​ce.​288.​5475.​2379. mitochondria in the morphogenesis and plasticity of spines and 34. Gao S, Lane MD. Effect of the anorectic fatty acid synthase inhibitor C75 synapses. Cell. 2004;119:873–87. https://​doi.​org/​10.​1016/J.​CELL.​2004.​ on neuronal activity in the hypothalamus and brainstem. Proc Natl Acad 11.​003. Sci U S A. 2003;100:5628–33. https://​doi.​org/​10.​1073/​pnas.​10316​98100. 55. Witte MM, Resuehr D, Chandler AR, et al. Female mice and rats exhibit 35. Esser V, Kuwajima M, Britton CH, et al. Inhibitors of mitochondrial carni‑ species-specific metabolic and behavioral responses to ovariectomy. Gen tine palmitoyltransferase I limit the action of proteases on the enzyme. Comp Endocrinol. 2010;166:520–8. https://​doi.​org/​10.​1016/j.​ygcen.​2010.​ Isolation and partial amino acid analysis of a truncated form of the rat 01.​006. liver isozyme. J Biol Chem. 1993;268:5810–6. 56. Asarian L, Geary N. Modulation of appetite by gonadal steroid hormones. 36. Mera P, Mir JF, Fabriàs G, et al. Long-term increased carnitine palmi‑ Philos Trans Royal Soc B: Biol Sci. 2006;361:1251–63. https://​doi.​org/​10.​ toyltransferase 1A expression in ventromedial hypothalamus causes 1098/​rstb.​2006.​1860. hyperphagia and alters the hypothalamic lipidomic profile. PLoS ONE. 57. Martínez de Morentin PB, González-García I, Martins L, et al. Estradiol 2014. https://​doi.​org/​10.​1371/​journ​al.​pone.​00971​95. regulates brown adipose tissue thermogenesis via hypothalamic AMPK. 37. Kim EK, Miller I, Aja S, et al. C75, a fatty acid synthase inhibitor, reduces Cell Metab. 2014;1:41–53. https://​doi.​org/​10.​1016/j.​cmet.​2014.​03.​031. food intake via hypothalamic AMP-activated protein kinase. J Biol 58. Chen X, Wang L, Loh DH, et al. Sex differences in diurnal rhythms of food Chem. 2004;279:19970–6. https://​doi.​org/​10.​1074/​jbc.​M4021​65200. intake in mice caused by gonadal hormones and complement of sex 38. Hu Z, Cha SH, Chohnan S, Lane MD. Hypothalamic malonyl-CoA as a chromosomes. Horm Behav. 2015;75:55–63. https://​doi.​org/​10.​1016/j.​ mediator of feeding behavior. Proc Natl Acad Sci. 2003;100:12624–9. yhbeh.​2015.​07.​020. https://​doi.​org/​10.​1073/​pnas.​18344​02100. 59. Yang Y, Atasoy D, Su HH, Sternson SM. Hunger states switch a flip-flop 39. Mera P, Bentebibel A, López-Viñas E, et al. C75 is converted to C75-CoA memory circuit via a synaptic AMPK-dependent positive feedback loop. in the hypothalamus, where it inhibits carnitine palmitoyltransferase Cell. 2011;146:992–1003. https://​doi.​org/​10.​1016/j.​cell.​2011.​07.​039. 1 and decreases food intake and body weight. Biochem Pharmacol. 60. Liu T, Kong D, Shah BP, et al. Fasting activation of AgRP neurons requires 2009. https://​doi.​org/​10.​1016/j.​bcp.​2008.​11.​020. NMDA receptors and involves spinogenesis and increased excitatory 40. Mir JF, Zagmutt S, Lichtenstein MP, et al. Ghrelin causes a decline in tone. Neuron. 2012;73:511–22. https://​doi.​org/​10.​1016/j.​neuron.​2011.​11.​ GABA release by reducing fatty acid oxidation in cortex. Mol Neurobiol. 027. 2018. https://​doi.​org/​10.​1007/​s12035-​018-​0921-3. 61. Reichenbach A, Stark R, Mequinion M, et al. AgRP neurons require 41. Wang Q, Liu C, Uchida A, et al. Arcuate AgRP neurons mediate orexi‑ carnitine acetyltransferase to regulate metabolic flexibility and peripheral genic and glucoregulatory actions of ghrelin. Mol Metab. 2014;3:64–72. nutrient partitioning. Cell Rep. 2018;22:1745–59. https://​doi.​org/​10.​ https://​doi.​org/​10.​1016/j.​molmet.​2013.​10.​001. 1016/J.​CELREP.​2018.​01.​067. 42. Sanz E, Yang L, Su T, et al. Cell-type-specific isolation of ribosome- 62. Chen SR, Chen H, Zhou JJ, et al. Ghrelin receptors mediate ghrelin- associated mRNA from complex tissues. Proc Natl Acad Sci. induced excitation of agouti-related protein/neuropeptide Y but not 2009;106:13939–44. https://​doi.​org/​10.​1073/​pnas.​09071​43106. pro-opiomelanocortin neurons. J Neurochem. 2017;142:512–20. https://​ 43. Wu Q, Lemus M, Stark R, et al. The temporal pattern of cfos activation in doi.​org/​10.​1111/​JNC.​14080. hypothalamic, cortical, and brainstem nuclei in response to fasting and 63. Wang Q, Liu C, Uchida A, et al. Arcuate AgRP neurons mediate orexigenic refeeding in male mice. Endocrinology. 2014;155:en20131831. https://​ and glucoregulatory actions of ghrelin. Mol Metab. 2013;3:64–72. https://​ doi.​org/​10.​1210/​en.​2013-​1831. doi.​org/​10.​1016/J.​MOLMET.​2013.​10.​001. 44. Sun Y, Wang P, Zheng H, Smith RG. Ghrelin stimulation of growth hor‑ 64. Wu CS, Bongmba OYN, Yue J, et al. Suppression of GHS-R in AgRP neurons mone release and appetite is mediated through the growth hormone mitigates diet-induced obesity by activating thermogenesis. Int J Mol Sci. secretagogue receptor. Proc Natl Acad Sci USA. 2004;101:4679–84. 2017. https://​doi.​org/​10.​3390/​IJMS1​80408​32. https://​doi.​org/​10.​1073/​pnas.​03059​30101. 65. Lage R, Diéguez C, Vidal-Puig A, López M. AMPK: a metabolic gauge regu‑ 45. Cañes L, Martí-Pàmies I, Ballester-Servera C, et al. High NOR-1 (neuron- lating whole-body energy homeostasis. Trends Mol Med. 2008;14:539–49. derived orphan receptor 1) expression strengthens the vascular wall https://​doi.​org/​10.​1016/j.​molmed.​2008.​09.​007. response to angiotensin II leading to aneurysm formation in mice. 66. López M, Lelliott CJ, Vidal-Puig A. Hypothalamic fatty acid metabo‑ Hypertension. 2021;77:557–70. https://​doi.​org/​10.​1161/​HYPER​TENSI​ lism: a housekeeping pathway that regulates food intake. BioEssays. ONAHA.​120.​16078. 2007;29:248–61. https://​doi.​org/​10.​1002/​bies.​20539. 46. Teilmann AC, Nygaard Madsen A, Holst B, et al. Physiological and 67. Andrews ZB, Liu Z-W, Walllingford N, et al. UCP2 mediates ghrelin’s action pathological impact of blood sampling by retro-bulbar sinus puncture on NPY/AgRP neurons by lowering free radicals. Nature. 2008;454:846–51. and facial vein phlebotomy in laboratory mice. PLoS ONE. 2014;9:1–19. https://​doi.​org/​10.​1038/​natur​e08132. https://​doi.​org/​10.​1371/​journ​al.​pone.​01132​25. 68. Bosy-Westphal A, Hägele FA, Müller MJ. What is the impact of energy 47. Paxinos G, Watson C. The Rat brain: in stereotaxic coordinates. 4th ed. expenditure on energy intake? Nutrients. 2021;13:3508. San Diego: Academic Press; 1998. 69. Steculorum SM, Ruud J, Karakasilioti I, et al. AgRP neurons control sys‑ 48. Lin X, Itoga CA, Taha S, et al. c-Fos mapping of brain regions activated temic insulin sensitivity via myostatin expression in brown adipose tissue. by multi-modal and electric foot shock stress. Neurobiol Stress. Cell. 2016;165:125–38. https://​doi.​org/​10.​1016/J.​CELL.​2016.​02.​044. 2018;8:92–102. https://​doi.​org/​10.​1016/j.​ynstr.​2018.​02.​001. 70. Shi YC, Lau J, Lin Z, et al. Arcuate NPY controls sympathetic output and 49. Galarraga M, Campión J, Muñoz-Barrutia A, et al. Adiposoft: automated BAT function via a relay of tyrosine hydroxylase neurons in the PVN. Cell software for the analysis of white adipose tissue cellularity in histologi‑ Metab. 2013;17:236–48. https://​doi.​org/​10.​1016/J.​CMET.​2013.​01.​006. cal sections. J Lipid Res. 2012;53:2791–6. https://​doi.​org/​10.​1194/​jlr.​ 71. Cano G, Passerin AM, Schiltz JC, et al. Anatomical substrates for the cen‑ D0237​88. tral control of sympathetic outflow to interscapular adipose tissue during 50. Gupta R, Ma Y, Wang M, Whim MD. AgRP-expressing adrenal chromaffin cold exposure. J Comp Neurol. 2003;460:303–26. https://​doi.​org/​10.​1002/​ cells are involved in the sympathetic response to fasting. Endocrinol‑ cne.​10643. ogy. 2017;158:2572–84. https://​doi.​org/​10.​1210/​en.​2016-​1268. 72. Cavalcanti-de-Albuquerque JP, Zimmer MR, Bober J, Dietrich MO. Rapid 51. Japundžić-Žigon N. Vasopressin and oxytocin in control of the cardio‑ shift in substrate utilization driven by hypothalamic Agrp neurons. vascular system. Curr Neuropharmacol. 2013;11:218–30. https://​doi.​ BioRxiv. 2016. https://​doi.​org/​10.​1101/​086348. org/​10.​2174/​15701​59X11​31102​0008. Zagmutt et al. Biology of Sex Differences (2023) 14:14 Page 26 of 26 73. Deng J, Yuan F, Guo Y, et al. Deletion of ATF4 in AgRP neurons promotes Publisher’s Note fat loss mainly via increasing energy expenditure. Diabetes. 2017;66:640– Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 50. https://​doi.​org/​10.​2337/​db16-​0954. lished maps and institutional affiliations. 74. Ruan H-B, Dietrich M, Liu Z-W, et al. O-GlcNAc transferase enables AgRP neurons to suppress browning of white fat. Cell. 2014;159:306–17. https://​doi.​org/​10.​1016/j.​cell.​2014.​09.​010. 75. Könner AC, Janoschek R, Plum L, et al. Insulin action in AgRP-expressing neurons is required for suppression of hepatic glucose production. Cell Metab. 2007;5:438–49. https://​doi.​org/​10.​1016/j.​cmet.​2007.​05.​004. 76. Brown JJ, Chinn RH, Lever AF, Robertson JI. Renin and angiotensin as a mechanism of diuretic-induced antidiuresis in diabetes insipidus. Lancet. 1969;293:237–9. https://​doi.​org/​10.​1016/​s0140-​6736(69)​91244-6. 77. Abdulla H, John E. Renin-angiotensin and sympathetic nervous system interactions in the control of blood pressure in fructose-induced meta‑ bolic syndrome. J Autacoids. 2012. https://​doi.​org/​10.​4172/​2161-​0479.​ 1000e​105. 78. Jiang J, Morgan DA, Cui H, Rahmouni K. Activation of hypothalamic AgRP and POMC neurons evokes disparate sympathetic and cardiovascular responses activation of hypothalamic AgRP and POMC neurons evokes disparate sympa-thetic and cardiovascular responses. Am J Physiol Heart Circ Physiol. 2020;319:1069–77. https://​doi.​org/​10.​1152/​ajphe​art.​00411.​ 2020.-​The. 79. Garabadu D, Agrawal N, Sharma A, Sharma S. Mitochondrial metabolism: a common link between neuroinflammation and neurodegeneration. Behav Pharmacol. 2019;30:641–51. https://​doi.​org/​10.​1097/​FBP.​00000​ 00000​000505. 80. Ventura-Clapier R, Moulin M, Piquereau J, et al. Mitochondria: a central target for sex differences in pathologies. Clin Sci (Lond). 2017;131:803–22. https://​doi.​org/​10.​1042/​CS201​60485. 81. Guevara R, Santandreu FM, Valle A, et al. Sex-dependent differences in aged rat brain mitochondrial function and oxidative stress. Free Radic Biol Med. 2009;46:169–75. https://​doi.​org/​10.​1016/J.​FREER​ADBIO​MED.​2008.​ 09.​035. 82. Jin S, Yoon NA, Liu ZW, et al. Drp1 is required for AgRP neuronal activity and feeding. Elife. 2021. https://​doi.​org/​10.​7554/​eLife.​64351. 83. Lu XY, Barsh GS, Akil H, Watson SJ. Interaction between α-melanocyte- stimulating hormone and corticotropin-releasing hormone in the regula‑ tion of feeding and hypothalamo-pituitary-adrenal responses. J Neurosci. 2003;23:7863–72. https://​doi.​org/​10.​1523/​JNEUR​OSCI.​23-​21-​07863.​2003. 84. Cansell C, Denis RGP, Joly-Amado A, et al. Arcuate AgRP neurons and the regulation of energy balance. Front Endocrinol. 2012. https://​doi.​org/​10.​ 3389/​FENDO.​2012.​00169. 85. Cui Z, Smith AS. In vivo measurement of enhanced agouti-related peptide release in the paraventricular nucleus of the hypothalamus through G s activation of agouti-related peptide neurons. J Biol Methods. 2019;6:e116. https://​doi.​org/​10.​14440/​JBM.​2019.​288. 86. Zhang C, Forlano PM, Cone RD. AgRP and POMC neurons are hypophysi‑ otropic and coordinately regulate multiple endocrine axes in a larval teleost. Cell Metab. 2012;15:256–64. https://​doi.​org/​10.​1016/J.​CMET.​2011.​ 12.​014. 87. Butler AA, Cone RD. The melanocortin receptors: lessons from knockout models. Neuropeptides. 2002;36:77–84. https://​doi.​org/​10.​1054/​NPEP.​ 2002.​0890. 88. Krashes MJ, Lowell BB, Garfield AS. Melanocortin-4 receptor-regulated energy homeostasis. Nat Neurosci. 2016;19:206–19. https://​doi.​org/​10.​ 1038/​NN.​4202. 89. Han Y, Xia G, Srisai D, et al. Deciphering an AgRP-serotoninergic neural cir‑ Ready to submit your research ? Choose BMC and benefit from: cuit in distinct control of energy metabolism from feeding. Nat Commun. 2021. https://​doi.​org/​10.​1038/​S41467-​021-​23846-X. • fast, convenient online submission 90. Wu Q, Boyle MP, Palmiter RD. Loss of GABAergic signaling by AgRP neu‑ rons to the parabrachial nucleus leads to starvation. Cell. 2009;137:1225– • thorough peer review by experienced researchers in your field 34. https://​doi.​org/​10.​1016/J.​CELL.​2009.​04.​022. • rapid publication on acceptance 91. Ohbuchi T, HaamTasker JJG. Regulation of neuronal activity in hypotha‑ • support for research data, including large and complex data types lamic vasopressin neurons. Interdiscip Inf Sci. 2015;21:225–34. https://​doi.​ org/​10.​4036/​IIS.​2015.B.​07. • gold Open Access which fosters wider collaboration and increased citations 92. Bé J-C, Lin D-T, Kang M-G, et al. Synapse-specific regulation of AMPA • maximum visibility for your research: over 100M website views per year receptor function by PSD-95. PNAS. 2006;103:19535–40. 93. Padilla SL, Qiu J, Nestor CC, et al. AgRP to Kiss1 neuron signaling links At BMC, research is always in progress. nutritional state and fertility. Proc Natl Acad Sci USA. 2017;114:2413–8. https://​doi.​org/​10.​1073/​pnas.​16210​65114. Learn more biomedcentral.com/submissions