DMH^Ppp1r17 neurons regulate aging and lifespan in mice through hypothalamic-adipose inter-tissue communication

SUMMARY

Recent studies have shown that the hypothalamus functions as a control center of aging in mammals that counteracts age-associated physiological decline through inter-tissue communications. We have identified a key neuronal subpopulation in the dorsomedial hypothalamus (DMH), marked by Ppp1r17 expression (DMH^Ppp1r17 neurons), that regulates aging and longevity in mice. DMH^Ppp1r17 neurons regulate physical activity and WAT function, including the secretion of extracellular nicotinamide phosphoribosyltransferase (eNAMPT), through sympathetic nervous stimulation. Within DMH^Ppp1r17 neurons, the phosphorylation and subsequent nuclear-cytoplasmic translocation of Ppp1r17, regulated by cGMP-dependent protein kinase (PKG; Prkg1), affect gene expression regulating synaptic function, causing synaptic transmission dysfunction and impaired WAT function. Both DMH-specific Prkg1-knockdown, which suppresses age-associated Ppp1r17 translocation, and the chemogenetic activation of DMH^Ppp1r17 neurons significantly ameliorate age-associated dysfunction in WAT, increase physical activity, and extend lifespan. Thus, these findings clearly demonstrate the importance of the inter-tissue communication between the hypothalamus and WAT in mammalian aging and longevity control.

Graphical Abstract

eTOC

Tokizane et al. demonstrate that DMH^Ppp1r17 neurons regulate the inter-tissue communication between the hypothalamus and white adipose tissue (WAT). Within DMH^Ppp1r17 neurons, age-associated Ppp1r17 translocation, regulated by protein kinase G (Prkg1), causes synaptic dysfunction and WAT impairment. DMH-specific Prkg1 knockdown and chemogenetic activation of DMH^Ppp1r17 neurons counteract aging and extend lifespan.

INTRODUCTION

Recent studies have revealed that the dysfunction of inter-tissue communications underlies age-associated physiological decline, inducing a variety of aging phenotypes and limiting healthspan and lifespan ^1–3. Indeed, a concept called “Interorgan Communication in Aging” was specifically introduced as a promising research direction in the September 2022 meeting of the National Advisory Council on Aging (NACA) (https://www.nia.nih.gov/approved-concepts#InterOrgan). In mammals, the hypothalamus functions as a high-order “control center of aging,” counteracting age-associated pathophysiological changes and thereby promoting longevity ^4,5. The hypothalamus is well known to regulate the production of a variety of hormones through the pituitary gland and control the autonomic nervous system directed to peripheral organs, including the liver, adipose tissue, skeletal muscle, intestine, and others. For example, gonadotropin-releasing hormone, whose gene expression is suppressed by NFκB activation in the aged mediobasal hypothalamus, counteracts age-associated functional decline ⁵, and the frequency of its pulsatile secretion is correlated with lifespan ⁶. We have previously demonstrated that the mammalian NAD⁺-dependent protein deacetylase SIRT1 in the hypothalamus, particularly in the dorsomedial and lateral hypothalamic nuclei (DMH and LH, respectively), plays a critical role in delaying aging and extending lifespan in mice ⁴. In the DMH and LH, SIRT1 interacts with and deacetylates Nk2 homeobox 1 (Nkx2-1), a transcription factor, and SIRT1 and Nkx2-1 together regulate key genes for neural activation, including orexin type 2 receptor and PR domain containing 13. The SIRT1/Nkx2-1-mediated signaling in the DMH and LH is critical to maintain youthful physiological traits, including physical activity, energy expenditure, body temperature, and sleep quality, and promote the sympathetic nervous tone directed to skeletal muscle through the maintenance of neuromuscular junction structures ^4,7. More recently, it has also been demonstrated that LH neurons expressing Slc12a8, a transporter specific to a key NAD⁺ intermediate, nicotinamide mononucleotide (NMN), regulate energy metabolism and skeletal muscle function through β2 adrenergic receptor-mediated sympathetic nervous stimulation ⁸. Remarkably, the restoration of Slc12a8 expression in the LH ameliorates sarcopenia and frailty-like symptoms in aged mice, further demonstrating that the maintenance of hypothalamic function is critical to counteract aging phenotypes and promote healthspan and lifespan.

On the other hand, factors secreted from peripheral tissues and organs play an important role in maintaining hypothalamic function and promoting lifespan. The expression and secretion of FGF21 are significantly enhanced by skeletal muscle-specific 4E-BP1 activation, providing protection against age-associated metabolic decline ⁹. It has also been reported that transgenic overexpression of FGF21 extends lifespan in mice ¹⁰ and that hepatic FGF21 is required for the metabolic and lifespan-extending effect of protein restriction in mice ^11,12. Another interesting secretory factor that controls aging and longevity in mice is extracellular nicotinamide phosphoribosyltransferase (eNAMPT). NAMPT is the rate-limiting enzyme in a major NAD⁺ biosynthetic pathway in mammals, converting nicotinamide and 5’-phosphoribose pyrophosphate to NMN. eNAMPT is actively secreted from adipose tissue through a process that is regulated by SIRT1/NAD⁺-mediated deacetylation, and enhances NAD⁺ biosynthesis remotely in the hypothalamus ^13,14, the nucleus accumbens ¹⁵, and other tissues ¹⁶. Intriguingly, eNAMPT is encapsulated into extracellular vesicles (eNAMPT-EVs), and the genetic and pharmacological increases in circulating eNAMPT-EVs delay aging and promote healthspan and lifespan in mice ¹⁶, suggesting that the inter-tissue communication between adipose tissue and the hypothalamus mediated by eNAMPT-EVs is critical for aging and longevity control in mammals.

In this study, we identified and characterized Protein Phosphatase 1 Regulatory Subunit 17 (Ppp1r17) neurons (DMH^Ppp1r17 neurons), a key neuronal subpopulation localized in the DMH, for the inter-tissue communication between the hypothalamus and white adipose tissue (WAT) and the regulation of aging and longevity in mammals. DMH^Ppp1r17 neurons belong to the previously identified SIRT1/Nkx2-1-double positive neuronal population (DMH^SIRT1/Nkx2-1 neurons) and are specifically localized in the compact region of the DMH (cDMH). Ppp1r17 was originally found in rabbit cerebellum as a substrate of protein kinase G (PKG) ¹⁷. Whereas Ppp1r17 has been speculated to function as an inhibitor of protein phosphatase (PP) 1 and PP2A ^18,19, its exact function remains poorly investigated. Ppp1r17 is also well conserved in various vertebrate species such as human, chimpanzee, monkey, rat, mouse, bovine, rabbit, chicken, xenopus, and zebra finch ²⁰, suggesting that Ppp1r17 carries some essential functions throughout evolution. Indeed, a recent study has revealed that Ppp1r17 is a gene likely involved in human-specific cortical neurodevelopment ²¹, and another study has suggested that Ppp1r17 is important to restrict overconsumption of food ²². Importantly, we found that DMH^Ppp1r17 neurons regulate WAT through stimulation of the sympathetic nervous system specifically directed to WAT and promote WAT functions, including the secretion of eNAMPT. Both DMH-specific knockdown of Prkg1 that encodes PKG, which suppresses age-associated nuclear-cytoplasmic translocation of Ppp1r17, and the chemogenetic activation of DMH^Ppp1r17 neurons significantly ameliorated age-associated dysfunction of WAT, increased circulating eNAMPT levels and physical activity, and extended lifespan in mice. Our findings clearly demonstrate that the inter-tissue communication between the hypothalamus and WAT mediated by DMH^Ppp1r17 neurons plays a critical role in counteracting age-associated physiological decline and promoting longevity in mammals.

RESULTS

Identification and characterization of DMH^Ppp1r17 neurons

In our previous study, we identified DMH^SIRT1/Nkx2-1 neurons as a key neuronal population enriched in the cDMH for mammalian aging and longevity control (Figure S1A) ⁴. Thus, we aimed to examine what neuronal subpopulations existed within DMH^SIRT1/Nkx2-1 neurons. Using Nkx2-1-CreERT2 mice crossed with Rosa-ZsGreen mice after tamoxifen induction, we collected ZsGreen-marked DMH^SIRT1/Nkx2-1 neurons and other Nkx2-1-positive neurons in the arcuate nucleus (Arc) and ventromedial hypothalamic nucleus (VMH) by laser microdissection (Figure S1B) and compared gene expression profiles between the Arc, VMH, and DMH. We were able to determine genes selectively expressed greater than or equal to 15-fold in DMH^SIRT1/Nkx2-1 neurons, compared to Nkx2-1-positive neurons from the Arc and VMH (Figure 1A). Among those DMH^SIRT1/Nkx2-1 neuron-specific genes, we found that Ppp1r17 was specifically expressed through the entire cDMH, from anterior to posterior parts of the DMH (Figure 1B and Movie S1). We also confirmed that DMH^Ppp1r17 neurons belonged to DMH^SIRT1/Nkx2-1 neurons in the cDMH (Figure S1C). Using vGluT2-ires-Cre or vGAT-ires-Cre mice crossed with Rosa-ZsGreen mice, in which glutamatergic and GABAergic neurons are labeled with ZsGreen, respectively, we found that the majority of DMH^Ppp1r17 neurons were glutamatergic (Figures 1C and D). These results were also confirmed by in situ hybridization (Figure S1D). Whereas the percentage of glutamatergic DMH^Ppp1r17 neurons decreased towards the most posterior part of the cDMH (Figure 1C), the percentage of GABAergic DMH^Ppp1r17 neurons increased (Figure 1D). Tyrosine hydroxylase (TH)-positive neurons also existed in this area ²³, but DMH^Ppp1r17 neurons did not overlap with those TH-positive neurons (Figure S1E).

Figure 1. Ppp1r17 is specifically expressed in a subset of DMH^SIRT1/Nkx2-1 neurons, and DMH^Ppp1r17 neurons are mainly glutamatergic.

(A) Heat map representation of genes enriched in DMH^SIRT1/Nkx2-1 neurons, compared to the Nkx2-1-positive neurons of the Arc and VMH. Genes expressed greater than or equal to 15-fold in each hypothalamic nucleus compared to other hypothalamic nuclei are shown in the heat map. Expression levels are indicated by a colored log2 scale of signal intensities (right), and corresponding symbol IDs are shown for each gene (left). (B) Localization of DMH^Ppp1r17 neurons throughout the entire DMH. The DMH is schematically shown from anterior (bregma −1.40 mm) to posterior (bregma −2.04 mm) parts of the hypothalamus (left). The orange areas indicate the compact region of the DMH. Immunohistochemical localization of Ppp1r17 through anterior to posterior parts of the DMH are shown (right). Scale bar, 200 μm. (C, D) Representative images of immunohistochemical localization of Ppp1r17 in the DMH region in vGluT2- or vGAT-Cre mice crossed with Rosa-ZsGreen mice. ZsGreen marks vGluT2 (C) or vGAT (D) positive neurons. Low (upper lanes) and high (lower lanes) magnification images for vGluT2 (C)- or vGAT (D)- and Ppp1r17-positive neurons in the DMH, and percentage of the double-positive neurons throughout the entire DMH (from bregma −1.56mm to −1.96mm) are shown. Scale bars indicate 200 μm in low magnification and 100 μm in high magnification images.

DMH^Ppp1r17 neurons project to the preoptic area (POA), the paraventricular nucleus (PVN), the LH, and the periaqueductal gray (PAG)

Next, we examined which parts of the brain DMH^Ppp1r17 neurons project to. For this purpose, we obtained a Ppp1r17-Cre mouse line, and Cre expression in DMH^Ppp1r17 neurons was confirmed by in situ hybridization (Figure S2A). We traced the projections of DMH^Ppp1r17 neurons by injecting the double-floxed EYFP adeno-associated virus (AAV) unilaterally to the cDMH of Ppp1r17-Cre mice (Figures 2A and B). EYFP-labeled axon terminals were detected mainly in the POA, the PVN, the LH, and the PAG (Figure 2C), but not in the SCN, the locus coeruleus (LC), and the raphe pallidus (RPa) (Figure S2B), which have been reported previously as projection sites from the DMH ^24–26. Because all these identified projection sites are involved in the homeostatic feedback regulation between the hypothalamus and peripheral tissues, such as the regulation of sympathetic outflow ²⁷, we suspected that DMH^Ppp1r17 neurons may signal to a specific tissue and regulate its function.

Figure 2. DMH^Ppp1r17 neurons project to the POA, PVN, LH, and PAG.

(A) A scheme of the AAV injection (AAV-Ef1a-DIO-EYFP) into the DMH of Ppp1r17-Cre mice for labeling axons. (B) Ppp1r17-Cre mice injected with AAV-Ef1a-DIO-EYFP into the DMH show the co-localization of Ppp1r17 and EYFP. A higher magnification image of the white box is shown at right. Scale bars indicate 200 μm in low magnification and 50 μm in high magnification image. (C) Fluorescent images of the POA, PVN, LH and PAG show EYFP-labeled terminals of DMH^Ppp1r17 neurons. POA, preoptic area; PVN, paraventricular nucleus; 3V, third ventricular; LH, lateral hypothalamus; PAG, periaqueductal grey; AQ, mesencephalic aqueduct. Scale bar, 200 μm.

DMH-specific Ppp1r17-knockdown increases fat mass and decreases wheel-running activity

To examine whether DMH^Ppp1r17 neurons indeed play an important role in any inter-tissue communication with the hypothalamus, we first decided to assess the functional significance of Ppp1r17 itself in these neurons. Whereas it has been reported that Ppp1r17 regulates neural progenitor cell-cycle progression ²¹, its cellular function in the adult brain remains poorly understood. Thus, we generated DMH-specific Ppp1r17-knockdown (Ppp1r17-KD) mice by stereotactically injecting a lentivirus expressing Ppp1r17 shRNA into the DMH of 3-month-old mice (Figure 3A). The knockdown efficiency was approximately 50% in the DMH (Figure S3A). Based on EGFP expression due to the lentivirus infection, Ppp1r17 shRNA virus-infected neurons remained intact (Figure S3B). Interestingly, DMH-specific Ppp1r17-KD mice gained significant body weight within 60 days after the injection, compared to fLuc shRNA-expressing AAV-injected control mice (Figure 3B). DMH-specific Ppp1r17-KD mice also showed increased fat mass, but no change in lean mass (Figure 3C). Whereas they showed no change in food intake during the dark time, they showed moderate but significant increases in food intake during the light time (Figure S3C). Furthermore, DMH-specific Ppp1r17-KD mice exhibited significant decreases in wheel-running activity during the dark time (Figures 3D and S3D). No differences in rectal body temperature were detected between DMH-specific Ppp1r17-KD and control mice (Figure S3E). To verify that these phenotypes were due to decreased DMH^Ppp1r17 neuronal function, we ablated DMH^Ppp1r17 neurons by injecting an AAV expressing diphtheria toxin A (dtA) in a Cre-dependent manner into the DMH of Ppp1r17-Cre mice (Figures S3F and G). Consistent with DMH-specific Ppp1r17-KD mice, body weight and fat mass were significantly increased in the DMH^Ppp1r17 neuron-ablated mice, compared to control wild-type mice to which the same dtA-expressing AAV was injected (Figures S3H and I). No differences were detected in lean mass and food intake between the groups (Figures S3J and K). These results suggest that Ppp1r17 is critical to maintain DMH^Ppp1r17 neuronal function, whose dysfunction results in increased fat mass and decreased physical activity.

Figure 3. DMH-specific Ppp1r17-KD mice exhibit increased fat mass and decreased wheel running activity.

(A) Schematic representation of viral injection of shRNA against fLuc (control) or Ppp1r17 into the DMH of 3-month-old male mice and the timeline for measurements of indicated parameters in DMH-specific Ppp1r17-KD mice. (B) Body weights of control fLuc-KD or Ppp1r17-KD mice were measured (n=11-15). Results are presented as mean ± SEM (***P < 0.001 Student’s t test). (C) Fat and lean masses of control fLuc-KD or Ppp1r17-KD mice were measured by EchoMRI 30 days after viral injection (n=11-15). Results are presented as mean ± SEM (****P < 0.0001, Student’s t test). (D) Wheel-running activities in DMH-specific control fLuc-KD or Ppp1r17-KD mice were examined (*P < 0.05, Wilcoxon matched-pairs signed-rank test, n=5-7). Gray areas indicate dark time.

DMH-specific Ppp1r17 KD decreases lipolysis and sympathetic nerve innervation in WAT

Because DMH^Ppp1r17 neurons were found to project to the PVN and PAG that are important for sympathetic outflow directed to adipose tissue, we suspected that DMH^Ppp1r17 neurons may directly regulate adipose tissue function. We first compared the histology of WAT between DMH-specific fLuc-KD control and Ppp1r17-KD mice. DMH-specific Ppp1r17-KD mice exhibited larger white adipocytes that appeared to accumulate more lipid (Figure 4A). Consistent with this observation, the expression of lipolytic genes, including Lipe, Pparα, and Adrb3, were reduced in both inguinal and epididymal WAT depots (iWAT and eWAT) from DMH-specific Ppp1r17-KD mice, compared to those from control mice (Figures 4B and C). The expression of Ucp1, a thermogenic gene, was also reduced significantly in iWAT (Figure 4B). Interestingly, the mRNA expression levels of nicotinamide phosphoribosyltransferase (Nampt), the rate-limiting enzyme for NAD⁺ biosynthesis in mammals, were significantly decreased in both iWAT and eWAT (Figures 4B and C). Contrarily to WAT, the gene expression profiles were not altered in brown adipose tissue (BAT), liver, and soleus muscle (Figures S4A–C), except for the decrease in Nampt expression in BAT and the increase in Fasn expression in the liver. These results provide compelling evidence supporting that DMH^Ppp1r17 neurons predominantly regulate WAT function. It has been well established that WAT is innervated by the sympathetic nervous system (SNS), inducing lipolytic response and reducing white adipose mass. One of the molecular markers for sympathetic nervous stimulation and lipolysis in WAT is the phosphorylation of hormone-sensitive lipase (HSL) ²⁸. Indeed, levels of phosphorylated HSL (p-HSL) were significantly reduced in the iWAT from DMH-specific Ppp1r17-KD mice, compared to fLuc-KD control mice (Figure 4D). This was also confirmed by immunostaining of p-HSL in iWAT sections (Figure S4D). These changes in p-HSL levels were not observed in BAT (Figure S4E), further supporting the selective effect of DMH^Ppp1r17 neurons on WAT function. Consistent with this notion, using a tissue-clearing technique called immunolabeling-enabled imaging of solvent-cleared organs (iDISCO) and 3D-image reconstruction ^29,30, we found that the sympathetic nerve innervation in iWAT was dramatically reduced in DMH-specific Ppp1r17-KD mice, compared to fLuc-KD control mice (Figure 4E). Indeed, the values of normalized TH fiber areas and the numbers of TH fiber branches were significantly decreased in DMH-specific Ppp1r17-KD mice, compared to fLuc-KD control mice (Figures 4F–G and S4F). Taken together, these findings strongly suggest that DMH^Ppp1r17 neurons directly regulate WAT functions through the activation of the SNS.

Figure 4. DMH-specific Ppp1r17-KD mice exhibit impairments of lipolysis and sympathetic nerve innervation in WAT.

(A) Paraffin-embedded iWAT sections of control fLuc-KD or Ppp1r17-KD mice were stained with hematoxylin and eosin. The quantification of the size of the adipocytes from control fLuc-KD or Ppp1r17-KD mice are shown at the bottom. Scale bar, 200 μm. (B, C) mRNA expression levels of genes related to lipolysis, adipocyte proliferation, and NAD⁺ biosynthesis in iWAT and eWAT of control fLuc-KD or Ppp1r17-KD mice were examined 60 days after virus injection. mRNA expression levels were normalized to those in control fLuc-KD mice. Results are presented as mean ± SEM (*P < 0.05, **P < 0.01, ***P < 0.001, two-way ANOVA with Šidák’s multiple comparisons test, n=5-7). (D) Phosphorylated HSL (p-HSL) and total HSL protein levels in iWAT from control fLuc-KD and Ppp1r17-KD mice were analyzed by Western blotting. p-HSL levels were normalized to total HSL levels. Results are presented as mean ± SEM (**P < 0.01, Student’s t test, n=5 per group). (E) Sympathetic nerve innervation visualized with tyrosine hydroxylase (TH) immunostaining in iDISCO-treated iWATs from control fLuc-KD and Ppp1r17-KD mice. (F, G) TH fiber areas and branches in iWAT from control fLuc-KD or Ppp1r17-KD mice were quantified by Imaris software (n=4). Results are presented as mean ± SEM (**P < 0.01, Student’s t test).

Ppp1r17 regulates synaptic function in DMH^Ppp1r17 neurons

How does Ppp1r17 mediate such DMH^Ppp1r17 neuronal function? To address this question, we performed RNA sequencing (RNA-seq) using DMH-specific Ppp1r17-KD and control fLuc-KD mice to determine how the gene expression profile is affected by Ppp1r17. Unbiased principal component analysis proved that DMH gene expression profiles in Ppp1r17-KD and fLuc-KD mice clustered separately (Figure S5A). Differential gene expression analysis showed that a total of 765 genes (6.9% of expressed genes) were differentially expressed in the DMH of Ppp1r17-KD mice (Figures S5B and C). Of these, 385 genes were up-regulated, and 380 were down-regulated. Because we happened to have two different groups in Ppp1r17-KD mice that exhibited moderately and highly decreased Ppp1r17 expression levels (Figure S5D), the Weighted Correlation Network Analysis (WGCNA) was able to be applied to define a hub gene set regulated by Ppp1r17 (Figure S5E). From this analysis, two gene modules that showed conspicuous significance, blue (up-regulated) and turquoise (down-regulated) modules, were identified. Gene ontology (GO) analysis revealed that whereas the blue module contained genes highly and significantly related to protein metabolism and catabolism processes and translation (Figures 5A and B), the turquoise module had genes related to synaptic signaling and function (Figures 5C and D). For example, neuroglobin (Ngb), synaptoporin (Synpr), neurensin 1 (Nrsn1), ras-related protein rab 1 (Rab1), glutamate metabotropic receptor 5 (Grm5), solute carrier family 12 member 5 (Slc12a5), and gephyrin (Gphn) were all associated with synaptic functions (Figure 5E). Interestingly, Synaptic GO (SynGO) analysis further proved that both pre- and post-synaptic cellular components were affected by Ppp1r17-KD (Figure S5F), implicating that Ppp1r17 may regulate synaptic transmission in DMH^Ppp1r17 neurons through the regulation of these genes. To further address this possibility, we employed the multiple- or micro-electrode array (MEA) system using primary embryonic hypothalamic neurons (Figure 5F). We confirmed that primary embryonic hypothalamic neurons highly expressed Ppp1r17, compared to primary neurons from other regions of the brain (Figures S5G and H). By knocking down Ppp1r17 in those primary embryonic hypothalamic neurons, mRNA expression levels of the same synaptic functional genes were significantly reduced, confirming the in vivo results from DMH-specific Ppp1r17-KD mice (Figure 5G). Importantly, Ppp1r17-KD primary embryonic hypothalamic neurons exhibited significant decreases in the network bursting frequency, the number of spikes per burst, and burst duration, measured by the MEA system, whereas their firing and bursting rates showed no significant differences, compared to those of fLuc-KD control neurons (Figures 5H–M). These results strongly suggest that Ppp1r17 regulates the transcription of genes related to synaptic function, promoting synaptic transmission in DMH^Ppp1r17 neurons. Interestingly, we identified two transcription factors, PEG3 (Paternally expressed 3) and Myt1l (Myelin transcription factor 1 like), a transcriptional repressor and a pan-neuronal transcriptional factor, respectively, whose binding sites are common among the genes in the turquoise module. Whereas knocking down PEG3 did not change the mRNA expression levels of the synaptic transmission genes, knocking down Myt1l moderately decreased the expression of Grm5, Slc12a5, Gphn, and Nrsn1 (Figures S5I–K). In particular, the expression levels of Slc12a5 and Gphn showed significant correlations to Myt1l expression levels (Figures S5J and K). These findings suggest that Ppp1r17 may target specific neuronal transcription factors, such as Myt1l, to regulate the expression of synaptic transmission genes.

Figure 5. Ppp1r17 regulates synaptic signaling in the DMH and primary embryonic hypothalamic neurons.

(A-D) Heat maps and gene ontology (GO) analyses are shown for a hub gene set regulated by Ppp1r17 from the Weighted Correlation Network Analysis (WGCNA). GO analyses in blue (A, B) and turquoise (C, D) modules identified from the WGCNA with RNA-seq data from the DMH of control fLuc-KD and Ppp1r17-KD mice. (E) mRNA expression levels of genes related to synaptic and neuronal functions in laser microdissected DMH of control fLuc-KD or Ppp1r17-KD mice were examined 60 days after viral injection. mRNA expression levels were normalized to those in control fLuc-KD mice. Results are presented as mean ± SEM (*P< 0.05, **P < 0.01, ****P < 0.0001, two-way ANOVA with Šidák’s multiple comparisons test, n=4). (F) Primary embryonic hypothalamic neurons on MEA plate after viral injection. Scale bar, 200 μm. (G) mRNA expression levels of genes related to synaptic and neuronal functions in control fLuc-KD or Ppp1r17-KD primary embryonic hypothalamic neurons were examined 7 days after viral injection. mRNA expression levels were normalized to those in control fLuc-KD neurons. Results are presented as mean ± SEM (****P < 0.0001, two-way ANOVA with Šidák’s multiple comparisons test, n=3). (H) Raster plot of single wells recorded at week 3 in control fLuc-KD or Ppp1r17-KD primary embryonic hypothalamic neurons. Purple boxes indicate network bursting as a functional network synchronization. (I-M) MEA readouts for network burst frequency (I), number of spikes per burst (J), burst duration (K), weighted mean firing rate (L), and mean burst frequency (M) in control fLuc-KO and Ppp1r17-KD primary embryonic hypothalamic neurons are shown (**P<0.01, ***P<0.001, ****P<0.0001, Student’s t test, n=12).

Translocation of Ppp1r17 is regulated by its PKG-dependent phosphorylation during aging

We next hypothesized that the function of Ppp1r17 was most likely reduced during aging, inducing the dysfunction of DMH^Ppp1r17 neurons. To address this hypothesis, we examined what happened to Ppp1r17 in DMH^Ppp1r17 neurons during aging. Interestingly, we found that the localization of Ppp1r17 was altered from nuclei to cytoplasm during aging in the entire cDMH (Figures 6A and B), accompanied with decreases in lipolytic and NAD⁺ biosynthetic gene expression and decreases in phosphorylated HSL in aged WAT (Figures S6A and B). Given that Ppp1r17 conveys significant impacts on gene transcription, it is conceivable that Ppp1r17 might primarily work in the nucleus so that its translocation from the nucleus to the cytoplasm causes the dysfunction of DMH^Ppp1r17 neurons. It has been reported that Ppp1r17 is phosphorylated by PKG at Thr72 and Thr123, which are located right before the putative nuclear export signals (NESs) 2 and 3 (Figures S6C and D) ²⁰. These putative NESs are homologous to the amino acid sequences known to bind to chromosomal region maintenance 1 (CRM1/exportin 1) ^31–34. We found that leptomycin B (LMB), a specific nuclear export inhibitor that directly blocks the interaction between CRM1 and NESs, induced a rapid nuclear accumulation of the Ppp1r17-mRFP fusion protein transfected to HEK293 cells (Figure S6E and Movie S2). Additionally, a point mutation of Thr123, but not Thr72, to alanine induced clear accumulation of Ppp1r17-mRFP in the nucleus (Figure S6F), confirming a critical role of phosphorylation in the regulation of Ppp1r17 subcellular localization. These results led us to the idea that the nuclear-cytoplasmic translocation of Ppp1r17, mediated by PKG, likely causes the dysfunction of DMH^Ppp1r17 neurons during aging.

Figure 6. DMH-specific Prkg1 knockdown restores youthful Ppp1r17 subcellular localization in DMH^Ppp1r17 neurons, increases circulating eNAMPT levels and wheel-running activity, and extends lifespan in aged mice.

(A) Representative confocal images and line scan analyses of Ppp1r17 subcellular localization in DMH^Ppp1r17 neurons in 5 month- and 25-month-old mice. Scale bar, 20 μm. (B) The quantification of Ppp1r17 nuclear localization using Imaris after 3D-reconstruction through the DMH from bregma −1.72 mm to −1.96 mm are shown. Results are presented as mean ± SEM (*P<0.05, **P<0,01, Student’s t test, n=3). (C) Schematic representation of viral injection of shRNA against control fLuc or Prkg1 into the DMH of 20-month-old male mice and the timeline for measurements of indicated parameters in DMH-specific Prkg1-KD mice. (D) RNAscope for Prkg1 mRNA and immunostaining for Ppp1r17 on DMH-containing hypothalamic sections of control fLuc-KD and Prkg1-KD mice. Higher magnification images are shown with DAPI staining (right). Scale bars are 100 μm in low magnification and 20 μm in high magnification image. (E) Plasma eNAMPT levels of 28-month-old DMH-specific control fLuc-KD and Prkg1-KD male mice. eNAMPT levels were normalized to transferrin levels (*P<0.05, Student’s t test, n=7). (F) Wheel-running activities in 22-month-old DMH-specific control fLuc-KD and Prkg1-KD mice were measured (*P < 0.05, Wilcoxon matched-pairs signed-rank test, n=8 in each). Gray areas indicate dark time. (G) Kaplan-Meier curves of DMH-specific control fLuc-KD and Prkg1-KD male mice (n=19 for each group in this cohort, log-rank test, P=0.026). Red arrow indicates the time of viral injection. (H) Age-associated mortality rates of DMH-specific control fLuc-KD and Prkg1-KD male mice (ANCOVA, P=0.011).

Reversing age-associated Ppp1r17 translocation by Prkg1-KD ameliorates age-associated physiological decline and extends lifespan in mice

Surprisingly, we found that the mRNA expression of Prkg1, which encodes PKG, was completely and specifically colocalized with Ppp1r17 in DMH^Ppp1r17 neurons (Figure S6G). Based on the results from cultured cells (Figures S6E and F), we hypothesized that knocking down Prkg1 in the DMH could restore the nuclear localization of Ppp1r17 in DMH^Ppp1r17 neurons and ameliorate age-associated physiological decline in vivo. To address this possibility, we generated DMH-specific Prkg1-KD mice by stereotactically injecting a lentivirus expressing Prkg1 shRNA into the DMH of 20-month-old mice (Figures 6C and S6H). As predicted, when knocking down Prkg1 specifically in the DMH, Ppp1r17 was translocated to the nucleus in DMH^Ppp1r17 neurons (Figure 6D), demonstrating that PKG is a critical regulator for the subcellular localization of Ppp1r17 in the DMH. Interestingly, from 21 months until 30 months of age, DMH-specific Prkg1-KD mice continuously slowed down increases in the frailty index designed for mice ^35,36 (Figures S6I and J). In their iWAT, lipolytic genes and Nampt tended to show increases, and Ucp1 also showed significant increases (Figure S6K), opposite to gene expression changes in the iWAT of DMH-specific Ppp1r17-KD mice. Consistent with these changes, iWAT of DMH-specific Prkg1-KD old mice also showed significant increases in TH protein levels (Figure S6L). Furthermore, DMH-specific Prkg1-KD old mice showed increased levels of circulating eNAMPT (Figure 6E), which is critical to support hypothalamic function and counteract age-associated functional decline ^13,16, and increased wheel-running activity during the first half of their active time (ZT12-18) (Figure 6F). Remarkably, DMH-specific Prkg1-KD mice showed significant extension of median and mean lifespans (fLuc-KD 879 days vs. Prkg1-KD 937 days, log rank test, χ² = 4.90, df = 1, P=0.026) and an extension of maximal lifespan (20% oldest fLuc-KD 993 days vs. Prkg1-KD 1054 days, t-test, P=0.0554) (Figures 6G and S6M, and Movie S3). Importantly, DMH-specific Prkg1-KD mice showed a significant delay in age-associated mortality (Figure 6H). The slope of age-associated mortality change appeared to decrease, but its difference between fLuc-KD and Prkg1-KD mice did not reach statistical significance (Figure 6H). The spectrum of cancers did not differ between DMH-specific fLuc-KD and Prkg1-KD mice (Figure S6N). Thus, these results clearly demonstrate that maintaining the nuclear localization of Ppp1r17 is critical to counteract age-associated physiological decline and promote lifespan in mice.

Chemogenetic activation of DMH^Ppp1r17 neurons increases wheel-running activity, lipolysis, and circulating eNAMPT and extends lifespan in aged mice

Consistent with the nuclear-cytoplasmic translocation of Ppp1r17 during aging, c-fos immunoreactivity, a marker of neuronal activity, was decreased in DMH^Ppp1r17 neurons during aging (Figures S7A and B). Thus, we predicted that restoring the neuronal activity of DMH^Ppp1r17 neurons could delay aging and promote lifespan. To address this prediction, we decided to employ a chemogenetic technology called Designer Receptors Exclusively Activated by Designer Drugs (DREADDs) to selectively activate DMH^Ppp1r17 neurons. An AAV carrying the double-floxed hM3D (Gq)-mCherry or control mCherry construct was injected stereotactically into the cDMH of Ppp1r17-Cre mice at 18 months of age (Figure 7A). Expressing hM3D (Gq)-mCherry selectively in DMH^Ppp1r17 neurons allowed us to stimulate these neurons by a DREADD agonist, Agonist 21 ³⁷ (Figure S7C). We administered Agonist 21 at the dose of 0.5 mg/kg through intraperitoneal injection (i.p.) to hM3D (Gq)-mCherry-expressing experimental (DREADD) and mCherry-expressing control (control) mice. Chemogenetic activation of DMH^Ppp1r17 neurons induced c-fos immunoreactivity in DMH^Ppp1r17 neurons of DREADD mice (Figure S7D), confirming that chemogenetic stimulation activated DMH^Ppp1r17 neurons. Indeed, chemogenetic activation of DMH^Ppp1r17 neurons increased wheel-running activity (Figures 7B–C and S7F), whereas control mice did not show any difference before and after Agonist 21 treatment (Figures S7E and F). Body weights were slightly decreased only in DREADD mice (Figure 7D), but food intake was unchanged in both control and DREADD mice after Agonist 21 treatment (Figure S7G). Plasma levels of free fatty acids and phosphorylated HSL levels in WAT were significantly increased after Agonist 21 treatment in DREADD mice, but not in control mice (Figures 7E and F). Plasma triglyceride and cholesterol levels were not altered after Agonist 21 treatment in both groups (Figures S7H and I). Importantly, circulating eNAMPT levels were significantly enhanced in DREADD mice, but not in control mice, after Agonist 21 treatment (Figures 7G). These results suggest that direct activation of DMH^Ppp1r17 neuronal activity stimulated WAT function through the SNS and independently increased wheel-running activity in aged mice. We then tested whether the long-term activation of DMH^Ppp1r17 neurons could extend the lifespan of aged mice. We started Agonist 21 treatment for control or DREADD mice at 22 months of age, administering it for four consecutive days per week. Remarkably, the continuous activation of DMH^Ppp1r17 neurons successfully extended lifespan (log rank test, χ² = 4.191, df= 1, P=0.0406) (Figure 7H). Consistent with DMH-specific Prkg1-KD mice, the DREADD mice also displayed a significant delay in age-associated mortality, compared to control mice (Figure 7I). Again, the slope of age-associated mortality change showed a trend of decrease, but its difference between DREADD and control mice did not reach statistical significance (Figure 7I). These findings clearly demonstrate that the neuronal activity of DMH^Ppp1r17 neurons plays a critical role in counteracting age-associated functional decline and promoting lifespan through the maintenance of the inter-tissue communication between the hypothalamus and WAT.

Figure 7. DREADD-mediated activation of DMH^Ppp1r17 neurons increases wheel-running activity, lipolysis, and circulating eNAMPT levels and extends lifespan in aged mice.

(A) Schematic representation of viral injection of mCherry- (control) or hM3D(Gq)-mCherry (DREADD) into the DMH of Ppp1r17-Cre mice and the timeline for measurements of indicated parameters in control and DREADD mice. Details are described in the Method section. (B) Wheel-running activity of 20-month-old DREADD male mice before (PBS) and after (Agonist 21) activation of DMH^Ppp1r17 neurons (*P < 0.05, Wilcoxon matched-pairs signed-rank test, n=5). Gray areas indicate dark time. (C-E) Total wheel-running activity counts over 3 days (C), body weights (D), and plasma free fatty acid levels (E) in 20-month-old control and DREADD male mice before (PBS) and after (Agonist 21) activation (*P<0.05, **P<0.01, two-way ANOVA with Šidák’s multiple comparisons test, n=5). (F) p-HSL and total HSL levels in iWAT from 20-month-old control and DREADD male mice after treatment with Agonist 21 were analyzed by Western blotting. p-HSL levels were normalized to total HSL levels. (*P<0.05, Student’s t test, n=5). (G) Plasma eNAMPT levels of 20-month-old control and DREADD male mice before (PBS) and after (Agonist 21) activation. eNAMPT levels were normalized to transferrin levels (*P<0.05, two-way ANOVA with Šidák’s multiple comparisons test, n=5). (H) Kaplan-Meier curves of control and DREADD male mice (n=13 for each group in this cohort, log-rank test, P=0.0406). Red arrow indicates the day when Agonist 21 treatment started. (I) Age-associated mortality rates of control and DREADD male mice (ANCOVA, P=0.0004).

DISCUSSION

Our present study has demonstrated that DMH^Ppp1r17 neurons, a subpopulation of DMH^SIRT1/Nkx2-1 neurons, specifically regulate WAT function, including lipolysis and eNAMPT-EV secretion, through the stimulation of the sympathetic nervous system directed to WAT. DMH^Ppp1r17 neurons also promoted wheel-running activity, whose maintenance is tightly associated with longevity in mice ³⁸. Importantly, genetic and chemogenetic activation of DMH^Ppp1r17 neuronal activity counteracted age-associated physiological decline and promoted lifespan in mice. This important function of DMH^Ppp1r17 neurons is most likely mediated by the stimulation of eNAMPT-EV secretion from WAT. Given that eNAMPT-EVs enhance NAD⁺ biosynthesis, SIRT1 activity, and neural activity in the hypothalamus and delay aging and promote lifespan ^13,16, this inter-tissue communication between DMH^Ppp1r17 neurons and WAT comprises one of the key machineries that regulate the pace of aging and determine lifespan in mammals.

DMH^Ppp1r17 neurons are a key component that controls mammalian aging and longevity

In worms and flies, it has been demonstrated that specific populations of sensory neurons are key to control aging and longevity. In worms, ciliated sensory neurons located in the head and tail sense environmental cues and regulate their lifespan ^39,40. Neuronal XBP-1s, the activated, spliced XBP-1, mediates the endoplasmic reticulum unfolded protein response (UPR^ER) in distal cell types, particularly intestinal cells, and promotes longevity in aged worms, suggesting that the communication between neurons and intestinal cells is important for aging and longevity control in C. elegans ⁴¹. The aging-delaying and lifespan-extending effect of dietary restriction is also mediated by skn-1-expressing ASI neurons in C. elegans, which communicate with peripheral tissues to enhance metabolism ⁴². In fruit flies, neurons producing the insulin-like peptide dilp-2, whose function is regulated by dFOXO, the single forkhead transcription factor, in pericerebral fat body, regulate lifespan through the insulin signaling in the peripheral fat body ⁴³, providing further support to the importance of intercellular or inter-tissue communication for aging and longevity control. It should be noted that intestinal cells in C. elegans and fat body in Drosophila have been suggested to have analogous functions to mammalian adipose tissue ⁴⁴. Nonetheless, whether a similar mechanism exists to control mammalian aging and longevity has long remained a big question.

Our finding that DMH^Ppp1r17 neurons communicate with WAT through the sympathetic nervous system and regulate aging and longevity through this communication exemplifies a remarkable conservation of the systemic framework for aging and longevity control throughout evolution. Furthermore, in C. elegans, the nicotinamidase PNC-1, a functionally equivalent enzyme to mammalian NAMPT, has an extracellular form and regulates the reproductive system ⁴⁵. In both worms and flies, it has been demonstrated that the nicotinamidase (PNC-1 in worms and D-NAAM in flies) regulates their lifespans in a SIR2-dependent manner ^46,47. Given this surprising functional similarity between PNC-1/D-NAAM and eNAMPT, it is conceivable that not only the systemic framework but also key components might be evolutionarily conserved for aging and longevity control. Therefore, the identification and characterization of DMH^Ppp1r17 neurons and their function has set the next stage to further elucidate the systemic regulatory mechanism of mammalian aging and longevity control.

DMH^Ppp1r17 neurons regulate WAT function and wheel-running activity

Mice with DMH-specific Ppp1r17-KD or the ablation of DMH^Ppp1r17 neurons showed WAT dysfunction through reduced SNS innervation and reduced wheel-running activity, two common age-associated phenotypes. On the other hand, mice with DMH-specific Prkg1-KD or continuous activation of DMH^Ppp1r17 neurons by DREADDs exhibited reciprocal phenotypes, significantly ameliorating WAT function and wheel-running activity. Because the DREADD activation of DMH^Ppp1r17 neurons simultaneously enhanced WAT function and wheel-running activity, the regulations of these functions are most likely mediated by independent projection sites of DMH^Ppp1r17 neurons. In this regard, while the PAG has been well known as a modulator of sympathetic responses, as well as its role in pain sensation ^48,49, the PVN has also been known to regulate sympathetic innervation of adipose tissue ⁵⁰ and physical activity ⁵¹. Therefore, both the PAG and PVN could be the primary sites that mediate DMH^Ppp1r17 neuronal signaling for the regulation of WAT function and wheel-running activity.

During the course of our study, a study by Caglar and Friedman also characterized DMH^Ppp1r17 neurons independently and showed that in young wild-type mice, chemogenetic activation of these neurons acutely decreased food intake during the dark time, whereas acute or chronic inhibition of DMH^Ppp1r17 neurons had no effects on food intake or body weight ²². Contrarily, our own results from the chemogenetic activation of DMH^Ppp1r17 neurons did not show any significant changes in food intake in aged mice. This could be explained simply by different experimental designs in each study. For example, we activated DMH^Ppp1r17 neurons only during the dark time to mimic natural DMH^Ppp1r17 neuronal activation and measured daily total food intake in aged mice, whereas they measured food intake for up to 4 hrs following the activation in young mice. Based on their chemogenetic results under specific conditions that cause intense hyperphagia, such as leptin deficiency and scheduled feeding, DMH^Ppp1r17 neurons appear to function to restrict overconsumption of food. However, this particular function of DMH^Ppp1r17 neurons does not seem to have a predominant role in the regulation of aging and longevity under a standard nutritional condition.

The nuclear-cytoplasmic translocation of Ppp1r17 induces the dysfunction of DMH^Ppp1r17 neurons during aging

The Ppp1r17 protein was first discovered in the cerebellum ⁵², and a phospho-selective serum antibody against Ppp1r17 was reported as the first anti-phosphoprotein antibody ⁵³. Ppp1r17 is evolutionarily conserved in various species and is speculated to be a phosphatase inhibitor similar to DARPP-32, which has a critical role in dopamine signaling ²⁰. The subcellular localization of DARPP-32 is regulated by phosphorylation, and nuclear DARPP-32 promotes phosphorylation of histone H3 due to inhibition of PP-1 ⁵⁴. Similarly, the subcellular localization of Ppp1r17 is also regulated by PKG in DMH^Ppp1r17 neurons. Our results clearly showed that phosphorylation of Thr123 on Ppp1r17 is key to mediate the nuclear-cytoplasmic translocation of Ppp1r17 through the binding to CRM1. Indeed, the nuclear-cytoplasmic translocation of Ppp1r17 occurs in DMH^Ppp1r17 neurons during aging, implicating that PKG must be activated in the DMH. Consistent with this notion, DMH-specific Prkg1-KD reversed the subcellular localization of Ppp1r17 back to the nucleus. At this moment, how PKG is activated in DMH^Ppp1r17 neurons during aging remains unclear, and future investigation will be necessary to elucidate the exact mechanism for such age-dependent PKG activation.

Similarly to DARPP-32, the nucleus is likely the primary place where Ppp1r17 functions. Ppp1r17 regulates the transcription of genes related to synaptic function. Consistently, the knockdown of Ppp1r17 in primary embryonic hypothalamic neurons caused defects in network bursting, suggesting a defect in synaptic transmission. Thus, when Ppp1r17 is translocated from the nucleus to the cytoplasm, it is expected to cause synaptic dysfunction in DMH^Ppp1r17 neurons and thereby trigger decreases in WAT function and wheel-running activity during aging. A significant body of evidence suggests that long-term synaptic plasticity requires the control of gene expression in the brain ⁵⁵. It is possible that Ppp1r17 is involved in transcriptional regulation required for the control of long-term synaptic plasticity in the target regions for DMH^Ppp1r17 neurons. Such function of Ppp1r17 may be important for DMH^Ppp1r17 neurons to integrate metabolic and environmental cues and fine-tune the sympathetic nervous tone to precisely regulate WAT function.

The inter-tissue communication between DMH^Ppp1r17 neurons and WAT is one of the core machineries for mammalian aging and longevity control

How do DMH^Ppp1r17 neurons mediate delaying aging and extending lifespan? The key is our finding that the genetic and chemogenetic activation of DMH^Ppp1r17 neurons significantly increased circulating eNAMPT levels in aged mice. We have previously demonstrated that maintaining circulating eNAMPT levels in aged mice mitigates age-associated pathophysiological changes and extends healthspan and lifespan ¹⁶. eNAMPT is secreted from adipose tissue, being encapsulated into extracellular vesicles (EVs). These eNAMPT-EVs appear to be targeted to specific tissues, including the hypothalamus, hippocampus, pancreas, and retina, and promote NMN/NAD⁺ biosynthesis in those target tissues ¹⁶. Most recently, we have demonstrated that administration of eNAMPT-EVs purified from young mice significantly increases NAD⁺ levels in the whole hypothalamus, particularly in the Arc and DMH, but not in the liver, in aged mice, indicating a unique tissue specificity of eNAMPT-EVs ¹⁴. Although the mechanism of specific eNAMPT-EV targeting still remains unknown, eNAMPT-EVs function as an essential agent for WAT to communicate with certain tissues, including the hypothalamus. Given that DMH^Ppp1r17 neurons regulate WAT function through the SNS, whereas WAT regulates DMH NAD⁺ levels through eNAMPT-EVs, these regulatory legs clearly comprise a critical feedback loop between the hypothalamus and WAT. It is conceivable that maintaining the robustness of this critical feedback loop is one of the core machineries that counteract aging and promote longevity.

If this is the case, aging could be triggered by a gradual breakdown of the robustness of this feedback system. How could it happen? It has been reported that SNS-mediated norepinephrine (NE) signaling in WAT is compromised due to NE degradation by resident immune cells, particularly adipose tissue macrophages (ATMs), leading to the dysfunction of lipolysis and energy expenditure ⁵⁶. Genetic or pharmacological inhibition of NE degradation by ATMs increases the sympathetic nervous tone in adipose tissue and restores adipose functions in aged or obese mice ^57,58. Additionally, accumulating senescent cells in WAT induce CD38 expression in ATMs during aging, contributing to the decrease in NAD⁺ and the acquisition of age-related pathologies ^59,60. All these events contribute to the induction of inflammation in WAT, and inflammatory cytokines, such as TNF-α, decrease NAMPT expression, causing a vicious cycle and reducing eNAMPT-EV secretion. In this scenario, administration of NMN and/or eNAMPT-EVs could halt such a vicious cycle and restore the robustness of the feedback loop between the hypothalamus and WAT. Thus, it will be of great importance to further elucidate the dynamics of this key inter-tissue communication between DMH^Ppp1r17 neurons and WAT through the SNS and eNAMPT-EVs. Such knowledge will allow us to develop effective anti-aging interventions based on our deeper understanding of the systemic regulatory network for mammalian aging and longevity control.

CONCLUSION

Our present study demonstrates the importance of DMH^Ppp1r17 neurons, a key neuronal subpopulation in the cDMH, in controlling the process of aging and determining lifespan in mice. Age-associated decline in DMH^Ppp1r17 neuronal activity leads to age-associated pathophysiological changes, including significant reduction in physical activity, lipolysis, and eNAMPT secretion from WAT. Genetic and chemogenetic activation of DMH^Ppp1r17 neurons in aged mice mitigates age-associated physiological decline and extends lifespan. These findings provide critical insights into the systemic regulatory network for mammalian aging and longevity control and the development of more effective anti-aging interventions.

Limitations of Study

Whereas it has been demonstrated that increasing eNAMPT-EVs delay aging and promote healthspan and lifespan in mice ¹⁶ and that eNAMPT-EVs significantly enhance NAD⁺ levels in the DMH ¹⁴, whether eNAMPT-EVs are directly taken up by DMH^Ppp1r17 neurons still remains unclear. The investigation of which cells exactly uptake eNAMPT-EVs in the brain is currently ongoing. Additionally, whether PKG has other substrates that could also affect the activity of DMH^Ppp1r17 neurons needs to be investigated, although Ppp1r17 must be a major substrate based on the reciprocal phenotypes of DMH-specific Ppp1r17-KD and Prkg1-KD mice. Lastly, how this feedback loop mediated by DMH^Ppp1r17 neurons and WAT interacts with other inter-tissue communications, such as the one between the hypothalamus and skeletal muscle ⁸, and whether there are any sex differences in these inter-tissue communications are a critical problem to answer in the near future. Further investigation will be necessary to dissect these multi-layered feedback loops between the hypothalamus and peripheral tissues in mammalian aging and longevity control.

STAR Methods

Contact for Reagent and Resource Sharing

Further information and requests for resources and reagents should be directed to the Lead Contact, Shin-ichiro Imai (imaishin@wustl.edu).

Material availability

This study did not generate new unique reagents. All mouse lines were obtained from NIA, MMRRC, Jackson Laboratories, and Charles River as listed in the key resources table. All AAVs were obtained from Addgene, except the AAV5-mCherry-flex-dtA which was custom-generated by the Viral Vectors Core at the Hope Center for Neurological Disorders (HCND) at Washington University. All lentiviruses were generated by the Viral Vectors Core at the HCND.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-Ppp1r17	Thermo Fisher	PA5-61599 RRID:AB_2645859
Rabbit anti-cFos	Synaptic Systems	226003 RRID:AB_2231974
Mouse anti-c-Fos	Abcam	ab208942 RRID:AB_2747772
Rabbit anti-Nkx2-1	Santa Cruz Biotechnology	sc-13040 RRID:AB_793532
Chicken anti-GFP	Abcam	ab13970 RRID:AB_300798
Rabbit anti-Tyrosine Hydroxylase	Millipore	AB152 RRID:AB_390204
Rabbit anti-HSL	Cell Signaling	4107 RRID:AB_2296900
Rabbit anti-p-HSL	Cell Signaling	4126S RRID:AB_490997
Mouse anti-Map2	Abcam	ab11267 RRID:AB_297885
Rabbit anti-mRFP	Rockland	600-401-379 RRID:AB_2209751
Rabbit anti-NAMPT	Bethyl	A700-058 RRID:AB_2891856
Rabbit anti-β-actin	Thermo Fisher	PA1-183 PRID:AB_2539914
Rabbit anti-transferrin	Abcam	ab82411 RRID:AB_1659060
Alexa 488 goat anti-rabbit	Molecular Probes	A-11008 RRID:AB_143165
Alexa 594 goat anti-rabbit	Molecular Probes	A-11012 RRID:AB_2534079
Alexa 488 goat anti-mouse	Molecular Probes	A-11001 RRID:AB_2534069
Alexa 594 goat anti-mouse	Molecular Probes	A-11032 RRID:AB_2534091
Alexa 647 goat anti-mouse	Jackson ImmunoResearch Laboratories	115-605-146 RRID:AB_2338912
Goat anti-chicken DyLight 488	Molecular Probes	SA5-10070 RRID:AB_2556650
HRP-conjugated goat anti-rabbit	Vector Laboratories	PI-1000-1 RRID:AB_2916034
Rabbit IgG HRP Linked F(ab’)2	Cytiva	NA9340 RRID:AB_772191
Bacterial and virus strains
AAV5-Ef1a-DIO-EYFP	Addgene	27056-AAV5
AAV5-hSyn-DIO-mCherry	Addgene	50459-AAV5
AAV5-hSyn-DIO-hM3D(Gq)-mCherry	Addgene	44361-AAV5
AAV5-mCherry-flex-dtA	Addgene	58536
Lentivirus-U6-PGK-ShRNA-Ppp1r17-GFP	This paper	N/A
Lentivirus-U6-PGK-ShRNA-Prkg1-GFP	This paper	N/A
Lentivirus-U6-PGK-ShRNA-fLuc-GFP	The Hope Center Viral Core	N/A
Chemicals, peptides, and recombinant proteins
Tamoxifen	Sigma-Aldrich	T5648
Agonist 21	HelloBio	HB6124
High-glucorse DMEM with red phenorl	Thermo Fisher	11965092
0.05% trypsin-EDTA	Gibco	25200-72
Antibiotics	Gibco	15240-062
Neurobasal medium	Gibco	21103-049
L-glutamine	Thermo Fisher	25030
Poly-D-lysine	Sigma-Aldrich	P6407
Laminin	Sigma-Aldrich	L2020
N2 supplement	Gibco	17502-048
B27 supplement	Gibco	17504-04
Ara-C	Sigma-Aldrich	C1768
Leptomycin B	Sigma-Aldrich	L2913
Paraformaldehyde (16%)	Electron Microscopy Sciences	15710
DAPI	Sigma-Aldrich	D9542
Triton-X	Sigma-Aldrich	T8787
Tween 20	Thermo Fisher	P7949
Glycine	Thermo Fisher	BP381-1
Methanol	Thermo Fisher	A412P-4
Dichloromethane	Sigma-Aldrich	L090000
Dibenzyl ether	Sigma-Aldrich	33630
HEPES	Sigma-Aldrich	H3375
NaCl	Sigma-Aldrich	746398
SDS	Thermo Fisher	BP166
EDTA	Sigma-Aldrich	E5134
DTT	Sigma-Aldrich	D9779
DNase	Qiagen	79254
Critical commercial assays
Agilent Mouse_v1 8 x 60K microarrays	Agilent Technologies	Design ID-028005
Illumina NovaSeq 6000	Illumina	https://www.illumina.com/systems/sequencing-platforms/novaseq.html
Infinity Kit for triglycerides and cholesterol measurements	Thermo Fisher	No. TR22421 and TR13421
HR series NEFA-HR kit	Wako Pure Chemical Industries	No. 999-34691
StepOnePlusTM Real-Time PCR System	Applied Biosystems	4376600
RNAscope Multiplex Fluorescent Reagent Kit v2	Advanced Cell Diagnostics Bio	323100
TSA Plus Fluorescence Palette Kit	PerkinElmer	NEL760001KT
EchoMRITM 3-in-1 analyzer	EchoMRI	http://www.echomri.com/Body_Composition_3_in_1.aspx
Maestro MEA plate reader	Axion Biosystems	https://www.axionbiosystems.com/products/mea/maestro-pro
Leica LMD 6000 system	Leica Microsystems	https://www.leica-microsystems.com/products/light-microscopes/p/leica-lmd7/
Deposited data
Microarray data	Illumina	GSE224543
RNAseq data	Agilent Technologies	GSE224509
Experimental models: Cell lines
HEK293 cell	ATCC	CRL-1573
Experimental models: Organisms/strains
C57BL/6J	In house breeding or NIA	N/A
Tg (Ppp1r17-cre) NL163Gsat/Mmucd	MMRRC	036188
B6.Cg-Gt(ROSA)26Sortm6(CAG-ZsGreen1)Hze/J	IMSR_JAX	007906
Nkx2-1tm1.1(cre/ERT2)Zjh/J	IMSR_JAX	014552
B6J.129S6(FVB)-Slc17a6tm2(cre)Lowl/MwarJ	IMSR_JAX	028863
Slc32a1tm2(cre)Lowl/J	IMSR_JAX	016962
CD1 IGS	Charles River	022
Oligonucleotides
Taqman primer: Gapdh	Thermo Fisher	Mm99999915_g1
Taqman primer: B2m	Thermo Fisher	Mm00437762_m1
Taqman primer: Lipe	Thermo Fisher	Mm00495359_m1
Taqman primer: Ppara	Thermo Fisher	Mm00440939_m1
Taqman primer: Pparr	Thermo Fisher	Mm01184322_m1
Taqman primer: Fasn	Thermo Fisher	Mm00662319_m1
Taqman primer: Dio2	Thermo Fisher	Mm00515664_m1
Taqman primer: Dio3	Thermo Fisher	Mm00548953_s1
Taqman primer: UCP1	Thermo Fisher	Mm01244861_m1
Taqman primer: Glut4	Thermo Fisher	Mm00436615_m1
Taqman primer: Glut2	Thermo Fisher	Mm00446229_m1
Taqman primer: Nampt	Thermo Fisher	Mm01293560_m1
Taqman primer: Srebp1c	Thermo Fisher	Mm00550338_m1
Taqman primer: Acly	Thermo Fisher	Mm 1302297_m1
Taqman primer: G6ase	Thermo Fisher	Mm00839363_m1
Taqman primer: Acc	Thermo Fisher	Mm01304277_m1
Taqman primer: Adrb1	Thermo Fisher	Mm00431701_s1
Taqman primer: Adrb2	Thermo Fisher	Mm02524224_s1
Taqman primer: Adrb3	Thermo Fisher	Mm00442669_m1
Taqman primer: Cytb	Thermo Fisher	Mm01208314_m1
Taqman primer: Idh3	Thermo Fisher	Mm00499674_m1
Taqman primer: Ppp1r17	Thermo Fisher	Mm00495458_m1
Taqman primer: ngb	Thermo Fisher	Mm00452101_m1
aqman primer: synpr	Thermo Fisher	Mm00511114_m1
Taqman primer: nrsn1	Thermo Fisher	Mm00494159_m1
Taqman primer: Rab1	Thermo Fisher	Mm00485433_m1
Taqman primer: Grm5	Thermo Fisher	Mm00690332_m1
Taqman primer: Slc12a5	Thermo Fisher	Mm00803929_m1
Taqman primer: Gphn	Thermo Fisher	Mm00556895_m1
Taqman primer: Myt1l	Thermo Fisher	Mm00485408_m1
mm-Ppp1r17	Advanced Cell Diagnostics Bio	497351
mm-FOS	Advanced Cell Diagnostics Bio	316921
Cre	Advanced Cell Diagnostics Bio	312281
mm-Prkg1	Advanced Cell Diagnostics Bio	497361
mm-vGlut2	Advanced Cell Diagnostics Bio	319171
mm-vGAT	Advanced Cell Diagnostics Bio	319191
mCherry	Advanced Cell Diagnostics Bio	431201
Recombinant DNA
pcDNA3-mRFP	Addgene	13032
pcDNA3-Ppp1r17 (WT)-mRFP	This paper	N/A
pcDNA3-Ppp1r17 (T72A)-mRFP	This paper	N/A
pcDNA3-Ppp1r17 (T123A)-mRFP	This paper	N/A
pcDNA3-Ppp1r17 (T72AT123A)-mRFP	This paper	N/A
Software and algorithms
Clocklab	Actimetrics	https://actimetrics.com/products/clocklab/
Image processing, cell counter module	ImageJ/Fiji	https://fiji.sc/
Prism 9	Graphpad	https://www.graphpad.com/
Feature Extraction software v11.5.1.1	Agilent Technologies	https://www.agilent.com/en/product/mirna-microarray-platform/mirna-microarray-software/feature-extraction-software-228496
Illumina DRAGEN Bio-IT	Illumina	https://www.illumina.com/products/by-type/informatics-products/dragen-bio-it-platform.html
R/Bioconductor package EdgeR	Robinson et al., 2010	N/A
R/Bioconductor package Limma	Ritchie et al., 2015	N/A
Limma’s voomWithQualityWeights	Liu et al., 2015	N/A
R/Bioconductor package GAGE	Luo et al., 2009	N/A
R/Bioconductor package heatmap3	Zhao et al., 2014	N/A
R/Bioconductor package WGCNA	Langfelder et al,. 2018	N/A
R/Bioconductor package clusterProfiler	Yu et al., 2012	N/A
SYNGO	Koopmans et al., 2019, Neuron	https://www.syngoportal.org
Imaris	Bitplane/Oxford Instruments	https://imaris.oxinst.com
ZEN Microscopy Software	Zeiss	https://www.zeiss.com/microscopy/en/products/software/zeiss-zen.html
Arivis Vision4D	Zeiss	https://www.arivis.com/solutions/vision4d
LAS X Life Science Microscope Software	Leica	https://www.leica-microsystems.com/products/microscope-software/p/leica-las-x-ls/
AxIS Navigator software	Axion Biosystems	https://www.axionbiosystems.com/products/software
NDP.view2	Hamamatsu Photonics	https://www.hamamatsu.com/content/dam/hamamatsu-photonics/sites/documents/99_SALES_LIBRARY/sys/SBIS0066E_NDPVIEW2.pdf
Other
Standard chow diet	LabDiet	5053
Microtherma 2 Type	ThermoWorks	TW2-193
SuperFect transfection reagent	Qiagen	301305
Microsyringe	Hamilton	7633-01
Needle for microsyringe	Hamilton	7803-03
Hamilton Neuros Syringe	Hamilton	65457-02
MembraneSlides	Leica	11505158
PicoPure™ RNA isolation kit	Applied Biosystems	12204-01
Pfu Ultra II	Agilent	600670
PrimeSTAR	Takara Bio	R050A
Nunc Glass Base Dish	Thermo Fisher	150680
Lipofectamine 3000	Invitrogen	L-3000
Superfrost™ Plus Adhesion Slides	Thermo Fisher	12-550-15
ImmEdge Hydrophobic Barrier PAP Pen	Vector Laboratories	H-4000
FluorSave Reagent	Millipore	345789
Goat serum	Vector Laboratories	S-1000
NanoZoomer HT model	Hamamatsu Photonics	https://nanozoomer.hamamatsu.com/jp/en.html
Zeiss Lightsheet 7	Zeiss	https://www.zeiss.com/microscopy/en/products/light-microscopes/light-sheet-microscopes/lightsheet-7.html
PureLink™ RNA Mini Kit	Thermo Fisher	12183025
High-Capacity cDNA Reverse Transcription Kit	Thermo Fisher	4368814
HybEZ™ oven	Advanced Cell Diagnostics Bio	PN321720
4-15% SDS-PAGE	BioRAD	4561084 or 4561086
Pierce BCA Protein Assay Kit	Thermo Fisher	23227
MEA plate	Axion Biosystems	M768-tMEA-48W
SiR-DNA	Cytoskeleton Inc.	CY-SC007
TrueBlack	Biotium	23007
Lenti-X	Takara Bio	631231

Data and code availability

DNA microarray and RNA-seq data have been deposited at GEO and are publicly available as of the date of publication. Accession numbers are listed in the key resources table. This paper does not report original code. Uncropped scans of all Western blots and all data used to generate graphs are included in Data S1. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL DETAILS

All mouse experiments and procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Washington University in St. Louis and performed following the National Institutes of Health guidelines. Wild-type C57BL/6J male mice were bred in our laboratory using mice purchased from Jackson Laboratories. Wild-type C57BL/6JN male mice were also obtained from the NIH aged rodent colony. All of the mice were housed at 22 °C in the Medical School and/or Danforth Campus animal facilities at Washington University on a 12:12 hr light/dark cycle in a group of 3-5 unless noted otherwise and were fed a standard chow diet (LabDiet 5053; LabDiet, St. Louis, MO) ad libitum throughout the study. Cages and bedding were changed once per week, and mice were monitored periodically for their health status. Young (3–6-month-old) and aged (18–27-month-old) male mice were age- and source-matched for each experiment. Ppp1r17-Cre [MMRRC:036188; Tg (Ppp1r17-Cre) NL163Gsat/Mmucd] mice were used to express Cre-recombinase in Ppp1r17-positive neurons, and the mice were backcrossed to wild-type C57BL/6J mice (Jackson Laboratories) for 6–7 generations before analysis. ROSA-loxSTOPlox-ZsGreen1 [IMSR_JAX:007906; B6.Cg-Gt(ROSA)26Sor^{tm6(CAG-ZsGreen1)Hze}/J] was used to express the reporter gene in the presence of Cre recombinase by using lox-STOP-lox sequence crossed with Nkx2-1-CreERT2 [IMSR_JAX:014552; Nkx2-1^{tm1.1(cre/ERT2)Zjh}/J], Vglut2-ires-Cre [IMSR_JAX:028863; B6J.129S6(FVB)-Slc17a6^tm2(cre)Lowl/MwarJ] or Vgat-ires-Cre [IMSR_JAX:016962; Slc32a1^tm2(cre)Lowl/J] mice. For the entire study, heterozygous Ppp1r17-Cre mice were generated by crossing heterozygous Ppp1r17-Cre and wild-type C57BL/6J mice. Mice were closely monitored and included daily body weight measurement during the experimental period. Male mice were used for all experiments except microarray analysis. No data were excluded, except for mice with deteriorating health issues due to the surgery or missing correct viral targeting.

METHOD DETAILS

Sample collection

All sample collection was done at Zeitgeber Time (ZT) 14.5-15.5 unless otherwise specified in each section of the methods. iWAT, eWAT and interscapular BAT were collected for adipose tissue analysis. For tissue preparation, the detailed methods are shown in each individual section.

Microarray analysis for DMH^SIRT1/Nkx2-1 and other Nkx2-1-positive neurons

DMH^SIRT1/Nkx2-1 neurons and other Nkx2-1-positive neurons in the Arc and VMH were collected from 2-month-old Nkx2-1-CreERT2 mice crossed with ROSA-loxSTOPlox-ZsGreen1 mice after tamoxifen injection (180 mg/kg i.p. once each day for 5 consecutive days) by laser micro-dissection using the Leica LMD 6000 system (Leica Microsystems) and PEN-membrane slides (Leica, #11505158). Approximately 500-1000 cells were pooled from each hypothalamic nucleus from individual mice, and RNA was isolated using the proteinase K/acid phenol method as described previously ⁶¹. The quality of total RNA was determined by Agilent 2100 Bioanalyzer (Agilent Technologies), and 2-3 samples from each hypothalamic nucleus were used for microarray analysis. The detailed procedure for microarray analysis was described previously ⁶². Briefly, total RNA was amplified using the WTA2 kit (Sigma-Aldrich) according to the manufacturer’s protocol. cDNAs were chemically labeled with Kreatech ULS RNA Labeling Kit (Kreatech Diagnostics) and purified with Qiagen PCR purification columns. cDNA samples were hybridized to Agilent Mouse_v1 8 x 60K microarrays (Design ID-028005, Agilent Technologies). Gridding and analysis of images were performed using Feature Extraction software (v11.5.1.1, Agilent Technologies). To identify differentially expressed genes in each sample, ratios of signal intensities to each sample were calculated, and genes whose expression levels were greater than or equal to 15-fold in each sample compared to the others were selected. The heat map for each log 2-transformed gene expression level was generated using GraphPad Prism V9.0 (GraphPad Prism Software, San Diego, CA). The microarray data were deposited into the NCBI GEO database (GEO Accession # GSE224543).

Immunohistochemistry (IHC) and Immunocytochemistry (ICC)

Mice were anesthetized using a ketamine/xylazine cocktail and perfused transcardially, first with 0.01 M phosphate-buffered saline (PBS) and then with 2% paraformaldehyde (PFA) and 0.2% picric acid in PBS [pH 7.4]. The brains were removed and post-fixed with 2% PFA and 0.2% picric acid in PBS overnight at 4 °C, immersed in 20% sucrose in PBS at 4 °C overnight until brains sank, and frozen by covering the brain with powdered dry ice. Fixed frozen brain tissue was sectioned using a cryostat at −20 °C. The sections obtained were attached on Superfrost^™ Plus Adhesion Slides (12-550-15, Thermo Fisher). Immunohistochemical analysis was performed using 20 μm-thick cryostat sections, and a hydrophobic barrier was created using an ImmEdge Hydrophobic Barrier PAP Pen (H-4000, Vector Laboratories). For young vs. old IHC, TrueBlack® (#23007, Biotium) was used in both young and old sections before blocking to reduce autofluorescence by lipofuscin accumulation in old mice. For white adipose tissue (WAT), harvested tissues were post-fixed with 4% PFA in 0.01 M PBS at least overnight, immersed in 70% ethanol for 24 hr at 4 °C and embedded in paraffin. Immunohistochemical analysis of WAT was performed using 5 μm-thick sections. For ICC, cells were fixed with 4% PFA for 10 min at room temperature (RT) and then washed with PBS. The sections or cells were blocked with 5% goat serum (S-1000, Vector Laboratories) and 0.3% Triton-X100 (T8787, Sigma-Aldrich) in PBS for 1 hr, and incubated with primary antibody including rabbit anti-Ppp1r17 (1:1,000; PA5-61599, Thermo Fisher), rabbit anti-c-fos (1:1,000; 226003, Synaptic Systems), mouse anti-c-fos (1:1,000; ab208942, Abcam), rabbit anti-Nkx2-1 (TTF-1, H-190) (1:200; D2412, Santa Cruz Biotechnology), chicken anti-GFP (1:1,000; ab13970, Abcam), rabbit anti-p-HSL (1:1,000; 4126S, Cell Signaling), mouse anti-Map2 (1:1,000; ab11267, Abcam), rabbit anti-mRFP (1:1,000; 600-401-379, Invitrogen) overnight at 4°C. The sections were then incubated with the secondary antibody, Alexa 488, 594 or 647 anti-rabbit or mouse immunoglobulin G (IgG; 1:1,000; Molecular Probes, Invitrogen, Carlsbad, CA) for 1 hr at RT. For GFP antibody, anti-chicken secondary antibody (1:1,000; SA5-10070, Invitrogen) was used. For the double staining of Ppp1r17 and c-fos or mCherry, samples were stained with lower concentration of anti-Ppp1r17 (1:5000) and HRP-conjugated goat anti-rabbit IgG (1:1000; PI-1000, Vector Laboratories) antibody with TSA Plus Fluorescence Palette Kit (1:200, NEL760001KT, PerkinElmer) first, and then c-fos or mRFP antibody was used. Nuclei were stained with DAPI solution (D9542, Sigma-Aldrich). Images were acquired using fluorescent microscopy (DMi8, Leica Microsystems) or confocal fluorescent microscopy (Leica SP5X white light laser confocal system, Leica Microsystems) depending on the experiment. Quantification was performed using the ImageJ software (NIH, Bethesda, MD, USA). For Ppp1r17 localization analysis, images were acquired using the confocal microscope and quantified by Imaris (Bitplane AG, Zurich, Switzerland) after 3D-reconstruction.

3D reconstruction of Ppp1r17 expression in the hypothalamus

Consecutive brain sections (12 μm-thick) were collected through the hypothalamus from Bregma −1.23 mm to −2.53 mm. All of the collected sections were stained simultaneously with the Ppp1r17 antibody as described above. Images were acquired using NanoZoomer (Hamamatsu Photonics, Hamamatsu City, Japan) and 3D-reconstructed by NDP.view2 software (Hamamatsu Photonics) and ImageJ (NIH, Bethesda, MD, USA).

Multiplex Fluorescent In Situ Hybridization (FISH, RNAscope^™)

Brains were collected and processed as described above for IHC. Multiplex FISH was then performed using the RNAscope Multiplex Fluorescent Reagent Kit v2 [Advanced Cell Diagnostics Bio (ACDBio)]. Probes for the following mRNAs were used (all from ACDBio): mm-Ppp1r17 (cat no. 497351), mm-FOS (cat no. 316921), Cre (cat no. 312281), mm-Prkg1 (cat no. 497361), mm-vGlut2 (cat no. 319171), mm-vGAT (cat no. 319191), and mCherry (cat no. 431201). In brief, RNAscope was performed according to the manufacturer’s protocol. After dehydration, the brain sections were treated with hydrogen peroxide for 10 min at RT and Protease III for 30 min at 40 °C using a HybEZ^™ oven. The probe hybridization was achieved by incubation of mRNA target probes for 2 h at 40 °C. The signal was amplified by subsequent incubation of Amp-1, Amp-2 and Amp-3 one drop each for 30, 30 and 15 min, respectively, at 40 °C using the oven. Each incubation step was followed by two 2-min washes using RNAscope washing buffer. Nuclei were stained with DAPI solution (D9542, Sigma-Aldrich) before adding coverslips. For Ppp1r17-IHC following RNAscope, sections were stained with anti-Ppp1r17 (1:1000) and HRP-conjugated goat anti-rabbit IgG (1:1000; PI-1000, Vector Laboratories) antibodies with TSA Plus Fluorescence Palette Kit (1:500, NEL760001KT, PerkinElmer). Images were acquired as described above for IHC experiments.

AAV and lentiviruses

AAV5-Ef1a-DIO-EYFP (#27056-AAV5) for tracing experiment, AAV5-hSyn-DIO-mCherry (#50459-AAV5: control virus) and AAV5-hSyn-DIO-hM3D(Gq)-mCherry (#44361-AAV5: activation) for DREADD experiments were obtained from Addgene. AAV5-mCherry-flex-dtA was generated by the Viral Vectors Core at the Hope Center for Neurological Disorders at Washington University (HCND) using pAAV-mCherry-flex-dtA plasmid (#58536) from Addgene for the ablation experiment. To generate shRNA-expressing lentiviral constructs for knockdown experiments, 56-bp double-stranded oligonucleotides, each of which contained a sense target sequence, a microRNA-based loop sequence (CTTCCTGTCA), an antisense sequence, a termination sequence of four thymidines and appropriate restriction enzyme sites at both ends, were generated for mouse Ppp1r17, Prkg1, Myt1l and firefly luciferase (fLuc) and cloned into the U6-PGK-GFP vector provided by the Viral Vectors Core at the HCND. The sense Ppp1r17, Prkg1, Myt1l and fLuc sequence are 5’-GGAGACCTACTACTCAGATTG-3’, 5’-CCGGAGAATCTCATCCTAGAT-3’, 5’-CGTGACTACTTTGACGGAAAT-3’ and 5’-TACGCGGAATACTTCGAAATG-3’, respectively. Culture supernatant was collected 48 h after transfection. Knockdown efficiencies were tested using primary hypothalamic neuron cultures. Large-scale lentiviral supernatant was produced by the Viral Vectors Core at the HCND and stored at −80 °C before use. Myt1l shRNA-expressing lentivirus was purified by Lenti-X^™ Concentrator (Takara, # 631231).

Stereotactic injection

Following anesthesia with 1.5-2% isoflurane gas depending on the age of mice, a mouse was placed in a three-point fixation stereotactic frame. Under aseptic technique, a midline incision was made, and the periosteum was dissected from the underlying cranium. Bregma was identified, and appropriate coordinates for the stereotactic injection were registered. A 2-mm hole was made utilizing a dental drill, and mice were then injected unilaterally or bilaterally, depending on the experimental paradigm using a microsyringe (7633-01, Hamilton) with a needle (7803-03, Hamilton) for lentivirus or a Hamilton Neuros Syringe (65457-02, Hamilton) for AAV. The type of virus, titers, injection volume, speed, and coordination for each experiment are listed in Table S1. The incision was closed with 4-0 nylon sutures. All injected mice had at least two weeks to fully recover before being used for any experiments.

Anterograde tracing experiment

3-month-old Ppp1r17-Cre male mice received unilateral injections of AAV5-Ef1a-DIO-EYFP into the DMH. Brains were collected 2 weeks after virus injection. To amplify the fluorescence of axon fibers expressing EYFP, we performed IHC for GFP (ab13970, abcam, 1:1,000). Tissue preparation and the GFP staining protocol are described in detail in the IHC method section.

DMH-specific Ppp1r17-KD in young mice

Lentiviruses carrying shRNAs against Ppp1r17 or fLuc were injected bilaterally into the DMH of 3-month-old wild-type males. Body weight, body temperature, food intake and body composition were measured 4 weeks after viral injection, and then the mice were placed in individual cages with a running wheel. The mice were weighed again 8 weeks after viral injection, and then brain, WAT, BAT, liver and soleus muscle were collected and processed for IHC, Western blot analysis and gene expression analysis.

DTA Ablation experiment

AAV5-mCherry-flex-dtA was injected bilaterally into the DMH of 3-month-old Ppp1r17-Cre male mice using a Hamilton Neuros Syringe. Wild-type littermates received the same treatment as the negative control for DMH^Ppp1r17 neuronal ablation. Brains were harvested 2 months after the viral injection, and IHC was performed to determine the extent of the ablation. WAT was collected at the same time for p-HSL measurement and gene expression analysis.

Wheel-Running Assay

Mice were placed in individual cages with a running wheel and housed under 12:12 light-dark cycle. Mice were habituated at least 10 days before analyzing wheel-running activity measurements. Running-wheel revolutions were recorded and averaged into counts/min using Clocklab software (Actimetrics, Evanston, IL).

Rectal body temperature measurement

Rectal body temperature was measured at ZT 3-3.5 and 15-15.5 using a Microtherma 2 Type (TW2-193) thermometer with a mouse rectal probe (model RET-3).

Metabolic Assessments

Body composition and plasma biochemical assays were performed at the Diabetes Models Phenotyping Core of the Diabetes Research Center at Washington University. Body composition was measured by EchoMRI^™ 3-in-1 analyzer. Total triglycerides and cholesterol in plasma were determined using Infinity Kit No. TR22421 and TR13421 (Thermo Fisher). In a 96-well plate format, 100 μl of each reagent was aliquoted. To the reagent, 1 μl of plasma sample or standard was added and mixed. The reactions were incubated for 30 minutes at RT. Total triglyceride level was determined by reading the absorbance at 540 nm, while total cholesterol was read at 490 nm. Sample concentrations were calculated from a standard curve correcting for blanks and a secondary wavelength at 660 nm. Free fatty acids in plasma were measured using the HR series NEFA-HR kit No. 999-34691 (Wako Pure Chemical Industries, Osaka, Japan). First, 2 volumes (50 μl) of reagent A were aliquoted to wells, 1 μl of plasma or standard was added, and the plate was incubated for 15 min at RT. Then, color was developed with the addition of 1 volume (25 μl) of reagent B to all wells. The plate was incubated for additional 15 min at RT and read at 540 nm, and corrected for blanks and a secondary wavelength at 660 nm.

Gene expression analysis

For WAT, BAT, liver, soleus and primary neurons, total RNA was extracted using the PureLink^™ RNA Mini Kit (12183025, Thermo Fisher) with DNase (79254, Qiagen) and converted to cDNA by High-Capacity cDNA Reverse Transcription Kit (4368814, Thermo Fisher). For the DMH samples, the compact region of the DMH was collected by laser microdissection using the Leica LMD 6000 system (Leica Microsystems). The detailed procedure for sample preparation was described previously ⁶³. Total RNA was extracted following laser microdissection using the Arcturus PicoPure^™ RNA isolation kit (12204-01, Applied Biosystems). For all samples, cDNA was synthesized using the Applied Biosystems High Capacity cDNA Reverse Transcription Kit (4368814, Thermo Fisher). Quantitative real-time RT-PCR was conducted with the StepOnePlus^™ Real-Time PCR System (Applied Biosystems) and Taqman^™ primers (Thermo Fisher), and relative expression levels were calculated for each gene by normalizing to B2m levels in WAT and BAT or Gapdh levels in laser microdissected samples, liver, and soleus muscle and then to one of the averages of the control mice. Primers used in this study were: Mm99999915_g1 (Gapdh), Mm00437762_m1 (B2m), Mm00495359_m1 (Lipe), Mm00440939_m1 (Pparα), Mm01184322_m1 (Pparγ), Mm00662319_m1 (Fasn), Mm00515664_m1 (Dio2), Mm00548953_s1 (Dio3), Mm01244861_m1 (UCP1), Mm00436615_m1 (Glut4), Mm00446229_m1 (Glut2), Mm01293560_m1 (Nampt), Mm00550338_m1 (Srebp1c), Mm1302297_m1 (Acly), Mm00839363_m1 (G6Pase), Mm01304277_m1 (Acc), Mm00431701_s1 (Adrb1), Mm02524224_s1 (Adrb2), Mm00442669_m1 (Adrb3), Mm01208314_m1 (Cytb), Mm00499674_m1 (Idh3), Mm00495458_m1 (Ppp1r17), Mm00452101_m1 (ngb), Mm00511114_m1 (synpr), Mm00494159_m1 (Nrsn1), Mm00485433_m1 (Rab1), Mm00690332_m1 (Grm5), Mm00803929_m1 (Slc12a5), Mm00556895_m1 (Gphn), Mm00485408_m1 (Myt1l).

Western blot analysis of eNAMPT, p-HSL and TH

To measure eNAMPT, blood was collected from the tail vein using a capillary coated with heparin sulfate. Blood was spun down at 3000 x g for 10 min at 4 °C. 1 μl of freshly collected plasma was incubated with 100 μl of 1x sample buffer at 95 °C for 10 min before being stored at −30 °C until its use. Right before the analysis, 5 μl of each sample was added to 35 μl of 1x sample buffer and further incubated at 95 °C for 30 min. This 30 min boiling was necessary to make eNAMPT bands discrete and quantifiable. 10 μl of the final mixture was separated on a 4-15% SDS-PAGE gel (#4561084 or 4561086, BioRAD) and analyzed by Western blotting with rabbit anti-NAMPT antibody (1:1000; A700-058, Bethyl). Rabbit anti-transferrin antibody (1:1000; ab82411, Abcam) was used for loading control. For p-HSL and TH blotting, WAT and BAT were dissected and stored at −80 °C after PBS perfusion. Protein from WAT and BAT was prepared using homogenization buffer (50 mM HEPES, 100 mM NaCl, 5% SDS, 2mM EDTA, 0.5mM DTT, with protease and phosphatase inhibitors), quantified by Pierce BCA Protein Assay Kit (#23227, Thermo Fisher), separated on 4-15% SDS-PAGE gels (8 μg/lane), and analyzed by Western blotting with rabbit anti-HSL (1:2000; 4107S, Cell Signaling) and p-HSL (1:2000; 4126S, Cell Signaling) or rabbit anti-TH antibody (1:400; AB152, Millipore). For TH blotting, rabbit anti-β-actin antibody (1:2000; PA1-183, Thermo Fisher) was used for loading control.

Clearing and immunofluorescence for WAT

WAT sample collection, delipidation, permeabilization, immunolabelling and clearing were performed as described previously with some modifications ⁵⁰. Briefly, samples were collected from PBS and 4% PFA perfused mice and post-fixed in 4% PFA at 4 °C for 24 hr. Fixed samples were washed in 20%, 40%, 60%, 80% methanol in B1N buffer (H₂O, 0.1% Triton X-100 and 0.3 M glycine, pH 7), and 100% methanol for 1 hr each. Samples were then delipidated with 100% dichloromethane (DCM; Sigma-Aldrich) overnight. After delipidation, samples were washed in 100% methanol for 1 hr twice, then in 80%, 60%, 40%, 20% methanol in B1N buffer for 1 hr each step. Samples were then washed in B1N buffer for 1 hr twice, followed by PTxwH buffer (PBS, 0.1% Triton X-100, 0.05% Tween 20 and 2 μg/ml heparin) for 1 hr twice before beginning the staining procedure. Samples were incubated in PTxwH buffer with a rabbit anti-TH antibody (1:400; AB152, Millipore) for 4 days with shaking. After primary antibody incubation, samples were washed in PTxwH buffer overnight, and then incubated in secondary antibody dilutions (Alexa-488; 1:400, Invitrogen) in PTxwH buffer for 4 days. Samples were finally washed in PTxwH buffer overnight and dehydrated in 25%, 50%, 75%, 100%, 100% methanol/H₂O series for 1 hr at each step at RT. The samples were incubated in 66% DCM/33% Methanol for 3 hr at RT with shaking and washed by 100% DCM for 15 min twice. The samples were incubated in dibenzyl ether (DBE; Sigma-Aldrich) to complete the clearing step. Samples were stored at RT in the dark until imaging. All whole-tissue samples were imaged on a light-sheet microscope (Zeiss Lightsheet 7, Zeiss) equipped with x5 objective lens, and 3D-reconstruction was done by Arivis Vision4D at the Washington University Center for Cellular Imaging. For tracing SNS nerve fibers, the FilamentTracer tool was applied to automatically reconstruct the neurites in 3D in iDISCO 3D images generated by Imaris software (Bitplane). 3D pictures were generated using the ‘snapshot’ tool. TH fiber areas and branches were automatically calculated by the FilamentTracer tool.

RNA sequencing and analysis

The DMH was collected by laser microdissection of brain sections from 5-month-old DMH-specific Ppp1r17-KD male mice 2 months after lentiviral injection. RNA was isolated from each DMH sample using the Arcturus PicoPure^™ RNA isolation kit (12204-01, Applied Biosystems) ⁶². RNA sequencing and analysis were performed at the Genome Technology Access Center at Washington University. The quality of the input RNA was assessed using Bioanalyzer 2100 and the RNA nano chip (Agilent Technologies, Santa Clara, CA). Samples were prepared according to library kit manufacturer’s protocol, indexed, pooled, and sequenced on an Illumina NovaSeq 6000. Basecalls and demultiplexing were performed with Illumina’s bcl2fastq2 software. RNA-seq reads were then aligned and quantitated to the Ensembl release 101 primary assembly with an Illumina DRAGEN Bio-IT on-premises server running version 3.9.3-8 software. All gene counts were then imported into the R/Bioconductor package EdgeR ⁶⁴, and TMM normalization size factors were calculated to adjust for samples for differences in library size. Ribosomal genes and genes not expressed in the smallest group size minus one sample greater than one count-per-million were excluded from further analysis. The TMM size factors and the matrix of counts were then imported into the R/Bioconductor package Limma ⁶⁵. Weighted likelihoods based on the observed mean-variance relationship of every gene and sample were then calculated for all samples, and the count matrix was transformed to moderated log 2 counts-per-million with Limma’s voomWithQualityWeights ⁶⁶. The performance of all genes was assessed with plots of the residual standard deviation of every gene to their average log-count with a robustly fitted trend line of the residuals. Differential expression analysis was then performed to analyze for differences between conditions, and the results were filtered for only those genes with Benjamini-Hochberg false-discovery rate adjusted p-values less than or equal to 0.05. For each contrast extracted with Limma, global perturbations in known Gene Ontology (GO) terms were detected using the R/Bioconductor package GAGE ⁶⁷ to test for changes in expression of the reported log 2-fold-changes reported by Limma in each term versus the background log 2-fold-changes of all genes found outside the respective term. The R/Bioconductor package heatmap3 ⁶⁸ was used to display heatmaps across groups of samples for each GO term with a Benjamini-Hochberg false-discovery rate adjusted p-value less than or equal to 0.05. To find the most critical genes, the Limma voomWithQuality Weights transformed log 2 counts-per-million expression data was then analyzed via weighted gene correlation network analysis with the R/Bioconductor package WGCNA ⁶⁹. Briefly, all genes were correlated across each other by Pearson correlations and clustered by expression similarity into unsigned modules using a power threshold empirically determined from the data. An eigengene was then created for each de novo cluster, and its expression profile was then correlated across all coefficients of the model matrix. Because these clusters of genes were created by expression profile rather than known functional similarity, the clustered modules were given the names of random colors where grey is the only module that has any pre-existing definition of containing genes that do not cluster well with others. These de novo clustered genes were then tested for functional enrichment of known GO terms with hypergeometric tests available in the R/Bioconductor package clusterProfiler ⁷⁰. Significant terms with Benjamini-Hochberg adjusted p-values less than 0.05 were then collapsed by similarity into ClusterProfiler category network plots to display the most significant terms for each module of hub genes in order to interpolate the function of each significant module. The information for all clustered genes for each module were then combined with their respective statistical significance results from Limma to determine whether those features were also found to be significantly differentially expressed. To analyze the enrichment and the number of mapped synaptic genes among WGCNA turquois module, the SynGO analysis was conducted (https://syngoportal.org/index.html) ⁷¹. Analyses for GO terms (cellular components) was performed. In the Gene Set Enrichment Analysis setting, the minimum gene count per term for ontology terms was set to three, and the level was expressed as Q-value. The RNA-seq data were deposited into the NCBI GEO database (GEO Accession # GSE224509).

Primary neuronal culture

Primary neuronal culture was performed similarly as described previously ⁷² with modifications. Cortical, hypothalamic, and cerebellar neurons were prepared from each brain region of embryonic day 16 (E16) CD1 mice. Brain regions were dissected and placed on ice in DMEM-High glucose (12800-017, Gibco) with antibiotics (15240-062, Gibco). Each region was digested by incubating in 0.05% trypsin-EDTA (25200-72, Gibco) at 37 °C for 10 min with shaking every two min in a microcentrifuge tube. Trypsin activity was neutralized by adding an equal volume of fetal bovine serum (FBS), and Trypsin-FBS media was removed from the tube. 1 ml of plating media comprised of Neurobasal medium (21103-049, Gibco), 10% FBS, 2 mM L-glutamine (25030, Gibco), and antibiotics was added into the tube and gently pipetted up-and-down 10 times with a flamed glass pasture pipette. The tube was kept at RT for 1-2 min to allow the chunky, un-dissociated tissues to settle out of solution, and then the upper phase containing the fully dissociated neurons was transferred to a new tube. Neurons were seeded on 100 μg/ml poly-D-lysine (P6407, Sigma-Aldrich)-coated 24-well culture dishes (1.5 × 10⁵ cells/well). The cells were incubated for 3-4 h at 37 °C, and the plating medium was replaced to complete culture medium containing Neurobasal medium, N2 supplement (17502-048, Gibco), B27 supplement (17504-04, Gibco), 2 mM L-glutamine, and antibiotics. Neurons started to grow neurites when they attached properly and were treated with 1 μM Ara-C (C1768, Sigma-Aldrich) after 1 day in vitro (DIV). To maintain the culture, half the medium was changed every 3 days. Neurons were harvested after 14 DIV for extracting RNA.

Multiple/micro electrode array (MEA)

Primary hypothalamic neuronal cultures were prepared as described above with the following modifications described below to optimize for Microelectrode array (MEA). After dissection and digestion of hypothalami, 2.5 x 10⁴ cells were plated onto each well of an MEA plate (M768-tMEA-48W, Axion Biosystems) coated with 100 μg/ml poly-D-lysine (P6407, Sigma-Aldrich) and 10 μg/ml Laminin (L2020, Sigma-Aldrich). To knockdown Ppp1r17, shRNA lentivirus (2 x 10⁶ IU/ml as a final concentration) was added after 8 DIV, and neuronal activity was recorded after 21 DIV on the Maestro MEA plate reader (Axion Biosystems) and analyzed with the AxIS Navigator software. The neural metric tool was used to generate raster plots for neurons. In these plots, each line is a spike that represents firing of neurons, and a group of five continuous spikes makes up a burst firing that is represented by a blue line. A series of bursts from neighboring neurons make up a network burst represented by the activity in the purple box in each raster plot.

Plasmid construction for Ppp1r17 mutants

Mus musculus Ppp1r17 cDNA was amplified by PCR with Pfu Ultra II (#600670, Agilent) and cloned in the KpnI/XhoI sites of pcDNA3-mRFP (#13032, Addgene), and subjected to site-directed mutagenesis for Ppp1r17 T72A, T123A and T72AT123A (PrimeSTAR; R050A, Takara Bio). Constructs were verified by DNA sequencing.

HEK293 cell culture and time-lapse imaging analysis

HEK293 were maintained in DMEM (11965092, Thermo Fisher) supplemented with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin and maintained at 37 °C and 5% CO₂. For cell imaging, cells were grown in a Nunc Glass Base Dish (150680, Thermo Fisher) and transfected using Lipofectamine 3000 (L-3000, Invitrogen) with pcDNA3-mRFP (#13032, Addgene) or pcDNA3-Ppp1r17-mRFP (WT, T72A, T123A or T72AT123A) plasmid for 24 h. For time-lapse imaging, transfected HEK293 cells underwent leptomycin B (20 ng/ml; L2913, Sigma-Aldrich) treatment or DMSO treatment with SiR-DNA (CY-SC007, Cytoskeleton Inc.) treatment for nuclear labeling. Imaging of Ppp1r17-mRFP was performed using a Leica SP5X white light laser confocal system with Leica Application Suite (LAS) Advanced Fluorescence and an incubation system where cells were maintained in a humidified incubator (37 °C and 5% CO₂) throughout the experiment. Time-lapse images were acquired at 3 min intervals for 1 h. The time-lapse images were used for quantitative analysis of Ppp1r17-mRFP translocation by ImageJ (NIH, Bethesda, MD, USA).

DMH-specific Prkg1-KD in aged mice

Lentivirus carrying shRNA against Prkg1 or fLuc was injected bilaterally into the DMH of 20-month-old wild-type males. Their body weights were measured once a week through this study. Tissues and blood collections were done at ZT 15, 12 weeks after virus injection. Nine weeks after the virus injection, the mice were placed in individual cages with a running wheel. The mice were perfused with PBS 12 weeks after virus injection, and brains and WATs were collected and processed for IHC and Western blot analysis, respectively.

Frailty Index (FI)

Frailty was assessed longitudinally by the same researcher (K.T.), as modified from the original mouse clinical FI ³⁵ and previous study ³⁶. Body weight and body temperature were not assessed in the current study, and thus, an FI of 29 total items was used. DMH-specific fLuc-KD control and Prkg1-KD male mice were assessed one at a time for multiple characteristics within 2-3 min per mouse. Briefly, mice were scored either 0, 0.5, or 1 for the degree of deficit that they showed in each of these items with 0 representing no deficit, 0.5 representing a mild deficit, and 1 representing a severe deficit ³⁶. FI testing sheet with all items can be found in Supplementary Materials.

Chemogenetics (DREADD)

18-month-old Ppp1r17-Cre male mice received bilateral injections of AAV5-hSyn-DIO-hM3D(Gq)-mCherry into the DMH. The mice were placed in individual cages with a running wheel 7 days after virus injection for habituation. Their body weights and food intake were measured every day through this study. PBS injections for habituation were given intraperitoneally (i.p.) once a day at ZT 11.5-12 for 9 days. Water-soluble DREADD Agonist 21 (HelloBio, HB6124) was dissolved in PBS and aliquots were stored at −30 °C. Then, the mice were injected with Agonist 21 by i.p. at a dose of 0.5 mg kg⁻¹ at ZT 11.5-12 for 8 days. Blood collections were done at the last day of each PBS and Agonist 21 injection around ZT 15. To see the disappearance of the Agonist 21 effects in the wheel-running activity, PBS injections were performed for 4 days after Agonist 21 injections. Before collection of tissues from the mice, Agonist 21 injection was performed for 4 days. The mice were perfused with PBS, and brains and WATs were collected and processed for IHC and Western blot analysis, respectively.

Life Span Analysis

All animals were kept in our animal facility with standard laboratory diet and water ad libitum throughout the study. No mice used for the longevity studies were used for any other physiological or metabolic tests. All mice in the aging cohorts were carefully inspected once a day. Their body weights were measured every Wednesday (once a week). The endpoint of life was the time when each mouse was found dead during daily inspection. Moribund mice were euthanized according to our institutional animal care guidelines, and the time at euthanasia was its endpoint. Necropsy was performed immediately after death or euthanasia by the Research Animal Diagnostic Laboratory at Washington University in St. Louis. Survival data of each cohort were analyzed by plotting the Kaplan-Meier curve and performing the log rank test using GraphPad Prism 9. Mean life span and maximum 80% life span, the age at which 20% of the mice in each cohort remain alive, were compared using Student’s t test. Age-associated mortality rate (qx) was estimated as the number of animals alive at the end of the interval over the number of animals at the start of the interval. The hazard rate (hz) was estimated by hz = 2qx/ (2-qx) ⁷³, natural logarithm of hz was plotted, and statistical analysis was performed using ANCOVA.

DMH specific Prkg1-KD for lifespan study:

Lentivirus carrying shRNA against Prkg1 or fLuc was injected bilaterally into the DMH of 20-month-old wild-type males (n=19 in each group). There was no treatment for this experiment.

Chemogenetic Experiment for lifespan study:

22-month-old Ppp1r17-Cre male mice received bilateral injections of AAV5-hSyn-DIO-hM3D(Gq)-mCherry into the DMH (n=13 in each group). Six weeks after virus injection, both groups of the mice were injected with Agonist 21 by i.p. at a dose of 0.25 mg kg⁻¹ at ZT 11.5-12 everyday Monday through Thursday (four consecutive days in each week) until their endpoint.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data Analysis

Results are presented as mean ± SEM. All statistical tests were performed using GraphPad Prism 9. Significance between two groups was assessed by Student’s t test or two-way ANOVA with Šidák’s test if it is repeatedly measured. Comparison of wheel-running activities was performed by Wilcoxon matched-pairs signed-rank test. Statistical comparison of wheel-running activities, body weight, food intake, lipid measurements and eNAMPT levels in plasma between pre- and post-treatments with Agonist 21 injection was performed by two-way ANOVA with with Šidák’s multiple comparisons test. Log-rank (Mantel-Cox) test was used for the statistical analysis of lifespan. Generally, normal data distribution was assumed and was not tested in these experiments. Sample sizes and other statistical parameters are indicated in the figures and text. *P< 0.05, **P< 0.01, ***P< 0.001, ****P<0.0001. Significance was concluded at P< 0.05.

Supplementary Material

Data S1

Data S1. Unprocessed data underlying the display items in the manuscript, related to Figures 1–7 and S1–S7.

Movie S1

Movie S1. Related to Figure 1; 3D-reconstructed IHC image of Ppp1r17 in the posterior hypothalamus.

Immunohistochemical localizations of Ppp1r17 (green) and DAPI (blue) in the posterior hypothalamus (Bregma −1.23 mm to −2.53 mm) are 3D-reconstructed.

Movie S2

Movie S2. Related to Figure S6; Live imaging of Ppp1r17 translocation in HEK293 cells by LMB.

Representative confocal live image of Ppp1r17-mRFP (red) and the nuclear label SiR-DNA (blue) in HEK293 with LMB treatment (20ng/ml). LMB treatment was started at the beginning of the Movie S2.

Movie S3

Movie S3. Related to Figures 6G and S6I; A video clip of 34-month-old fLuc-KD control and Prkg1-KD male mice.

Highlights.

• DMH^Ppp1r17 neurons control WAT function, eNAMPT secretion, and physical activity

• Ppp1r17 is localized to cytoplasm by PKG with age, causing synaptic dysfunction

• DMH-specific Prkg1 knockdown counteracts aging and extends lifespan

• Chemogenetic activation of DMH^Ppp1r17 neurons delays aging and extends lifespan