Learn more: PMC Disclaimer | PMC Copyright Notice
Neural Control
Separate and shared sympathetic outflow to white and brown fat coordinately regulates thermoregulation and beige adipocyte recruitment
1,2,3 and Timothy J. Bartness1,2,3,†Associated Data
Abstract
White adipose tissue (WAT) and brown adipose tissue (BAT) are innervated and regulated by the sympathetic nervous system (SNS). It is not clear, however, whether there are shared or separate central SNS outflows to WAT and BAT that regulate their function. We injected two isogenic strains of pseudorabies virus, a retrograde transneuronal viral tract tracer, with unique fluorescent reporters into interscapular BAT (IBAT) and inguinal WAT (IWAT) of the same Siberian hamsters to define SNS pathways to both. To test the functional importance of SNS coordinated control of BAT and WAT, we exposed hamsters with denervated SNS nerves to IBAT to 4°C for 16–24 h and measured core and fat temperatures and norepinephrine turnover (NETO) and uncoupling protein 1 (UCP1) expression in fat tissues. Overall, there were more SNS neurons innervating IBAT than IWAT across the neuroaxis. However, there was a greater percentage of singly labeled IWAT neurons in midbrain reticular nuclei than singly labeled IBAT neurons. The hindbrain had ~30–40% of doubly labeled neurons while the forebrain had ~25% suggesting shared SNS circuitry to BAT and WAT across the brain. The raphe nucleus, a key region in thermoregulation, had ~40% doubly labeled neurons. Hamsters with IBAT SNS denervation maintained core body temperature during acute cold challenge and had increased beige adipocyte formation in IWAT. They also had increased IWAT NETO, temperature, and UCP1 expression compared with intact hamsters. These data provide strong neuroanatomical and functional evidence of WAT and BAT SNS cross talk for thermoregulation and beige adipocyte formation.
the brain regulates body fat through sympathetic nervous system (SNS) nerves directly innervating peripheral fat depots. In humans and rodent models, fat is stored in white adipose tissue (WAT) and brown adipose tissue (BAT). WAT stores energy, while BAT oxidizes fatty acids and generates heat through nonshivering thermogenesis, which decreases fat mass and maintains core body temperature (for reviews see Refs. 27, 30, 60). When stimulated, SNS nerve terminals will release norepinephrine (NE) and initiate lipolysis in WAT and nonshivering thermogenesis in BAT (for reviews see Refs. 5, 7, 8). When SNS nerves projecting to WAT and BAT are either chemically or surgically denervated, lipid mobilization (9, 16, 20, 72) and nonshivering thermogenesis (21, 28, 43) are respectively blocked. Uncoupling protein 1 (UCP1) in BAT mitochondria uncouples the electron transport during ATP synthesis and generates heat when SNS outflow to BAT is increased (14, 51). UCP1-dependent nonshivering thermogenesis is necessary for thermoregulation as demonstrated by the cold intolerance of UCP1 knockout (KO) mice during acute (18–24 h) cold exposure without previous cold acclimation (22, 31).
Although WAT and BAT have distinct functional differences, it has been observed that WAT can remodel itself to possess BAT-like characteristics (18, 32, 39, 54, 74). BAT-like or beige adipocytes express functional UCP1 in WAT and have BAT-like multilocular lipid droplet morphology when stimulated by β-adrenergic agonism (18, 39, 74). Interestingly, while UCP1 KO mice are unable to perform BAT nonshivering thermogenesis, their inguinal WAT (IWAT) shows recruitment of multilocular brown-like adipocytes (38, 65) and increased expression of the BAT marker cell death-inducing DFFA-like effector A and cytochrome c oxidase activity (41). These data are indicative of WAT remodeling in the absence of functional BAT, thus suggesting a potential cross talk that is an interaction between WAT and BAT SNS for coordinated thermoregulatory responses. However, the specific contribution of induced beige adipocytes in WAT to whole body thermoregulation and fat metabolism has not been tested. In addition, the neuroanatomical basis of the functional cross talk between WAT and BAT in thermoregulation is unclear.
In previous studies investigating the SNS outflow to WAT and BAT, the central SNS neurons projecting to WAT (58, 62, 63) or to BAT were independently labeled (3, 59, 64) with pseudorabies virus (PRV), a retrograde transsynaptic tract tracer. Similarly labeled brain regions were observed in independent WAT and BAT neuroanatomical studies suggestive of shared SNS outflow to both BAT and WAT. However, a specific neuroanatomical investigation of whether WAT and BAT have shared or separate SNS circuitries that may potentially participate in the coordinated control of thermoregulation has not been made. The use of isogenic strains of PRV each with unique fluorescent reporter protein provides an advantage to define the presence of separate or more importantly shared central SNS innervations across the neuroaxis from two fat depots within the same animal (1, 50).
Thus the present study is designed to investigate whether there are shared or separate central SNS outflows to WAT and BAT that may coordinately regulate their functions. We have labeled the central SNS circuitries between WAT and BAT for the first time using isogenic strains of PRVs tagged with unique fluorescent reporter proteins [either green fluorescent protein (GFP) or red fluorescent protein (RFP)]. We have also tested whether there is coordinated SNS control of WAT and BAT thermogenic function in acute (16–24 h) cold-exposed hamsters with SNS denervation of BAT. For those hamsters, we assessed WAT beigeing with physiological, molecular, and histological parameters of temperature measurements, UCP1 expression, and adipocyte morphology, respectively.
MATERIALS AND METHODS
Animals
Adult male Siberian hamsters (Phodopus sungorus, 3 mo old) were singly housed under a long day photoperiod light cycle (16-h light:8-h dark) and an ambient temperature of 22 ± 2°C. Hamsters were given ad libitum regular rodent chow (Purina Rodent Chow, St. Louis, MO) and tap water for all studies. We used Siberian hamsters, because they are a natural model of obesity reversal due to their seasonal obesity in long day photoperiod and its reversal in short day photoperiod (20). The adiposity of long day Siberian hamsters is comparable to rats and mice genetically manipulated for obesity or given a high-fat diet, and this adiposity can be manipulated without dietary intervention but simply with photoperiod. In addition, the SNS outflow to WAT and BAT in Siberian hamsters has been extensively characterized (5, 8) and the SNS innervation of their fat tissues is comparable to other mammals with more similarities than differences in the general pattern (2, 42). All procedures were approved by the Georgia State University Institutional Animal Care and Use Committee and were in accordance with Public Health Service and United States Department of Agriculture guidelines.
Experiment 1: Do IWAT and IBAT Have Shared or Separate Central SNS Innervations?
Preliminary PRV injections tests.
It was necessary to verify that the two isogenic PRV strain-infected neurons at a similar rate and that one was not more virulent than the other. As a control to ensure that they were functionally equivalent, a 1:1 equal volume mix (50) of PRV 152 [3 × 108 plaque-forming units (pfu)/ml] with GFP and PRV 614 (1.8 × 108 pfu/ml) with RFP was unilaterally injected into either IWAT (0.15 µl/locus, total volume of 1.5 µl) or IBAT (0.15 µl/locus, total volume of 0.75 µl) to ensure similar virulence between the two viruses and to decrease the possibility for the phenomenon of exclusion, which occurs in a dual virus tract tracing study when a virus infects a neuron and consequently decreases the ability of another virus from infecting it (34). There were similar percentages of PRV 152 and PRV 614 SNS-labeled neurons in the brains of these hamsters, thereby demonstrating functional equivalence of PRV 152 at 3 × 108 pfu/ml and PRV 614 at 1.8 × 108 pfu/ml (Nguyen NT, Bartness TJ, unpublished observations). Therefore, we used these PRV 152 and 614 titers in our dual tract tracing study with IWAT and IBAT.
Dual PRV tract tracing of IWAT and IBAT.
All virus injections were executed according to Biosafety Level 2 standards. Based on the proximity of IWAT and IBAT to the brain, we injected PRV into IWAT first and injected isogenic PRV with a different reporter tag into IBAT 24 h later to obtain the optimal time for the progression of both viruses across the neuroaxis (Nguyen NT, Bartness TJ, unpublished observations). In another cohort, the fat depots in which the isogenic PRVs were injected were reversed to account for fat depot differences in viral uptake. The number of fluorescently labeled SNS neurons across the neuroaxis (i.e., SNS ganglia, spinal cord, and brain) from the two cohorts was averaged for the calculation of singly and doubly labeled neurons. Because it was not technically possible to ensure that the PRVs simultaneously reached every brain site to label SNS outflow to IWAT and IBAT, there may be more doubly labeled neurons than we quantified despite our controls to minimize this risk.
In brief for PRV injections into IWAT, hamsters (n = 7) were anesthetized. An incision was made around the inguinal region to expose the right IWAT. Hamsters were unilaterally injected with 0.15 µl of PRV 152 or PRV 614 at 10 loci along the fat depot to evenly distribute the virus. As we previously observed no side preference of PRV infection between right and left fat depots (Song CK, Bartness TJ, unpublished observations), we performed unilateral injections of the PRV into the right IWAT depot. The Hamilton syringe was held in place for 1 min after each injection to prevent efflux when removing the syringe. The incision site was closed with sterile wound clips, and hamsters were given subcutaneous injections of Ketofen (5 mg/kg; Fort Dodge Animal Health, Fort Dodge, IA), an analgesic, daily for 3 days postsurgery. Hamsters were transferred to clean cages. For PRV injections into IBAT, the same hamsters were reanesthetized and shaved followed by a small interscapular incision. A unilateral injection of isogenic PRV counterpart with a different reporter tag was made into the right IBAT at five loci for a total of 0.75 µl 24 h after IWAT injections. A reduced volume of PRV was injected into IBAT to accommodate its smaller size (36, 59). We followed the same surgical procedures previously used for closing the incision site and Ketofen administration. We have previously tested PRV injections for potential viral leakage by placing the virus on the surface of an exposed fat depot and found no positive labeling in the SNS ganglia, spinal cord, and brain as opposed to intrafat injections of the virus (59). We have also examined the specificity of viral transport by injecting PRV into a SNS-denervated fat depot and observed no labeling, confirming that PRV requires SNS nerves from fat for transsynaptic labeling (26).
Tissue fixation.
Hamsters were euthanized 6 days after PRV injections into IWAT and perfused with 0.9% heparin saline (100 ml) followed by 4% paraformaldehyde (150 ml). Brains, spinal cords, and SNS ganglia were harvested and postfixed in the same fixative for 4 h (brains and spinal cords) or 15 min (SNS ganglia) at 4°C after which brains and spinal cords were transferred to a 30% sucrose in 0.1 M phosphate-buffered saline (PBS; pH 7.4) solution and SNS ganglia to an 18% sucrose solution. Brains were sectioned at 35-µm thickness on a freezing stage sliding microtome. Spinal cords and SNS ganglia (T1-T3, T12-L3) were sectioned at 40 and 16 µm, respectively, on a Leica cryostat and directly transferred to slides (Superfrost Plus; VWR International, Arlington, IL). Sections of brains, spinal cords, and SNS ganglia were stored in 0.1 M PBS with 0.1% sodium azide (NaN3) at 4°C until double fluorescent immunohistochemistry (IHC) was performed.
Double fluorescent IHC.
Double fluorescent IHC was used to amplify GFP and RFP labeling in every fourth brain, spinal cord, and SNS ganglia section to eliminate counting neurons twice as we previously described (50). We have previously tested the primary and secondary antibodies used in this protocol to ensure specificity and absence of cross reactivity (50). All steps were performed at room temperature (RT). Sections were blocked in 20% normal goat serum (NGS; Vector Laboratories, Burlingame, CA) in 0.4% PBS Triton (PBTx) for 30 min followed by incubation in rabbit anti-RFP (1:2,000; Rockland Immunochemicals, Gilbertsville, PA) and mouse anti-GFP (1:500; Abcam, Cambridge, MA) antibodies in 0.4% PBTx with 2% NGS and 0.1% NaN3 for 24 h. Then sections were incubated with goat anti-rabbit CY3 (1:800; Jackson Immunoresearch, West Grove, PA) and goat anti-mouse Alexa 488 (1:500; Jackson Immunoresearch) antibodies in 0.4% PBTx with 2% NGS for 2 h. Double fluorescent IHC for spinal cords and SNS ganglia followed the same steps except the mouse anti-GFP and rabbit anti-RFP antibody concentrations were increased to 1:400 and 1:600, respectively, and the incubation times were increased to 2 days because the sections were slide mounted. The concentrations of goat anti-mouse Alexa 488 and goat anti-rabbit CY3 antibodies were also increased to 1:350 and 1:550, respectively. For negative controls, primary antibodies were omitted resulting in a notable absence of amplified staining.
Quantification analysis.
Quantification of PRV-labeled SNS neurons was performed as previously described (50). PRV-labeled SNS neurons were considered positive based on cell size, morphology, and fluorescent intensity. They were analyzed using an Olympus BX41 microscope with appropriate filters for GFP and RFP. Images were acquired at ×10 and ×20 magnifications using an Olympus DP73 camera and were adjusted for brightness and contrast using Adobe Photoshop CS5 (Adobe Systems, San Jose, CA). Adobe Photoshop was also used to merge GFP and RFP images to visualize yellow-orange doubly labeled neurons. Positively labeled neurons singly or doubly projecting to IWAT and IBAT were quantified using the manual tag feature of Adobe Photoshop thus ensuring each neuron was counted once. A mouse brain atlas (53) was used to identify brain sites as we and others have previously done (10, 50, 58, 59, 63, 64), because no commercially available Siberian hamster brain atlas exists. Absolute values of PRV-labeled SNS neurons found in each nucleus or region were either kept as absolute values or converted into a percentage of total PRV-labeled SNS neurons. We also collapsed quantifications of PRV-labeled neurons across the whole brain and in nuclei found in the hindbrain, midbrain, or forebrain to better examine the distribution of labeled SNS neurons to IWAT and/or IBAT across these large brain divisions. Distinct and collapsed neuronal quantifications were averaged across the number of hamsters. The absolute values or percentages of singly labeled IWAT and IBAT and doubly labeled neurons are represented as the mean absolute or percentage of total PRV-labeled SNS neurons. For emphasis of their potential contribution and importance to thermoregulation and beigeing, we compared labeled SNS neurons across hypothalamic nuclei and classic SNS and thermoregulatory nuclei in bar graphs in addition to our supplementary table of labeled cell counts. We quantified the labeling from the ipsilateral side of PRV injections into IWAT and IBAT in all spinal cords and SNS ganglia and presented the data in the same manner as the brain sections.
Experiment 2: Does SNS Denervation of IBAT Increase SNS Drive to and Beige Adipocyte Formation in IWAT?
Norepinephrine turnover during 16-h cold challenge.
IBAT SNS denervation was performed using a modified method described in previous publications for IWAT SNS denervation (23, 29, 55). Briefly, age- and weight-matched male hamsters (n = 52) were divided into two groups: vehicle control (VC) group and IBAT SNS-denervated group for norepinephrine turnover (NETO), a direct neurochemical measure of SNS drive, in IBAT and WAT tissues. IBAT SNS denervation was achieved by microinjections of 6-hydroxydopamine (6OHDA), a selective neurotoxin to SNS nerves (55). Hamsters were anesthetized and an incision was made around the interscapular region to expose the lobes of IBAT. Saline or 6OHDA (10 mg/ml, 2 µl/locus for a total of 30 µl per IBAT lobe; Sigma-Aldrich, St. Louis, MO) was bilaterally injected into IBAT using a Hamilton syringe. The syringe was held in place for 1 min to prevent efflux. The skin was closed with sterile wound clips and Ketofen was administered.
NETO was measured 2 wk after intra-IBAT injections of 6OHDA or saline in hamsters housed either RT or 4°C as described (12). In brief, hamsters were handled daily for 2 wk before NETO measures to adapt them to the handling associated with the procedure and to decrease stress-induced NE release. On the day that NETO was measured, hamsters were placed either in 4°C or RT for 16 h and NETO was measured during the last 4 h of the experiment (12). Hamsters were injected intraperitoneally with α-methyl-p-tyrosine (250 mg/kg α-MPT; Sigma-Aldrich), an active competitive inhibitor for tyrosine hydroxylase (TH), which is the rate-limiting enzyme for NE production, thus preventing the synthesis of catecholamine. A supplemental dose of α-MPT (125 mg/kg) was given 2 h after the initial dose to ensure the inhibition of catecholamine synthesis. Four hours after the first α-MPT injection, hamsters were weighed and then decapitated. IBAT, IWAT, epididymal WAT (EWAT), and retroperitoneal WAT (RWAT) were quickly harvested, weighed, frozen in liquid nitrogen, and stored at −80°C until NE extraction. To obtain baseline NE values for between animal calculations of NETO, one-half of hamsters from each treatment and temperature groups were euthanized without receiving α-MPT injections 4 h before the conclusion of the study (12, 70). The adipose tissues were processed and extracted for NE with dihydroxybenzylamine (Sigma-Aldrich) as an internal control for extraction efficiency. NE content and NETO were measured as described previously (for review see Ref. 70) and following our methods and the modification of the method of Mefford (40). Calculations were made according to the following formula: k = (lg[NE]0 − lg[NE]4)/(0.434 × 4) and K = k[NE]0, where k is the constant rate of NE efflux, [NE]0 is the initial NE concentration, [NE]4 is the final NE concentration, and K = NETO.
IBAT SNS denervation and iButton and transponder implantation.
VC or 6OHDA were injected into the IBAT of a separate cohort of hamsters (n = 32). IBAT and IWAT temperatures of those same hamsters were measured using implantable programmable temperature transponders (IPTT) 300 (serial no. 2144251; Bio Medic Data Systems, Seaford, DE) placed under the fat depot as previously described (11, 36, 59, 64, 68, 69). In brief, an IPTT was carefully placed under IBAT and was secured with sterile sutures (Ethicon; Johnson & Johnson, Somerville, NJ) after intra-IBAT injections of 6OHDA or saline. A small incision was made at the inguinal region to expose IWAT and another IPTT was placed and secured under it similarly to IBAT. Temperature readings on IPTTs were detected with a portable handheld reader programmer (model no. DAS-1001R; Bio Medic Data Systems). The portable handheld reader for IPTTs was placed on the ventral side around the inguinal region of IWAT to detect IWAT temperature and on the dorsal side above the interscapular region to detect IBAT temperature (36, 64, 68). Each IPTT had a unique ID number and was programmed with the reader such that when the reader was placed on either ventral or dorsal side, it only read the corresponding IPTT placed under IWAT or IBAT, respectively. Core temperature was measured in the same hamsters using iButton data loggers (Maxim Integrated, DS1922L, 8KB) placed intraperitoneally as previously described (68). iButtons were calibrated using One Wire Viewer (version 3.17.44) and programmed to automatically record core body temperature every 30 min during the 24-h cold challenge. iButtons were first dipped in paraffin wax and gas sterilized. A small lateral incision was made on the ventral side of the hamster through the skin and peritoneal wall for iButton insertion into the abdominal cavity. The peritoneal wall was closed with sterile sutures and skin incision sites were closed with sterile wound clips and Ketofen was administered.
Twenty-four hour cold challenge.
Two weeks postsurgeries, hamsters in VC and 6OHDA groups were further divided into either 4°C or RT groups. Body mass was recorded before and after the challenge. IWAT and IBAT temperatures were recorded every hour, and core temperature was also recorded with the iButton data logger system every 30 min during 24 h of the cold challenge. At the end of the 24-h cold challenge, hamsters were decapitated. IBAT, IWAT, EWAT, and RWAT were quickly dissected, weighed, frozen in liquid nitrogen, and subsequently stored at −80°C until further analysis by Western blot for TH, a marker of SNS nerves, and UCP1, a marker of brown and beige adipocytes. Portions of IWAT and IBAT were placed into tissue cassettes and stored in neutral buffered 10% Formalin for fixation before paraffin embedding for fat UCP1 IHC. Remaining portions of IBAT and IWAT were used to measure calcitonin gene-related peptide (CGRP), a marker of sensory nerves, to verify 6OHDA chemical specificity for SNS nerves and to measure the degree of sensory innervation in IBAT and IWAT.
Western blot and fat UCP1 IHC.
TH and UCP1 protein expression in IWAT and IBAT were measured by Western blot as described previously (19). Primary antibodies used were rabbit anti-UCP1 (1:500; Abcam), rabbit anti-TH (1:500; EMD Millipore, Temecula, CA), and rabbit anti-α tubulin (1:500; Cell Signaling, Danvers, MA) as a loading control. Membranes were blocked with 5% nonfat dry milk in Tris-buffered saline (TBS) and then incubated with primary antibodies at 4°C for 24 h. They were washed with TBS with 0.1% Tween 20 followed by a 2-h incubation with goat anti-rabbit Alexa Fluor 680 nm antibody (ThermoFisher Scientific, Carlsbad, CA) RT before visualization of protein bands using Odyssey FC Imaging System (Li-Cor Biotechnology, Lincoln, NE). TH and UCP1 protein expression was normalized to α tubulin.
Paraffin-embedded IWAT and IBAT tissues of 24-h cold-exposed hamsters were sliced at 6 µm on a rotary microtome, mounted onto slides, and dried on a warming plate (37°C). Sections were deparaffinized and rehydrated followed by an antigen retrieval step with target retrieval solution (Dako, Carpinteria, CA). They were then incubated with rabbit anti-UCP1 antibody (1:150; Abcam) overnight at 4°C followed by a 20-min incubation with biotinylated donkey anti-rabbit antibody (1:100; Jackson Immunoresearch, West Grove, PA), then avidin-biotin complex (Vector Laboratories) for 30 min per kit direction. Lastly, a 10-min exposure to diaminobenzidine (Vector Laboratories) was used for visualization of fat UCP1. Bright-field photomicrographs were acquired at ×20 and ×40 magnification using an Olympus DP73 camera on an Olympus BX41 microscope.
CGRP enzyme immunoassay.
CGRP levels in IBAT and IWAT were measured using an enzyme-linked immunosorbent assay kit (SPI Bio, Massy, France) according to the manufacturer’s directions as we have previously done (68). The correlation coefficient was 99.9% for CGRP assays.
Statistical Analyses
Results are expressed as means ± SE. Statistical analyses were carried out using Systat Software (version 11.0; Systat San Jose, CA). The percentage of PRV-labeled SNS neurons projecting to IWAT and/or IBAT was transformed for analysis using square root transformation and tested for normality with Shapiro-Wilk test. A one way ANOVA with Student-Newman-Keuls post hoc test was used to compare the absolute values and percentages of doubly labeled neurons with singly labeled IWAT and IBAT neurons. Percentage data of labeled neurons that did not fit normal distribution and violated equal variance test were analyzed using Kruskal-Wallis test for nonparametric data. Placement of transponders under IBAT and IWAT was visually verified during fat tissue extraction after cold challenge. Fat temperature recordings of hamsters that had misplaced transponders were omitted from analysis. IBAT and IWAT temperatures were analyzed by one-way repeated-measures ANOVA. NETO and protein values were statistically analyzed by two-way ANOVA (treatment × temperature) and core temperature data were analyzed by two-way repeated-measures ANOVA. The Student-Newman-Keuls post hoc test was used when differences within or between groups were obtained.
RESULTS
Experiment 1: Do IWAT and IBAT Have Shared or Separate Central SNS Innervations?
SNS ganglia.
A regional distribution of PRV-labeled SNS neurons to IWAT and IBAT was found across T1-T3 and T12-L3 vertebral levels of the SNS ganglia (Fig. 1, A–D). There was a higher percentage of singly labeled IBAT neurons at the thoracic level compared with singly labeled IWAT and doubly labeled neurons (P < 0.05), and this pattern was reversed at the lumbar level (Fig. 1, C and D). T3 SNS ganglion had the highest percentage of singly labeled IBAT neurons (~60%) compared with other vertebral levels (~30%) although these values did not reach statistical differences (Fig. 1, A and C). The SNS ganglia from T13-L2 had a higher percentage of PRV-labeled neurons projecting to IWAT compared with doubly labeled neurons (P < 0.05; Fig. 1C). This is consistent with our previous observation that there is a pattern for SNS ganglia at the lumbar level to innervate IWAT (Fig. 1D) (58, 59).
Distribution of sympathetic nervous system (SNS) neurons to interscapular brown adipose tissure (IBAT) and inquinal white adipose tissue (IWAT) across thoracic and lumbar regions of the SNS ganglia. A and B: representative pseudorabies virus (PRV) labeling of SNS neurons to IBAT (red) and IWAT (green) in T3 (A) and L3 (B) SNS ganglia. Doubly labeled neurons are indicated by white arrows. C: percentage of total PRV-labeled SNS neurons across each thoracic (T1–3, 12–13) and lumbar (L1–3) regions. D: percentage of total PRV-labeled SNS neurons across collapsed thoracic and lumbar regions. Scale bar = 50 µm. Values that do not share a common superscript are significantly different at P < 0.05.
Spinal cord.
There was a higher percentage of singly labeled SNS neurons projecting to IBAT (~40%) and to IWAT (~50%) compared with the population of doubly labeled neurons (~10%) at all levels of the spinal cord, particularly the lumbar region (P < 0.05; Fig. 2, A–D). When PRV-labeled neurons in the spinal regions were collapsed, the numbers of singly labeled IWAT and singly labeled IBAT neurons were greater than doubly labeled ones (P < 0.05; Fig. 2E). However, unlike the SNS ganglia, there was no difference in the percentage of singly labeled neurons projecting to either IBAT or IWAT in the spinal cord (Fig. 1, C and D).
Distribution of SNS neurons to IBAT and IWAT in the spinal cord. A–C: representative PRV labeling of SNS neurons to IBAT (red) and IWAT (green) in lumbar region. Doubly labeled neurons are indicated by white arrows. D: percentage of total PRV-labeled SNS neurons across cervical, thoracic, and lumbar regions. E: mean total PRV-labeled SNS neurons in whole spinal cord. Scale bar = 50 µm. IML, intermediolateral horn. Values that do not share a common superscript are significantly different at P < 0.05.
Brain.
The central SNS neurons ultimately projecting to IBAT and IWAT were labeled in a caudal to rostral manner in the brain (Figs. 3, ,4,4, ,5,5, and and6).6). Bilateral infection in the hindbrain, midbrain, and forebrain was observed when isogenic PRV-labeled SNS neurons bilaterally innervating IWAT (63) and IBAT (64). Within the whole brain and in the hindbrain, midbrain, and forebrain, there was a consistent pattern for a higher percentage of singly labeled IBAT neurons compared with singly labeled IWAT and doubly labeled neurons (Figs. 3–5 and and6,6, A and B). Percentages of doubly labeled neurons vary across many distinct areas in the hindbrain, midbrain, and forebrain demonstrating evidence of greater separate SNS circuitry in some brain sites but greater shared SNS circuitry in others (Figs. 3–5 and and6,6, D and E; Supplemental Table 1).
Photomicrographs illustrating central SNS neurons to IBAT and IWAT in the hindbrain. A and E: schematics [adapted from Paxinos and Franklin (53) with permission] illustrating areas of PRV labeling highlighted in orange. B–D and F–H: low (×10) magnification images of PRV-labeled SNS neurons in ROb and GiV (B), Sol (C), RVL and LPGi (D), GiA (F), A5 and rs (G), and RPa (H). Dashed lines indicate regions of high magnification (×20). SNS neurons projecting to IBAT (red), to IWAT (green), and to both fat depots (yellow-orange; white arrows). Scale bar = 100 µm. 4V, fourth ventricle; A5, A5 noradrenaline cells; GiA, gigantocellular reticular nucleus, anterior part; GiV, gigantocellular reticular nucleus, ventral part; LPGi, lateral paragigantocellular nucleus; py, pyramidal tract; ROb, raphe obscurus nucleus; RPa, raphe pallidus nucleus; rs, rubrospinal tract; RVL, rostroventrolateral reticular nucleus; Sol, nucleus of the solitary tract.
Photomicrographs illustrating central SNS neurons to IBAT and IWAT in the midbrain. A: schematic [adapted from Paxinos and Franklin (53) with permission] illustrating areas of PRV labeling highlighted in orange. B–D: low (×10) magnification images of PRV-labeled SNS neurons in DMPAG (B), DR (C), and VLTg (D). Dashed lines indicate regions of high magnification (×20). SNS neurons projecting to IBAT (red), to IWAT (green), and to both fat depots (yellow-orange; white arrows). Scale bar = 100 µm. 4V, fourth ventricle; DMPAG, dorsomedial periaqueductal gray; DR, dorsal raphe; VLTg; ventral lateral tegmental area.
Photomicrographs illustrating central SNS neurons to IBAT and IWAT in the forebrain. A, E, and I: schematics [adapted from Paxinos and Franklin (53) with permission] illustrating areas of PRV labeling highlighted in orange. B–D, F–H, and J–L: low (×10) magnification images of PRV-labeled SNS neurons in DMH (B), Arc (C), VMHVL (D), PVH (F), subZI (G), LH (H), SCh (J), AH (K), and MPO (L). Dashed lines indicate regions of high magnification (×20). SNS neurons projecting to IBAT (red), to IWAT (green), and to both fat depots (yellow-orange; white arrows). Scale bar = 100 µm. 3V, third ventricle; AH, anterior hypothalamus; Arc, arcuate nucleus; DMH, dorsomedial hypothalamic nucleus; LH, lateral hypothalamus; MPO, medial preopotic nucleus; opt, optic tract; PVH, paraventricular nucleus of the hypothalamus; SCh, suprachiasmatic nucleus; subZI, sub zona incerta; VMHVL, ventromedial hypothalamic nucleus, ventral lateral part.
Distribution of central SNS neurons to IBAT and IWAT. A: percentage of total PRV-labeled SNS neurons in whole brain. B: percentage of total PRV-labeled SNS neurons across hindbrain, midbrain, and forebrain. C: comparison of hindbrain, midbrain, and forebrain total mean PRV-labeled SNS neurons. D: percentage of total PRV-labeled SNS neurons across hypothalamic nuclei. E: percentage of total PRV-labeled SNS neurons across brain sites involved in SNS outflow and thermoregulation. PH and Thal Regions, posterior hypothalamus and thalamic regions; LH, lateral hypothalamus; VMH, ventromedial hypothalamus; DMH, dorsomedial hypothalamus; AH, anterior hypothalamus; PVH, paraventricular hypothalamus; NTS, nucleus of the solitary tract; RPa, raphe pallidus; ROb, raphe obscurus; POA, preoptic area. Values that do not share a common superscript are significantly different at P < 0.05.
The hindbrain consistently had more PRV-labeled SNS neurons (i.e., singly and doubly labeled neurons) than the midbrain and forebrain regions (P < 0.05; Fig. 6C). The percentage of singly labeled IBAT neurons tended to be higher than that of singly labeled IWAT and doubly labeled neurons within the hindbrain (Fig. 6B). The raphe pallidus nucleus (RPa) was the only hindbrain nucleus that had a greater percentage of singly labeled IBAT neurons (~40%) compared with IWAT neurons (~20%; P < 0.05) (Figs. 3H and and6E;6E; Supplemental Table 1). It was also the only nucleus we observed that had a higher percentage of doubly labeled neurons (~40%) compared with singly labeled IWAT neurons (P < 0.05; Figs. 3H and and6E;6E; Supplemental Table 1). Other hindbrain sites such as the lateral paragigantocellular nucleus (LPGi), Kolliker-Fuse nucleus (KF), medullary reticular nucleus dorsal part (MdD), motor trigeminal nucleus (Mo5), solitary nucleus ventrolateral part (SolVL), and ventral spinocerebellar tract (vsc) had a general tendency to have higher percentages of singly labeled IBAT neurons compared with IWAT neurons (Fig. 3, C and D; Supplemental Table 1). There was a higher percentage of singly labeled IWAT and IBAT neurons than doubly labeled ones in several hindbrain regions, including the hypoglossal nucleus (12N), gigantocellular reticular nucleus (Gi), dorsal paragigantocellular nucleus (DPGi), intermediate reticular nucleus (IRt), linear nucleus of the medulla (Li), medullary reticular nucleus ventral part (MdV), and parvicellular reticular nucleus/-alpha part (PCRt/A) (P < 0.05; Supplemental Table 1).
The midbrain had ~60% fewer singly and doubly labeled neurons compared with the hindbrain (P < 0.05; Fig. 6C). The number of singly labeled neurons innervating either IWAT or IBAT was greater than that of doubly labeled neurons in many midbrain sites including dorsomedial tegmental area (DMTg); pontine reticular nucleus caudal, oral, and ventral parts (PnC, PnO, and PnV); red nucleus magnocellular part/superior cerebellar peduncle (brachium conjunctivum) (RMC/scp); red nucleus parvicellular part/scp (RPC/scp); retroparafascicular nucleus (RPF); reticulotegmental nucleus of the pons (RtTg), supraoculomotor periaqueductal gray/supraoculomotor cap (Su3/C); and supratrigeminal nucleus (Su5) (P < 0.05; Supplemental Table 1). There was a higher percentage of singly labeled IBAT than IWAT neurons in the DMTg and Su5 (P < 0.05; Supplemental Table 1). The medial longitudinal fasciculus (mlf) and ventral tegmental area (VTA) had tendencies for greater percentages of singly labeled IBAT neurons than IWAT neurons as well (Supplemental Table 1). In contrast, the PnC and RtTg had ~20–40% more singly labeled IWAT neurons than IBAT neurons (P < 0.05; Supplemental Table 1).
Similar to the midbrain, the forebrain had less singly and doubly labeled neurons than the hindbrain (P < 0.05; Fig. 6C). There was a greater percentage of singly labeled neurons to either IBAT or IWAT than doubly labeled ones in the posterior hypothalamic area (PH), paraventricular nucleus of the hypothalamus (PVH) medial magnocellular part (PaMM), PVH posterior part (PaPo), PVH ventral part (PaV), subparaventricular zone of the hypothalamus (SPa), parasubthalamic nucleus (PSTh), xiphoid thalamic nucleus (Xi), and zona incerta (ZI) (P < 0.05; Fig. 5F; Supplemental Table 1). Brain sites such as the arcuate nucleus (Arc), dorsomedial hypothalamic nucleus (DM), lateral hypothalamus (LH), medial preoptic area (MPA), medial preoptic nucleus medial part (MPOM), and PVH anterior parvicellular part (PaAP) had tendencies for higher percentages of singly labeled IBAT neurons compared with IWAT neurons, but this did not reach statistical significance (Fig. 5, B–C, F, H, and L; Supplemental Table 1). The hypothalamus, located within the forebrain, has a large population of SNS neurons involved in the control of fat metabolism. We found a pattern of more singly labeled SNS neurons projecting to IBAT and IWAT in caudal hypothalamic nuclei than rostral ones such as the anterior and paraventricular hypothalamus (Figs. 5, B, D, F, H, and K, and and6D6D).
We also examined brain sites that are classically involved in SNS outflow and essential for thermoregulation to determine if there were differences in SNS circuitries in those sites (Fig. 6E). There were similar percentages of singly labeled IWAT and IBAT neurons in most of those regions except for the RPa, which had a higher percentage of singly labeled IBAT and doubly labeled neurons compared with singly labeled IWAT neurons (P < 0.05; Fig. 6E). The RPa in addition to the raphe obscurus nucleus (ROb) and A5 region (A5) had a greater percentage of doubly labeled neurons compared with the nucleus of the solitary tract (NTS), PVH, and preoptic area (POA) (P < 0.05; Figs. 3, B, G, and H, and and6E6E).
Experiment 2: Does SNS Denervation of IBAT Increase SNS Drive to and Beige Adipocyte Formation in IWAT?
IBAT SNS denervation impaired BAT nonshivering thermogenesis.
All hamsters survived the acute cold challenge. There were no differences in body mass across groups before the cold challenge, nor were there differences in body mass between VC and 6OHDA-treated groups at either RT or after cold exposure (Fig. 7A). All cold-exposed hamsters lost weight during the challenge (P < 0.05; Fig. 7A). There were no changes in IBAT CGRP levels between VC- and 6OHDA-treated hamsters (Fig. 7B). IBAT NETO tended to decrease in 6OHDA-treated hamsters compared with VC-treated ones at RT (Fig. 7C). Acute cold exposure significantly increased IBAT NETO in VC hamsters but not in 6OHDA-treated hamsters (P < 0.05; Fig. 7C). Consistent with this, IBAT TH expression was lower in 6OHDA-treated hamsters compared with their VCs in both cold and RT conditions (P < 0.05; Fig. 7, D and E). In addition, UCP1 protein levels were reduced in IBAT of 6OHDA-treated hamsters even at RT (P < 0.05; Fig. 7, D and F). Twenty-four-hour cold exposure upregulated UCP1 expression in IBAT of VC but not of 6OHDA hamsters (P < 0.05; Fig. 7, D and F). As expected, IBAT temperature of 6OHDA-treated hamsters did not increase during cold exposure as it did in VC hamsters (P < 0.05; Fig. 7G). Specifically, IBAT temperature was lower in 6OHDA-treated than VC hamsters after 4, 7–8, 11, 13, 18, and 21–22 h of cold exposure (P < 0.05) while there were no differences in IBAT temperatures between 6OHDA- and VC-treated hamsters at RT (Fig. 7, G and H). IBAT from 6OHDA-treated hamsters in both cold and RT conditions had a WAT-like unilocular appearance with lower UCP1 immunoreactive (IR) staining compared with VC hamsters (Fig. 7I). These data and the absence of change in IBAT CGRP levels in 6OHDA-treated hamsters (Fig. 7B) demonstrated that SNS denervation of IBAT with 6OHDA selectively impaired IBAT SNS activity without affecting sensory innervation.
SNS denervation impairs IBAT function in 16- to 24-h cold challenge. A: body mass decreases after cold challenge. B: IBAT calcitonin gene-related peptide (CGRP) levels. C: IBAT SNS drive measurement [norepinephrine turnover (NETO)] during 16-h cold exposure. D: IBAT TH and uncoupling protein 1 (UCP1) protein expression. Relative IBAT TH (E) and UCP1 (F) protein expression normalized to α tubulin. IBAT temperatures of cold-exposed (G) and room temperature (H) hamsters. I: UCP1 immunostained IBAT. Lights on from 0 to 13 and 21–24 h (white bars) of the cold challenge; lights off from 13 to 21 h (black bar). VC, vehicle control; RT, room temperature. Values that do not share a common superscript are significantly different at P < 0.05. *P < 0.05, VC vs. 6-hydroxydopamine (6OHDA).
Thermoregulatory function and beigeing in IWAT of IBAT SNS-denervated hamsters.
There was no difference in IWAT NETO between 6OHDA- and VC-treated hamsters at RT. As expected, cold exposure significantly increased IWAT NETO in VC hamsters, but the increase was exaggerated in 6OHDA-treated hamsters (P < 0.05; Fig. 8A). Cold exposure also increased NETO in EWAT and RWAT of VC hamsters (P < 0.05), confirming previous observations (12); however, this response was absent in EWAT and attenuated in RWAT of cold-exposed 6OHDA hamsters (Fig. 8A). There were no differences in IWAT UCP1 expression for cold-exposed VC hamsters or in either treatment group housed at RT (Fig. 8B). IWAT UCP1 was increased in cold-exposed 6OHDA-treated hamsters, and this was significantly different (P < 0.05) from cold-exposed VC hamsters but did not reach significance compared with hamsters housed at RT (Fig. 8B). There were no differences in UCP1 protein expression in EWAT and RWAT for any of the groups (data not shown).
IWAT of IBAT SNS-denervated hamsters increase beigeing and thermoregulatory functions during 16- to 24-h cold challenge. A: WAT SNS drive measurement (NETO) during 16-h cold exposure. B: relative IWAT UCP1 protein expression normalized to α tubulin. C: core body temperatures of all groups recorded every 30 min. IWAT temperatures of cold-exposed (D) and room temperature (E) hamsters. F: UCP1-immunostained IWAT. G: CGRP increases in IWAT of 6OHDA-treated hamsters. #P < 0.05, Cold vs. RT; ‡P < 0.05, all vs. VC RT; *P < 0.05, 6OHDA cold vs. VC RT. Lights on from 0 to 13 and 21 to 24 h (white bars) of cold challenge; lights off from 13 to 21 h (black bar). VC, vehicle control; RT, room temperature. Values that do not share a common superscript are significantly different at P < 0.05. *P < 0.05, VC vs. 6OHDA.
IBAT SNS-denervated hamsters were able to maintain their core body temperatures despite impaired IBAT thermogenic function, as there were largely no differences in core body temperatures across groups either at RT or during 24-h cold exposure (Fig. 8C). By contrast, there was a consistent trend for IWAT temperatures of cold-exposed IBAT SNS-denervated hamsters to be increased compared with VC-treated hamsters (Fig. 8D). This reached significance after 5 and 8 h (P < 0.05) and tended to be increased after 6 (P = 0.069) and 23 (P = 0.071) hours of cold exposure (Fig. 8D). There were no differences in IWAT temperatures between 6OHDA- and VC-treated groups housed at RT except at the 13th hour during the experiment, when IWAT temperature was higher in 6OHDA than VC hamsters (P < 0.05; Fig. 8E). Adipocytes in IWAT from 6OHDA-treated hamsters were small and had UCP1-positive, BAT-like multilocular lipid droplet morphology (Fig. 8F). 6OHDA cold-exposed hamsters had the greatest amount of UCP1 IR staining compared with other groups (Fig. 8F).
Finally, CGRP, a marker of sensory nerves, was increased in IWAT of 6OHDA-treated hamsters vs. VC-treated hamsters at RT (P < 0.05). A similar difference was observed in cold-exposed hamsters but did not reach statistical significance (P = 0.069; Fig. 8G).
DISCUSSION
The present study characterized central SNS innervations to IWAT and IBAT for the first time using dual tract tracing with isogenic strains of PRV containing either GFP or RFP within the same animal. Throughout the brain, there was a consistent pattern of greater SNS circuitry to IBAT than IWAT, but shared SNS innervation to IWAT and IBAT was less frequent in the forebrain and midbrain compared with the hindbrain. The RPa, ROb, and A5 region had notably high percentages of doubly labeled neurons, whereas the PnC and RtTg, midbrain nuclei had higher separate SNS circuitry to IWAT than IBAT. The shared and separate central SNS outflow to IWAT and IBAT may work together to maintain homeostasis by coordinating thermoregulation and beige adipocyte recruitment, as demonstrated by increased IWAT SNS drive, temperature, and UCP1 expression during 16- to 24-h cold exposure in hamsters with impaired IBAT function due to SNS denervation.
We performed an extensive characterization of IWAT and IBAT SNS circuitries across the neuroaxis and found differential SNS innervation patterns in the SNS ganglia, spinal cord, and brain. Singly labeled SNS neurons projecting to IBAT were mostly found in the thoracic region of the SNS ganglia, while singly labeled SNS neurons projecting to IWAT were mostly found in the lumbar region. These data are consistent with prior studies showing regional differences in the SNS ganglia innervating IWAT and EWAT (75) or IWAT, mesenteric WAT, and RWAT (1, 50). There were no differences in SNS innervation from the intermediolateral horn (IML) of the spinal cord to IBAT and IWAT, suggesting that differences in SNS drive to IBAT and IWAT originate from the SNS ganglia. In addition, the spinal cord had low levels of doubly labeled neurons, suggesting that postganglionic projections to peripheral tissues, such as IWAT and IBAT, are distinct. In this situation, differential SNS drive to IWAT and IBAT during the acute cold challenge may rely on the innervation from the lumbar and thoracic regions of the SNS ganglia, respectively, but not the spinal cord.
We found divergent and convergent central SNS outflow to IWAT and IBAT across the brain. The hindbrain had a greater percentage of singly and doubly labeled SNS neurons than the midbrain and forebrain due to projections from rostral nuclei and the presence of caudal nuclei essential for integrative control of thermoregulation and potentially of beigeing. The relative sufficiency of the hindbrain for thermoregulation has been demonstrated in chronic decerebrate rats in which the brain has been transected at the supracollicular level (49). These rats were able to maintain core body temperature when housed at 8°C and 10°C, and there were comparable increases in IBAT and WAT NETO in intact and decerebrate cold-exposed rats (49). Forebrain projections, however, are still necessary for complete and stable thermoregulation, because chronic decerebrate rats, unlike intact rats, could not maintain core body temperature at 4°C (49).
We found notable increases of doubly labeled neurons in the RPa, ROb, and A5 region. These nuclei regulate thermoregulatory responses and SNS outflow (2, 3, 45, 46). RPa neurons are necessary for adaptive thermoregulation during cold exposure including populations of thyrotropin-releasing hormone neurons, which have been shown to be activated during 24-h cold exposure (13, 15, 47, 52). The ROb contains serotonergic neurons that increase firing with a cold stimulus, but are inhibited by a warm one (66). The thermoregulatory role of the A5 region has not been fully tested as the raphe, but it contains noradrenergic neurons that have frequently been labeled in tract tracing studies from fat (2, 3, 44, 50, 58, 59, 63, 64). Singly labeled neurons to IWAT also contribute to IWAT thermogenic functions and beige adipocyte formation, and it has been shown that midbrain neurons in the reticular formation respond to cooling of the skin (48). We found significantly higher percentages of SNS neurons projecting to IWAT than IBAT in the PnC and RtTg midbrain nuclei; therefore, these sites may participate in both thermoregulation and beigeing of IWAT.
It is possible that some of our PRV tract tracing included SNS blood vessel innervation. The consideration of SNS input to fat vasculature is important, but it does not create interpretational issues of our virus tract tracing and denervation studies, because of the direct relation between increases in blood flow and tissue metabolic activity that includes glucose uptake into the brain and peripheral tissues such as BAT. It has been suggested that the function of SNS innervation of fat vasculature, particularly of WAT, is to increase capillary permeability so that liberated free fatty acids can leave the interstitial space and reduce their extracellular concentration to promote lipolysis by decreasing end product inhibition (24, 56, 57). This would enhance the function of WAT SNS as the primary driver of WAT lipolysis (5).
The present neuroanatomical evidence reveals the reality of a SNS cross talk between IWAT and IBAT for functional coordination of SNS outflow. This cross talk is most likely mediated by sensory nerves from IWAT not IBAT, because we found increased CGRP in IWAT depots of 6OHDA-treated hamsters (Figs. 7B and and8G).8G). Further support of IWAT and IBAT SNS cross talk mediated by IWAT sensory nerves is demonstrated by intra-IWAT injection of CL316,243 increases IBAT temperature, an effect that is lost when sensory nerves from IWAT are denervated (25). Thus sensory nerves from IWAT projecting to brain sites of SNS outflow (58) can separately increase SNS drive to only IWAT or IBAT or to both fat depots via the SNS ganglia at a mechanistic level. IWAT and IBAT SNS cross talk is not easily detected in intact hamsters due to functional IBAT depots. Intact hamsters cold exposed for 16–24 h do, however, have increased IWAT NETO (12) and BAT-like adipocyte morphology compared with RT controls some of which may be due to increased IWAT SNS drive. This cross talk in intact cold-exposed hamsters may be more apparent with a longer cold challenge.
The impetus for thermogenic function and beigeing in IWAT of IBAT SNS-denervated hamsters most likely occurs through increased SNS outflow, as demonstrated by enhanced SNS drive in IWAT of these hamsters. A change from long day to short day photoperiod in Siberian hamsters naturally increase SNS outflow triggering beigeing in RWAT, IWAT, and EWAT and increased UCP1 expression (Ryu V, Zarebidaki E, Albers E, Xue B, Bartness TJ, unpublished observations). In addition, knockdown of DMH neuropeptide Y neurons in rats is sufficient for IWAT beigeing, and when SNS nerves to IWAT were denervated, there was a significant decrease in IWAT NE content (a surrogate used for SNS drive measurement, but not a replacement for NETO) and UCP1 immunostaining compared with intact contralateral control (17). Finally, WAT of β-adrenoreceptor KO mice, which do not express necessary receptors for SNS stimulation, has decreased UCP1, BAT-like appearance, and reduced peroxisome proliferator-activated receptor-γ coactivator 1-α and CIDEA expression (4, 33, 37). These data suggest that SNS innervation and particularly increased SNS drive to WAT contribute to beigeing followed by a secondary increase in UCP1 expression and temperature (6). SNS triggering WAT UCP1 has been observed at the WAT depot level in intact mice given chronic cold exposure (10 days) (71). Those mice had increased multilocular adipocytes and UCP1 IR staining in WAT depots.
WAT SNS drive, as measured by NETO and UCP1 expression, was significantly increased in IWAT, but not in EWAT or RWAT, in 6OHDA-treated hamsters. This correlation between NETO and UCP1 expression further supports our hypothesis that increased SNS drive to IWAT is the driving force triggering beigeing and enhanced thermogenic function in IWAT. The presence of differential SNS drive among fat depots to a variety of energy challenges has been previously documented (11, 12, 50), and this selective response to both energy challenges and IBAT 6OHDA treatment may reflect differential neuronal circuitries to individual fat depots (50, 75). Beigeing of all WAT depots and greater increases in UCP1 mRNA and protein expression and morphological changes are typically observed at ~7 days of chronic cold exposure in intact mice (73). We chose 16- to 24-h cold exposure for our study, because it is likely that SNS- coordinated control of IWAT and IBAT would occur during the early, critical period of the cold challenge, but it is possible that increased SNS drive to and beigeing of EWAT and RWAT would occur in 6OHDA-treated hamsters exposed to cold for an extended period of time.
The metabolic relevance of beige adipocyte contribution to nonshivering thermogenesis has been the subject of debate. Mitochondria from IWAT of cold acclimated mice have comparable UCP1 expression to mitochondria found in IBAT, and they are functionally thermogenic (61). The thermogenic density of recruited beige adipocytes in IWAT, however, is only one-fifth that of IBAT (61). A recent study exposing intact, wild-type mice to 10°C for 14 days found that although cold exposure induced IWAT beigeing, it did not result in large increases in mitochondrial oxidative activity (35), suggesting that beige adipocytes did not make an appreciable contribution to thermoregulation. However, in our study, we found hamsters with impaired IBAT thermogenic function were able to maintain core body temperature during a 24-h cold challenge with increased IWAT thermogenesis and beigeing, which implied a contribution of IWAT to maintenance of core temperature. A limitation of our model is that IBAT was the only BAT depot that was denervated and we did not exclude the thermogenic contribution of remaining, intact BAT depots. Therefore, it is possible that intact mediastinal, axillary, perirenal, and cervical BAT contributed to the maintenance of core body temperature in the 6OHDA cold-exposed hamsters in this study. This may be the reason as to why UCP1 KO mice are cold intolerant, because all BAT depots do not express UCP1. A model of SNS impairment in all BAT depots is needed to further explore whether beige adipocyte recruitment is sufficient for long-term maintenance of thermoregulation when neural activation of BAT is absent.
Here we have defined the separate and shared SNS innervations to IWAT and IBAT across the neuroaxis. We have found high percentages of doubly labeled neurons in the hindbrain (RPa, ROb, and A5 region) and significantly greater separate SNS circuitry to IWAT that IBAT in the midbrain (PnC and RtTg). We have also physiologically tested the coordinated SNS control for thermoregulation and beige adipocyte recruitment. Hamsters that had impaired IBAT function (6OHDA SNS denervation) survived 16- to 24-h cold challenge and this was associated with increased SNS drive to IWAT resulting in signs of beigeing (i.e., increased IWAT temperature and UCP1 expression and immunostaining) and maintenance of body temperature.
Perspectives and Significance
Our study demonstrates a high order of SNS regulation and coordination for thermoregulation and beigeing via a functional cross talk between IWAT and IBAT SNS most likely regulated by brain sites involved in SNS outflow and mediated, at least in part, by IWAT sensory nerves. This coordination between IWAT and IBAT is evolutionarily plausible as the maintenance of core body temperature is crucial for survival in extreme temperatures. A distributed control of body temperature has the potential to allow compensation for an event that causes BAT function to be impaired. It is clear that fat depots are not created equal but have unique characteristics and functions. This has been demonstrated by differential SNS outflow to WAT and BAT depots during energetic challenges of cold exposure, food deprivation, and glucoprivation (12). The differential SNS drive could be attributed to cross talk between different fat depots and feedback from sensory nerves. At present understanding the mechanism and regulation of beigeing provides an alternative strategy for treating and reversing obesity since obese individuals have significantly decreased BAT but excess WAT depots (67). However, the physiological contribution of WAT beigeing for maintaining body temperature and its overall contribution to whole body energy expenditure are still largely unknown and require further investigation of WAT and BAT SNS control.
GRANTS
We thank Dr. Lynn Enquist for generous gift of PRV 152 (Princeton University, Princeton, NJ; Virus Center Grant P40RR-018604). This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R37-DK-035254 (to T. J. Bartness) and R01-DK-35254 (to B. Xue).
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.
AUTHOR CONTRIBUTIONS
N.L.T.N. and T.J.B. conception and design of research; N.L.T.N., C.L.B., V.R., and Q.C. performed experiments; N.L.T.N., C.L.B., V.R., Q.C., and B.X. analyzed data; N.L.T.N. and B.X. interpreted results of experiments; N.L.T.N. prepared figures; N.L.T.N., C.L.B., V.R., and B.X. drafted manuscript; N.L.T.N. and V.R. edited and revised manuscript; B.X. approved final version of manuscript.
ACKNOWLEDGMENTS
We thank Drs. J. Christopher Ehlen, Xin Cui, and Emily Bruggeman for technical assistance with the chromatograph, UCP1 fat IHC, and paraffin embedding, respectively. We also thank Drs. Ruth Harris, Aaron Roseberry, and Hang Shi for insightful comments on the manuscript. Finally, we thank Vaibhav Maheswari (an undergraduate research assistant) for help with tissue collection and Georgia State University Department of Animal Resources for husbandry care.