Leptin Receptor Signaling in Sim1-Expressing Neurons Regulates Body Temperature and Adaptive Thermogenesis

Abstract

Leptin signals to regulate food intake and energy expenditure under conditions of normative energy homeostasis. The central expression and function of leptin receptor B (LepRb) have been extensively studied during the past two decades; however, the mechanisms by which LepRb signaling dysregulation contributes to the pathophysiology of obesity remains unclear. The paraventricular nucleus of the hypothalamus (PVN) plays a crucial role in regulating energy balance as well as the neuroendocrine axes. The role of LepRb expression in the PVN in regard to the regulation of physiological function of leptin has been controversial. The single-minded homolog 1 gene (Sim1) is densely expressed in the PVN and in parts of the amygdala, making Sim1-Cre mice a useful model for examining molecular mechanisms regulating PVN function. In this study, we characterized the physiological role of LepRb in Sim1-expressing neurons using LepRb-floxed × Sim1-Cre mice. Sim1-specific LepRb-deficient mice were surprisingly hypophagic on regular chow but gained more weight upon exposure to a high-fat diet than did their control littermates. We show that Sim1-specific deletion of a single LepRb gene copy caused decreased surface and core body temperatures as well as decreased energy expenditure in ambient room temperatures in both female and male mice. Furthermore, cold-induced adaptive (nonshivering) thermogenesis is disrupted in homozygous knockout mice. A defective thermoregulatory response was associated with defective cold-induced upregulation of uncoupling protein 1 in brown adipose tissue and reduced serum T4. Our study provides novel functional evidence supporting LepRb signaling in Sim1 neurons in the regulation of body weight, core body temperature, and cold-induced adaptive thermogenesis.

The obesity epidemic has caused an ongoing public health crisis throughout the world. Nearly two thirds of the US population are estimated to be overweight or obese (1, 2). Obesity is associated with a variety of comorbidities, including diabetes, cardiovascular diseases, and osteoarthritis (3). Taken together, these factors reduce the lifespan and contribute to a lower quality of life. The underlying cause of obesity is multifactorial and involves gene–environment interaction. A variety of circulating factors play a fundamental role in regulating body weight. Perhaps the most well-studied relevant factor is leptin. Leptin is an adipose-secreted hormone that suppresses food intake and increases energy expenditure, contributing to optimal energy homeostasis. During obesity, despite elevated levels of leptin, no net negative energy balance is observed, suggesting that dysregulation of leptin receptor (LepRb) signaling could contribute to the pathogenesis of obesity (4–9). Thus, it is imperative to understand region-specific function of the LepRb signaling in the brain to elucidate the pathophysiology of obesity and provide targets for development of therapeutics.

Previous studies have sought to characterize the key central nervous system (CNS) sites responsible for various physiological roles of LepRb signaling. By employing site-specific deletion of LepRb using Cre-mediated recombination, these studies have investigated physiological characteristics of mice with LepRb deleted from various brain regions. For example, deletion of LepRb from proopiomelanocortin (POMC) neurons results in mild obesity, hyperleptinemia, and glucose intolerance with no significant difference in food intake, energy expenditure, or reproductive phenotypes (10–15). The LepRb signaling in the ventromedial nucleus of the hypothalamus (VMH) neurons plays a role in mediating weight-reducing effects of leptin by both suppressing food intake and increasing energy expenditure (16). Leptin action on LepRb-expressing neurons in the lateral hypothalamus area reduces body weight by suppressing food intake (17). Chemogenetic activation of LepRb-expressing neurons in the dorsomedial nucleus of the hypothalamus (DMH) neurons increases brown adipose tissue (BAT) thermogenesis, locomotion, and energy expenditure, and ablation of LepRb signaling from the DMH neurons results in dramatic weight gain by reducing energy expenditure and locomotor activity, without affecting food intake (18). Interestingly, LepRb signaling in the hindbrain also functions to suppress food intake and body weight with inconclusive effects on energy expenditure (19). Importantly, the role of LepRb signaling in paraventricular nucleus of the hypothalamus (PVN) neurons in mediating any of the known actions of leptin has been controversial (20). The PVN functions as the hypothalamic “hub” receiving inputs from other hypothalamic nuclei such as the ARC and DMH and sending signals to regulate food intake, energy expenditure, and autonomic functions through the brainstem as well as endocrine axes through the median eminence (21).

Leptin also regulates several neuroendocrine axes, such as the hypothalamic–pituitary–thyroid (HPT) axis to regulate energy homeostasis in response to the environmental challenges (22). Leptin-deficient ob/ob mice are unable to upregulate the HPT axis upon refeeding or by cold exposure, an effect that can be corrected by leptin replacement (23, 24). However, the role of LepRb in the PVN in mediating these effects has been controversial. For example, several ex vivo brain slice recordings and in situ hybridization studies have demonstrated the effect of leptin on firing rates and expression of LepRb on various subtypes of PVN neurons, including TRH-expressing neurons in rodents (25–29). However, anatomical evidence obtained from the LepRb-GFP mouse line does not support a dense expression of the receptor in the PVN (30). We have previously shown that fasting increases the firing activity of the PVN neurons and that this effect is reversed during ad libitum feeding or by IP injections of leptin, indicating the ability of PVN neurons to sense peripheral energy states. Furthermore, leptin can directly increase the firing activity of the TRH-expressing neurons in the PVN at physiological concentrations ex vivo (28, 31, 32). Acting on PVN as well as ARC neurons, leptin increases the activity of the HPT axis, resulting in increased plasma thyroid hormone levels to indirectly regulate energy homeostasis (22, 26, 29, 32). Energy expenditure is directly upregulated by leptin as well as by thyroid hormones in response to food intake or cold exposure.

In rodents, energy expenditure is comprised of early nonshivering and late shivering thermogenic pathways. The early nonshivering response is constitutively active in sub–thermo-neutral temperatures and can also be triggered rapidly (within minutes) by cold exposure. This would activate the sympathetic pathway through β₃ adrenergic receptors in mice to upregulate uncoupling protein 1 (UCP1 or thermogenin)–mediated heat production in the BAT. UCP1 allows influx of H⁺ into the mitochondrial matrix and thus uncouples oxidative phosphorylation while the energy from the H⁺ motive efflux dissipates as heat (33). The late shivering response to cold is triggered more slowly (minutes to hours) and involves conversion of ATP to ADP to generate kinetic energy and produce heat through muscle contraction (34). Increases in local muscle T3 increases sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase and myosin heavy chain synthesis, thereby promoting muscle contraction (23). Furthermore, leptin can regulate both the UCP1-dependent nonshivering and ATP-dependent shivering responses to cold by directly acting on LepRb in hypothalamic neurons and indirectly elevating the thyroid hormones. Leptin also directly affect thermoregulation by increasing O₂ consumption in phosphatidylinositol 3‐kinase–dependent and LepRb-mediated pathways in skeletal muscle, hence increasing the core body temperature (8). Even though the neuronal pathways conveying signals of central thermosensors regulating early nonshivering and late shivering responses have been fairly defined (35), the anatomical details of the LepRb-expressing neurons that can regulate thermogenesis to adapt to the environmental challenges are unclear.

In this study, we investigated the role of LepRb signaling in the Sim1 neurons on energy homeostasis and leptin-regulated neuroendocrine functions using a mouse model with LepRb deleted from neurons expressing the transcription factor single-minded homolog 1 gene (Sim1) (36, 37), which is selectively expressed in the PVN, the supraoptic nucleus (SON), the posterior hypothalamus, and parts of amygdala. This was achieved by crossing mice expressing the Cre recombinase gene under the Sim1 promoter (Sim1-Cre) with mice where the LepRb gene is flanked by loxP sites (LepRb^loxP mice). We then investigated energy homeostasis, thermoregulation, and the endocrine axis of the offspring with LepRb deleted from Sim1-expressing neurons in comparison with control groups in both sexes.

Materials and Methods

Mouse breeding and housing

A transgenic C57BL/6 mouse expressing Cre recombinase under the control of the Sim1 promoter [Tg(Sim1-cre)1Lowl/J, catalog no. 006395, The Jackson Laboratory] was crossed to a mouse strain containing leptin receptor flanked by two loxP sites in direct orientation (LepRb^flox/flox, B6.129P2-Lepr^tm1Rck/J, or ObR^flox, catalog no. 008327; The Jackson Laboratory, Bar Harbor, ME). Offspring containing a single Sim1-Cre allele and a LepRb^flox/+ allele were then crossed to obtain mice used in this study. The offspring of this cross included 25% wild-type (LepRb^+/+;Sim1-Cre), 50% heterozygous (LepRb^flox/+;Sim1-Cre), and 25% Sim1-specific LepRb knockout (KO) (LepRb^flox/flox;Sim1-Cre) mice. Two subsequent longitudinal cohorts of mice were generated. The study of body weight and normal chow and fat-enriched food intake was conducted in both cohorts. Seventy-two and 78 experimental mice were generated for the first the second cohorts, respectively. The findings observed in the first cohort were reproduced in the second cohort. These mice were also used for infrared (IR) thermal study as well as interscapular thermal probes. Twenty-six other mice were used to examine effects of leptin on PVN neuron electrical properties. Additionally, 29 and 28 other naive mice were used for the UCP1 BAT and the total T4 (TT4) studies, respectively. In all experiments, parameters examined were compared between siblings with various genotypes. At an ambient temperature of 22°C, two or three male or female mice were housed in a cage to prevent thermal stress.

In all cohorts, mice were maintained on a standard chow diet, unless indicated otherwise, and kept on a 12-hour light/12-hour dark cycle. Mice had free access to normal chow food during cold-exposure experiments. Mice were handled according to procedures previously approved by the Institutional Animal Care and Use Committee of Vanderbilt University Medical Center.

Validation of recombination of the LepRb gene and deletion of the LepRb protein in hypothalamic neurons

The deletion of LepRb was validated first by confirming the recombination of the LepRb gene in hypothalami of KO mice using a PCR assay from hypothalamic genomic DNA (38). Genomic DNA was isolated from TRIzol extracts of whole hypothalami according to the manufacturer’s instructions. Briefly, 100% ethanol was used to precipitate DNA pellet by centrifuging at 2000g at 4°C. The pellet was washed twice in 0.1 M sodium citrate (pH 8.5) in 10% ethanol. The DNA pellet was air dried, dissolved in 8 mM NaOH, and the pH was neutralized with 1 M Tris (pH 8.0). PCR reactions were run using forward primer, 5′-GTC ACC TAG GTT AAT GTA TTC-3′, and reverse primer, 5′-GCA ATT CAT ATC AAA ACG CC-3′, to amplify the recombined allele of ∼400 bp. The reaction proceeds with 94°C for 5 minutes, 40 cycles of 94°C for 1 minute, 61°C for 1 minute, and 68°C for 2 minutes, followed by 72°C for 5 minutes. Furthermore, deletion of the LepRb protein in PVN neurons was validated by examining their responses to applications of leptin in brain slices by electrophysiology (31, 32, 39).

Indirect calorimetry: measurements of energy expenditure, O₂ consumption, water intake, and locomotion

Mice were individually placed in metabolic cages (identical to home cages with bedding) in a 12-hour light/12-hour dark cycle, temperature/humidity-controlled dedicated room located in the Vanderbilt Mouse Metabolic Phenotype Core. Energy expenditure measures were obtained by indirect calorimetry (Promethion; Sable Systems, Las Vegas, NV). The calorimetry system consists of cages identical to home cages with bedding equipped with water bottles and food hoppers connected to load cells for food and water intake monitoring. All animals had ad libitum access to standard rodent chow and water. The air within the cages is sampled through microperforated stainless steel sampling tubes that ensure uniform cage air sampling. Promethion utilizes a pull-mode, negative pressure system with an excurrent flow rate set at 2000 mL/min. Water vapor is continuously measured and its dilution effect on O₂ and CO₂ are mathematically compensated for in the analysis stream (40). O₂ consumption and CO₂ production are measured for each mouse at 5-minute intervals for 30 seconds. Incurrent air reference values are determined after measuring every four cages. Energy expenditure is calculated using the Weir equation: kcal/h = 60[0.003941(Vo₂) + 0.001106(Vco2)], where Vo₂ is O₂ consumption and Vco₂ is CO₂ output (41). Ambulatory activity is determined simultaneously every second with the collection of the calorimetry data. Ambulatory activity and position are detected with XYZ beam arrays (BXYZ-R; Sable Systems, Las Vegas, NV) with a beam spacing of 1.0 cm interpolated to a centroid resolution of 0.25 cm. Consecutive adjacent IR beam breaks are counted with a minimum movement threshold set at 1 cm. Both allmeters and pedometers are conversions of the beam breaks into meters traveled. Allmeters represent movement that only breaks one single beam (repeatedly). This includes all types of movement, including back and forth in front of a single beam (such as what could happen when scratching), whereas pedmeters select for movements that are active, directional, and that cross more than one beam. Subtracting pedmeters from allmeters results in an index of fine movements such as scratching or grooming. Data acquisition and instrument control were coordinated by MetaScreen v2.2.18, and the raw data were processed using ExpeData v1.7.30 (Sable Systems).

Thermal imaging

For the IR imaging studies, mice were shaved on the dorsal side, excluding the head, to accurately measure the true heat emissivity of the surface of the skin. The hair coat would insulate the heat dispersed by the body and would thus interfere with accurate measurement of heat produced by the interscapular BAT (iBAT). To perform the IR imaging, the animals were transported from the housing facility to the examination area. Images were obtained from fully awake (unanesthetized) mice to avoid anesthetic interference in the iBAT activation as described by Zhang and Scarpace (4). Each mouse was removed from the cage, placed on the laboratory bench, and held by the tail to image the mouse five times using a Fluke thermal camera (model no. TiR125; Fluke Corporation, Everett, WA). The imaging studies started with the collection of the first set of baseline images following exposure to room temperature (25°C). After acquiring five baseline images, mice were placed back in the original cage and moved to the housing facility. An automated imaging-processing object recognition pipeline in MATLAB language utilizing mathematical morphology was developed and yielded a fast and robust approach useful for many applications. In our case, when applied to a large number of noncontact IR thermometry images (total of five images per mouse per condition) to determine surface body temperature in live unanesthetized mice, the method made the data analysis convenient, fast, and reliable.

Data processing

The original thermal raw images were stored in .IS2 format. To analyze the images, the files were converted to comma-separated (.CSV) files using Fluke thermal imaging software (Fluke SmartView^® 3.5). CSV image files were loaded using a custom MATLAB script. The script automatically selected each one of the images, detected the object edges, with the object being the mouse, and used mathematical morphology erosion and dilation operations iteratively to create a mask to discard the background temperatures not pertinent to the body of each mouse. For each image the average surface body temperatures within the animal were calculated (42).

Measurements of body temperature using subcutaneous probes

Mice had ad libitum access to normal chow as well as water throughout all studies that involved exposure to 4°C. Thermal telemetry was performed using an implantable programmable temperature transponder (IPTT-300; BioMedic Data Systems, Inc., Seaford, DE). Probes were implanted by briefly anesthetizing the mouse with isoflurane and removing fur to expose the skin in the areas of interest for implantation of temperature transponders. Individual transponders were injected in BAT by inserting a presterilized disposable needle subcutaneously between scapular bones. The incision was closed with sterile sutures to minimize the risk of infection. The animals were allowed to recover for 7 days before the cold challenge experiments. Animals were singly housed in the housing facility at least for a week before being transported to the examination area for measurement of iBAT temperatures. The iBAT temperature measurements were performed in unanesthetized, freely moving mice in their original cage by telemetry using a detachable mobile RSP-6004 data acquisition probe. The iBAT temperature was measured in triplicate within 30 seconds and averaged. The study started with collection of the first set of baseline measurements in ambient temperature (25°C). Mice in the original cage were moved to a temperature-controlled cold room that was maintained at 4°C. They remained in the cold room for 4 to 6 hours to induce BAT activation. The iBAT temperature was measured in triplicate within 30 seconds at 0, 1, 2, 4, and 6 hours.

BAT UCP1 measurements by Western blot analysis

At 8:00 am, mice fed with normal chow ad libitum were placed in 4°C. At 12:00 am, mice were euthanized and BAT was collected and frozen in −80°C. BAT samples were homogenized in radioimmunoprecipitation assay buffer and spun to remove fat and debris. The resulting supernatant was then used for subsequent analysis. A bicinchoninic acid assay was performed to measure protein contents of samples. Protein samples were run on 15% polyacrylamide (SDS-PAGE, Mini-PROTEAN^® TGX™ precast) gels (Bio-Rad Laboratories, Hercules, CA). Gel proteins were transferred to a nitrocellulose membrane for antibody bindings. For detection of UCP1 protein, the primary UCP1 antibody (Abcam, ab10983) was applied, followed by secondary horseradish peroxidase–conjugated rabbit antibody. Chemiluminescent substrate horseradish peroxidase was applied and then films were exposed for luminescence detection. For detection of glyceraldehyde-3-phosphate dehydrogenase (GAPDH), membranes were stripped with Restore™ Western blot stripping buffer (Thermo Fisher Scientific, Waltham, MA). Primary GAPDH Ab was applied and then secondary IR fluorescent (LI-COR)–conjugated rabbit was applied. The membranes were then scanned for intensity of bands reflecting the protein concentrations of UCP1 over GAPDH.

Total T4 measurements

Mice were exposed to 5 hours of 4°C before decapitation for blood withdrawal and measurements of serum total T4 levels using an ELISA kit as recommended by the manufacturer (AccuDiag T4 ELISA kit, catalog no. 3149-18; Diagnostic Automation/Cortez Diagnostics Inc., Woodland Hills, CA).

Statistical analysis

All statistical analyses were performed using the MATLAB 2013b statistics toolbox and verified with R Studio, R version 3.1.0 (2014-04-10) and Prizm software (version 6.05). Two-way ANOVA was used to compare repeated measurements among two or more groups of data sets. An unpaired t test was performed to compare two groups of data sets, unless otherwise indicated. All reported P values are two-tailed with a P value ≤0.05 indicating statistical significance.

Results

Body weight and food intake in LepRb^loxP/loxP;Sim1-Cre mice

To study a putative role of PVN LepRb neurons on energy homeostasis and adiposity, we generated a mouse line with LepRb deleted in the Sim1 neurons by crossing LepRb^loxP/loxP mice with Sim1-Cre mice (see “Materials and Methods” for mouse validation). We then investigated parameters of energy homeostasis in LepRb^loxP/loxP;Sim1-Cre male and female offspring by monitoring food intake under normal chow or a high-fat (HF) diet (HFD) and body weight up to 42 weeks of age. Male LepRb^loxP/loxP; Sim1-Cre mice (n = 15) displayed a trend to weigh less compared with LepRb^+/+;Sim1-Cre mice (n = 7) and LepRb^loxP/+;Sim1-Cre mice (n = 11) during 26 weeks of chow feeding (P > 0.05, two-way ANOVA; Fig. 1A). At week 27, mice were given ad libitum access to a 60% HFD as their only source of food. On the HFD, LepRb^loxP/loxP;Sim1-Cre mice gained weight more rapidly and became heavier than did their wild-type littermates beginning at 33 weeks of age (P < 0.05, weeks 33 to 41, two-way ANOVA). LepRb^loxP/loxP;Sim1-Cre mice gained more weight on the HFD (approximately twofold) compared with the control groups (1.6- to 1.7-fold, P < 0.005; one-way ANOVA; Fig. 1A and 1B; Table 1). In female mice, no significant difference in the body weight of LepRb^loxP/loxP;Sim1-Cre mice (n = 17) was observed when compared with LepRb^+/+;Sim1-Cre mice (n = 5) or LepRb^loxP/+;Sim1-Cre mice (n = 7) during 27 weeks of normal chow feeding (P > 0.05, two-way ANOVA; Fig. 1C). At week 27, female mice were placed on a 60% HFD. Under these conditions the LepRb^loxP/loxP;Sim1-Cre mice gained weight at a faster rate than did their controls. The LepRb^loxP/loxP;Sim1-Cre mice became significantly heavier than did their Cre controls as well as LepRb^loxP/+;Sim1-Cre mice at 40 weeks of age (P < 0.05, weeks 40 to 41, two-way ANOVA, Tukey multiple comparisons test). The female LepRb^loxP/loxP;Sim1-Cre mice gained significantly more weight on an HFD (∼1.9-fold) compared with the control groups (∼1.4-fold, P < 0.005, one-way ANOVA; Fig. 1C and 1D; Table 2).

Figure 1.

Body weight and food intake of a cohort of LepRb^loxP/loxP;Sim1-Cre mice and their control littermates. The growth curves of body weights of (A) male and (C) female LepRb^loxP/loxP;Sim1-Cre mice fed ad libitum with normal chow (NC) until 27 wk of age followed by an HFD (blue shade) are shown. (A) Graph of the body weight of male LepRb^loxP/loxP;Sim1-Cre mice (n = 15), LepRb^+/+;Sim1-Cre mice (n = 7), and LepRb^loxP/+;Sim1-Cre mice (n = 11) during NC and HF feeding (two-way ANOVA). (B) Bar graph of body weight gain of male mice during the last week of HF feeding compared with their body weight during the last week of NC diet (one-way ANOVA). (C) Graph of the body weight of female LepRb^loxP/loxP;Sim1-Cre mice (n = 17), LepRb^+/+;Sim1-Cre mice (n = 5), and LepRb^loxP/+;Sim1-Cre mice (n = 7) during NC and HF feeding (two-way ANOVA). (D) Bar graph of body weight gain of female mice during the last week of HF feeding compared with their body weight during the last week of an NC diet (one-way ANOVA). (E–H) Graphs of grams of food consumed by (E and F) male and (G and H) female mice on NC feeding (two-way ANOVA) followed by an HFD (blue shade, two-way ANOVA). (F and H) Bar graphs of grams of food consumed by (F) male and (H) female mice averaged for 1 wk during NC feeding and 1 wk during HF feeding (blue shade) at the steady-state. (I–N) Graphs of calories taken by (I and J) male and (L and M) female mice on NC feeding (two-way ANOVA) and an HF diet (blue shade). (J and M) Bar graphs indicate calories taken by (J) male and (M) female mice averaged for 1 wk during NC feeding and 1 wk during HF feeding (blue shade) at the steady-state (two-way ANOVA). (K and N) Bar graphs indicate the ratios of calories taken on HF over NC feeding obtained from (J) and (M) by (K) male and (N) female mice. Data are presented as mean ± SEM. *P < 0.05; ***P < 0.001; ****P < 0.0001. ns, not significant.

Table 1.

Body Weight of Male Mice on an HFD Presented From the Age of 31 to 41 wk, as Described in Fig. 1A and 1B

	LepRb+/+;Sim1-Cre Mice (n = 7) (g ± SEM)	LepRbloxP/+;Sim1-Cre Mice (n = 11) (g ± SEM)	LepRbloxP/loxP;Sim1-Cre Mice (n = 15) (g ± SEM)	P Value: c vs b and a (Two-Way ANOVA)
	a	b	c
Week 31	34.23 ± 0.68	33.95 ± 1.14	36.93 ± 0.62	ns
Week 32	34.9 ± 0.5491	34.75 ± 1.19	37.58 ± 0.67	ns
Week 33	35.51 ± 0.88	36.13 ± 1.26	39.08 ± 0.64	<0.05
Week 34	36.4 ± 0.83	36.74 ± 1.58	40.49 ± 0.75	<0.01
Week 35	37.22 ± 0.84	37.99 ± 1.65	41.5 ± 0.73	<0.01
Week 36	37.65 ± 1.01	38.96 ± 1.67	42.68 ± 0.78	<0.001
Week 37	39.31 ± 1.061	39.86 ± 1.88	43.79 ± 0.81	<0.001
Week 38	40.71 ± 1.11	41.81 ± 1.43	44.63 ± 0.75	<0.01
Week 39	40.89 ± 1.25	42.23 ± 1.38	45.43 ± 0.75	<0.01
Week 40	41.83 ± 1.27	43.04 ± 1.33	45.84 ± 0.71	<0.05
Week 41	41.74 ± 1.19	42.96 ± 1.35	46.0 ± 0.70	<0.001

The body weight of various genotypes of these mice diverge beginning from 33 wk of age.

Abbreviation: ns, not significant.

Table 2.

The Body Weight of Female Mice on an HFD Presented From the Age of 33 to 41 wk, as Described in Fig. 1C and 1D

	LepRb+/+;Sim1-Cre Mice (n = 5) (g ± SEM)	LepRbloxP/+;Sim1-Cre Mice (n = 7) (g ± SEM)	LepRbloxP/loxP;Sim1-Cre Mice (n = 17) (g ± SEM)	P Value: c vs b and a (Two-Way ANOVA)
	a	b	c
Week 33	29.28 ± 1.32	29.5 ± 0.99	30.8 ± 0.89	ns
Week 34	30.08 ± 1.49	30.0 ± 1.12	32.16 ± 0.96	ns
Week 35	30.03 ± 1.43	31.2 ± 0.96	32.68 ± 0.99	<0.05 (only a)
Week 36	30.78 ± 1.29	32.8 ± 1.07	33.85 ± 1.00	<0.05 (only a)
Week 37	31.36 ± 1.41	33.6 ± 1.14	35.37 ± 1.01	<0.01 (only a)
Week 38	32.06 ± 1.68	34.6 ± 0.90	36.72 ± 1.08	<0.01 (only a)
Week 39	32.28 ± 1.68	34.1 ± 0.85	37.04 ± 1.03	<0.05 (only a)
Week 40	32.74 ± 1.35	35.3 ± 0.61	38.16 ± 1.19	<0.05 (a and b)
Week 41	32.63 ± 1.51	35.3 ± 0.56	38.67 ± 1.15	<0.001 (a and b)

The body weight of various genotypes of these mice diverge beginning from 35 wk of age.

Abbreviation: ns, not significant.

We also measured food intake in 24-week-old male and female mice for up to 3 weeks while on normal chow feeding and for several weeks while under HFD conditions. LepRb^loxP/loxP;Sim1-Cre mice fed with normal chow ate significantly less food than did their control counterparts (P < 0.05 for males on days 1 to 7 and for females on days 1 to 17, two-way ANOVA; Fig. 1E–1H;Table 3 and 4). Upon switching to an HFD, the previously described transient HF-induced hyperphagia was observed (43). After this initial period, however, there were no significant differences in food intake of male or female LepRb^loxP/loxP;Sim1-Cre mice compared with their control groups (P > 0.05, two-way ANOVA; Fig. 1E–1H; Tables 3 and 5). We investigated food intake per calories taken to examine whether the source of diet has effects on calorie intake in LepRb^loxP/loxP;Sim1-Cre mice compared with control animals. When fed with normal chow, both male and female LepRb^loxP/loxP;Sim1-Cre mice consumed significantly less calories than did their control groups (P < 0.05, two-way ANOVA). This difference disappeared during HF feeding. Furthermore, the caloric intake from HF food over normal chow was significantly higher in both male and female LepRb^loxP/loxP; Sim1-Cre mice (P < 0.05), but not in control siblings (P > 0.05, two-way ANOVA; Fig. 1I–1N, Tables 4 and 6). Taken together, these results indicate a contribution of LepRb expressed in Sim1-Cre neurons in the regulation of steady-state energy homeostasis under standard and HF chow feeding.

Table 3.

Food Intake (Mass and Calories) of Normal Chow (During Week 26) and HF Food (During Week 29) in Male Mice From 26 to 29 wk of Age

		LepRb+/+;Sim1-Cre Mice (n = 3)	LepRbloxP/+;Sim1-Cre Mice (n = 5)	LepRbloxP/loxP;Sim1-Cre Mice (n = 7)	P Value: c vs b and a (One-Way ANOVA)
		a	b	c
Normal chow diet (days 1 to 7)	g	4.06 ± 0.057	3.99 ± 0.032	3.54 ± 0.043	<0.0001
	kcal	13.51 ± 0.23	13.26 ± 0.075	11.71 ± 0.028	<0.0001
HFD (days 20 to 27)	g	2.28 ± 0.06	2.47 ± 0.09	2.37 ± 0.06	ns
	kcal	12.33 ± 0.24	13.37 ± 0.47	12.95 ± 0.61	ns

Abbreviation: ns, not significant.

Table 4.

The Ratio of Caloric Intake From Normal Chow (During Week 26) Over HF Food (During Week 29) Consumed by Male Mice From 26 to 29 wk of Age

		LepRb+/+;Sim1-Cre Mice (n = 3)	LepRbloxP/+;Sim1-Cre Mice (n = 5)	LepRbloxP/loxP;Sim1-Cre Mice (n = 7)	P Value: c vs b and a (One-Way ANOVA)
		a	b	c
Normal chow/HF diet	kcal	1.01 ± 0.023	0.91 ± 0.019	1.113 ± 0.025	<0.05

Table 5.

Food Intake (Mass and Calories) of Normal Chow (During Weeks 25 and 26) and HF Food (During Weeks 31 to 33) in Female Mice From 25 to 33 wk of Age

		LepRb+/+;Sim1-Cre Mice (n = 3)	LepRbloxP/+;Sim1-Cre Mice (n = 3)	LepRbloxP/loxP;Sim1-Cre Mice (n = 7)	P Value: c vs b and a (One-Way ANOVA)
		a	b	c
Normal chow (days 1 to 17)	g	3.52 ± 0.055	3.76 ± 0.082	2.84 ± 0.040	<0.0005
	kcal	11.79 ± 0.183	12.61 ± 0.28	9.51 ± 0.131	<0.0005
HF food (days 43 to 60)	g	2.17 ± 0.072	2.17 ± 0.082	2.20 ± 0.032	ns
	kcal	11.32 ± 0.29	11.95 ± 0.57	12.04 ± 0.38	ns

Abbreviation: ns, not significant.

Table 6.

The Ratio of Caloric Intake From Normal Chow (During Weeks 25 to 27) Over HF Food (During Weeks 31 to 33) Consumed by Female Mice From 25 to 33 wk of Age

		LepRb+/+;Sim1-Cre Mice (n = 3)	LepRbloxP/+;Sim1-Cre Mice (n = 5)	LepRbloxP/loxP;Sim1-Cre Mice (n = 7)	P Value: c vs b and a (One-Way ANOVA)
		a	b	c
Normal chow/HF diet	kcal	0.94 ± 0.043	0.96 ± 0.055	1.28 ± 0.068	<0.05

LepRb Sim1-null mice have lower energy expenditure and water intake

Leptin receptor–deficient mice are hypothermic and have reduced energy expenditure. We thus examined whether mice lacking leptin receptor in Sim1 cells displayed lower energy expenditure. Using indirect calorimetry, we measured energy expenditure of all genotypes of 24-week-old male normal chow-fed mice by placing them in metabolic chambers (Promethion SABLE systems) at ambient temperature for 4 days. The energy expenditure averaged over two cycles was significantly lower in LepRb^loxP/loxP;Sim1-Cre mice (0.37 ± 0.01 kcal/h per mouse, n = 6) than LepRb^loxP/+;Sim1-Cre mice (0.45 ± 0.01 kcal/h per mouse, n = 7, P < 0.001) or LepRb^+/+;Sim1-Cre mice (0.44 ± 0.02 kcal/h per mouse, n = 4, P < 0.01), whereas there were no differences between LepRb^loxP/+;Sim1-Cre and LepRb^+/+;Sim1-Cre animals during the light cycle (P > 0.05). This parameter, however, was not significantly different among any of the genotypes examined during the dark cycle (0.52 ± 0.02 kcal/h per mouse in LepRb^+/+;Sim1-Cre mice, 0.53 ± 0.01 kcal/h per mouse in LepRb^loxP/+;Sim1-Cre mice, and 0.51 ± 0.003 kcal/h per mouse in LepRb^loxP/loxP;Sim1-Cre mice (P > 0.05, two-way ANOVA, Sidak multiple comparisons test; Fig. 2A and 2B). The lower energy expenditure in LepRb^loxP/loxP;Sim1-Cre mice was associated with lower Vo₂ in the light cycle (1.55 ± 0.06 in LepRb^+/+;Sim1-Cre mice, 1.51 ± 0.04 in LepRb^loxP/+;Sim1-Cre mice, and 1.38 ± 0.03 in LepRb^loxP/loxP;Sim1-Cre mice, P < 0.05), but not during the dark cycle (1.77 ± 0.05 in LepRb^+/+;Sim1-Cre mice, 1.76 ± 0.03 in LepRb^loxP/+;Sim1-Cre mice, to 1.70 ± 0.01 in LepRb^loxP/loxP; Sim1-Cre mice, P > 0.05, two-way ANOVA, Sidak multiple comparisons test; Fig. 2C). Because water intake and locomotion are affected by body temperature and food intake, we also examined whether water intake was also altered in these mice. We observed that LepRb^loxP/loxP; Sim1-Cre mice drank significantly less water compared with LepRb^loxP/+;Sim1-Cre or LepRb^+/+;Sim1-Cre littermates both in the light (1.66 ± 0.16 mL in LepRb^loxP/loxP;Sim1-Cre mice, 2.35 ± 0.15 mL in LepRb^loxP/+;Sim1-Cre mice, and 2.41 ± 0.19 mL in LepRb^+/+;Sim1-Cre mice) and dark cycles (2.1 ± 0.16 in LepRb^loxP/loxP;Sim1-Cre mice, 2.88 ± 0.18 mL in LepRb^loxP/+;Sim1-Cre mice, and 2.97 ± 0.22 mL in LepRb^+/+;Sim1-Cre mice, P < 0.05), whereas no significant difference in water intake was observed between LepRb^loxP/+;Sim1-Cre and LepRb^+/+;Sim1-Cre mice in any cycle (P > 0.05, two-way ANOVA, Sidak multiple comparisons test; Fig. 2D and 2E).

Figure 2.

Mice lacking the PVN leptin receptor displayed lower energy expenditure and water intake and higher locomotion at ambient room temperatures. (A) Graph represents energy expenditure of 24-wk-old male LepRb^loxP/loxP;Sim1-Cre mice (n = 6), LepRb^loxP/+;Sim1-Cre mice (n = 7), and LepRb^+/+;Sim1-Cre mice (n = 4) under normal chow feeding measured by indirect calorimetry (Promethion SABLE systems) at ambient room temperatures during 24 h. (B) Bar graph of energy expenditure averaged from two light and two dark (gray shade) cycles obtained from the experiment in (A) (two-way ANOVA). (C) Bar graph of Vo₂ of mice from the experiment in (A) during two light and two dark (gray shade) cycles (two-way ANOVA). (D and E) Graphs of (D) cumulative and (E) daily water intake in milliliters per cycle obtained from the experiment in (A) in both light and dark cycles (gray shade) (two-way ANOVA). (F–H) Bar graphs of locomotion indices in the (F) x-, (G) y-, and (H) z-axes of the mice obtained from the experiment in (A) in the light and dark (shaded boxes) cycles (two-way ANOVA, Sidak multiple comparisons test). Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001. ns, not significant.

Because shivering thermogenesis is associated with locomotion, we investigated whether locomotive activities are different among various genotypes. We thus examined locomotion indices, including x-, x-, and z-axes beam breaks, inside the Promethion cages. Results indicate that the total distance traveled by a mouse indicated by x- and y-axes beam breaks were not significantly different among any genotypes in either the light or dark cycle (P > 0.05, two-way ANOVA, Sidak multiple comparisons test). However, LepRb^loxP/loxP;Sim1-Cre mice displayed significant increases in rearing indicated by increased z-axis beam breaks in both the light (18,214 ± 967 breaks per cycle) and dark cycles (21,267 ± 1252 breaks per cycle) than did LepRb^loxP/+;Sim1-Cre or LepRb^+/+;Sim1-Cre mice in each corresponding cycle (10,847 ± 1412 in LepRb^+/+;Sim1-Cre mice or 12,626 ± 1496 in LepRb^loxP/+;Sim1-Cre mice in the light cycle, and 12,404 ± 1612 in LepRb^+/+;Sim1-Cre mice in the dark cycle, or 15,160 ± 1797 in LepRb^loxP/+;Sim1-Cre mice, P < 0.05). No significant difference in z-axis beam breaks was observed between LepRb^loxP/+;Sim1-Cre and LepRb^+/+;Sim1-Cre animals (P > 0.05, two-way ANOVA, Sidak multiple comparisons test; Fig. 2F–2H). Despite an increase in rearing, detailed analysis of locomotive activities indicate that fine movements such as grooming or scratching were not significantly different between the KO and control mice (P > 0.05; see “Materials and Methods”).

Deletion of LepRb in Sim1 neurons results in decreased surface body temperature in ambient temperatures

LepRb^loxP/loxP;Sim1-Cre mice displayed lower energy expenditure as measured by indirect calorimetry (see Fig. 2), and central leptin receptor–deficient mice are hypothermic and have defective thermoregulatory responses. We therefore investigated whether LepRb^loxP/loxP; Sim1-Cre mice are defective in maintaining basal body temperatures by measuring dorsal surface temperatures of 12-week-old male and female mice fed with normal chow by using an IR Fluke sensor at ambient room temperatures. An automated image-processing object recognition pipeline in MATLAB language utilizing mathematical morphology was developed. A large number of noncontact IR thermometry images were processed to determine surface body temperature in live nonanesthetized mice. The method made the data analysis convenient, fast, objective, and consistent. The skin body temperature of male LepRb^loxP/loxP;Sim1-Cre mice (32.76 ± 0.059°C, n = 15) was significantly lower than that of LepRb^loxP/+;Sim1-Cre mice (33.37 ± 0.078°C, n = 11, P < 0.001) and LepRb^+/+;Sim1-Cre animals (33.85 ± 0.073°C, n = 7, P < 0.0001). LepRb^loxP/+;Sim1-Cre mice also had lower body temperatures than did LepRb^+/+;Sim1-Cre animals in ambient temperatures (P < 0.01, one-way ANOVA, Tukey multiple comparisons test; Fig. 3A and 3B). Data obtained from female mice also indicate a degree of hypothermia in LepRb^loxP/loxP;Sim1-Cre animals: the skin temperature of LepRb^loxP/loxP;Sim1-Cre female mice (31.94 ± 0.10°C, n = 17) was significantly lower than that of LepRb^loxP/+;Sim1-Cre mice (32.34 ± 0.18°C, n = 7, P < 0.05) and LepRb^+/+;Sim1-Cre mice (32.75 ± 0.055°C, n = 5, P < 0.001). Furthermore, LepRb^loxP/+;Sim1-Cre animals displayed a reduced body temperature when compared with LepRb^+/+;Sim1-Cre mice (P < 0.05, one-way ANOVA, Tukey multiple comparisons test; Fig. 3C). Taken together, these data suggest a role for LepRb signaling in Sim1 neurons in the regulation of body temperature and indicate a gene-dosage effect for LepRb in thermogenesis.

Figure 3.

Deletion of LepRb in the PVN results in decreased surface body temperature in a gene dose–dependent manner in chow-fed male and female mice. (A) Representative dorsal skin temperature image obtained by an IR (Fluke) sensor from a 12-wk-old male mouse displays the heat produced from the animal at ambient room temperatures. (B) Bar graph of IR intensities obtained from thermal images of the dorsal skin of 12-wk-old normal chow-fed male LepRb^loxP/loxP;Sim1-Cre mice (n = 15), LepRb^loxP/+;Sim1-Cre mice (n = 11), and LepRb^+/+;Sim1-Cre mice (n = 7) at ambient room temperatures (one-way ANOVA). (C) Bar graph of IR intensities obtained from thermal images of dorsal skin of normal chow-fed female LepRb^loxP/loxP;Sim1-Cre mice (n = 17), LepRb^loxP/+;Sim1-Cre mice (n = 7), and LepRb^+/+;Sim1-Cre mice (n = 5) at ambient room temperatures (one-way ANOVA). Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Deletion of LepRb from Sim1 neurons results in decreased iBAT temperature and thermoregulatory response to cold

The results of the skin temperature measurements by the IR sensor and indirect calorimetry collectively indicate that mice with LepRb deleted from Sim1 neurons are hypothermic in ambient temperatures. We thus decided to investigate their iBAT temperatures by using interscapular thermal probes that measure and report the values via telemetry (see “Materials and Methods”). First, the iBAT temperatures were measured in various groups of 24-week-old male normal chow-fed mice in ambient temperatures. LepRb^loxP/loxP;Sim1-Cre mice (36.33 ± 0.128°C, n = 9) had significantly lower iBAT temperatures than did both LepRb^loxP/+;Sim1-Cre mice (36.86 ± 0.159°C, n = 8, P < 0.05) and LepRb^+/+;Sim1-Cre mice (37.57 ± 0.129°C, n = 7, P < 0.0001). LepRb^loxP/+;Sim1-Cre mice also had lower iBAT temperatures than did their LepRb^+/+;Sim1-Cre counterparts (P < 0.05, one-way ANOVA, Tukey multiple comparisons test; Fig. 4A and 4B). Mice deficient in leptin receptor signaling are unable to upregulate their body temperature when exposed to acute cold (4°C) due to defects in adaptive thermogenesis (24). We thus examined whether the leptin receptor in Sim1 neurons plays a role in adaptive thermogenesis. To address this, male mice were exposed to a cold challenge (4°C) and monitored for changes in iBAT temperature. After exposure to 4°C, the iBAT temperature of all groups dropped (Fig. 4). However, 4 hours after exposure to cold, the iBAT temperatures of LepR^loxP/loxP;Sim1-Cre mice (n = 9) dropped 1.6°C to 34.84 ± 0.197°C (P < 0.0001, two-way ANOVA), whereas the iBAT temperature of LepR^+/+;Sim1-Cre mice (n = 7) only dropped 1.3°C to 36.23 ± 0.276°C (P < 0.001, two-way ANOVA). Similarly, the iBAT temperature of LepR^loxP/+;Sim1-Cre mice (n = 8) also dropped after cold exposure to 36.18 ± 0.299°C (P < 0.05, two-way ANOVA, Sidak multiple comparisons test) and was not significantly lower than for the LepR^+/+;Sim1-Cre mice (P > 0.05, two-way ANOVA, Tukey comparison test). Furthermore, the iBAT temperatures of LepR^loxP/loxP;Sim1-Cre mice at 4 hours after exposure to cold were significantly different from the other groups (P < 0.001, two-way ANOVA, Tukey comparison test; Fig. 4A and 4B).

Figure 4.

Deletion of LepRb in the PVN results in hypothermia at ambient temperature and defective cold-induced thermogenesis in both male and female mice. The iBAT temperatures were interrogated using interscapular thermal probes in various groups of mice on normal chow feeding at ambient room temperatures followed by exposure to 4°C for 5 h. (A and B) Graphs of iBAT temperatures of male LepRb^loxP/loxP;Sim1-Cre mice (n = 9), LepRb^loxP/+;Sim1-Cre mice (n = 8), and LepRb^+/+;Sim1-Cre mice (n = 7) at ambient room temperatures (one-way ANOVA), followed by 5 h exposure to 4°C (blue shade, two-way ANOVA, Tukey comparison test). (C and D) Graphs of iBAT temperatures of female LepRb^loxP/loxP;Sim1-Cre mice (n = 8), LepR^+/+;Sim1-Cre mice (n = 5), and LepRb^loxP/+;Sim1-Cre mice (n = 7) at ambient room temperatures (one-way ANOVA), followed by 5 h exposure to 4°C (blue shade, two-way ANOVA). Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. ns, not significant.

We also examined the iBAT temperatures of groups of 24-week-old female normal chow-fed mice in ambient temperatures using interscapular thermal probes. The body temperature of LepR^+/+;Sim1-Cre mice (n = 5, 36.96 ± 0.12°C) and LepRb^loxP/+;Sim1-Cre mice (n = 7, 36.76 ± 0.16°C) were not significantly different (P > 0.05, one-way ANOVA). However, LepRb^loxP/loxP;Sim1-Cre mice (n = 8, 36.1 ± 0.156°C) had significantly lower body temperature than did both the Sim1-Cre (P < 0.05) and LepRb^loxP/+;Sim1-Cre mice (P < 0.01, one-way ANOVA, Tukey comparison test). After exposure to 4°C for 4 hours, the body temperature of the LepR^+/+;Sim1-Cre mice dropped about 0.9°C to 35.92 ± 0.1855°C, P < 0.01, two-way ANOVA). The body temperature of the LepRb^loxP/+;Sim1-Cre mice also dropped to 35.52 ± 0.2212°C (P < 0.0001, two-way ANOVA). The body temperature of the LepRb^loxP/loxP;Sim1-Cre mice dropped ∼1.7°C to 34.38 ± 0.22°C (P < 0.0001, two-way ANOVA, Sidak multiple comparisons test; Fig. 4C and 4D). When we compared their iBAT temperatures 4 hours after exposure to cold, the body temperature of LepRb^loxP/+;Sim1-Cre mice was not significantly lower than that of the Sim1-Cre mice (P > 0.05, two-way ANOVA). However, the body temperature of LepRb^loxP/loxP;Sim1-Cre mice was significantly lower than the Sim1-Cre (P < 0.0001) and the LepRb^loxP/+; Sim1-Cre mice (P < 0.001, two-way ANOVA, Tukey comparison test).

LepRb Sim1-null mice are defective in cold-induced upregulation of BAT UCP1

Because we observed a defect in adaptive thermogenesis in mice lacking LepRb in Sim1 neurons, we examined Ucp1 expression in BAT, a main regulator of nonshivering thermogenesis in response to 4-hour exposure to 4°C. At 8:00 am, 24-week-old normal chow-fed male mice were placed in 4°C with access to food. At 12:00 pm, mice were euthanized and BAT was collected for UCP1 protein quantification. Results indicate that 4-hour cold exposure was able to upregulate BAT UCP1 protein in LepRb^+/+;Sim1-Cre mice (1.19 ± 0.102, n = 7) and LepRb^loxP/+;Sim1-Cre mice (1.21 ± 0.121, n = 8, P > 0.05) to a similar extent. However, LepRb^loxP/loxP;Sim1-Cre mice failed to upregulate BAT Ucp1 in response cold (0.728 ± 0.0725, n = 14) compared with both LepRb^loxP/+; Sim1-Cre and LepRb^+/+;Sim1-Cre mice (P < 0.01, one-way ANOVA, Tukey multiple comparisons test; Fig. 5A and 5B).

Figure 5.

Mechanisms involved in hypothermia and defective adaptive thermogenesis in mice lacking LepRb in PVN neurons. (A) Western blot analysis of UCP1 and GAPDH expression in the BAT of chow-fed male mice after 4°C exposure for 4 h. The membranes were stripped of UCP1 antibody and reblotted with GAPDH antibody. (B) Bar graph of data in (A) indicate effects of cold exposure on BAT UCP1 in LepRb^loxP/loxP;Sim1-Cre mice (n = 14), LepRb^loxP/+;Sim1-Cre mice (n = 8), and LepRb^+/+;Sim1-Cre mice (n= 7, P < 0.001, one-way ANOVA). (C) Male chow-fed mice lacking LepRb in the PVN display lower serum total T4 after 4°C exposure. Bar graphs indicate that serum levels of TT4 were significantly lower in LepRb^loxP/loxP;Sim1-Cre mice (n = 14) than in LepRb^+/+;Sim1-Cre mice (n = 7) and LepRb^loxP/+;Sim1-Cre mice (n = 7, P < 0.05, one-way ANOVA). Data are presented as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001. ns, not significant.

Mice lacking LepRb in Sim1 neurons display lower serum total T4 after 4°C exposure

Leptin-deficient mice and humans display overt hypometabolism in part due to the dysregulated HPT axis (44). Furthermore, leptin-deficient mice display disrupted thermoregulatory responses and inability to upregulate thyroid hormone levels after exposure to cold (45–47). We thus hypothesized that the cold-induced upregulation of thyroid hormone levels are defective in mice lacking LepR in Sim1 neurons. Twenty-four-week-old normal chow-fed male mice were exposed to 5 hours of 4°C before blood withdrawal for serum thyroid level measurements. Results indicate that the serum TT4 levels were not different between the LepRb^loxP/+;Sim1-Cre mice (4.27 ± 0.25 μg/dL, n = 7) and LepRb^+/+;Sim1-Cre mice (4.72 ± 0.42 μg/dL, n = 7, P > 0.05). Serum levels of TT4 was significantly lower in LepRb^loxP/loxP;Sim1-Cre mice (3.57 ± 0.19 μg/dL, n = 14) than in LepRb^+/+;Sim1-Cre mice (P < 0.01) and LepRb^loxP/+;Sim1-Cre mice (P < 0.05, one-way ANOVA, Tukey multiple comparisons test; Fig. 5C). These data suggest a role of LepRb in the PVN in regulation of serum thyroid hormone levels (26, 45–48).

Discussion

Our data show that mice with LepRb deleted from Sim1 neurons display near normal growth rate under normal chow and mild obesity when fed with an HFD. Furthermore, deletion of the LepRb gene was associated with decreased body temperature and energy expenditure in ambient temperatures in a gene dose–dependent manner. The obesity accompanied by lower energy expenditure is indicative of improved feeding efficiency. Sim1-specific LepRb-deficient mice also showed defective adaptive thermoregulatory response when exposed to cold temperatures associated with the inability to upregulate UCP1 in the BAT and TT4 in serum. These data shed light on the role of PVN LepRb function in the regulation of energy homeostasis and adaptive thermoregulatory neuroendocrine responses.

Previous studies have sought to divulge the key CNS sites responsible for various physiological roles of LepRb signaling by employing site-specific deletion of LepRb using Cre-mediated recombination. These mutations resulted in mild obesity associated with hyperphagia and/or suppression of energy expenditure, if any, and collectively suggest that the central sensors of leptin responsible for conveying leptin receptor signaling are distributed throughout CNS centers. In this study we present findings describing the functional role of LepRb expressed in the Sim1 neurons in the regulation of energy homeostasis using cell type–specific KO of LepRb gene from Sim1-expressing neurons by Cre recombinase in previously characterized transgenic mice (36, 37). This study demonstrates that LepRb Sim1 neurons play an important role in the regulation of energy expenditure and thermogenesis. Mice with LepRb deleted from Sim1-expressing neurons display near normal growth rate when fed with normal chow. Despite normal body weight, Sim1-LepRb–deficient mice consumed slightly less normal chow compared with the control groups. This apparently counterintuitive finding could be explained by the fact that the KO mice also had lower energy expenditure, because at homeostasis, food intake can be adjusted in response to energy expenditure (49, 50). Alternatively, because leptin decreases the firing rate of non-TRH PVN neurons (31), deletion of LepRb likely results in increases in activity of these anorexigenic neurons, resulting in the hypophagia observed in the KO mice. When we challenged their homeostatic stability by placing them on a 60% HFD, both male and female KO mice gained weight rapidly and became heavier at a greater rate than did their control groups. Interestingly, despite their increased body weight under HF feeding, the KO mice ate similar calories of food compared with control counterparts. The increased body weight and/or adiposity in KO mice under HF feeding could be due to their relative lower energy expenditure. Alternatively, these findings may support the hypothesis that various LepRb-expressing neuron populations in Sim1 neurons regulate food intake or energy expenditure, as reported for melanocortin 4 receptor pathways (36). Additionally, the obesity under HF feeding observed in KO mice indicates an increased feed efficiency. In support of this notion, the ratio of calorie intake from an HFD to normal chow food was greater in KO mice than in Cre control or heterozygous mice of either sex (Fig. 1K and 1N).

Role of LepRb signaling in Sim1 neurons contributes to core body temperature and adaptive thermogenic responses

Among a wide array of phenotypes, leptin-deficient obese (ob/ob) and leptin receptor mutant (db/db) mice are hypothermic at ambient temperatures and are unable to initiate adaptive thermoregulatory responses (nonshivering thermogenesis) when exposed to 4°C (51, 52). The neuronal pathways transmitting signals of temperature-sensitive neurons (central thermosensors) to the brainstem to regulate thermogenesis have been fairly defined (35). For example, warm-sensitive γ-aminobutyric acid–ergic neurons in preoptic anterior hypothalamus inhibit the cold-sensitive neurons in the DMH by releasing γ-aminobutyric acid to suppress BAT-induced thermogenesis (53). The cold-sensitive DMH neurons activate the Raphe nuclei in the brainstem and intermediolateral nucleus (IML) to activate thermogenesis in BAT through activation of β₃ adrenergic receptor (54). This would trigger thermogenesis by upregulating UCP1 expression as well as O₂ consumption in the BAT (33). The VMH also relays the peripheral inputs to the BAT to regulate thermogenesis, as electrical stimulation of the VMH causes an increase in iBAT thermogenesis (55). However, how leptin signaling modulates these pathways to regulate thermogenesis has been less clear.

In this study, we show that deletion of LepRb from Sim1 neurons in both male and female mice results in hypothermia defined by lower core body temperature in ambient room temperatures as well as decreased energy expenditure during the light cycle. Most notably, lines of evidence presented in this study indicate gene-dosage effects of LepRb on the regulation of body temperature. Dorsal skin IR images indicate that the BAT temperature dose-dependently deceases by deletion of each allele of the LepR gene in Sim1 neurons in both male and female mice. Furthermore, data obtained from interscapular iBAT temperatures by using temperature probes support the gene-dosage role of PVN LepRb in regulation of basal body temperature. These findings are in agreement with the previously described correlation between CNS allelic expression of LepRb signaling and the obesity phenotype in mice (56). Additionally, our study demonstrates that deletion of LepRb signaling from Sim1 neurons mediates the adaptive thermoregulatory response to exposure to 4°C temperatures. We show this effect by directly interrogating interscapular BAT temperatures using temperature probes via telemetry. Furthermore, a defective adaptive thermoregulatory response to cold in KO mice was associated with the inability to upregulate BAT UCP1 expression compared with their controls. Despite the gene-dosage effect of LepRb deletion on body temperatures, no difference in BAT UCP1 expression was observed in heterozygous compared with Cre control mice.

Role of LepRb signaling in PVN TRH in regulation of the HPT axis: cold-induced elevation in thyroid hormone

In normal rodents, TSH levels significantly increase within 10 minutes after exposure to 4°C cold. The serum levels of TSH remain significantly high up to 24 hours. Serum T3 and T4 levels begin rising accordingly within minutes and remain high up to 2 days. The increases in levels of thyroid hormones are associated with the adaptive response through upregulating body temperature (45–47). Owing to a lack of leptin regulation of the HPT axis, leptin-deficient obese (ob/ob) mice have lower body temperature during both light and dark cycles and also have low thyroid hormone levels. This hypothermia can be corrected by injections of thyroid hormones (44). Furthermore, leptin injections can raise both thyroid hormones and body temperature through BAT UCP1 activity (23, 57). In the present study, we show that mice with homozygous deletion of LepRb signaling from Sim1-expressing neurons fail to upregulate serum total T4 and UCP1 upon cold exposure when compared with control mice.

In this study we were unable to detect a significant difference in BAT UCP1 or serum total T4 levels between heterozygous and Cre control mice. This observation is consistent with the notion that LepRb signaling in PVN neurons regulates TRH release (22, 58, 59) and is involved in fasting-induced suppression of thyroid hormones levels (57). These findings provide evidence supporting involvement of LepRb signaling in Sim1-expressing neurons in regulation of the HPT axis.

Mechanisms underlying the hypothermia observed in KO mice

In addition to the basal metabolic rate, the nonshivering BAT-mediated adaptive thermogenesis as well as the exercise muscle-induced thermogenesis contribute to maintaining the core body temperatures and total thermogenesis measured by O₂ consumption in wild-type mice. Which of these thermogenic components are defective to results in hypothermia observed in PVN LepRb-deleted mice is unclear. Because the BAT-mediated adaptive thermoregulatory response to cold is activated in ambient temperatures below thermoneutrality (∼30°C for mice), we conclude that a defect in this component contributes to the hypothermia observed in KO mice. However, it is unclear whether defective exercise muscle-induced thermogenesis also contributes to hypothermia observed in KO mice.

In addition to the nonshivering BAT-mediated thermogenesis, the shivering muscle-mediated thermogenesis is perturbed in leptin-deficient mice (24), and the shivering thermogenic response to cold or leptin is directly correlated with locomotion (60). Therefore, we examined the locomotive activity in LepRb/Sim1-deleted mice. Analysis of our results based on x- and y-axis beam breaks reveals that the distance traveled by mice is not different between KO and control groups. Furthermore, fine movements such as grooming and scratching obtained by subtraction of allmeters from pedometer indices were also not different. However, the z-axis beam breaks indicating vertical movements were significantly higher in KO mice than in their control animals. We interpret these results as that first the shivering thermogenesis is not compromised as a result of LepRb deletion in Sim1 neurons. Second, the increased rearing could indicate greater exploration and attempts to access the feeder or entries into the habitat. Because no evidence supporting increased exploration or food intake was observed in the KO mice, we postulate that the KO mice made greater attempts to enter into the habitat to avoid heat loss to compensate for lower core body temperatures and basal metabolic rate (18). Furthermore, lack of changes in pedestrian locomotion in Sim1 LepR deletion is in contrast to POMC LepRb deletion that results in significant decreases in locomotive activity and energy expenditure associated with hyperglycemia and obesity in mice (15). Thus, we postulate that defective BAT-mediated thermogenesis per se contributes to hypothermia observed in the Sim1/LepRb-deleted mice in this study (61).

Mechanisms underlying decreased water intake in KO mice

In addition to lower energy expenditure and O₂ consumption, we observed that mice with homozygous deletion of LepRb from Sim1 neurons displayed significantly lower water intake than did heterozygous or Cre control mice measured under normal ambient temperatures. In this study we analyzed the notion that the reduced water intake in KO mice could be due to reduced core body temperatures and/or food intake, because the interconnections between food and water intake are supported by anatomical evidence (62, 63).

Sim1 is expressed in SON neurons as well as in magnocellular PVN neurons. These neurons, which also express LepRb, synthesize and release arginine vasopressin (AVP) to regulate water excretion (64). Interestingly, AVP and TRH mRNA levels are all decreased by fasting and restored to fed levels with leptin treatment (65). As its levels increase after feeding, leptin stimulates AVP release to suppress water excretion (66). Thus, deletion of LepR from AVP-secreting neurons in nondiabetic mice is expected to result in decreased AVP levels and increased water excretion and hence compensatory water intake. In this study we observed that LepRb-deleted mice have lower water intake than did their control groups. Thus, deletion of LepRb from Sim1-expressing AVP-releasing neurons is unlikely to be involved in mediating decreased water intake, and we thus surmise that lower body temperature and lower food intake may contribute to the lower water intake observed in KO mice.

Subtypes of Sim1 neurons underlying hypothermia and HPT dysregulation

In addition to PVN, Sim1 is expressed in the supraoptic nucleus, the posterior hypothalamic nuclei, the mediobasal amygdala, and the nucleus of the lateral olfactory tract. It is also sparsely expressed in the lateral hypothalamus, the medial preoptic nucleus, the bed nucleus of the stria terminalis, and the periaqueductal gray neurons (36). Thus, involvement of other hypothalamic nuclei in phenotypes observed in LepRb-Sim1-Cre KO mice cannot be ruled out.

Leptin receptor signaling regulates the activity of the HPT axis through the “direct pathway” on TRH-expressing neurons in the PVN and the “indirect pathway” in POMC and AgRP neurons of the arcuate nucleus [e.g., Nillni et al. (22) and Ghamari-Langroudi et al. (32)]. Because the expression of Sim1 in the ARC is negligible, we thus surmise that deletion of LepRb in the TRH-expressing PVN neurons is responsible for disrupting the HPT observed in KO mice. TRH-expressing neurons are also present in the medullary raphae nucleus. Their downstream neurons in the hindbrain express LepRb, and their activation promotes thermogenesis through sympathetic output and BAT (67, 68). However, Sim1 expression is negligible in the hindbrain. Thus, a role of hindbrain TRH/LepRb neurons in generating phenotypes observed in KO mice is unlikely.

Thermoregulation can also occur independent of the HPT axis through the sympathetic nervous system and BAT activation. Neurons in the magnocellular preoptic area, lateral and posterior hypothalamus, and SON coexpress LepRb and Sim1 and are shown to mediate thermoregulation in response to changes in ambient temperatures. LepRb in any of these nuclei can mediate thermoregulatory responses and BAT activation observed in phenotypes in KO mice. Comprehensive mapping and gene expression profiles of neurons controlling thermogenesis will require further research.

Therapeutic application of targeting thermoregulation

The presence and function of BAT in adult humans has been elucidated in recent years. For example, 1-month-long exposure to overnight mild (18 to 19°C) cold increases UCP1 expression, energy expenditure, and “browning” of white adipose tissue (beige adipocytes), hence improving glucose homeostasis in humans (69, 70). There has been intense research to harness the underlying signaling pathways in the CNS to control energy expenditure to develop potential drugs to alleviate morbidity in patients with obesity (71, 72). Thus, activation of central thermogenic circuits to increase the UCP1-induced thermogenesis may provide a therapeutic approach to treat obesity (73). In this context, our data suggest that LepRb signaling expressed in this pathway could be potentially targeted by positive allosteric modulators to increase signaling in the presence of endogenous agonists (74).

Glossary

Abbreviations:

AVP

arginine vasopressin

BAT

brown adipose tissue

CNS

central nervous system

DMH

dorsomedial nucleus of the hypothalamus

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

HF

high-fat

HFD

high-fat diet

HPT

hypothalamic–pituitary–thyroid

iBAT

interscapular brown adipose tissue

IR

infrared

KO

knockout

LepRb

leptin receptor B

POMC

proopiomelanocortin

PVN

paraventricular nucleus of the hypothalamus

Sim1

single-minded homolog 1

SON

supraoptic nucleus

TT4

total T4

UCP1

uncoupling protein 1

VMH

ventromedial nucleus of the hypothalamus

Vo₂

O₂ consumption