GCGR agonism requires GABAergic signaling in the medial basal hypothalamus to promote weight loss in obese mice

Abstract

Glucagon receptor (GCGR)-mediated thermogenesis is a key component for the next-generation of obesity therapeutics. Herein, we investigated the central and peripheral mechanism by which activation of the GCGR augments metabolic rate to promote weight loss. Chronic treatment of obese mice with a long-acting GCGR agonist (LAGCGRA) reduced body weight and fat mass at both room temperature and thermoneutrality. Metabolic cage studies highlight that whilst GCGR agonism induces a negative energy balance via effects on both sides of energy balance, weight loss is primarily due to augmented metabolic rate in obese mice. Mechanistically, we report for the first time that GCGR agonism recruits GABAergic signaling in the medial basal hypothalamus to promote uncoupling protein 1(UCP1)-dependent thermogenesis in adipose tissue, stimulate caloric expenditure, and drive a negative energy balance in obese mice. Our preclinical findings provide insight in to how multi-receptor agonists engaging the GCGR may function to improve the weight loss efficacy of anorectic agents. Collectively, our results point to a liver→brain→fat axis activated by GCGR agonism for weight loss in obesity. Future studies are required to validate our findings in the clinic.

Graphical abstract

Highlights

• •Chronic GCGR agonism requires beta-adrenergic signaling to fully reduce body weight and induce brown adipose tissue thermogenesis.
• •Reduced adiposity by GCGR agonism requires GABAergic, but not glutamatergic, signaling by the medial basal hypothalamus.
• •For the chronic induction of body weight loss by GCGR agonism, UCP1 is required.
• •Thus, there are mechanisms independent from futile cycling that chronic GCGR agonism operates through to reduce adiposity.

1. Introduction

Treating excess adiposity costs a staggering $1.96 trillion USD annually, with 38% of the world population classified clinically as being overweight or obese [1]. Obesity is a major risk factor for many chronic diseases (e.g. cancer, cardiovascular disease, and type 2 diabetes (T2D)) [2,3]. The good news is that recent discoveries have transformed the treatment paradigm of patients with obesity and T2D (i.e. incretin-based therapeutics) by delivering clinically relevant weight loss. However, current obesity medications are unable to increase energy expenditure, a shortfall which the next generation of incretin-based strategies will address [[4], [5], [6]]. Currently, in clinical trials, novel therapeutics combine glucagon receptor (GCGR) engagement with incretin receptor modulation to simultaneously activate energy expenditure and suppress appetite for durable weight loss [[6], [7], [8], [9]].

While modulation of the GCGR has been classically connected with the liberation of glucose during such events as hypoglycemia and fasting [10], pharmacologic studies identified the ability of GCGR activation to increase energy expenditure and reduce food intake, which together induces rapid weight loss [11,12], effects that have been shown to translate to humans [13,14]. However, the tissues and mechanism(s) by which GCGR agonism acts to promote a negative energy balance are still under investigation [15]. Indeed, the GCGR is expressed by several metabolically relevant organs including the liver where activation increases gluconeogenesis, glycogenolysis, lipid oxidation, futile cycling and FGF21 production [16,17], white and brown adipose tissue, and the central nervous system [18]. However, brain GCGR action has been controversial, due to inconsistent findings [19,20]. In addition to effects in the liver [21], the GCGR is suggested to induce hepatokines that target the brain and adipose tissue to increase metabolic rate [[22], [23], [24]]. Mechanistic studies have revealed that acute modulation of the GCGR may not require adipose tissue thermogenesis for the induction of energy expenditure [16]. However, whether chronic GCGR activation requires classic neuronal circuitry and thermogenic pathways in the periphery to stimulate energy wasting remains to be examined. The objective of the current investigation was to better understand the potential mechanism(s) engaged by GCGR agonism to augment whole-body metabolism and stimulate weight loss. Our report is the first to demonstrate that hypothalamic inhibitory circuits are required for GCGR activation to stimulate energy wasting in adipose tissue and increase whole-body metabolic rate and protect from obesity. This signal is at the nexus of a liver→brain→ fat axis that is activated by GCGR agonism on a chronic timescale for weight loss in obesity.

2. Results

2.1. Chronic GCGR activation promotes weight loss via the augmentation of whole-body metabolic rate in obese mice

Chronic treatment of obese mice with GCGR agonists has been shown to promote weight loss via the suppression of appetite and stimulation of whole-body energy expenditure [25]. Indeed, we found that administration of a LAGCGRA dose-dependently reduced body weight in obese mice via both a drop in caloric intake and augmented caloric expenditure in animals housed at thermoneutrality (Figure 1A,B,E,F). Previous findings suggest that GCGR agonism stimulates whole-body metabolic rate via the liver, adipose tissue, and/or the brain [22,[26], [27], [28], [29]]. While glucagon can also act at the GLP-1 receptor and exert effects on beta cell function [30,31] the LAGCGRA was developed to maximize affinity for GCGR by reducing potency at GLP-1R [27]. While GCGR knockout mice in liver show a diminished response, we still wanted to verify that effects due to the LAGCGRA were mediated by the GCGR in the liver. Thus, we treated diet-induced obese male mice with a liver-specific GCGR siRNA to remove the GCGR and test the efficacy of the LAGCGRA. To knockdown the GCGR specifically in the liver, we used an siRNA conjugated to N-acetylgalactosamine (GalNAc), a ligand that binds to asialoglycoprotein receptor (ASGPR) specifically in the hepatocyte [32] (Supplemental Fig. 5). In line with previously published reports GCGR siRNA lowered body weight in vehicle treated mice [28,33,34]. As anticipated LAGCGRA reduced bodyweight in control mice. However, the weight loss efficacy of the LAGCGRA was ablated in animals treated with the liver-specific GCGR siRNA, confirming the requirement of hepatic GCGR engagement for the weight loss efficacy of GCGR activation (Figure 1C,D).

Figure 1.

LAGCGRA Stimulates Weight Loss and Progressively Increases Energy-Expenditure. Diet induced obese mice housed at thermoneutrality were treated with vehicle or LAGCGRA at 3,10, and 30 nmol/kg for 7 days, approximately 4 h into the light cycle during which we measured body weight (A) and food intake (B). Diet induced obese mice housed at thermoneutrality were treated with vehicle, GCGR siRNA (10 mg/kg), LAGCGRA (10 nmol/kg), or LAGCGRA + GCGR siRNA for 7 days, approximately 4 h into the light cycle during which we measured body weight (C) and food intake (D). Diet induced obese mice housed at thermoneutrality were treated with vehicle or LAGCGRA (10 nmol/kg) for 7 days, approximately 4 h into the light cycle during which we used metabolic chambers to measure energy expenditure (E) and food intake (F). n = 6 male mice in each group. Data expressed as mean ± SEM and analyzed by two-way ANOVA (A–F). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Together, our findings clearly indicated that GCGR activation is a highly effective thermogenic agent that induces dramatic effects on body weight via the GCGR in the liver. However, it is still not well understood how activation of the GCGR augments thermogenesis for anti-obesity results. Thus, next we conducted a series of studies to determine how GCGR agonism elevates whole-body metabolism.

2.2. UCP1 contributes to the thermogenic effect of GCGR activation

While the whole-body metabolic effects of GCGR agonism have been reported, whether adipose tissue thermogenesis contributes to the whole-body effect has not been established. Thus, we determined whether the LAGCGRA activates adipose tissue thermogenesis to increase whole-body metabolic rate by monitoring adipose tissue temperature in animals housed at thermoneutrality. Strikingly, in accordance with the elevated whole-body metabolism, we found that the LAGCGRA time-dependently increased white and brown adipose tissue temperature (Figure 2B,C). Further, in support of UCP1-dependent adipose tissue thermogenesis, we found that treatment of obese animals with the LAGCGRA stimulated the mRNA expression and protein abundance of the key thermogenic protein, uncoupling protein 1 (UCP1) in adipose tissue following both acute (∼ the onset of the increased whole-body thermogenesis) and chronic treatment (Figure 2D–E).

Figure 2.

LAGCGRA Reduces BW and Increases Energy Expenditure in Both a UCP1 Dependent and Independent Manner. We treated C57BL/6J mice that were housed at thermoneutrality with LAGCGRA (10 nmol/kg) approximately 4 h into the light cycle (A). During LAGCGRA administration, iBAT & iWAT temperature was recorded (B,C). Following 3 and 7 days of LAGCGRA treatment, the iBAT was removed and UCP1 protein levels were measured with immunoblotting and quantified (D,E). UCP1KO mice housed at thermoneutrality were treated with LAGCGRA (10 nmol/kg) for 14 days (F) during which bodyweight (G–I) and energy expenditure (J–M) were recorded. n = 6–8 per group. Data expressed as mean ± SEM and analyzed by 2-way Anova (B,C,G,H,K,M), on-way Anova (I), unpaired two-tailed t-test (D,E). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Therefore, to determine whether UCP1 contributes to the observed effects the GCGR agonist plays on metabolic rate and body weight, we utilized obese UCP1KO mice housed at thermoneutrality (Figure 2F, Supplement Fig. 1B). Following 14-days of treatment, the LAGCGRA reduced body weight in both WT and UCP1KO, when compared to vehicle treated animals (Figure 2G–I). However, after the seventh day of treatment, weight loss in the UCP1KO mice plateaued (Fig. 2H), whilst in WT mice treated with the LAGCGRA it continued to promote body weight loss, leading to a significant difference in weight loss between WT and UCP1KO animals follow 14-days of treatment (Fig. 2I). Furthermore, energy expenditure was significantly increased in the control treated mice starting on the first day of treatment, an effect that was enhanced with subsequent days of treatment (Figure 2J,K). However, the UCP1KO mice failed to show an increase in whole body energy expenditure in response to the LAGCGRA (Figure 2L,M). Thus, these findings suggest that the LAGCGRA augments whole-body thermogenesis to promote weight loss via UCP1-dependent thermogenesis in obese mice housed at thermoneutrality (Fig. 2). Thus, despite confirming that the anti-obesity efficacy of GCGR activation starts in the liver, a key novel finding of our studies is that UCP1-dependent adipose tissue thermogenesis is required for GCGR activation to increase metabolic rate. However, a key question that still needed to be examined was whether GCGR agonism stimulates UCP1-depentment thermogenesis via a direct effect in adipose tissue and/or a central mode of action.

2.3. Classic beta-adrenergic receptors are required for GCGR agonism to induce adipose tissue thermogenesis

Since a key driver of adipose tissue thermogenesis is adrenergic signaling [35,36], we determined whether the adrenergic system played a role in the effects of GCGR activation on energy balance. Here, we utilized Adrb1, 2, 3 KO (BARKO) mice which lack the beta-1, 2, 3-adrenergic receptors [37]. Due to the mixed background of the BARKO mice, age and weight-matched mice were used as positive controls. Consistent with our previous findings (Fig. 2G), the LAGCGRA stimulated ∼10% body weight loss in WT animals following 8 days of treatment (Figure 3A) and increased adipose tissue temperature (Fig. 3B). By contrast, there was no effect of the LAGCGRA on body weight, or adipose tissue temperature in the BARKO mice (Figure 3C,D). Next, we utilized BARKO mice that were provided a high fat diet to induce obesity. Consistent with our previous findings (Figure 3A,C) the LAGCGRA stimulated ∼30% body weight loss in WT animals following 14 days of treatment (Fig. 3E) and ∼10% body weight loss in the BARKO mice (Fig. 3G). LAGCGRA treatment increased iBAT temperature in both the WT and BARKO mice (Figure 3F,H), however the increase in iBAT temperature was significantly higher in the WT mice than the BARKO mice, when compared to each groups respective vehicle treated counterparts (Figure 3I,J). In summary, our findings suggest that GCGR activation recruits the classic adrenergic system to promote whole-body thermogenesis and reduce body weight in mice.

Figure 3.

Beta-Adrenergic Receptors are Required for LAGCGRA to Reduce BW, FI, and Increase Tissue Thermogenesis. WT and Adrb1,2,3 KO (BARKO) male mice housed at thermoneutrality, were implanted with temperature probes above the iBAT. Animals were pre-dosed for 7 days, then treated with vehicle or LAGCGRA (10 nmol/kg) for 8 days during which body weight (A,C) and iBAT temperature (B,D) were measured. n = 5. WT and Adrb1,2,3 KO (BARKO) male mice housed at thermoneutrality were provided a high fat diet, were implanted with temperature probes above the iBAT. Animals were pre-dosed for 7 days, then treated with vehicle or IUB288 (10 nmol/kg) for 14 days during which body weight (E,G) and iBAT temperature (F,H) were measured. IUB288-induced change in iBAT temperature was calculated (I) as well as the AUC of the change in iBAT temperatures (J). n = 10–11. Data expressed as mean ± SEM and analyzed by two-way ANOVA (A–I) and two-tailed unpaired t-test (J) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

2.4. GCGR agonism promotes neural activity in the arcuate nucleus of the hypothalamus

While GCGR-based therapeutics require signaling cascades in peripheral tissues, such as the liver and the adipose tissue, to promote energy wasting, a key question in the field is whether the anti-obesity efficacy of GCGR activation also requires the central nervous system (CNS) to drive a negative energy balance. Therefore, in light of our findings suggesting that activation of the GCGR may stimulate metabolic rate via the classic adrenergic to UCP1-dependent thermogenesis pathway, we evaluated whether the LAGCGRA stimulated neuronal activity (cFOS) in hypothalamic areas of the central nervous system (CNS) known to drive adipose tissue thermogenesis [[38], [39], [40]] (Figure 4A). Importantly, we found that while there was no effect in other hypothalamic centers critical to energy balance (Figure 4E–I), chronic treatment of obese mice with the LAGCGRA, stimulated neuronal activity (cFOS levels) within the arcuate nucleus (Figure 4B–D), an area of the medial basal hypothalamus (MBH) known to respond to peripheral signals to control both feeding behavior and energy expenditure.

Figure 4.

LAGCGRA Induces cFOS in the Arcuate Nucleus of the Hypothalamus. Following 2 weeks of LAGCGRA or Vehicle treatment, brains were collected and prepared for cFOS counting. The paraventricular (PVN), anterior (AH), lateral (LH), dorsomedial (DMH), ventromedial (VMN), and arcuate (ARC) nuclei of the hypothalamus were identified and defined within three distinct bregma regions (A). cFOS + cells were labelled and counted within their defined regions (B–C). cFOS + cells were counted and expressed as a % of the average vehicle cFOS within each nuclei (D–I). n = 16–20. Data expressed as mean ± SEM and analyzed by two-tailed unpaired t test (D–I). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

2.5. GCGR agonism does not require MBH glutamatergic signaling to promote weight loss

Since our findings suggested that LAGCGRA leads to the activation of the arcuate nucleus, we evaluated whether glutamatergic and/or GABAergic signaling in the MBH was required for the LAGCGRA to mediate its effects on energy balance. To ablate glutamatergic signaling, we utilized transgenic Vglut2^Flox (gene: slc17a6) mice developed by Bradford Lowell [41] and provided them a HFD, until they were obese. At this point, we performed intracranial injections of AAV^Cre targeted to the VMN (central location that is surrounded by the lateral hypothalamus, arcuate nucleus, and dorsomedial nucleus) but encompassed the MBH (Supplemental Fig. 2, 4D-F), ultimately generating Vglut2^MBHKO mice (Figure 5A–B). Importantly, treatment of the Vglut2^MBHKO mice with the LAGCGRA promoted similar body weight and fat mass loss to that of WT animals (Figure 5C–H). Together, our findings suggest that GCGR activation does not require glutamatergic signaling in the MBH to stimulate body weight loss.

Figure 5.

LAGCGRA Reduces BW Independent of MBH Glutamatergic Signaling. Vglut2^flox male mice were provided a HFD until they reached 35g of BW. The mice underwent intracranial injections of AAV^Cre targeting the VMH but encompassing the MBH (A). Cre expression mediated the KO of Vglut2 (B). Animals were kept at thermoneutrality and treated 14 days with vehicle or LAGCGRA (10 nmol/kg) during the light cycle. During this time, we measured bodyweight (C–E) and body composition (F–H). n = 10–19. Data expressed as mean ± SEM and analyzed by two-way mixed-effects ANOVA (C–H). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

2.6. GABAergic signaling in the MBH is required for GCGR agonism to reduce body weight

Next, we examined the role of GABAergic signaling in mediating the weight loss efficacy of the GCGR agonist by utilizing transgenic Vgat^Flox (gene: slc32a1) mice [42]. Similar to Vglut2^MBHKO mice, animals were housed at thermoneutrality and were provided with a HFD until they became obese. We then performed intracranial injections of AAV^Cre targeted to the MBH (Supplemental Fig. 2, 4A-C), to generate Vgat^MBHKO mice (Figure 6A–B). Consistent with our previous studies, treatment of obese WT animals with the LAGCGRA reduced body weight and fat mass. Strikingly, these effects were significantly attenuated in the Vgat^MBHKO mice, with only a third of the weight lost in the GFP control mice occurring in the Vgat^MBHKO mice, and no fat mass loss being observed in the absence of GABAergic signaling in the MBH (Figure 6C–H). Taken together, our findings suggest that treatment of obese mice with the LAGCGRA promotes weight loss via GABA signaling in the MBH.

Figure 6.

MBH GABAergic Signaling is required for LAGCGRA to reduce BW and adiposity. Vgat^flox male mice were provided a HFD until they reached 35g of BW. The mice underwent intracranial injections of AAV^Cre targeting the VMH but encompassing the MBH (A). Cre expression mediated the KO of Vgat (B). Animals were kept at thermoneutrality and treated 14 days with vehicle or LAGCGRA (10 nmol/kg) during the light cycle. During this time, we measured bodyweight (C–E) and body composition (F–H). n = 10–19. Data expressed as mean ± SEM and analyzed by two-way mixed-effects ANOVA (C–H). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

2.7. GCGR activation requires GABAergic signaling by MBH neurons to induce adipose tissue thermogenesis

Treatment of mice with the LAGCGRA increases adipose tissue thermogenesis, which could depend on central signaling to adipose tissue (Figure 7A). To test this potential mechanism, we utilized the Vgat and Vglut2 MBH KO mice (Figure 5, Figure 6) and treated them with either vehicle or the LAGCGRA for seven days, while continuously recording adipose tissue temperature (an indirect marker of thermogenesis). In accordance with our energy balance findings, the LAGCGRA significantly increased adipose tissue temperature (∼1 C) in WT and the Vglut2^MBHKO when compared to vehicle treated controls (Figure 7B,D). By contrast, the LAGCGRA failed to increase brown fat temperature in the Vgat^MBHKO mice (Figure 7C,E), despite the ability to mount a thermogenic response to a β3-adrenergic receptor agonist (CL-316,243 1 mg/kg) (Supplemental Fig. 3). Together, our data provide support to the notion MBH GABAergic signaling mediates the thermogenic effects of chronic GCGR agonism. Specifically, our findings suggest that treatment of obese mice with the LAGCGRA may signal through MBH GABA neurons to activate UCP1-dependent thermogenesis in adipose tissue to augment metabolic rate and promote weight loss (Figure 8).

Figure 7.

LAGCGRA requires GABAergic signaling by MBH neurons to induce Adipose Tissue Thermogenesis. To test the role of hypothalamic GABAergic and Glutamatergic signaling in tissue GCGR-induced tissue thermogenesis (A). Vglut2^flox and Vgat^flox male mice were provided a HFD until they reached 35g of BW. Animals underwent intracranial injections of AAV^Cre targeting the VMH but encompassing the MBH. Next, temperature sensing microchips were implanted above the iBAT. Animals were kept at thermoneutrality and pre dosed with vehicle for 7 days. The animals were then treated with vehicle for 7 days and iBAT temperature was recorded and averaged over the light cycle hours (B–E). Next, the animals were treated with LAGCGRA (10 nmol/kg) for 7 days and iBAT temperature was recorded and averaged over the light cycle hours (B–E). n = 15–18. Data expressed as mean ± SEM and analyzed by two-way mixed-effects ANOVA (B–E). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 8.

Overall model. Chronic treatment with a long-acting glucagon receptor agonist (LAGCGRA) progressively increases energy expenditure and adipose tissue thermogenesis, ultimately leading to a reduction in adipose mass. Loss of hepatic GCGR attenuates the observed weight loss effect from LAGCGRA, highlighting its importance in the overall LAGCGRA mechanism. Furthermore, GABAergic signaling in the MBH is required for the reduction in adipose mass and enhanced thermogenesis following treatment with a GCGR agonist. Additionally, adipocytes require both beta-adrenergic receptors and UCP1 for the full thermogenic and weight loss effects following LAGCGRA. The mechanism by which GCGR agonism signals through the MBH is unclear. While GCGR agonists can cross the blood brain barrier, it's likely the signal originates at the liver and then acts through MBH GABA signaling. This may occur by either vagal afferents or circulating FGF21 from the liver, future studies will address these possibilities.

3. Discussion

Multi-receptor agonists developed to modulate the GCGR are at the forefront of next-generation obesity therapeutics. Herein, we report that the treatment of obese mice with a LAGCGRA reduces body weight by accelerating whole-body metabolic rate. Mechanistically, the primary finding of our studies is that the LAGCGRA (e.g., GCGR agonism) requires GABAergic signaling in the MBH to promote UCP1-dependent thermogenesis in adipose tissue, whole-body metabolic rate, and drive weight loss in obese mice.

GCGR-based therapeutics are highly effective thermogenic agents shown to induce weight loss in preclinical models and man [25,43]. However, there is an ongoing need to identify how GCGR-based therapeutics function to augment whole-body metabolism. To better understand the unique biology by which the GCGR acts to stimulate energy wasting, we utilized genetically modified mice in combination with whole-body, tissue specific, and molecular metabolic measurements.

Because peripheral engagement of metabolic tissues could be driven by the CNS, we tested the role of hypothalamic effectors in GCGR biology. A major finding of our studies is that hypothalamic GABAergic, but not glutamatergic, signaling is required for GCGR agonism to induce a negative energy balance, and induce weight loss, presumably via the induction of adipose tissue thermogenesis and increased metabolic rate. It is not surprising that these collective effects by the LAGCGRA required intact hypothalamic GABAergic signaling, since GABA-containing cells exist throughout the hypothalamus. Indeed, feeding behavior and energy expenditure are controlled by separate neuronal circuits in the MBH, and there are sites that contain GABAergic cells essential for both processes, including the dorsomedial hypothalamus, lateral hypothalamus, and arcuate nucleus [[44], [45], [46]]. Our data suggest that the arcuate nucleus is a crucial area of the CNS leveraged by GCGR pharmacotherapies to promote weight loss. The arcuate receives many signals from the periphery that target the median eminence, as well as vagal afferent signals from the gut [[47], [48], [49], [50]]. Indeed, the most prominent population of arcuate GABAergic cells contain AGRP, however, these cells induce weight gain, not weight loss [50,51]. Alternatively, there is a sub-population of POMC cells that contain Vgat and non-AGRP GABAergic cells that are essential for energy balance [52,53]. These unique cells impact food intake and energy expenditure, offering another explanation for our findings. Together, our studies shed light on the metabolic benefits of GCGR activation and indicate that the thermogenic activity of the GCGR requires GABAergic neuronal circuits in the MBH.

While our findings suggest hypothalamic signaling may be critical for GCGR pharmacology, this does not confirm that the brain is the primary target site for GCGR agonists to mediate their anti-obesity efficacy. Although central administration of glucagon increases energy expenditure [29,54], expression of the GCGR within the CNS is controversial [20,55], with majority of studies suggesting that GCGR mimetics act indirectly to leverage pathways in the brain known to promote thermogenesis. Indeed, the GCGR is highly expressed in the liver [16], and hepatic activation of the GCGR induces gluconeogenesis, amino acid catabolism, FGF21 expression, the release of bile acids, and accelerates caloric expenditure [27,28,56]. Previous work has shown that hepatic GCGR is necessary for ∼70% of the body weight loss observed following GCGR agonism [28]. Interestingly, both GABAergic signaling and hepatic GCGR are required for a similar degree of body weight loss by GCGR agonism, indicating these systems may either be part of a shared system or drive parallel mechanisms.

As proposed by other groups [6,57], the LAGCGRA could induce liver signals that result in activation of core hypothalamic energy balance centers that reduce body weight. In support of this notion, hypothalamic centers can induce tissue thermogenesis and browning of adipose tissue [58,59]. Thus, our results point to the presence of a liver→brain→fat axis that is targeted by this group of treatments for weight loss in obesity, which has been proposed by other groups [[60], [61], [62]]. While we have identified key features in this system, more information is needed for the nature of how this system operates for therapeutic benefits. For instance, we don't yet know how the liver engages the brain following chronic periods of GCGR agonism. Previous findings demonstrated that severing the hepatic branch of the vagus nerve is required for the anorectic effects of glucagon [[63], [64], [65]]. Such vagal signals project through the NTS and reach hypothalamic regions [66,67]. Thus, GABAergic signaling in the MBH could have been activated by the LAGCGRA in our studies via vagal hepatic-brain signaling. Alternatively, the liver may drive the release of hormones that impact hypothalamic function. For example, the hepatokine FGF21 is released from the liver in response to GCGR agonism and acts in the brain to induce weight loss [68].

In addition to acting indirectly via hepatokines, both brown and white adipose tissues express the GCGR [16], and may be direct targets of agents modulating the GCGR. Indeed, glucagon stimulates lipolysis, cellular respiration, and metabolic gene expression in rodent and human adipocytes [16,69]. Further, a key feature of chronic GCGR agonism is the progressive increase in adipose tissue thermogenesis, and energy expenditure. This feature is unique from other thermogenic activators such as β3AR mediated thermogenesis [70]. We found that with repeated β3AR agonist treatment, the magnitude of whole-body metabolism is diminished over time (Supplemental Fig. 1A). This is consistent with previous studies showing β3AR agonism desensitizes after repeated treatment [70]. However, this state of thermogenic desensitization does not occur with GCGR agonism, suggesting that the GCGR leverages adrenergic and additional signaling to stimulate the combustion of excess calories, but the underlying mechanism has not yet been revealed. Future studies are required to elucidate these novel thermogenic pathways.

In summary, we present for the first time that GCGR agonism requires GABAergic signaling in the medial basal hypothalamus to stimulate UCP1-dependent thermogenesis in adipose tissue, augment whole-body metabolic rate, and promote weight loss in obese mice.

4. Materials and methods

4.1. Animals

The procedures included in this manuscript were approved by the Committee on the Use and Care of Animals at Indiana University (Indianapolis, IN). All mice were provided with high fat diet (HFD) (D12451 Rodent Diet with 45 kCal% Fat) and water ad libitum, unless noted otherwise, and kept in a temperature-controlled (23 °C) room and or incubators (30 °C) on a 12-hour light–dark cycle. Vglut2^flox mice were generated by Bradford Lowell (Beth Israel deaconess Med Centre (Harvard)) (RRID:IMSR_JAX:012,898) [41], Vgat^flox mice were generated by Bradford Lowell (Beth Israel deaconess Med Centre (Harvard)) (RRID:IMSR_JAX:012,897) [42], Adrb1,2,3 KO mice were provided from our collaborator Dr. Hyun Roh and generated as previously described [36], and C57BL/6J mice were purchased from The Jackson Laboratory (RRID:IMSR_JAX:000664). All experiments were performed using male mice, as female mice have been shown to be resistant to diet-induced obesity. Prior to studies in Vglut2^flox, Vgat^flox, and BARKO male mice the animals were provided with a high fat diet regimen (HFD; 45% kcal from fat, high sugar, D12451, Research Diets) started from 3 to 5 weeks of age, until at least 35g bodyweight was reached. Mice that weighed less than 35g after (by 40 weeks of age) were removed from the studies. Mice were group-housed prior to stereotaxic surgery in all cohorts, but single-housed post-operatively for body weight and body composition studies, indirect calorimetry, and temperature measurements at adipose tissue depots. All mice were genotyped via polymerase chain reaction (PCR) across the genomic region of interest prior to study and only mice homozygous for Vglut2^flox, and Vgat^flox were included. In house colonies were used for breeding Vglut2^flox, and Vgat^flox mice. The animals were randomly assigned an ear tag at the time of tailing (prior to genotyping).

4.2. Viral constructs

All the AAVs were acquired from Addgene (Addgene HQ Addgene, Watertown, Massachusetts). All AAVs used include titer from stock. AAV-hSyn-Cre-GFP (3.1 × 10¹³ GC/ml), Addgene plasmid 105,540, serotype 9; AAV-CAG-GFP (2.4 × 10¹³ GC/ml), Addgene plasmid 37,825, serotype 9.

4.3. Stereotaxic injections of viral constructs

Intracranial injections were performed on Vglut2^flox, and Vgat^flox mice once they reached a weight of at least 35g, while being ≤40 weeks of age. Prior to craniotomy, mice were anesthetized with 1.5–2% isoflurane. A hind-limb pinch was used to ensure successful anesthesia. The mouse was placed on a water heated pad set to 37 °C, on the stereotaxic platform (Kopf), and the head was fixed using the ear-bars and the nose piece which provided a steady supply of 1.5–2% isoflurane. Eyes were covered with Artificial Tears ointment (Pivetal) to avoid drying of the eyes. The scalp was shaved, and the shaved area was disinfected with 10% iodine solution (PVP) (Medline) followed by sterile PBS, this disinfection was repeated three consecutive times. A pre-sterilized scalpel was used to create an incision that exposed the skull. Bregma and lambda were located, and skull was leveled in accordance with these landmarks. Holes were drilled bilaterally at anteroposterior (AP) −1.15, mediolateral (ML) + & - 0.45. A pulled pipette containing the appropriate viral vector was lowered to dorsoventral (DV) −5.75. The virus was injected at a rate of ∼10 nl/min until a total of 300 nL was injected. At least 5 min were allowed for the virus to diffuse into the brain, and the pipette was then raised slowly out of the hole in the skull. The holes in the skull were filled with bone wax, and the incision was closed with surgical glue (VetBond, 3M). Post operative analgesics (Carprofen, 5 mg/kg) were administered to prevent post-surgical pain or discomfort. The GFP fluorescent report was used to confirm proper targeting of the viral constructs after the completion of studies, and if GFP was not observed throughout the VMN, the animal was omitted from analyses.

4.4. Phenotypic studies

Body weight was monitored daily, starting 7 days after surgery. Body weights were assessed on a precision balance (OHAUS Corporation, model # SPX421). Body composition data was collected with assistance from Indiana University School of Medicine Mouse Metabolic Phenotyping Center from the EchoMRI™-100H Body Composition Analyzer. TSE Phenomaster cages were used for indirect calorimetry measurement, with assistance from the Indiana University School of Medicine Mouse Metabolic Phenotyping Center. Indirect calorimetry data was used to calculate fat oxidation and glucose oxidation via the Weir equation.

4.5. Pharmacologic treatments

C57BL/6J mice were injected (Intraperitoneally (IP)) with the β3 receptor agonist (CL-316,243 (2 mg/kg)) for 4 consecutive days approximately 4 h after the light cycle began. Prior to treatment all animals were acclimated to IP injections via IP injections of sterile PBS for 7 consecutive days. Prior to treatment, C57BL/6J and BARKO mice were injected (Subcutaneously (SQ)) with sterile PBS for 7 days. C57BL/6J and BARKO mice were injected SQ with either Vehicle or the long-acting glucagon receptor agonist (IUB288) (10 nmol/kg) for 8 or 14 days. Following a 2 week wash out, C57BL/6J and BARKO mice were treated IP with IUB288 (1 nmol/kg). Vglut2^flox, and Vgat^flox mice were injected (Subcutaneously (SQ)) with the long-acting glucagon receptor agonist (LAGCGRA (IUB288 10 nmol/kg)) for 14 consecutive days approximately 4 h after the light cycle began. Prior to treatment all animals were acclimated to SQ injections via SQ injections of Vehicle for 7 consecutive days.

4.6. Temperature at adipose depots

Immediately following stereotaxic surgeries, while still under anesthesia, C57BL/6J, BARKO, Vglut2^flox, and Vgat^flox mice were implanted with temperature programmable microchips (UCT-2112 Temperature Microchip), from UID (Lake Villa, Illinois) via the UPGI-Q Pistol Grip Injection from UID, under the skin, above the interscapular brown adipose tissue and the inguinal white adipose tissue (beige adipose tissue). Animals were singly housed 3 days after surgery. Temperatures were recorded using the automated UID home cage monitoring system. Data were collected using the “Activity Tracking Preset”. Each matrix plate is composed of 8 zones, all of which record when a temperature microchip is located above said zone. The “Activity Tracking Preset” samples with the following frequency, each zone 1–8 is sequentially active for 250 ms (1 full cycle), this cycle is immediately repeated followed by a 1s delay. This process is repeated until the competition of studies. Start Matrix→Cycle 1→Cycle 2→1s Delay→Cycle 3→ Cycle 4→ 1s Delay→(repeats). This provides approximately 1,440 readings taken per hour. After ending the program, data were then averaged (automated) and binned into 30min increments, the data are compiled into a timeline over the duration of the study and exported as an Excel file for further analysis. Data were analyzed by averaging all temperatures recorded during the light cycle hours, beginning immediately after the time of injection which was documented daily.

4.7. Tissue collection

Mice were euthanized with a high concentration of isoflurane. Blood was collected with a 1 ml insulin syringe via cardiac puncture, the blood was transferred to a 1 ml Eppendorf microcentrifuge tube containing 15uL EDTA, and stored on ice. The blood was then centrifuged at 4 °C for 15 min at 5,00×g, the plasma (supernatant) was then pipetted into a fresh 1 ml microcentrifuge tube and stored at −80 °C. Following the blood collection the intrascapular brown adipose tissue, inguinal white adipose tissue, quadricep, visceral adipose tissue, and liver were removed from the animal, placed in individual 1 ml microcentrifuge tubes with corresponding labels, and placed onto dry ice. These samples were then stored at −80 °C.

4.8. Perfusion and immunohistochemistry

Immediately following tissue collection, mice were perfused transcardially with phosphate buffered saline (PBS) followed by 10% neutral buffered formalin (NBF) (Fisher brand, catalog #245685). Brains were removed from the skull with rongeurs, placed in 10% NBF overnight, and then in 30% sucrose for at least 36 h. A freezing microtome was used to section the brains (30um) in four series, and stored in antifreeze solution (25% ethylene glycol, 25% glycerol). For immunohistochemistry staining, sections were washed in PBS for 5 min in 4 sequential washes. The washes were followed by an hour of blocking (PBS containing 2.5% triton, 3% Normal Donkey Serum [Lampire Biological Laboratories, catalog # 7332100]), followed by overnight incubation in blocking solution along with chicken anti-GFP (GFP-1020, RRID: AB_1000240, Aves, 1:1000). The next day the sections were washed in PBS for 5 min over 10 sequential washes, followed by 2 h of incubation in the fluorescent secondary Alexa Fluor® 488 AffiniPure Donkey Anti-Chicken IgY (IgG) (H + L) (703-545-155, RRID: AB_2340375, Jackson) with anti-GFP. This was followed by 4, 5 min PBS washes. Additionally, cFos staining was done to quantify neuronal activation. For this cFos staining was done in day 1 of staining using Rabbit anti-c-Fos antibody (2250S c-Fos 9F6, RRID: AB_, Cell Signaling) followed by Peroxidase AffiniPure Donkey Anti-Rabbit IgG (H + L) (711-035-152, RRID: AB_10015282, Jackson). This was followed by the GFP staining as described above. The sections, now stained, were mounted onto slides and cover slipped with Fluoromount-G (Southern Biotech). Images from 30um thick sections containing GFP alone were collected throughout the hypothalamus using an Echo Fluorescent microscope with a 4× objective. The target regions were identified by overlaying the corresponding bregma level from an anatomical reference atlas (Allen Mouse Brain Atlas) on microscope images. cFos counts in Figure 5 were done from 16 to 20 mice per group, counts from 4 representative sections were averaged into a single count for each animal within each region. Each region was counted by someone blind to treatment.

4.9. RNA isolation

Total RNA was extracted from brown adipose tissue and brain tissue using TRIzol™ Reagent (Thermo Fisher Scientific, Cat# [15596026]). Samples were homogenized in 1 mL of TRIzol. Samples were spun at 12,000×g for 10 min at 4 °C and then the lipid layer was removed from the top of the homogenized TRIzol mixture. Chloroform was then added (0.2 mL per 1 mL TRIzol), and samples were vortexed and incubated at room temperature for 2 min. Following centrifugation at 12,000×g for 15 min at 4 °C, the aqueous phase was transferred to a fresh tube, and RNA was isolated using a QIAcube Connect (QIAGEN, Cat# [9002864]) running the “RNeasy Mini” settings per manufacturer's protocol.

4.10. cDNA synthesis

cDNA was synthesized from 1 μg of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems™, Cat# [4368813]) according to the manufacturer's instructions. The reaction mix included random primers and reverse transcriptase in a final volume of 20 μL. Reverse transcription was carried out in a thermal cycler and synthesized cDNA was stored at −20 °C until further use.

4.11. Quantitative PCR (qPCR)

Quantitative real-time PCR was performed using the QuantStudio™ 7 Flex Real-Time PCR System (Thermo Fisher Scientific) with TaqMan™ gene expression assays. Each 10 μL reaction contained 5 μL of TaqMan™ Universal PCR Master Mix (Applied Biosystems™ Cat# [4326708]), 0.5 μL of mouse GCGR, UCP1, RPLP0, SLC32A1, SLC17A6, or GAPDH primer (Applied Biosystems™ [Mm00433546_m1, Mm01244861_m1, Mm01974474_gH, Mm00494138_m1, Mm01335985_m1, Mm99999915_g1]), and 4.5 μL of cDNA template. Thermal cycling conditions were as follows: initial activation at 50 °C for 2 min, denaturation at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15s and 60 °C for 1 min. Expression levels were normalized to RPLP0 or GAPDH & relative gene expression was calculated using the 2−ΔΔCt method.

4.12. Protein extraction

Brown adipose tissue (BAT) samples were homogenized in RIPA buffer (Thermo Fisher Scientific Cat# [89,901]) supplemented with Halt™ Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific, Cat# [78,430]) at a concentration of 10 μL per mL of RIPA buffer. Homogenates were centrifuged at 13,000×g for 10 min at 4 °C, and the intermediary aqueous phase containing total protein lysate was carefully collected and this was repeated 2 more times. Protein concentrations were determined using the Pierce™ BCA protein assay kit (Thermo Fisher Scientific, Cat# [23,227]), according to the manufacturer's instructions.

Protein samples were denatured by mixing with Laemmli sample buffer, prepared by combining 950 μL of 2x Laemmli buffer (Bio-Rad, Cat# [1610737]) with 50 μL of BME (Bio-Rad, Cat# [1610710]) to achieve a 1:1 final ratio with the protein lysate. Samples were boiled at 95 °C for 5 min to ensure complete denaturation.

4.13. Western blotting

Equal amounts of denatured protein lysates were loaded onto SDS-PAGE gels (Criterion™ Precast Gel 4–20% Tris–HCL, 1.0 mm 18 Well Comb, 30uL (Bio-Rad, Cat# [3450033])) and separated by electrophoresis. Proteins were then transferred to nitrocellulose membranes (Midi Format, 0.2 μM Nitrocellulose (Bio-Rad, Cat# [1704159]) using the Trans-Blot® Turbo™ Transfer System (Bio-Rad, Cat# [1704150]). Membranes were blocked in Intercept® Protein-Free Blocking Buffer (LICORbio, Cat# [927–80001]) for 1 h at room temperature and incubated overnight at 4 °C with primary antibodies against UCP1 (Abcam Cat# [ab209483], 1:5,000) and Actin (MilliporeSigma Cat# [MAB1501], 1:2,500) diluted in Intercept® Antibody Diluent (LICORbio, Cat# [927–65001])

After washing with TBS-T, membranes were incubated with species-appropriate secondary antibodies—either goat anti-rabbit IgG (LICORbio Cat# [926–68071]) or goat anti-mouse IgG (LICORbio Cat# [926–32210]) at 1:20,000—for 1 h at room temperature. Membranes were washed with TBS-T then stored in TBS and protein bands were detected using the LI-COR® Odyssey M Imaging System and quantified using Image Studio software (version 5.2.5, LICORbio).

4.14. Statistics

Data in Figure 1, Figure 2, Figure 3, Figure 5, Figure 6, Figure 7E, S1A, S1B, S3B, S3C, S4B, S4E were analyzed by two-way repeated measures ANOVA with Fisher's LSD post hoc test. Data in Figure 2I was analyzed by one-way ANOVA. Data in Figure 2, Figure 3J, 4D-4I, S4C, S4F, S5 were analyzed by student's unpaired two tailed t-test. Data in Fig. S4A and S4D were analyzed by paired one-tailed t-test. Data was expressed as mean +/− SEM. No data were removed unless the injection missed the targeted location, or the animal was sick or injured at the time of the experiment (loss of >10% body weight). Significance was determined at p ≤ 0.05. All data were analyzed using Prism (GraphPad, La Jolla, California).

CRediT authorship contribution statement

Andrew J. Elmendorf: Writing – review & editing, Writing – original draft, Validation, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ellen Conceição Furber: Writing – review & editing, Validation, Methodology, Investigation, Formal analysis. Betty Lorentz: Writing – review & editing, Validation, Methodology, Investigation, Formal analysis, Data curation. Connor A. Mahler: Writing – review & editing, Visualization, Validation, Supervision, Methodology, Investigation. Brian A. Droz: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Richard Cosgrove: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Jonquil Marie Poret: Writing – review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Patrick J. Knerr: Writing – review & editing, Validation, Supervision, Formal analysis, Data curation. Ricardo J. Samms: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Jonathan N. Flak: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Funding

This work was supported by NIH R01DK136897, ADA 17-INI-15, the Lilly Research Award Program, and the Indiana Biosciences Research Institute.

Declaration of competing interest

Jonathan N. Flak has received research support from Eli Lilly and Company for this project. Ellen Conceição Furber, Brian A. Droz, Richard Cosgrove, Jonquil Marie Poret, and Ricardo J. Samms are all employees at Eli Lilly. Patrick Knerr was previously an employee at Novo Nordisk, currently receives research funding from Eli Lilly and Company, and is a founder and shareholder in Volari Therapeutics, all unrelated to this project.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

Supplemental 1.

Validation of temperature sensing system and UCP1 KO. Lean C57/Bl6 mice housed at thermoneutrality were treated for 4 days (4 h into the light cycle) with the β3 agonist CL-316,243 (2 mg/kg), during which, iBAT temperature was recorded (A). UCP1 WT and KO mice were treated with LAGCGRA (10 nmol/kg), iBAT tissue was removed and UCP1 protein levels were measured with immunoblotting and quantified (B). n = 7–8 per group (A), n = 6 per group (B). Data expressed as mean ± SEM and analyzed by two-way ANOVA (A,B). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Supplemental 2.

Viral spread of VMN targeted, MBH encompassing AAV^Cre injection. Following the 15th day of LAGCGRA treatment, brains were removed and prepared for immunohistochemistry. GFP staining was used to identify viral infection, the viral spread in each animal was identified and drawn. The spread in each mouse was overlayed with the others in each group. n = 10–19.

Supplemental 3.

Vglut2^MBH and Vgat^MBH KO mice have increased adipose tissue temperature following β3 agonism. Vglut2^flox and Vgat^flox male mice were provided a HFD until they reached 35g of BW. Animals underwent intracranial injections of AAV^Cre targeting the VMH but encompassing the MBH. Next, temperature sensing microchips were implanted above the iBAT. Animals were kept at thermoneutrality and pre dosed with vehicle for 7 days. The animals were then treated with the β3 agonist CL-316,243 (1 mg/kg) for 6 days (B–C), iBAT temperature was recorded and averaged over the light cycle hours (B–C). n = 7–11 per group. Data expressed as mean ± SEM and analyzed by two-way mixed-effects ANOVA (B–C). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Supplemental 4.

Vgat^MBH and Vglut2^MBH KO validation. Vgat^flox and Vglut2^flox male mice underwent intracranial injections of AAV^Cre targeting the VMH but encompassing the MBH. Injections were unilateral for gene expression studies and bilateral for bodyweight and body composition studies. Three weeks after AAV^Cre injections, the brains were removed, the MBH was dissected out and separated into injected and non-injected sides. RNA was then extracted and qPCR was performed for the genes SLC32A1 (A) and SLC17A6 (D) Mice that were injected bilaterally with AAV^Cre were studied for 2 weeks following injection during this time we measured bodyweight (B,E) and body compositions at day 14 (C,F) n = 5–37. Data expressed as mean ± SEM and analyzed by paired one-tailed t-test (A,D), two-way ANOVA (B,E), and unpaired two-tailed t-test (C,F) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Supplemental 5.

Validation of siRNA mediated liver GCGR knockdown. To knockdown the GCGR specifically in the liver, we used an siRNA conjugated to N-acetylgalactosamine (GalNAc), a ligand that binds to asialoglycoprotein receptor (ASGPR) specifically in the hepatocyte. Mice were treated with either vehicle or GCGR siRNA (10 mg/kg). RNA was extracted from the liver and qPCR was performed for GCGR expression and normalized to RPLP0 expression. n = 6. Data expressed as mean ± SEM and analyzed by unpaired two-tailed t-test ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Data availability

Data will be made available on request.