Role of Hypothalamic Reactive Astrocytes in Diet-Induced Obesity

Abstract

INTRODUCTION

The major causes of metabolic disorders such as obesity are associated with the hypothalamus, which is crucial for bodyweight control (Thaler et al., 2012; Thorburn and Proietto, 1998). Hypothalamus consists of several small nuclei with various functions that communicate via different neurotransmitters and neuropeptides or hormones (Horvath et al., 2004; Quadt et al., 2018). It has been established that arcuate nucleus (ARC) of the hypothalamus firstly senses metabolic signals and hormones, which are then conveyed to the second-order hypothalamic regions of paraventricular nucleus (PVN) and lateral hypothalamic area (LHA) of the hypothalamus (Bouret et al., 2004; Yaswen et al., 1999). ARC, PVN, LHA, ventromedial hypothalamus (VMH), and dorsomedial hypothalamus (DMH) contain most of the neurons involved in feeding and body weight control (Gold, 1973; Gooley et al., 2006; King, 2006; Schneeberger et al., 2014; Timper and Bruning, 2017; Waterson and Horvath, 2015).

In these hypothalamic regions, numerous types of neurons can be classified using various cellular markers with corresponding functional properties as follow: orexigenic (promoting appetite) properties of neuropeptide Y (NPY)- (Tatemoto et al., 1982) and Agouti-related peptide (AgRP)- positive neurons (Miltenberger et al., 1997) and anorexigenic (suppressing appetite) properties of pro-opiomelanocortin (POMC) (Cowley et al., 2001) and cocaine- and amphetamine-regulated transcript (CART) (Kristensen et al., 1998). In addition, CART neurons in the ARC decrease adiposity and increase energy expenditure (Lau et al., 2018). POMC/CART neurons secrete an α-melanocyte-stimulating hormone (α-MSH), which suppresses appetite via melanocortin 4 receptor (MC4R) (Baltatzi et al., 2008; Fan et al., 1997; Gantz et al., 1993; Huszar et al., 1997). In addition to ARC, PVN mediates many diverse functions via secretion of corticotropin-releasing hormone (CRH), thyrotropin-releasing hormone (TRH), oxytocin, and vasopressin (Konturek et al., 2005). Inputs from NPY neurons in ARC regulate energy intake via TRH secretion (Beck, 2006; Konturek et al., 2005; Nillni, 2010). Interestingly, ARC projections affect appetite through MC4R-expressing oxytocin neurons (Qin et al., 2018). Also, vasopressin cells are involved in brown fat thermogenesis (Hill, 2012). LHA occupies a large portion of the hypothalamus and contains melanin-concentrating hormone (MCH) which is an orexinergic peptide (Barson et al., 2013), and hypocretin/orexin. Hypocretin/orexin neurons in LHA increase food intake while regulating brown adipose tissue activity to enhance energy expenditure (Martins et al., 2016; Tupone et al., 2011). VMH also contains many distinct neuronal populations. Among them, steroidogenic factor 1 (SF-1) neurons have anorexigenic properties (Zhang et al., 2020). Lastly, DMH cholinergic neurons increase food intake (Jeong et al., 2017), whereas TrkB-expressing neurons suppress food intake (Liao et al., 2019). When animals are challenged by high-fat diet (HFD) feeding, neurons that are homeostatically regulated in the hypothalamus lose metabolic control (Moraes et al., 2009) and start to show dysfunctional firing patterns and distorted regulation by neurotransmitters and neurohormones (Beutler et al., 2020; Sa et al., 2022). It has been further reported that with chronic feeding of 20 weeks of HFD, markers of neuronal injury become evident in the neurons of the hypothalamus (Thaler et al., 2012).

Before hypothalamic neurons show dysfunctional firing patterns and markers of neuronal injury, hypertrophic and hyperplasic astrocytes are concurrently observed in the hypothalamus even 1 day after HFD feeding (Buckman et al., 2013; Thaler et al., 2012). Astrocytes normally participate in brain energy metabolism by controlling glycogen storage, sensing glucose, and supplying fuel to neurons under physiological conditions (Belanger et al., 2011; Choi et al., 2012; Fuente-Martin et al., 2012; Garcia-Caceres et al., 2016; Timper et al., 2020). In addition, astrocytes secrete chemokines, cytokines, and neurotrophic factors to promote neuronal development, neuroplasticity, and synaptic plasticity (Casse et al., 2018; Jo et al., 2014). Reactive astrocytes appearing in short-term HFD act to reduce HFD overload and return to normal (Thaler et al., 2012), whereas reactive astrocytes appearing in chronic HFD produce inflammatory factors and elevate inflammatory signals such as IκB kinase-β (IKKβ)/nuclear factor κB (NF-κB) signaling (Douglass et al., 2017; Zhang et al., 2008; 2017). Furthermore, reactive astrocytes have been reported to release gliotransmitters such as vascular endothelial growth factor (VEGF) and γ-aminobutyric acid (GABA) under DIO (Gruber et al., 2021; Zhang et al., 2017). VEGF inducese hypothalamic angiopathy and systemic hypertension. However, little is known about the effect of GABA from reactive astrocytes on the surrounding hypothalamic neurons in DIO. Increasing lines of evidence suggests that elevated levels of GABA from reactive astrocytes act as a common molecular mechanism in other neuroinflammatory diseases, such as Alzheimer’s disease, Parkinson’s disease, white matter stroke, inflammation-induced anxiety, and epilepsy (Chun et al., 2020; Heo et al., 2020; Jo et al., 2014; Nam et al., 2020; Pandit et al., 2020; Shim et al., 2019). Recently, GABA from hypothalamic reactive astrocytes under chronic HFD has been shown to contribute to the pathology of obesity by inhibiting the excitability of neighboring neurons facilitating energy expenditure (Sa et al., 2022). In this review, we summarize recent discoveries related to the causes of reactive astrogliosis and the consequences of reactive astrocytes in the hypothalamus in an attempt to gain insights into the interplay between reactive astrocytes and neurons under DIO.

REACTIVE ASTROCYTES IN THE HYPOTHALAMUS

CAUSES OF REACTIVE ASTROCYTES

What causes reactive astrogliosis? One obvious candidate is the fatty acid itself in HFD, which acts as a triggering molecule. It is well known that chronic overnutrition increases peripheral fat levels, as well as free fatty acids from the plasma to the brain (Karmi et al., 2010; Wang et al., 1994). High levels of circulating saturated free fatty acids have been found to activate inflammatory signaling in cultured astrocytes (Gupta et al., 2012). In addition, enrichment of saturated fatty acids causes lipid accumulation in the hypothalamus (Borg et al., 2012; Giles et al., 2016; Posey et al., 2009). Among the various types of saturated fatty acids, palmitic acid (16:0), which is the predominant saturated fatty acid in the circulatory system and tissues, is the most common free fatty acid accounting for 21%-30% of human deposited fat (Bysted et al., 2005; de Almeida et al., 2002; Firl et al., 2013; Kingsbury et al., 1961; Liu et al., 2015). Notably, it has been demonstrated that palmitic acid treatment by intracerebroventricular cannulation induces reactive gliosis in the hypothalamus (Jin et al., 2020). Interestingly, long-term treatment of fatty acids increases GABA production in cultured hypothalamic astrocytes, indicating that astrocytes turn into reactive astrocytes (Lee et al., 2018). These studies raise a strong possibility that fatty acids in HFD can cause reactive astrogliosis.

What is the triggering mechanism of reactive astrocytes? In our previous studies, we have demonstrated how common molecular pathways such as MAOB-dependent putrescine degradation and GABA production are shared, even though the triggering factors are different (Chun and Lee, 2018; Chun et al., 2020; Heo et al., 2020; Jo et al., 2014; Nam et al., 2020; Pandit et al., 2020; Shim et al., 2019). We have reported that autophagic degradation pathway is commonly triggered by pathogenic molecules such as diphtheria toxin, Aβ, cytokines, damaged tissue debris, and viral infections, which usually accompany neuroinflammation (Chun et al., 2020; Ju et al., 2021). Under pathological conditions, astrocytes take up or internalize these toxic molecules to degrade them and subsequently turn on the urea cycle to convert the accumulating toxic ammonia to less toxic urea (Cohen, 1981; Ju et al., 2021; Meijer et al., 1990; Morris, 2002). The net consequence of this degradation and turning-on of the urea cycle is the production of putrescine, which further turns on MAOB-dependent production of GABA (Ju et al., 2021). Of note, a recent study has reported that MAOB, which is mainly expressed in astrocytes, is elevated in transcriptionally profiled hypothalamic cells of DIO mice (Rossi et al., 2019). In addition to GABA, excessive hydrogen peroxide (H₂O₂), a reactive oxygen species (ROS) originating from MAOB in reactive astrocytes, has been shown to cause glial activation and neuronal death (Chun et al., 2020). It is highly likely that the same common molecular mechanism working in the hypothalamus causes reactive astrogliosis and the production of GABA and H₂O₂ in a MAOB-dependent manner (Chun et al., 2020; Sa et al., 2022). Therefore, elevated free fatty acids, acting as pathogenic molecules during chronic HFD, may produce putrescine by activating the urea cycle of astrocytes and MAOB produces GABA and H₂O₂ in astrocytes, the process by which astrocytes turn into reactive astrocytes (Fig. 2). This interesting hypothesis needs to be further investigated and validated in the future.

CONSEQUENCES OF REACTIVE ASTROCYTES

What are the functional consequences of reactive astrocytes? It is well known that chronic HFD induces metabolic damages in hypothalamic neurons. For example, POMC neurons in ARC have shown elevated autophagy, an independent marker of neuronal stress or injury, in mice fed HFD for 8 months (Thaler et al., 2012). In other hypothalamic regions, signs of neuronal injury, including long-lasting desensitization of neurons and a reduction in synaptic inputs, have also been observed in DIO (Beutler et al., 2020; Moraes et al., 2009). In response to chronic HFD, homeostatic circuits of neuronal activity are disrupted (Beutler et al., 2020). Moreover, chronic HFD impairs neuronal responses to nutrients and hormones in a way that is expected to promote weight gain, which lasts for weeks after mice have been returned to a low-fat diet and lost weight (Beutler et al., 2020).

Long-lasting damage to hypothalamic neurons could be a direct consequence of reactive astrocytes. Reactive astrocytes produce and release inflammatory factors including tumor necrosis factor-α (TNF-α) and Interleukin-6 (IL-6) in response to fatty acid treatment (Gupta et al., 2012). However, the consequences of these pro-inflammatory factors are rather contradictory. It has been reported that high concentrations of TNF-α injection lead to weight loss in mice, whereas low concentrations of TNF-α injection result in sufficient changes to cause obesity (Arruda et al., 2011). Along with these conflicting results, mice with astrocyte-specific overexpression of IL-6 exhibit reactive astrocytes, but no weight change in response to HFD (Hidalgo et al., 2010). From these reports, it is possible to conclude that pro-inflammatory factors from reactive astrocytes might not directly cause long-lasting neuronal damage.

Intriguingly, mice with astrocytic IKKβ/NF-κB activation after chronic HFD feeding not only show increased expression of several pro-inflammatory genes such as TNF-α and IL-6 (Liu et al., 2017), but also a significantly elevated level of extracellular GABA in the hypothalamus (Gonzalez-Garcia and Garcia-Caceres, 2021; Zhang et al., 2008; 2017). Astrocytes release numerous neuroactive molecules, including the classical inhibitory neurotransmitter GABA (Garcia-Caceres et al., 2019; Lee et al., 2010; Yoon and Lee, 2014). Accumulating evidence suggest that elevated levels of GABA in reactive astrocytes act as a common molecular mechanism in various neuroinflammatory diseases (Chun and Lee, 2018), such as Alzheimer’s disease (Chun et al., 2020; Jo et al., 2014), Parkinson’s disease (Heo et al., 2020), recovery after stroke (Nam et al., 2020), stab-wound injury (Chun et al., 2022), epileptic seizure (Pandit et al., 2020) and inflammation-induced anxiety (Shim et al., 2019). Consistently, it has been reported that extracellular GABA in mediobasal hypothalamus (ARC and VMH) is elevated in DIO mice (Zhang et al., 2017). In our recent study, we have demonstrated that GABA from reactive astrocytes in LHA after chronic HFD feeding contributes to exacerbation of obesity (Sa et al., 2022). Mechanistically, MAOB-dependent GABA from reactive astrocytes tonically and strongly inhibits the excitability of the newly identified GABRA5-positive GABAergic pacemaker firing neurons that facilitate energy expenditure, resulting in increased fat mass and body weight in DIO mice (Sa et al., 2022). Along with elevated GABA production, MAOB-dependent production of H₂O₂ is the key common molecular switch that turns on a distinct state of severe reactive astrocytes, and the toxic level of H₂O₂ is sufficient for neurodegeneration (Chun et al., 2020). Taken together, these recent discoveries strongly suggest that the accumulation of toxic H₂O₂ in hypothalamic reactive astrocytes might directly contribute to the long-lasting impairment and loss of neighboring neurons in the hypothalamus (Fig. 2). This exciting possibility awaits future investigation.

CONCLUSIONS AND FUTURE QUESTIONS

We have comprehensively reviewed the role of hypothalamic reactive astrocytes (Table 1), the regional distribution of reactive astrocytes in the hypothalamus (Fig. 1), the triggering mechanisms of reactive astrocytes, and the consequences of reactive astrocytes in DIO (Fig. 2). Indeed, reactive astrocytes are actively involved in the pathogenesis of DIO. Mechanistically, elevated free fatty acids during chronic HFD feeding activate inflammatory signals via IKKβ/NF-κB pathway and produce putrescine possibly by turning on the urea cycle in reactive astrocytes. The putrescine produced is then degraded by MAOB to GABA in reactive astrocytes. This GABA is tonically released and inhibits the excitability of neighboring neurons, especially those of the newly identified GABRA5-positive neurons, which facilitate energy expenditure. This cascade of events leads to the exacerbation of obesity in DIO. In addition, the role of MAOB-dependent H₂O₂ is still unknown and further investigation is needed to clarify the relationship between toxic H₂O₂ from reactive astrocytes and the loss of hypothalamic neurons.

Although the presence of reactive astrocytes in the hypothalamus after HFD has been observed for a long time, the role of reactive astrocytes in DIO has only recently been investigated. Moreover, most studies have focused on reactive astrocytes in ARC in response to HFD. Since overall metabolic signals from ARC are conveyed to and act on other hypothalamic regions, it is necessary to deeply understand the role of reactive astrocytes in other regions of the hypothalamus, as has been recently discovered in LHA. These newly developed approaches and tools will be very useful for developing potentially effective therapeutic strategies to fight against obesity with minimal side effects.

Article information

References