Brain–gut–adipose-tissue communication pathways at a glance

For survival and reproduction, the body has developed an intricately balanced system to efficiently control energy homeostasis at multiple levels. In short, the brain continually monitors the systemic metabolic state and adjusts behavior, as well as humoral and neuronal outputs to peripheral effector organs, to ensure an appropriate energy supply. If the central nervous system (CNS) senses a caloric shortage or surplus, the brain orchestrates responses that alter food intake, nutrient partitioning and physiological functions such as hepatic glucose production, adiposity and thermogenesis. Efficient maintenance of the delicate homeostatic balance of energy, glucose and lipid metabolism largely depends on system-wide synchronicity of metabolic processes that can only be achieved by central regulatory influences and master circuits in the brain. Disruption of such synchronicity, or failure of any of the key components of this system, are common pathophysiological causes of metabolic disorders such as obesity and diabetes.

The most well-studied metabolic sensing region in the forebrain is the hypothalamus, where a number of nuclei such as the arcuate nucleus (ARC) and ventromedial hypothalamus (VMH) express high levels of receptors that bind adipokines (cell signaling molecules secreted by adipose tissue, such as leptin) and gut hormones. Central leptin resistance, one of the major causes of obesity, is caused by defective leptin sensing in these brain regions (Gautron and Elmquist, 2011). In the hindbrain, the nucleus tractus solitarius (NTS)–dorsal motor nucleus of the vagus (DMV) complex is the best-studied brain area with respect to detection of metabolic feedback, especially from the gastrointestinal system via vagal afferents or the circulation (Berthoud et al., 2006).

In this article and the accompanying poster, we provide an overview of current knowledge regarding coordinated system-wide connections that maintain energy homeostasis, highlighting how disruptions in metabolic sensing and integration between the brain, white adipose tissue (WAT) and the gastrointestinal system can contribute to metabolic disorders. Although we focus here on brain–gut–adipose-tissue communication pathways, it is important to acknowledge that other important CNS-peripheral communications (such as those involving muscle and brown adipose tissue) also contribute to energy homeostasis.

During embryogenesis, the parasympathetic ganglia of both the gut and the brain originate from the neural crest. This common origin could explain why many gut-secreted hormones and peptides and their receptors are also found in the brain; indeed, many hormones produced by the gastrointestinal tract are referred to as brain-gut hormones.

Under CNS guidance, the gut is in charge of energy intake. Energy homeostasis can be well maintained only when the brain-gut circuit is synchronized. This synchronization is achieved by neuronal and humoral interaction between the brain and gut. In the gastrointestinal tract, nutrient digestion and absorption trigger the secretion of gut satiety signals such as cholecystokinin (CCK), peptide YY (PYY) and glucagon like peptide-1 (GLP-1) for stimulation of vagal sensory nerves to provide feedback to the brain (Berthoud, 2008). Vagal sensory terminals also express receptors for other gut hormones and peptides that are involved in metabolic regulation. The NTS in the hindbrain is one of the major processors of vagal afferent signals, as evidenced by the finding that upper intestinal lipids inhibit hepatic glucose production through a vagal-NTS circuit (Wang et al., 2008). Additionally, gut hormones and nutrients can influence several specific brain regions and neurons via the circulation. These factors reach these regions and neurons via transporters across either the blood-brain barrier (BBB) and/or less BBB-protected circumventricular organs where neurons can sense blood-borne signals more directly (Cottrell and Ferguson, 2004; Gross, 1992). Among these regions, the NTS–area postrema (AP) complex in the hindbrain and the ARC in the forebrain are two of the major targets of gut hormones; insulin, produced in the pancreas, and ghrelin, produced in the stomach, are among the best studied gut hormones acting in these brain areas (Kojima et al., 1999; Obici et al., 2002; Pocai et al., 2005; Tschöp et al., 2000). In response to gut hormones and nutrient signals, these brain areas relay the information to other key regulatory areas, such as the paraventricular nucleus (PVH) of the hypothalamus (to maintain systemic metabolic homeostasis by modulating neuroendocrine and autonomic outputs) and the suprachiasmatic nuclei [SCN; where the central biological clock is located (Yi et al., 2006)] (to synchronize the behavior and physiology of the whole body and balance the sympathetic and parasympathetic autonomic outflow). Defective or weakened signal feedback by the brain in response to satiety and nutrients from the gut could cause over-feeding and dis-inhibition of liver glucose production, and thereby promote metabolic diseases (Berthoud, 2002).

The autonomic system uses sympathetic and parasympathetic outflow to balance peripheral energy production and expenditure. Pre-autonomic neurons located in the PVH project to the lower autonomic centers in the hindbrain and spinal cord and mediate autonomic outflow to different parts of the gut system. Good examples of these autonomic-center–gut connections are the parasympathetic outflow conveying the central insulin effect on liver glucose production (Obici et al., 2002; Pocai et al., 2005), and the central glucose-sensing effect on liver lipoprotein secretion (Lam et al., 2007). In contrast, activation of the sympathetic nervous system normally antagonizes peripheral insulin sensitivity. In WAT, this antagonism triggers free fatty acid (FFA) release; in liver, it stimulates glucose production via gluconeogenesis and glycogenolysis; in muscle, it increases FFA oxidation, decreases glucose oxidation and glycogen synthesis, and results in reduced glucose uptake. Abnormal elevation of sympathetic outflow could be caused by overactivation of orexin neurons under conditions of sleep deprivation (Yi et al., 2009), lack of inhibition of neuropeptide Y neurons in the ARC in the mediobasal hypothalamus (van den Hoek et al., 2008), or dysregulation of the melanocortin receptor in diverse populations of cells in the forebrain and hindbrain (Skibicka and Grill, 2009).

Feedback from gut hormones to the hypothalamic autonomic control center in the PVH can also influence neuroendocrine output. Most of the neuroendocrine-derived hormones (e.g. epinephrine, glucocorticoids, thyroid hormones and growth hormone) also counteract insulin’s action in the liver and other tissues (Yi et al., 2010).

Adipose tissue not only dynamically accumulates and releases lipids, but also acts as an endocrine organ. Adipose tissue produces adipokines that inform the brain about whole-body long-term energy-storage status, and the tissue drives the brain’s control of energy balance and the long-term regulation of body weight. In terms of adipose tissue crosstalk with the CNS, the best studied adipokine is leptin (Halaas et al., 1995; Maffei et al., 1995), although adiponectin, resistin and apelin are also well studied with respect to their specific signaling pathways and functionalities in the CNS (Banerjee et al., 2004; Kadowaki et al., 2006; Pope et al., 2012; Rajala et al., 2004; Singhal et al., 2007; Steppan et al., 2001). It is not yet clear whether other adipokines, including those recently identified [such as retinol binding protein 4 (RBP4), visfatin and omentin], are involved in adipose-tissue–brain communication.

Leptin sensing in the CNS is necessary for feeding and energy expenditure regulation. It also acts as a natural antagonist of the gut hormone ghrelin. In addition to adipokines, WAT also communicates with the brain via neuronal inputs (Bartness et al., 2010b). The inability to sense leptin signals (leptin resistance) in key areas in the hypothalamus and hindbrain is thought to be one of the underlying mechanisms in the development of obesity. Generally, leptin resistance can cause the brain to misunderstand long-term energy stores and, as a result, initiate extra energy production – i.e. overfeeding (hyperphagia) and glucose production (hyperglycemia), which are typical symptoms of metabolic disorders.

The dominant autonomic-WAT connections are neuronal sympathetic outputs that initiate lipid mobilization in WAT. This sympathetic outflow can be generated from the higher autonomic center in the PVH or the lower centers in the hindbrain and spinal cord. The rate of lipolysis stimulated by this ‘hardwired’ sympathetic outflow (mainly by releasing norepinephrine) is greater than that stimulated by epinephrine released by the adrenal medulla. Loss of autonomic control from the PVH [for instance by lesion of the nuclei (Foster et al., 2010)] can increase the accumulation of lipids in WAT without influencing lipid mobilization. This indicates that obesity could be caused by excess energy deposition owing to malfunction of the sympathetic autonomic nervous system.

Although the CNS as a whole has been well studied with respect to its role in detecting and integrating metabolic signals, the roles of specific cell types, such as astrocytes, have not. Astrocytes are a major class of macroglial cells that occupy up to 50% of the total brain volume. Beyond their classical roles in neuronal proliferation and differentiation, modulation of synaptic efficacy, and maintenance of ionic and neurotransmitter homeostasis, astrocytes also express adipokine receptors, including leptin receptors (Pan et al., 2008). Thus, a defect in adipokine signals to the brain would affect both neurons and astrocytes. For example, in response to a high-fat diet, astrocytes display increased expression of leptin receptors (Hsuchou et al., 2009). In addition, a high-fat diet induces astrocytosis, causing an increase in astrocytic components between the blood vessels and the anorexigenic proopiomelanocortin (POMC) neurons in the ARC of the mediobasal hypothalamus (Horvath et al., 2010). Notably, leptin’s role in directing the inhibition of food intake occurs in part through signaling in POMC neurons (Seeley et al., 1997). In obesity, although leptin concentration is increased in the circulation (hyperleptinemia), the reactive astrocytes with their enlarged ensheathments hamper the POMC neurons’ ability to sense leptin in blood vessels, which might contribute to leptin resistance.

In addition to astrocytes, another important non-neuronal cell type involved in metabolic sensing is the tanycyte, which resides along the wall of the third ventricle. Owing to their expression of glucose transporter 2 (García et al., 2003), tanycytes are thought to be involved in glucosensing from the cerebral spinal fluid in the brain ventricle system.

For accurate and timely monitoring of the peripheral metabolic state, certain areas are developed by the brain to more directly sense the metabolic feedback from the energy input site (the gastrointestinal system) and the energy storing site (WAT). These short-term and long-term inputs are integrated at different levels within the CNS, and are incorporated into neuroendocrine and autonomic outputs to balance energy intake and expenditure. A malfunction at any level in this input-output circuit could contribute to a homeostasic imbalance and contribute to metabolic disorders.

Recent progress has been made in the treatment of metabolic syndrome through a combinational strategy, integrating different gut hormones and/or adipokines, while eliminating their potential side effects (Day et al., 2009). However, there are still many unknowns that make therapeutic strategies problematic. For example, before we completely understand the mechanism of leptin resistance (for which cell specificity and intracellular signaling have yet to be clarified), leptin cannot be used to treat obesity. Another important unanswered question is how we can target the autonomic system to modulate energy homeostasis. In the past, efforts have been made to determine whether there are specific circuits between autonomic neurons and peripheral organs and tissues. For instance, are the liver and pancreas controlled by the same or different groups of pre-sympathetic neurons (Kreier et al., 2006)? Are there neurochemical characteristics specific to each subtype of pre-autonomic neuron that enable their connections to peripheral organs or tissues (Bartness et al., 2010a; Oldfield et al., 2007; Oldfield et al., 2002)? Is it possible to target specific subtypes of pre-autonomic neurons to manipulate their output to specific peripheral organs or tissues? Improved neuroanatomical and neurophysiological methods, combined with genetic tools, will be helpful in addressing these questions and will improve the outlook for anti-obesity therapy.