Drosophila as a model to study obesity and metabolic disease

The regulation of lipid homeostasis is crucial for animals in a wide variety of contexts. Fat reserves are typically tightly regulated to meet energy needs without exceeding a maximum adiposity threshold. In humans, obesity is defined as increased adipose fat accumulation, and typically presents an increased risk to health (WHO fact sheet: Obesity and overweight 2016; www.who.int/mediacentre/factsheets/fs311/en/). Patients with a genetic or environmentally induced excess of fat storage often exhibit hyperglycemia, insulin resistance and cardiovascular disease, hallmarks of metabolic syndrome. Non-alcoholic fatty liver, retinopathy, neuropathy, nephropathy and susceptibility to infection are also increased in obese patients, whereas lifespan is reduced. Even though obesity is an escalating global public health problem, the interactions of genetic predisposition with environmental and lifestyle factors in the etiology of obesity and obesity-related co-morbidities are still poorly understood. Given the inherent genetic heterogeneity of human populations, animal models such as mice and flies are of particular value to disentangle the roles of nurture and nature in fat accumulation and homeostasis.

Researchers began to develop an interest in employing Drosophila melanogaster (hereafter called simply Drosophila) as a model system for obesity research in the early 1960s after the pioneering work of Dr Winifred Doane, who isolated the first obese fly mutant called adipose from a Nigerian wild population. adipose mutant flies suffer from excessive fat storage but reduced carbohydrate reserves (Doane, 1961, 1960a,b). However, it was not until 2003 that the affected adipose gene was identified using a positional cloning approach (Häder et al., 2003). Excitingly, the Adipose protein proved to be structurally and functionally conserved in mice (Suh et al., 2007), humans (Lai et al., 2009) and even plants (Ducos et al., 2017). A recent study of the mammalian Adipose homolog WDTC1 provided the first mechanistic insight into how Adipose might control body fat accumulation. Groh and colleagues showed that WDTC1 functions as a substrate receptor in the Cullin really interesting new gene (RING) E3 ligase complex (Groh et al., 2016). This complex is proposed to mediate the anti-obesity function of WDTC1 by epigenetic silencing of target genes via monoubiquitylation of histone 2A (Groh et al., 2016). Interestingly, Baumbach et al. had previously identified Drosophila Cullin 4 as an anti-obesity gene using a reverse genetic approach (Baumbach et al., 2014a). The finding that the impairment of Cullin 4 in the fly causes an adipose-like mutant phenotype suggests that the molecular function of Adipose/WDTC1 might be conserved between Drosophila and mammals. Thus, adipose/WDTC1 is not only the first example of a human obesity-related gene to be discovered in the fly but also serves as a prime example of how Drosophila research and mammalian studies complement each other in the discovery and mechanistic understanding of obesity and related metabolic disorders.

Drosophila is a particularly useful model for obesity and metabolic disease for a number of reasons. First, flies contain tissues, organs and systems analogous to all those involved in human obesity and associated metabolic diseases (Fig. 1, Table 1). In addition, Drosophila develop obesity and its associated complications during caloric overload, similar to humans. Moreover, most genes and gene families known to function in metabolic disease are conserved between flies and humans (Reiter, 2001). Flies represent an ideal model in which to study the ever-expanding group of complications associated with obesity and metabolic disease owing to their vast range of genetic resources, which can be investigated using cutting-edge approaches. In this Review, we first present organism-level studies of obesity in which flies exhibit a number of phenotypes consistent with a diagnosis of metabolic disease. Then, we discuss the Drosophila tissue and organ systems and tools that researchers use to assess the sequelae of obesity and obesity-related metabolic disease. In closing, we highlight emerging strategies that are being used to study the complications of human obesity in Drosophila, including endocrinology, reproductive and cancer biology.

Triacylglycerols (TAGs) are the main lipid storage form in the fly, as in humans. Quantification of TAG content has therefore been used to define obesity in flies. TAGs can be measured in extracted lipids or crude fly homogenates using thin layer chromatography (Al-Anzi et al., 2009) or mass spectrometry (Carvalho et al., 2012), or estimated by performing an enzymatic assay that measures glycerol content after lipolysis (Hildebrandt et al., 2011; Williams et al., 2011, after McGowan et al., 1983). At the cellular level, obesity can be characterized via quantification of the fat body lipid droplet (LD) size and number (Grönke et al., 2005; Musselman et al., 2011). Some researchers have taken advantage of the altered ratio of fat-to-lean body mass to identify genotypes with increased fat content using buoyancy-based genetic screens (Reis et al., 2010; Tsuda-Sakurai et al., 2015). Lipogenesis can be further characterized by the incorporation of labeled carbons from the diet into various lipid classes using mass spectrometry (Musselman et al., 2013).

One strength of the fly is its established genetic tool kit. Collections of laboratory-generated and naturally occurring genetic variants isolated from all over the world make Drosophila an ideal model in which to test the effects of genes on obesity (Reed et al., 2010; Reis et al., 2010, among others). Genetic screens using these and other resources have identified genes that confer obesity in Drosophila via TAG quantification or buoyancy screening (Baumbach et al., 2014a; Lee et al., 2014; Mosher et al., 2015; Reis et al., 2010). Microscopy-based genome-wide in vitro screens have identified genes that modulate cellular lipid storage in Drosophila tissue culture cells (Beller et al., 2008; Guo et al., 2008). Several conditional in vivo gene expression systems, among which the bipartite GAL4-UAS system is the most popular (Brand and Perrimon, 1993), enable overexpression or ribonucleic acid interference (RNAi)-mediated transgenic gene knock down in a spatially or temporarily restricted manner. Conditional strategies have been used to rapidly identify specific regulators of obesity in a single gene or genes using various reverse genetics paradigms (exemplified by Lee et al., 2014; Pospisilik et al., 2010). Finally, knock-in by homologous recombination (Rong et al., 2002) and gene editing technologies such as transcription activator-like effector nucleases (TALENs) (Beumer et al., 2008) and clustered regularly interspaced short palindromic repeats (CRISPR) (Bassett et al., 2013; Gratz et al., 2013) allow in vivo genome editing with single base resolution and have been successfully employed in genetic studies of Drosophila obesity research (Allen et al., 2017; Gáliková et al., 2015; Sajwan et al., 2015).

Chronic feeding of high-carbohydrate and high-fat diets produces obesity in flies, as in humans, along with a host of pathophysiological complications. Diet-induced obesity can be generated by several strategies in both larvae and adult Drosophila. For high-carbohydrate feeding, high-sucrose diets are the most common (Buescher et al., 2013; Garrido et al., 2015; Havula et al., 2013; Navrotskaya et al., 2016; Musselman et al., 2011; Pasco and Léopold, 2012; Reis, 2016; Rovenko et al., 2015a); however, high-glucose and high-fructose diets have also been used (Rovenko et al., 2015b). For high-fat feeding, coconut oil supplementation is the most common way to elicit diet-induced obesity in flies (Birse et al., 2010; Heinrichsen et al., 2014; Hong et al., 2016; Reed et al., 2010), although lard and the hydrogenated soybean and palm oil product known as Crisco have also been used (Lee et al., 2017; Musselman et al., 2011; Woodcock et al., 2015). The formulation of these diets can be tricky and care must be taken to avoid reduced survival owing to unusually sticky or dry food conditions. A small filter paper wick (1 cm×6 cm) can be used to stabilize many diets, providing a water source and a place to rest. Some investigators have recently begun using an obesogenic high-sugar, high-fat diet in Drosophila obesity studies that is similar to that used in many rodent studies and is probably closest to the typical ‘Western’ obesogenic diet (G. Melkani, personal communication). Diet-induced obesity in flies is associated with many of the pathophysiological consequences found in humans, including hyperglycemia, insulin resistance, cardiac arrhythmia and fibrosis, reduced longevity (Birse et al., 2010; Na et al., 2013) and nephrosis (Na et al., 2015). One can also assess the ability to survive on and process high-calorie obesogenic diets, or the degree of ‘obesity tolerance’. Like humans, some obese flies can be quite healthy (Musselman et al., 2013), whereas other genotypes develop severe effects with only modest levels of overfeeding (Garrido et al., 2015; Nakagami et al., 2003; Teesalu et al., 2017). A collection of wild-caught Drosophila strains exhibited various degrees of diet-dependent obesity, consistent with a model where fat storage represents a balance between optimizing storage depots and overloading the animal (Reed et al., 2014, 2010).

The Drosophila alimentary system is composed of foregut, midgut and hindgut, which are specialized for various functions in both the larva and adult (reviewed in Lemaitre and Miguel-Aliaga, 2013). Food and water enter the foregut after ingestion and travel toward the posterior of the animal. In both developmental stages, the proventriculus represents the beginning of the midgut and stores a striking amount of TAG. The gastric cecae (and crop in adults) contribute to nutrient digestion along the anterior and middle midgut with most absorption occurring with the help of the microvilli lining the posterior midgut (Buchon and Osman, 2015). Specialized sub-regions of the Drosophila alimentary tract exist, as in humans, and have been characterized by biochemical and gene expression studies, including regional TAG accumulation (Buchon and Osman, 2015; Gutierrez et al., 2007; Harrop et al., 2014, and references therein). Kidney-like Malpighian tubules link to the intestinal tract at the midgut/hindgut junction and function in filtration and excretion, whereas the hindgut reabsorbs ions and water. With respect to obesity, the gut lumen is of importance owing to two interconnected functions: as the exclusive route for energy intake and as the interaction site of the intestinal microbiome with the fly host.

The gut absorbs dietary macronutrients, including sugars, proteins and fats. The fly midgut serves as the major site of dietary lipid absorption and also metabolizes both glucose and lipids into metabolic intermediates, which are loaded into the hemolymph for use in other tissues and organs. Therefore, the gut is crucial for peripheral body fat storage, as shown in several studies. Lipoprotein complexes containing highly conserved apolipoproteins (called apolipophorins in Drosophila) carry sterols and diacylglycerols from the gut to other tissues (Palm et al., 2012). Fly lipoproteins also contain the signaling molecule Hedgehog, a cholesterol-linked, gut-derived ligand that binds the transmembrane receptor Patched on fat body target cells to promote lipolysis during larval starvation (Palm et al., 2013; Rodenfels et al., 2014). Other gut-expressed factors also contribute to the control of systemic lipid homeostasis. Gut enteroendocrine cells secrete tachykinins, a group of peptide hormones that stimulate gut contraction and lipid catabolism during starvation (Song et al., 2014). The nuclear receptor DHR96 promotes midgut absorption and fat body storage of dietary fat, at least in part via transcriptional activation of the gastric lipase encoded by the magro gene (Sieber and Thummel, 2009). This study also demonstrated that the human anti-obesity drug orlistat, a gastric lipase inhibitor, is able to reduce body fat accumulation in adult flies. Supporting a crucial role for lipolysis, midgut lipid accumulation and global fat storage are reduced by the insulin signaling pathway inhibitor Foxo in enterocytes, via reducing the expression of magro as flies age (Karpac et al., 2013). Excessive lipid accumulation in the fly gut and fat body is also a feature of ‘humanized’ flies upon cross-species expression of the human peptide neurotensin in Drosophila midgut enteroendocrine cells. Obesity in these flies is triggered by an evolutionarily conserved mechanism acting via the cellular energy sensor 5′ adenosine monophosphate (AMP)-activated protein kinase (Li et al., 2016a). In addition, the acidic pH of the gastric lumen may be important for fly obesity given that both global vacuolar-type H⁺-adenosinetriphosphatase (ATPase) mutants and flies treated with pharmacological inhibitors of alimentary acidity accumulate extra fat (Lin et al., 2015b). This effect could be mediated via the gut microbiome, which both shapes and depends upon the acidity of the gut (Overend et al., 2016). Collectively, these data emphasize the importance of gut physiology for fat homeostasis in Drosophila and highlight the intricate interaction between the gut epithelium and the gut microbiome.

As in humans, the Drosophila gut hosts a complex microbiome enriched in Lactobacillus and Acetobacter species. Both diet and immunity play a role in controlling the composition of the microbiome, affecting fly health and metabolism (reviewed in Wong et al., 2016). Adult axenic flies overstore fats under various dietary conditions compared with flies with natural gut microbiota (Wong et al., 2014). Lactobacillus sp. abundance promotes co-colonization by Acetobacter sp. in the adult gut, which in turn negatively correlates with the fat storage level of the fly (Newell and Douglas, 2014). The composition of the gut microbiome of adult Drosophila corresponds to body fat content and depends on the host genotype (Chaston et al., 2016). The fly diet impacts the composition of the microbiota in the gut because a high-sugar diet shifts the gut microbiome to uracil-producing species, which promote fat storage and growth in Drosophila larvae (Whon et al., 2017). The availability of dietary glucose to the adult fly depends on the microbiome because flies with commensal Acetobacter tropicalis eat more than axenic flies but store less TAG owing to the consumption of dietary sugar by the bacteria (Huang and Douglas, 2015). Collectively, these data demonstrate that the gut microbiota and its metabolism modulate fat storage in the fly.

A range of tools exists for studies of the Drosophila gastrointestinal tract. The fly gut is beautifully suited to whole-mount immunohistochemistry and easy to isolate for quantitative biochemistry techniques. Likewise, the GAL4/UAS binary expression system described in whole organism studies enables overexpression, including RNAi, in specific gut cell subpopulations (Table 1). The fly eubacterial microbiome is often characterized by performing 16S ribosomal RNA sequencing and can be manipulated experimentally using axenic culture or culture with a defined microbial composition (Chaston et al., 2016).

The insect fat body represents the central metabolic hub in body fat storage control. The fat body accumulates the vast majority of body fat during development and caloric overload, and executes lipolysis when energy from stored nutrients is needed for survival during metamorphosis, starvation or egg production. The fat body takes up lipids from the fly blood or hemolymph and esterifies them as stored TAGs and cholesterol esters when nutrients are abundant. In addition, the fat body carries out glycolysis and lipogenesis using carbohydrates in the blood (reviewed in Arrese and Soulages, 2010).

Fly obesity research centers on fat body studies at two different developmental stages: late larval stages and the adult fat body. The larval and the adult fat body are composed of two different cell lineages. The larval fat body represents a contiguous organ of a fairly constant number of postmitotic, large endoreplicative cells, which undergo histolysis shortly after the adult fly ecloses (Aguila et al., 2013, 2007). In contrast, the adult fat body is believed to be composed of diploid cells derived from cell clusters in the larval body wall and from adepithelial cells of the imaginal discs (Hoshizaki et al., 1995). The adult fat body cells are allocated to various unconnected adipose tissue depots such as the abdominal subcuticular fat body, the head fat body and the pericardial fat body. The ontogenetic origin and the cellular composition of adult adipose tissue depots are fairly ill-defined; however, evidence is accumulating that both hypertrophy and hyperplasia contribute to the plasticity of adult adipose tissue with respect to fat storage capacity in obese flies (DiAngelo and Birnbaum, 2009). Nonetheless, fat bodies at both stages play the same crucial role in fat storage.

The main cellular characteristics of both the larval and the adult fat bodies are the LDs, the universal organelles of intracellular fat storage, which are routinely stained by lipophilic dyes such as Oil Red O, BODIPY493/503 or Nile Red for microscopic imaging. These dyes are also used, albeit less frequently, to detect neutral lipids in other tissues. Fly fat body cells are multilocular, meaning that they contain LD populations of different sizes that belong to at least two distinct functional classes (Wilfling et al., 2013), which differ in their properties owing to differential association with particular sets of LD-associated proteins. Mass spectrometry-based characterization of the fly LD proteome at different developmental stages (Beller et al., 2006; Cermelli et al., 2006; Krahmer et al., 2013) revealed a surprising complexity that is consistent with functional heterogeneity among LDs that varies in relation to the metabolic needs of the animal (reviewed in Farese and Walther, 2009; Walther and Farese, 2012).

Storage lipid mobilization from the LDs depends on the adipokinetic hormone (Akh), which circulates in the hemolymph. The Akh pathway is functionally analogous to mammalian glucagon signaling and represents the most important lipocatabolic pathway required during fasting conditions (Fig. 2). Akh operates via transcriptional and posttranslational mechanisms. The binding of the Akh peptide to its cognate G protein-coupled receptor (Akh receptor; AkhR) (Bharucha et al., 2008; Grönke et al., 2007) on the fat body cell surface triggers canonical cyclic AMP (cAMP)/protein kinase A (PKA) signaling (Fig. 2, green components). Increased cAMP promotes the activity of the cAMP-responsive element binding (CREB) transcription factor (Fig. 2). Consistently, targeted downregulation of CREB specifically in the fat body causes overeating and obesity in flies (Iijima et al., 2009). The cAMP-dependent PKA activation orchestrates at least two different pro-lipolytic responses. On the one hand, PKA-dependent phosphorylation of the LD-associated protein Perilipin 1 (Plin1) is believed to promote the access of TAG lipases to the LD surface (Patel et al., 2006, 2005), thereby allowing storage lipid mobilization (Fig. 2). On the other hand, phosphorylation by PKA inhibits salt-induced kinase 3 (SIK3) (Wang et al., 2011), which results in hypophosphorylation and the consequent nuclear translocation of histone deacetylase 4 (HDAC4) (Wang et al., 2011; Choi et al., 2015). In the nucleus, HDAC4 deacetylates the starvation-induced transcription factor Foxo to boost its transcriptional activity. An important transcriptional target of Foxo is brummer (bmm) (Grönke et al., 2005), which encodes the fly homolog of adipose triglyceride lipase (ATGL), the major mammalian TAG lipase. Transcriptional regulation of bmm adjusts body fat content in a dosage-dependent manner given that bmm gain-of-function flies are lean, whereas bmm mutants represent a fly obesity model (Fig. 2). Given that the nuclear localization of Foxo is also associated with reduced insulin/insulin-like growth factor signaling (IIS), the activated Akh and the downregulated IIS pathways mechanistically converge under fasting conditions to promote Foxo-dependent transcription and lipocatabolism.

In addition to canonical cAMP/PKA signal transduction, the Akh pathway signals using intracellular calcium (iCa²⁺) as a second messenger to control fly body fat storage by using a mechanism called store-operated calcium entry (SOCE) (Fig. 2; red components). Akh binding to AkhR triggers an inositol trisphosphate (IP3) second messenger response, which is relayed by small Gg and Gq protein subunits and phospholipase C (Plc21C) (Baumbach et al., 2014b). At the endoplasmic reticulum (ER) membrane, IP3 binding to the IP3 receptor (IP3R; called Itp-r83A in the fly) causes Ca²⁺ efflux, which is sensed by an ER calcium sensor known as the Stromal interaction molecule (Stim, Fig. 2). Stim, in turn, interacts with the plasma membrane calcium channel Orai (called Olf186-F in the fly) to elevate iCa²⁺ levels by extracellular calcium influx. Elevated iCa²⁺ concentrations are translated to a plethora of cellular responses by cytoplasmic Ca²⁺-dependent proteins such as Calmodulin (Cam). At the ER membrane, the Sarco/ER Ca²⁺-ATPase (SERCA; also called Ca-P60A in the fly) antagonizes the Ca²⁺ efflux mediated by the action of IP3R and of the related Ryanodine receptor (RyR) to balance iCa²⁺ homeostasis and to aid the termination of G protein-coupled receptor-triggered iCa²⁺ signaling (Fig. 2). Accordingly, genetic manipulation of these factors modulates fat body TAG content in adult flies (Subramanian et al., 2013; Baumbach et al., 2014a,b; Pospisilik et al., 2010). SERCA appears to be a central control element of body fat storage control because impairment of dSeipin, a positive regulator of SERCA encoded by the fly homolog of the human gene, which is affected in Berardinelli–Seip congenital lipodystrophy (Magré et al., 2001), causes lean flies (Bi et al., 2014). By contrast, flies lacking THADA, a SERCA uncoupler encoded by the fly homolog of the human Thyroid Adenoma Associated gene, are obese (Moraru et al., 2017). Impairment of iCa²⁺ homeostasis is associated with various tissue-autonomous phenotypes such as reduced lipogenesis and mitochondrial dysfunction. Given the pleiotropic roles of calcium in metabolism (reviewed in Arruda and Hotamisligil, 2015), further research is required to achieve a comprehensive understanding of how iCa²⁺ modulates lipid storage. Remarkably, the role of SOCE in fat storage control has been recently recapitulated in mammals (Maus et al., 2017).

The catabolic Akh pathway is antagonized by the anabolic IIS pathway. Increasing IIS signaling via insulin-like peptides (Ilps) and sugar in the hemolymph promotes fatty acid biosynthesis and esterification by the fat body. Insulin resistance in flies also leads to obesity in some contexts. For example, loss of the insulin receptor substrate Chico or the Chico binding protein dSH2B reduces IIS signaling and increases obesity in flies and mammals (Böhni et al., 1999; Song et al., 2010). Lipogenesis is regulated by a number of factors and substrates other than IIS, including acetyl-CoA and the activated carrier NADPH (Teesalu et al., 2017). The conserved transcription factors sterol regulatory element-binding protein (SREBP) and carbohydrate-responsive element-binding protein (ChREBP; also called Mondo and Mio in the fly) promote the expression of lipogenic enzymes such as fatty acid synthase, stearoyl-CoA desaturase and diacylglycerol O-acyltransferase 1 (DGAT1; also called Midway in flies), which all increase triglyceride storage (Buszczak et al., 2002; Garrido et al., 2015; Havula et al., 2013; Kunte et al., 2006; Musselman et al., 2013; Sassu et al., 2012). Other genes required for lipogenesis in the fat body include those encoding the phosphatidate phosphatase dLipin (Schmitt et al., 2015; Ugrankar et al., 2011), pantothenate kinase and phosphopantothenoylcysteine synthase (Musselman et al., 2016) and Myc (Parisi et al., 2013). These represent potential targets for the development of anti-obesity therapeutic strategies, which can be modeled in the fly.

In addition to fat body-autonomous mechanisms of body fat storage control, the communication between the fat body and other fly organs/tissues is a research area that has seen recent progress (Fig. 1). For example, Unpaired 2 was identified as a leptin-related adipokine, which signals from the fat body to the central brain by modulating Ilp secretion from insulin-like peptide-producing cells (see below) in response to dietary carbohydrate or lipid (Rajan and Perrimon, 2012). Similarly, the Stunted peptide acts as an insulinotropic adipokine in response to dietary amino acids (Delanoue et al., 2016), and the Drosophila tumor necrosis factor-α adipokine Eiger represses Ilp production in response to starvation (Agrawal et al., 2016). Yet, adipokine-mediated communication emanating from the fat body is not restricted to the brain. Upon food withdrawal, the adipose secretes Ilp6, which causes neutral lipid accumulation and processing in the so-called oenocytes (Chatterjee et al., 2014).

Oenocytes (‘wine cells’) are goblet-shaped fat body-associated hepatocyte-like cells that are found segmentally alongside both the larval and adult cuticle (Makki et al., 2014). Larval oenocytes control storage lipid turnover and energy homeostasis in peripheral tissues during fasting (Gutierrez et al., 2007). Adult oenocytes carry out the synthesis of very long chain fatty acids, which fuel the formation of cuticular hydrocarbons that protect the fly from desiccation and act as pheromones (Wicker-Thomas et al., 2015). Interestingly, these authors found that an obesogenic diet decreased cuticular hydrocarbon production whereas fatty acid synthesis and lipophorin receptor expression in the fat body promoted cuticular hydrocarbon synthesis.

Tools to study the fat body by tissue-specific modulation of gene expression are described in Table 1. The larval fat body is also particularly well-suited for clonal analysis, which provides a powerful method to discriminate between cell-autonomous and non-autonomous gene functions with respect to obesity (reviewed in Griffin et al., 2014). Drosophila uses the flippase/flippase recognition target (Flp/FRT) recombination system to generate single homozygous mutant clones in otherwise heterozygous, chimeric fat bodies. Clonal analysis has been successfully employed to detect autonomous roles of Plin1 (Beller et al., 2010), dSeipin (Bi et al., 2014) and the lipogenesis enzyme encoded by the dLipin gene (Schmitt et al., 2015) in fat body lipid storage droplets. Given the role of the fat body as the central hub in inter-organ communication, in vivo monitoring of signaling pathway activities is particularly valuable. Examples are the tGPH sensor (a fusion protein of GFP with the pleckstrin homology domain of the Drosophila homolog of the general receptor for phosphoinositides-1 expressed under the control of the Drosophila β-tubulin promoter), used to monitor phosphoinositide 3-kinase signaling (Britton et al., 2002), or the calcium-dependent nuclear import of LexA (CaLexA) system, used to monitor iCa²⁺ second messenger signaling (Masuyama et al., 2012). Lipid turnover in larval oenocytes and the fat body has been monitored by coherent anti-Stokes Raman scattering (CARS) microscopy (Chien et al., 2012). This label-free method of LD analysis shows great promise because it is capable of qualitative as well as quantitative lipid composition analyses.

The fly heart (also called the dorsal vessel, Fig. 1) can exhibit arrhythmic beating, fibrosis and cardiac failure, which is exacerbated during obesity, similar to the diseased human heart (Diop et al., 2015; Hardy et al., 2015; Heinrichsen et al., 2014). Although their ontogeny differs, the Drosophila heart is an innervated multichambered tube that circulates the hemolymph via an open circulatory system. Hemolymph, like blood, contains a variety of important hormones and metabolites needed by peripheral tissues and, therefore, the heart plays a key role in metabolic homeostasis. Heart function can be measured by high-speed video analysis of the heart rate in vivo, also known as Semi-automated Optical Heartbeat Analysis (SOHA), enabling quantification of systolic and diastolic strength, heart rate, heart period and arrhythmia (Cammarato et al., 2015). Such studies of heart function have been used to identify protective roles for cardiac proliferator-activated receptor-γ coactivator-1 (PGC-1; also called Spargel in the fly) and Bmm (Diop et al., 2015) and a deleterious role for the cardiac target of rapamycin (TOR) pathway (Birse et al., 2010) in flies with diet-induced obesity and cardiovascular disease.

Obese Drosophila exhibited increased cardiac steatosis and fibrosis compared with control flies, supporting an analogous burden to the insect and human hearts (Birse et al., 2010; Diop et al., 2015; Hardy et al., 2015; Na et al., 2013). Heart failure rate is highest in obese flies fed high-calorie diets and lowest in lean low-calorie-fed flies (Bazzell et al., 2013). The accumulation of fat in the fly heart is associated with obesity and an increased heart failure rate in fatty acid transport protein mutants (Sujkowski et al., 2012). Finally, the heart itself can control circulating and stored lipids in a non-autonomous manner via the regulation of lipoprotein metabolism. Cardiomyocyte-specific knockdown of the microsomal triglyceride transfer protein or its target lipoprotein, Lpp, prevents fat body TAG accumulation during high-fat diet-induced obesity (Lee et al., 2017).

Other tools to study heart structure and function aside from SOHA and the lipid assays described above have been developed over recent years. Fibrosis can be visualized in whole-mount hearts using confocal microscopy with actin-phalloidin staining or immunohistochemistry for extracellular matrix proteins such as collagens (especially pericardin), perlecan and the ADAMTS (a disintegrin and metalloprotease with thrombospondin repeats) homolog Lonely heart (Chartier et al., 2002; Drechsler et al., 2013). Stimulation-induced heart failure can also be measured in whole flies using external electrical pacing (Wessells et al., 2004).

Although Drosophila skeletal muscle and flight muscle have both been used extensively to understand muscle physiology, the role of muscles in lipid homeostasis is less well understood. High-calorie obesogenic diets reduced endurance in a repetitive climbing assay (Bazzell et al., 2013), considered to be a measure of both heart and muscle function. Increased muscle activity using a gentle exercise paradigm (approximately 2 h exercise per day over 5 days) led to reduced TAG content, enabling researchers to compare the responses of a range of genotypes to exercise-induced reduction in obesity (Mendez et al., 2016). These so-called ‘gene by environment’ interaction studies (with environment including parameters such as diet and exercise) are a particularly attractive feature of Drosophila research.

The fly muscle also plays a functional role in lipid homeostasis and obesity. Interestingly, muscle Foxo is able to promote fat body lipogenesis remotely, inhibiting Akh expression via the secreted myokine Unpaired 2 (Zhao and Karpac, 2017). Similarly, the muscle acts as an endocrine regulator of lipid storage in the fat body by secreting activin, a member of the transforming growth factor beta (TGF-β) cytokine family (Song et al., 2017). Using a muscle-targeted driver to express transgenic RNAi, a genome-wide screen for obesity genes revealed 149 additional putative muscle-specific regulators of systemic TAG content in adult flies, including nicotinamide adenine dinucleotide (NADH)-ubiquinone oxidoreductase, pyrroline 5-carboxylate reductase, inositol-requiring enzyme-1, nucleosome remodeling factor-38, the octopamine receptor, ornithine decarboxylase and vascular endothelial growth factor-related factor 1 (Pospisilik et al., 2010). A smaller RNAi screen targeting secreted proteins in the muscle identified Wingless, Ilp4, Angiotensin-converting enzyme, Unpaired 3 and Diptericin B as genes conferring reduced systemic TAG content, possibly by endocrine regulation of the fat body (Lee et al., 2014). Muscle AMP-activated protein kinase promotes adiposity in Drosophila larvae by phosphorylation of the myosin regulatory light chain, enabling peristalsis by the gut musculature and, presumably, improved nutrient uptake (Bland et al., 2010).

Tools to study the structure and function of muscle during obesity include immunohistochemistry, exercise endurance, electrophysiology of the flight muscle and the neuromuscular junction, and tissue-specific modulation of gene expression (Table 1).

Feeding (and thus energy intake) is a contributor to obesity that depends on the fly nervous system (Branch and Shen, 2017). Larvae feed almost constantly and live immersed in the laboratory diet until they leave the food in preparation for metamorphosis. In contrast, adults feed intermittently, with mated females eating more than males to support oogenesis, and adult feeding cycles under the control of the circadian clock (Xu et al., 2008). The Drosophila central nervous system (CNS) executes food intake control using various mechanisms, which can be studied using a wide range of methods. Food labeling by dyes or radioactivity, food uptake quantification by the so-called capillary feeder assay (Ja et al., 2007), proboscis extension reflex assay (Shiraiwa and Carlson, 2007), and automated food intake monitoring using the flyPAD (Itskov et al., 2014) or Fly Liquid-Food Interaction Counter (Ro et al., 2014) can be used to quantify feeding. Food preference assays are used to determine important feeding parameters such as total intake, meal size, meal frequency and feeding motivation (reviewed in Branch and Shen, 2017; Deshpande et al., 2014). Collectively, these parameters determine energy intake, which is crucial for the development of obesity.

A genetic screen approach combined the silencing of various neuronal populations with body fat measurements to locate obesity control centers in the adult fly brain (Al-Anzi et al., 2009). Importantly, this study identified groups of neurons that control body fat levels via feeding and/or metabolic rate control, demonstrating coordinated regulation of energy intake and expenditure by the brain. At the molecular level, a number of signaling molecules act to control feeding in the brain. One such cytokine is Unpaired 1 (Upd1), a leptin-related protein that controls food intake and body weight gain in response to a high fat diet (Beshel et al., 2017). Upd1-expressing neurons suppress the secretion of neuropeptide F (NPF), the fly analog of the mammalian orexigenic neuropeptide Y. A recent study suggests that inhibition of the neuronal orexigenic NPF in the CNS is also mediated by the adipose-derived metabolite tetrahydrobiopterin (Kim et al., 2017). Close communication between the fat body and the CNS in feeding control is also supported by the neuronal upregulation of the orexigenic small NPF, which is associated with the hyperphagia of obese flies caused by reduced iCa²⁺ signaling in the fat body (Baumbach et al., 2014a). In addition to these neuropeptides (reviewed in Schoofs et al., 2017) and metabolites, neurotransmitters/neurohormones such as octopamine (the fly equivalent of norepinephrine) are instrumental for body fat control in the CNS. Lack of octopamine causes a reduced metabolic rate and increased obesity in flies (Li et al., 2016b).

Two groups of key neurosecretory cells play important roles in the systemic control of fat and carbohydrate storage: corpora cardiaca (CC) cells, which are also known as adipokinetic hormone producing cells (APCs), and median neurosecretory cells (MNCs), known as insulin-like peptide producing cells (IPCs) (Fig. 1). The APCs belong to the Drosophila neuroendocrine system and have central catabolic functions in fly physiology similar to those of pancreatic α-cells in human physiology. APCs originate from the embryonic dorsal neuroectoderm (Wang et al., 2007), become part of the ring gland during larval stages, and eventually relocate to the posterior foregut in the adult fly (Gruber et al., 2013). Genetic ablation of APCs by targeted expression of pro-apoptotic genes causes lipid and sugar homeostasis defects in larvae (Kim and Rulifson, 2004; Lee and Park, 2004) and in adult flies (Grönke et al., 2007; Isabel, 2004). This is consistent with APCs being the exclusive source of Akh, a fly glucagon-like neuropeptide that is secreted in response to a drop in circulating sugars (Kim and Rulifson, 2004). Conditional blocking of APC secretion by targeted expression of the tetanus toxin light chain (Sweeney et al., 1995) inhibits the Akh-dependent release of Ilp3 from larval IPCs (Kim and Neufeld, 2015). Conversely, APC hyperpolarization caused by targeted expression of the heat-sensitive TrpA1 channel (Hamada et al., 2008) triggers Akh peptide secretion under permissive temperature conditions, which in turn causes storage fat depletion in flies (Waterson et al., 2014). However, Akh mutant flies are obese and hypoglycemic but lack a pre-adult metabolic phenotype (Gáliková et al., 2015), which suggests some caution is advisable concerning the functional specificity of cell ablation or global manipulation of neurosecretory cells. In line with this, APC cells were reported to also express and secrete Limostatin, which represses Ilp in IPCs (Alfa et al., 2015), illustrating the close interaction between APC and IPC cells in Drosophila metabolic control.

IIS signaling is highly conserved between flies and humans. Approximately 14 MNCs are known to be IPCs. IPCs are found in the anterior brain and consist of seven cell bodies in each brain lobe, positioned bilaterally in both the larva and the adult. Ilps, which exhibit structural similarity to human insulin, bind and activate a highly conserved insulin receptor that can also be activated by insulin (Nässel et al., 2015; Sajid et al., 2011; Sekine et al., 2010). Accordingly, IPCs and Ilps regulate metabolic homeostasis and obesity in a manner analogous to that of the pancreatic β-cell; thus, IPC ablation leads to hyperglycemia (Rulifson et al., 2002). IPCs project long axons onto the gut, heart and APCs to deliver their cargo, and receive input from a number of other cells in addition to their primary input of increased circulating nutrient concentration. As is the case for insulin secreted by pancreatic β-cells, feeding promotes Ilp expression and secretion from brain IPCs, although larval IPCs respond to elevated concentrations of protein (in particular the amino acid leucine) rather than sugar (Géminard et al., 2009; Ikeya et al., 2002; Manière et al., 2016). Early genetic ablation studies targeting IPCs by expression of cell death effectors did not reveal major effects on obesity (Broughton et al., 2005; Rulifson et al., 2002). Recently, more nuanced studies have shown that these cells play roles in controlling lipid storage. Conditional ablation of IPCs during adulthood elicits modest increases in stored and circulating TAGs (Haselton et al., 2010). The IPC-expressed gene nudt3 is a human obesity susceptibility gene that is required in IPCs for lipolysis of stored fat during starvation (Williams et al., 2015). Similarly, the serotonin receptor 5-HT1A is required in IPCs for the catabolism of stored fat to supply energy during starvation (Luo et al., 2012). Reduction of the fly adiponectin receptor in IPCs also led to increased TAG content, consistent with a role for these cells in the control of obesity (Kwak et al., 2013). IPC expression of RNAi targeting the lipogenic transcription factor ChREBP increased the feeding rate, although there was no effect observed on obesity (Docherty et al., 2015). Little is known about whether IPCs produce other peptides, which makes ablation studies difficult to interpret. Drosulfakinin is a cholecystokinin-like satiety hormone that is expressed in IPCs and several other neurons, consistent with a complex role for these cells in the regulation of obesity (Söderberg et al., 2012).

Future studies on fly obesity should greatly benefit from up-and-coming techniques that allow acute and reversible organ-specific genetic manipulation to resolve causality and to limit compensatory regulation underlying energy homeostasis. Particularly promising in this respect are in vivo CRISPR transcriptional activation (CRISPRa) and interference (CRISPRi) approaches (Ghosh et al., 2016; Lin et al., 2015a), which allow tightly regulated and reversible promoter activation and blocking, respectively.

Small molecule screens using whole Drosophila (Gasque et al., 2013; Men et al., 2016b) or cells (Tschapalda et al., 2016) to identify anti-obesity agents or anorectic drugs have been successful. Moreover, advances in analytical technology increasingly compensate for the disadvantage of the fly being a small model organism. Some examples of these analytical technologies are direct analysis in real time mass spectrometry in whole flies (Chiang et al., 2016), tissue- and hemolymph-specific mass spectrometry analyses (Chintapalli et al., 2013; Musselman et al., 2016), in situ lipid profiling in Drosophila sections using matrix-assisted laser desorption/ionization mass spectrometry imaging for comprehensive lipid composition analysis of individual fly organs (Niehoff et al., 2014), and the use of optical coherence tomography for non-invasive monitoring of Drosophila heart function (Men et al., 2016a).

Interestingly, an increasing number of obesity-related complications are being modeled in the fly, such as obesity-associated bipolar disorder (Williams et al., 2016) and diet-dependent metastasis of transformed cells (Hirabayashi et al., 2013; Hirabayashi and Cagan, 2015). Drosophila is also being used to study the pathophysiological consequences of aberrant lipid accumulation in non-storage tissues. For example, glial cells accumulate an increased number of LDs in response to neuronal mitochondrial dysfunction, which promotes neurodegeneration in both the fly and the mammalian brain (Liu et al., 2015). Reproduction also requires tight control of lipid homeostasis. Although females must increase oocyte and whole-animal lipogenesis to support the production of offspring (Sieber and Spradling, 2015), morbid maternal obesity adversely affects egg size, glycogen and TAG content, adult mass and gene expression in offspring and even in the F2 generation (Buescher et al., 2013; Dew-Budd et al., 2016; Matzkin et al., 2013). Highly obesogenic diets reduce fecundity (Brookheart et al., 2017; Matzkin et al., 2013, 2011) as does genetically induced obesity (Palu et al., 2017). Interestingly, obesity can affect the metabolic state of the offspring through the male Drosophila germline via effects on histone H3 methylation, a mechanism that appears to be conserved in humans (Öst et al., 2014).

Collectively, more than 100 years after its first use as a laboratory model organism, Drosophila is ready to answer current basic research questions and to help in fighting the pandemic human health threats of the 21st century that comprise obesity and obesity-related metabolic diseases.

The authors thank Hartmut Sebesse (Max Planck Institute for Biophysical Chemistry) for providing graphical design templates, and artist Ty Whitbeck for anatomical drawings.