Genetics of Fat Storage in Flies and Worms: What Went Wrong?

Body weight and fat storage are strongly influenced by an individual’s genetic makeup. In humans, genetic polymorphisms have been identified that have effects on body mass index (BMI) and fat content (Meyre et al., 2009; Speliotes et al., 2010; Choquet and Meyre, 2011a), and studies of monogenic rodent models of obesity have defined a variety of genes and signaling pathways that control fat storage and metabolism (Barsh and Schwartz, 2002). However, many other genes that regulate these processes undoubtedly remain to be discovered. Although forward genetic screens in the mouse have the potential to identify new obesity genes, such screens are expensive and lengthy endeavors. The fruit fly Drosophila melanogaster and the nematode (roundworm) Caenorhabditis elegans have provided powerful genetic models for establishing the genetic foundations of many biological phenomena. Forward genetic screens in flies and worms can be performed quickly and inexpensively, so the use of these model systems for obesity research has the potential to accelerate the rate of gene discovery. At least two fat-regulating genes first identified in Drosophila were also shown to affect fat content in mice (Suh et al., 2007; Pospisilik et al., 2010). These genes have not yet been examined in the context of human obesity. In Drosophila, the most frequently used method for measurement of fat is a coupled colorimetric triglyceride assay kit (CCA). In C. elegans, the typical methodology for high-throughput screening involves mixing Nile Red, a vital dye that stains lipid droplets, with the E. coli bacteria on which the worms feed. Large RNAi screens have been performed using these assays. In total, such screens have identified about 500 Drosophila and 400 C. elegans genes for which RNAi perturbation affects fat storage (Ashrafi et al., 2003; O’Rourke et al., 2009; Pospisilik et al., 2010). Among the conserved molecules that would be expected to be involved in fat storage based on mammalian studies, the Drosophila screen identified members of the TOR pathway and mediators of glucose and lipid mobilization, and the first C. elegans screen identified a serotonin receptor and the Tubby homolog. However, the vast majority of the genes found in these screens had not been previously implicated in control of fat storage and metabolism, and therefore represent new targets for future analysis. Unfortunately, recently published results show that the high-throughput assays used in these screens do not provide accurate assessments of stored fat content. Thus, it is unclear whether the new genes identified in the screens are genuine regulators of fat deposition and metabolism. The CCA relies on hydrolysis of triglycerides by a lipoprotein lipase to generate glycerol, which is then phosphorylated and oxidized to produce hydrogen peroxide. The amount of hydrogen peroxide generated from glycerol is read out using a peroxidase reaction to generate products called quinoneimines that absorb light in the visible range. Although CCA provides accurate readings when used to measure serum triglycerides, which are components of soluble lipoprotein complexes, it is unsuitable for measuring stored fat in tissue samples. First, it fails to give accurate readings when tested on insoluble triglyceride mixtures that resemble stored fat within cells, or on purified fly fat (Van Veldhoven et al., 1997; Al-Anzi and Zinn, 2010). This is because the lipases used in the kits cannot efficiently cleave the fatty acid–glycerol bonds in insoluble lipid mixtures. (Reference Van Veldhoven et al., 1997 shows that resuspension of stored fat samples from mammalian cells with a specific detergent can improve cleavage and allow more accurate assessment of insoluble fat; this method has not been applied to Drosophila.) Second, CCA actually measures the phosphorylatable glycerol backbones present in the samples being evaluated. Cleavage of non-fat glycerides such as monoand di-glycerides by the assay kit lipases also generates glycerol. Thus, when CCA is used to evaluate tissue extracts, the observed signal actually represents the amount of free glycerols in the tissue together with all types of glycerides (Van Veldhoven et al., 1997; Al-Anzi and Zinn, 2010). Third, when CCA is used on whole crushed Drosophila samples, the signal-tonoise ratio is influenced by compound eye pigments, whose absorption spectra overlap with those of the peroxidase reaction products that are measured as the final output of the assay (Al-Anzi and Zinn, 2010). Very recently, a response to Al-Anzi and Zinn (2010) was published that defended the use of CCA for evaluation of fat content in Drosophila (Hildebrandt et al., 2011). These authors showed that strains that are higher in fat content can generate larger CCA signals, and noted that CCA readings can be corrected for eye pigment absorption and free glycerol differences by measuring blank reactions in which lipase and/or peroxidase substrates are omitted (this was not done in the published highthroughput screen (Pospisilik et al., 2010).) However, even when these corrections are performed, CCA still cannot accurately measure insoluble fat and cannot distinguish between triglycerides, diglycerides, and monoglycerides (Van Veldhoven et al., 1997; Al-Anzi and Zinn, 2010). In any large-scale CCA screen, there will be both false positives (strains that give different CCA readings from controls but do not actually differ in triglyceride content) and false negatives (strains that do not differ from controls when assessed by CCA, but which actually are different in triglyceride content). False positives can be eliminated using other assays for secondary screening, but false negatives cannot be detected, and these genes will simply be missed. For the vital Nile Red assay, four papers showed that when Nile Red is fed to live C. elegans, it stains lysosome-related organelles (LROs), not lipid droplets (Brooks et al., 2009; O’Rourke et al., 2009; Yen et al., 2010; Zhang et al., 2010). (However, a recent chemical screening paper argued that live Nile Red staining does report on