Atherosclerosis remains the leading cause of death (Lozano et al., 2012). The major driver of atherosclerosis is increased circulating cholesterol-rich lipoprotein particles (Stamler et al., 1986). Statin drugs, which decrease hepatic cholesterol synthesis and thereby promote hepatic cholesterol-rich lipoprotein particle clearance, are potent reducers of the risk of death from atherosclerosis (Cholesterol Treatment Trialists' (CTT) Collaboration et al., 2015). Nevertheless, these drugs do not fully ameliorate the risk of ischemic cardiovascular events, ultimately reflecting gaps in understanding of the drivers of atherosclerotic progression.

The persistence of TAG-rich lipoproteins in the post-prandial state has emerged as a major risk factor for atherosclerosis (Jørgensen et al., 2014; The TG and HDL Working Group of the Exome Sequencing Project et al., 2014). Namely, in large prospective cohort studies, risk of atherosclerosis was found to be attributable to the non-fasting circulating TAG levels, which are a proxy measurement of atherogenic lipoprotein particles (chylomicron remnants, intermediary density lipoprotein, IDL; and small, dense low density lipoprotein particles) that should normally not accumulate in the circulation. Indeed, this population-level investigation has pointed to hereditable defects in vascular clearance of TAG-rich lipoproteins, and driven drug development for Mendelian causes of severely elevated TAG levels (Gaudet et al., 2014). For the general population, there are no effective therapies available to blunt post-prandial hypertriglyceridemia (Nordestgaard and Varbo, 2014).

A better understanding of the molecular cues governing intestinal fat absorption, storage and lipoprotein production may lead to new therapies to decrease post-prandial hyperlipidemia (Dash et al., 2015; Giammanco et al., 2015). A detailed gene-regulatory scheme governing intestinal lipid physiology is still lacking, although much progress has been made in defining the machinery of lipid absorption and lipoprotein production (Iqbal and Hussain, 2009; Abumrad and Davidson, 2012; Dash et al., 2015). The most apparent gaps in our understand of intestinal lipid handling relate to how and why enterocytes store a substantial fraction of absorbed lipids in cytoplasmic lipid droplets, and the cues that control this storage and subsequent release (Robertson et al., 2003; Zhu et al., 2009; Douglass et al., 2012; Dash et al., 2015; Giammanco et al., 2015).

Lxrs are nuclear hormone receptors whose major endogenous ligands are oxysterols, cholesterol catabolites that accumulate in proportion to cholesterol excess (Janowski et al., 1999), and select sterane intermediates of cholesterol synthesis and bile acids (Song et al., 2000; Yang et al., 2006). When activated by ligands, Lxrs drive a multi-organ gene transcriptional program that induces cholesterol elimination (Calkin and Tontonoz, 2012). Genetic activation of Lxrα in the intestine blunts cholesterol absorption (Lo Sasso et al., 2010) and fatty acid absorption (Cruz-Garcia and Schlegel, 2014); and drives the absorbed lipids into a cytoplasmic lipid droplet pool (Cruz-Garcia and Schlegel, 2014). Pharmacological activation of intestinal Lxrs also blunts fatty acid absorption and transport (Briand et al., 2016). Conversely, global deletion of mouse Lxrα increases the fraction of dietary cholesterol that is absorbed in the presence of a non-sterol Lxr agonist, an effect not seen when Lxrβ is deleted (Hu et al., 2012). Since oxysterols are excreted in bile (Mutemberezi et al., 2016), these studies argue that the intestine is exposed to multiple Lxr-activating signals that could dampen lipid transport with every meal. Unfortunately, Lxrα activation with synthetic non-sterol ligands up-regulates hepatic lipogenesis and Very Low Density Lipoprotein (VLDL) particle secretion (Repa et al., 2000; Schultz et al., 2000; Grefhorst et al., 2002; Bradley et al., 2007). This property has impeded Lxr-based drug development (Kirchgessner et al., 2016).

Zebrafish models of dyslipidemia have emerged in recent years as a powerful system for studying intestinal lipid transport, vascular lipoprotein metabolism, and the early steps of atherosclerosis (Fang et al., 2014; Schlegel, 2016). Most notably, zebrafish have similar abundance and distribution of circulating lipoproteins as humans (Stoletov et al., 2009; Liu et al., 2015), unlike numerous preclinical models (Yin et al., 2012). Additionally, examination of intestinal lipid handling has been studied extensively in zebrafish with results concordant with findings seen in Mendelian Diseases (Schlegel and Stainier, 2006; Avraham-Davidi et al., 2012; Levic et al., 2015).

It is important to stress that zebrafish carry a singly Lxrα ortholog, whose encoding gene is syntenic to human LXRα (Reschly et al., 2008; Cruz-Garcia et al., 2009; Fonseca et al., 2017). The Lxrβ locus was lost in the fish phylum (Fonseca et al., 2017). Furthermore, zebrafish Lxrα binds and is activated by endogenous and synthetic ligands in a manner similar to human LXRα (Archer et al., 2008; Reschly et al., 2008). Finally, human intestines express LXRα only (Uhlén et al., 2015). In this study we examine the effect of modulating zebrafish intestinal Lxrα activation on intestinal lipid transport and on the accumulation of vascular and hepatic lipids using new live imaging and physiological methods on our previously described Lxrα deletion and intestinal over-expression lines. We show intestinal Lxrα activation delays transport of ingested lipids, decreasing the post-prandial plasma TAG excursion. This effect protects animals from HCD-induced hepatic lipid accumulation. The methods we have optimized will be of broad use to others interested in studying aspects of enterocyte handling of lipids.

The intestine can serve as a reservoir for absorbed dietary lipids in metazoans ranging from insects to fish, rodents, and humans (Robertson et al., 2003; Sieber and Thummel, 2009; Zhu et al., 2009; Douglass et al., 2012). The molecular signals controlling this storage function are beginning to emerge. We previously demonstrated that zebrafish Lxrα cell-autonomously regulates intestinal handling of absorbed lipids, with over-expression of Lxrα (i.e., increased activation relying on endogenous ligands only) diverting absorbed lipids to an enterocyte cytoplasmic storage pool by inducing expression of the long-chain acyl-CoA ligase gene acsl3, which encodes a lipid droplet-targeted enzyme that funnels CoA thioesters of fatty acids to neutral and phospholipids (Cruz-Garcia and Schlegel, 2014).

Here we developed novel methods for live imaging of larvae and biochemical assessment in adults to monitor lipid transport in zebrafish. This work was motivated by our observation that whole mount ORO staining was not sufficiently sensitive to discern differences between WT and nr1h3^−/− animals. In larvae we found using a more sensitive live imaging approach that intestinal over-expression of Lxrα delays and deletion of Lxrα increases the rate of transport of orally gavaged fluorescent lipid tracers. This relatively facile approach in larvae should be of use in future genetic and pharmacological studies to explore not only the consequences of modulating Lxr target gene dose, but in studying the effects of drugs on intestinal lipid handling. Since there is modest larval manipulation and serial measurement required (i.e., gavage, and repeated live imaging, with periods of free swimming at the euthermal temperature and normal light cycle conditions), a preclinical validation platform could emerge from the use of our methods. A moderate-scale chemical screen could also be based on this workflow with zebrafish larvae (Clifton et al., 2010).

In adults, we observed a sexually dimorphic role for Lxrα in regulating the rate of transport of absorbed lipids. Specifically, we found that genetic activation of intestinal Lxrα delayed the appearance of lipids in the circulation acutely in males. Differences in intestinal sterol transporters do not appear to account for these differences. Rather, there appears to be sexually dimorphic, intestinal Lxrα-directed differences in plasma Lpl activity. The molecular cues regulating the higher Lpl activity and relatively blunted post-prandial lipemic excursion in female zebrafish will require further study; however, it is important to note that women also have higher Lpl activity than men (Mittendorfer et al., 2003). This relative protection from post-prandial hyperlipidemia is reminiscent of the human epidemiological observations of relative cardiac protection in premenopausal women (Lozano et al., 2012). It is also reminiscent of the long-standing observation that 17-β-estradiol protects ovariectomized female mice from atherosclerotic progression in Apoe^−/− mice (Bourassa et al., 1996). Restoration of this physiological estrogen with subcutaneous pellets reduces total plasma cholesterol, VLDL/IDL cholesterol, and TAGs (Bourassa et al., 1996). Remarkably, 17-β-estradiol protects Ldlr^−/− mice from atherosclerosis; however, the effect does not appear to be related to changes in circulating lipoproteins, as measured in the non-fasting state (Marsh et al., 1999). Thus, there appear to be context-dependent atheroprotective roles for 17-β-estradiol in mice. In future studies, we will determine whether the sexually dimorphic post-prandial lipid excursion in zebrafish impacts atherogenesis. In both sexes, intestinal over-expression of Lxrα in the intestine blunted the development of hepatic cholesterol accumulation during a long HCD challenge, a feeding paradigm in which Lxrα deletion caused severe hypercholesterolemia in both sexes. Since our experimental approach fully suppressed Lpl activity in females and males, there are probably additional sexual modifiers of postprandial lipid excursion that merit additional study.

More generally, our results in a zebrafish model of Lxrα deletion and cell type-limited over-expression are important because postprandial dyslipidemia is not amenable to most currently available lipid-lowering therapies (Nordestgaard and Varbo, 2014). Our results suggest that selective activation of Lxrα in the intestine might serve to treat this condition. The atherosclerosis that emerges from repeated, prolonged bouts of exposure to TAG-rich lipoprotein particles in the post-prandial state might be ameliorated by intestine-limited Lxrα activation. Similar to HDCA, cholane and cholestane Lxr agonists might be useful for this purpose: while such ligands induce transcription of the master transcription factor driving de novo lipogenesis Srebf1, they simultaneously block Srebf1 proteolytic maturation by stabilizing the precursor in the endoplasmic reticulum, thereby avoiding the induction of hypertriglyceridemia and hepatic steatosis (Kaneko et al., 2003; Quinet et al., 2004; Peng et al., 2008, 2011; Kratzer et al., 2009). Non-sterol ligands lack the biophysical properties to arrest Srebf1 maturation, and their administration is marked by hepatic steatosis and hypertriglyceridemia (Schultz et al., 2000; Grefhorst et al., 2002; Bradley et al., 2007; Kirchgessner et al., 2016). Full dissection of the transcriptional program controlled by intestinal Lxrα, as well as dissection the reasons for the sexually dimorphic traits seen in our study, will provide a thorough mechanistic basis for developing new therapies that leverage the capacity of the intestine to store (and safely oxidize) absorbed fatty acids, while promoting net cholesterol elimination.

Finally, we note that the severe hypercholesterolemia seen in the nr1h3^−/− mutant animals fed the HCD provides a unique system for studying atherosclerosis: this degree of non-HDL cholesterol increase was achieved without additional genetic manipulations, reflecting conservation of aspects of lipoprotein metabolism other established models lack (Yin et al., 2012), such as the retention of an ortholog of the human Cholesteryl Ester Transfer Protein gene (Schlegel, 2016). We anticipate the live-imagine and adult physiology methods described here will allow the zebrafish model to be useful for preclinical testing of a range of lipid-lowering therapies.

TB and AS designed the study, analyzed the data and wrote the paper. TB and SH performed experiments. All authors discussed the results and commented on the manuscript.

This work was supported by a Grant In Aid from the Western States Affiliate of the American Heart Association to AS. (15GRNT24670009).