Two related nuclear receptors mediate circadian fat metabolism in two different tissues using a lipid messenger as an intermediary. This signalling pathway might be relevant to the understanding of metabolic disorders. See Letter p.550
In the minuet, a popular court dance of the baroque era, couples exchange partners in recurring patterns. This elaborately choreographed exercise comes to mind when reading Liu and colleagues' paper1 on page 550 of this issue. In this study, the nuclear receptors PPARα and PPARδ are two of the three stars in a metabolic minuet that promotes appropriate fat utilization.
PPARα drives fat use in muscle and liver and is a well-known target of the fibrate class of lipid-lowering drugs. By contrast, PPARγ is essential for the development of white-fat tissue, mediating fat storage. PPARδ is more broadly expressed than its two brothers and is more enigmatic, having functions that overlap with both. In muscle it promotes fatty-acid breakdown and increases muscle endurance2,3. And in the liver, it stimulates fatty-acid synthesis, or lipogenesis, as Liu and co-workers have previously demonstrated4. This lipogenic activity is now shown to generate a 'dancing partner' for PPARα.
The circular pattern for this dance comes from the circadian activity of PPARδ in the liver (Fig. 1). Mice eat at night, storing excess calories as fat. During the day, Rev-erbα and Rev-erbβ, two nuclear receptors that also have circadian activity, repress lipogenesis in this organ5. Liu et al. report that nocturnal expression of at least a subset of key lipogenic enzymes in the liver depends on PPARδ. They also make the surprising observation that mice lacking PPARδ in the liver have defective fat uptake in muscle, but only at night. The authors deduce that the night-time liver could be synthesizing a signalling molecule that, when secreted, promotes fat uptake by the muscle. Indeed, they find that blood serum collected from normal mice in the dark phase of the day can promote fat uptake by cultured muscle cells, but that serum from mice lacking PPARδ in the liver cannot.
Extensive analysis narrowed down the factors transmitting the effects of PPARδ through the blood to a handful of lipid candidates, and Liu et al. focused on a phosphatidylcholine dubbed PC(18:0/18:1), demonstrating that treatment with this phospholipid, but not with other closely related phosphatidylcholine species, induces fatty-acid uptake into muscle cells both in vitro and in vivo. This is a hallmark of PPARα activation, and, consistently, PC(18:0/18:1)-mediated fatty-acid uptake was diminished in PPARα-deficient muscle cells and in mice.
Thus, this dance starts at night when liver PPARδ is activated, increasing PC(18:0/18:1) production. In an exchange of partners, PC(18:0/18:1) crosses from the liver to muscle, where it joins with PPARα in the next step, promoting fat uptake and fatty-acid oxidation. The cycle is completed as the levels or activities of all three partners fall during the day, setting up the next round.
Now that they have been worked out, these dance-like steps might seem relatively simple. Yet their potential importance is highlighted by the authors' observations that circadian production of PC(18:0/18:1) is dampened in mice fed a high-fat diet, and that PC(18:0/18:1) treatment improves metabolic parameters in diabetic mice, modestly decreasing blood levels of triglycerides and improving glucose homeostasis. Overall, these results are consistent with the beneficial effects of PPARα-activating fibrate drugs. They also suggest that the time of day at which fibrate treatment is given might be important, and that a drug that specifically targets PPARδ could still have PPARα-mediated side effects.
The new data also raise a host of difficult but intriguing questions. For instance, why does fatty-acid production in the liver promote the opposite process of fatty-acid oxidation in skeletal muscle? A more tractable question is whether PC(18:0/18:1) directly activates muscle PPARα. The answer is probably yes, given previous observations6 that other phosphatidylcholines can also activate PPARα and that the nearly identical PC(16:0/18:1) is a highly specific ligand for PPARα in the liver7. However, Liu et al. report that PC(16:0/18:1) does not activate PPARα in muscle cells. The reason for this apparent discrepancy is not obvious, and the nature of the endogenous functional activators of all three PPARs remains unclear. Extensive functional, biochemical and structural studies are needed to fully address this long-standing question.
Both PC(18:0/18:1) and PC(16:0/18:1) are abundant components of cell membranes. This raises the broader question of how such common molecules could function as specific metabolic signals. It could be that cellular compartmentalization is involved, such that the phospholipids that signal in the nucleus are somehow separated from the same molecular species in the cell membrane.
Several studies from another lab7,8,9 have suggested a specific compartmentalization pathway in which the enzyme fatty-acid synthase is required for production of the endogenous PPARα ligand in the liver. In response to nutrient signals, this pathway channels lipid synthesis through specific subcellular compartments to generate nuclear PC(16:0/18:1); only newly minted phosphatidylcholine is active in this scenario. The lipogenic component of the PC(18:0/18:1) story provides an intriguing parallel. Unfortunately, however, the idea that only newly produced intracellular phosphatidylcholine is active is not consistent with the biological effects of exogenously added PC(16:0/18:1) described previously7, nor with those of PC(18:0/18:1) in the current study. It is not clear how PC(18:0/18:1) exerts its effects in skeletal muscle, nor how it avoids PPARα activation in the liver, which would counteract the effects of PPARδ in a futile cycle of coincident fat synthesis and oxidation.
And a final question concerns generality. If this PPAR dance is the minuet, what about the gavottes and rigadoons, to say nothing of the square dances? The intracellular regulatory effects of lipid signalling molecules such as diacylyglycerol and ceramides are well known, and release of the specific lipid-controlling hormone C16:1n7-palmitoleate from adipose tissue promotes insulin action in muscle and suppresses fat accumulation in the liver10. More in line with the PPARδ–PC(18:0/18:1)–PPARα interchange, the nuclear receptors SF-1 and LRH-1 respond to phospholipid ligands11,12,13 to exert direct metabolic effects. Clearly, we don't know all the steps that the dance master has choreographed.
Liu, S. et al. Nature 502, 550–553 (2013).
Wang, Y.-X. et al. Cell 113, 159–170 (2003).
Narkar, V. A. et al. Cell 134, 405–415 (2008).
Liu, S. et al. J. Biol. Chem. 286, 1237–1247 (2011).
Feng, D. et al. Science 331, 1315–1319 (2011).
Lee, H. et al. Circ. Res. 87, 516–521 (2000).
Chakravarthy, M. V. et al. Cell 138, 476–488 (2009).
Chakravarthy, M. V. et al. Cell Metab. 1, 309–322 (2005).
Jensen-Urstad, A. P. L. et al. J. Lipid Res. 54, 1848–1859 (2013).
Cao, H. et al. Cell 134, 933–944 (2008).
Urs, A. N., Dammer, E. & Sewer, M. B. Endocrinology 147, 5249–5258 (2006).
Lee, J. M. et al. Nature 474, 506–510 (2011).
Blind, R. D., Suzawa, M. & Ingraham, H. A. Sci. Signal. 5, ra44 (2012).
About this article
Gastroenterology Clinics of North America (2016)