Fatty acid availability controls autophagy and associated cell functions

ABSTRACT

Macroautophagy/autophagy requires enormous membrane expansions during concerted actions of transient autophagic vesicles and lysosomes, yet the source of the membrane lipids is poorly understood. Recent work in adipocytes has now pinpointed the de novo lipogenesis pathway as the preferred source of fatty acids for phospholipid in autophagic membrane synthesis, as loss of FASN (fatty acid synthase) disrupts autophagic flux and lysosome function in vivo and in vitro. These data indicate fatty acid synthesis channels lipid for membrane expansions, whereas fatty acids from circulating lipoproteins provide for adipose lipid storage. Importantly, autophagy blockade upon loss of fatty acids promotes a strong thermogenic phenotype in adipocytes, another striking example whereby autophagy controls cell behavior.

Autophagy is commonly known as a degradation and recycling system, activated in response to starvation to release nutrients, and maintaining cellular homeostasis by degrading surplus or dysfunctional proteins and organelles. However, a less explored function of autophagy is the regulation of cellular behavior by controlling the turnover of bio-active, regulatory macromolecules, i.e. proteins, RNAs, signaling mediators, in the process of autophagosomal degradation. For example, autophagy is recruited during development to facilitate transitions to new states or differentiation. Autophagy has also been implicated in driving hepatic stellate cells from a quiescent state to an activated, fibrogenic state, and in adipose tissue, autophagy has been linked to the transitions between fat-storing white adipocytes and thermogenic beige adipocytes.

The capture of cargo and the kinetics of autophagic flux is enabled by a double-membraned compartment termed a phagophore, which undergoes massive expansion during autophagy, subsequently maturing into an autophagosome. Further complex membrane dynamics are required for autophagosome fusion with lysosomes, which enables degradation of the autophagosomal content and recycling of degradation products into the cytosol. Due to the large numbers of autophagosomes that must be synthesized following autophagy activation, autophagy is a highly lipid-demanding process. In an instant, upon activation, a pool of fatty acids/phospholipids must be made available to synthesize each autophagosome membrane. It is estimated that the phagophore membrane expansions require 4000 new phospholipids per second. In addition, new membranes are needed for autophagic lysosomal reformation, which is required to repopulate functional lysosomes consumed during autophagic degradation. Despite this remarkable demand for membrane lipid in active autophagy, the source of fatty acids for each of the steps described above is not well understood.

Cells of tissues in vivo obtain fatty acids via two independent pathways: uptake from exogenous sources through hydrolysis of circulating lipoproteins, and synthesis via de novo lipogenesis (DNL), a pathway that converts carbohydrates and amino acids into the fatty acid palmitate (C16:0) (Figure 1). How cells might partition these two sources of fatty acids for specific cellular functions has not been fully elucidated. By studying adipocytes, a cell type that takes up a large amount of exogenous fatty acids from lipoproteins but is also abundant in DNL enzymes, we were able to uncover an indispensable role specifically for DNL in the autophagic degradation process [1]. Adipocytes deficient in FASN (fatty acid synthase), the rate limiting enzyme of DNL, exhibit extensive autophagosome accumulation and a massive buildup of SQSTM1/p62, a major receptor protein present within autophagosomes. Autophagic flux assays revealed the defect in FASN-deficient adipocytes occurs at late stages of autophagy, rather than the autophagosome initiation and formation steps. Lysosomal activity and lysosomal CTSB (cathepsin B) maturation are also significantly impaired by FASN deficiency in adipocytes. These results suggest that fatty acids produced by DNL play a key requisite role in autophagosome-lysosome membrane fusion steps and lysosome function.

Figure 1.

In adipocytes, fatty acids from de novo lipogenesis (DNL) are directed to synthesis of phospholipids for phagophore and lysosome membrane expansion and fusion, while fatty acids from circulating lipoproteins are primarily stored in lipid droplets. This hypothesis is derived in part from experiments showing that inhibition of adipocyte ACACA/ACC1 or FASN in the DNL pathway (denoted 1 in the Figure) causes autophagosome accumulation and blocks autophagy flux in vitro and in vivo [1]. Fatty acids produced in this pathway are postulated to be channeled to phospholipid synthesis to accommodate phagophore expansion and autophagosome fusion with lysosomes. In contrast, other studies show that fatty acids derived from hydrolysis of circulating lipoproteins are the prime source of fatty acids in triglycerides stored in lipid droplets in these cells. Autophagy is thought to not only degrade complex molecules into basic constituents as nutrients or for resynthesis, but also to control the levels of key cell regulators and secreted bioactive factors. Thus, availability of fatty acids from DNL is likely a pathway that has major roles in cellular and systemic regulation through its control of autophagic flux. Figure created with Biorender. PLs, phospholipids; TGs, triglycerides.

A fascinating result is that restoration of cellular fatty acid levels by exogenous fatty acid administration does not fully restore autophagic degradation in the absence of DNL. Supplementation of FASN-deficient adipocytes with palmitate, the product of FASN, reduces SQSTM1 accumulation but is insufficient to fully restore autophagic flux. These data indicate exogenous fatty acids or mobilization of stored fatty acids can compensate in part for a lack of DNL, but are inadequate for optimal rates of autophagy. Thus, local, on-demand fatty acid synthesis at sites of autophagosome or lysosome synthesis may be required. Consistent with this hypothesis, FASN is observed to preferentially colocalize with the autophagosome marker LC3B during autophagy activation in cultured adipocytes.

Lipidomics analysis of the FASN-deficient adipocytes revealed no substantial changes to the major membrane lipid classes, indicating the absence of DNL does not result in a reduction in one or more phospholipids, which could have accounted for the autophagy defect. However, DNL deficiency results in a bulk remodeling of the fatty acyl chain composition of all lipid classes. Thus, FASN-deficient adipocytes harbor lipids with longer (≥ C18) and more unsaturated fatty acyl chains. The specific fatty acyl composition of many cellular lipids, including membrane and signaling lipids has not been thoroughly investigated. These data suggest DNL and fatty acyl chain composition may be a critical determinant of membrane lipid functionality in autophagosome dynamics.

Based on these data, we suggest a new model whereby DNL is the preferred source of fatty acids for phagophore and lysosomal membrane synthesis (Figure 1). Under certain conditions of decreased DNL, cells may utilize stored lipids or exogenous fatty acids to synthesize membranes; however, these autophagosome and lysosome membranes can be dysfunctional. By using DNL as a source of fatty acids, cells can precisely control the fatty acyl and phospholipid composition of cellular membranes, an important determinant of organelle function. Of note, studies in yeast have previously linked DNL to autophagy, and fatty acid synthesis and phospholipid synthesis enzymes, such as fatty acid CoA synthetase and phosphatidylinositol synthase have been shown to localize to sites of autophagosome synthesis. It will be important to determine which other fatty acid modifying and phospholipid synthesis enzymes are present at sites of autophagosome or lysosome synthesis. Crucially, the link between DNL and autophagy is not restricted to adipocytes, as inhibition of DNL in human hepatoblastoma cells (HepG2 cells) also impairs autophagic degradation. These data indicate that the hypothesis displayed in Figure 1 may be a general paradigm of cell biology.

A most exciting aspect of this work is the implication that autophagy in adipocytes is directly linked to downstream changes in adipose cell behavior. Adipose tissue depots from adipocyte-specific FASN-deficient mice have minor reductions in fat storage but acquire “beige” adipocytes, characterized by high mitochondrial content, uncoupling protein UCP1, and high energy expenditure. At the whole animal level, adipose FASN-deficient mice are more glucose tolerant and display higher energy expenditure than wild-type controls. Thus, the inhibition of autophagy in FASN-deficient adipocytes is inextricably linked to the initiation of the adipocyte beiging program. Data by others also suggest this is a causative relationship. In theory, a reduction in autophagic flux could suppress the degradation of a positive regulator of adipocyte beiging (“Thermogenic regulator” in Figure 1). Alternatively, autophagy could promote the release of negative regulators of beiging. There is also evidence that turnover of UCP1 itself is under control of autophagy. How fine-tuning autophagy can affect adipocytes and other cell types as well as whole-body phenotypes, and which macromolecules are involved are intriguing questions to be explored.

Funding Statement

The work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases [Grant no. DK116056].

Disclosure statement

No potential conflict of interest was reported by the author(s).