The FATP family is considered the primary transporter in the putative protein-mediated FA transport mechanism. There are six isoforms in mammals, and each displays a unique pattern of tissue expression. They are members of the solute carrier protein superfamily, classified specifically as Slc27a1-6. FATP1 was the first family member discovered in 1994 and is most highly expressed in skeletal muscle and adipose tissue. FATP1 was shown to localize to the plasma membrane and significantly increased radiolabeled oleic acid uptake when overexpressed in the adipocyte 3T3-L1 cell line (Schaffer and Lodish, 1994). The structure of FATPs is unknown but based on sequence and protease protection assays of epitope tagged FATP1, FATPs are oriented with an extracellular N-terminus and a cytoplasmic C-terminus (Lewis et al., 2001; Stahl, 2004). FATPs contain one or more membrane-spanning regions, multiple membrane-associated regions, and an AMP-binding motif in the intracellular region. FA transport through FATP is ATP-dependent, and the sequence of the AMP-binding region is required for transport. Loss-of-function mutations in the AMP-binding domain prevent FA uptake, suggesting AMP-binding is directly involved in or coupled to the transport mechanism (Stuhlsatz-Krouper et al., 1998). Further, overexpression of FATP1 increased intracellular FA acylation, thus suggesting a role for FATPs in cellular acyl-coA synthetase activity (Steinberg et al., 1999). These results, along with sequence similarity to acyl-CoA synthetases, suggested that FATPs function to transport FAs through the plasma membrane and esterify them as they enter the cell. Evidence has emerged suggesting some FATPs do not actually transport FAs, but rather enhance FA uptake through esterification at the ER, thus lowering intracellular levels of non-esterified FAs and encouraging FA uptake (Milger et al., 2006). Whether the transport and esterification functions of FATP are coupled or independent is widely debated and may depend on the cellular context.

FATP1 is the primary member expressed in adipocytes. Translocation to the plasma membrane is stimulated by insulin and causes an increase in FA uptake (Kim et al., 2004). This insulin stimulation is consistent with the current model that inhibition of lipolysis through the intracellular hormone sensitive lipase (HSL) signals a shift toward glucose utilization and storage of TGs. FATP1 KO mice showed delayed clearance of serum FAs following insulin injection. FA uptake in primary brown adipocytes from these mice was diminished following insulin treatment, suggesting FATP1 is needed to transport FA from the circulation into tissue. Radiolabeled oleic acid uptake into WAT in FATP1 KO mice was diminished, while uptake by the liver and heart was increased, demonstrating a systemic compensation for disruption of FA transport by specific FATPs (Wu et al., 2006). Single nucleotide polymorphisms (SNPs) in FATP1 are linked to increased plasma TG levels in humans, and high expression of other FATPs in peripheral tissue correlates with obesity and insulin resistance owing to an accumulation of intracellular FAs (Lobo et al., 2007).

FATP1 in BAT is necessary for maintaining core body temperature after prolonged cold exposure (Wu et al., 2006). It is primarily localized to the plasma membrane upon cold activation, but a small pool is detected on intracellular membranes as well. Besides uptake of exogenous FAs, FATP1 may aid in lipid transport between organelles in brown adipocytes or esterify incoming FAs for quick turnover into β-oxidation. FATP1 expression is significantly induced in cold-activated BAT, and FATP1 KO mice exhibit a significant drop in body temperature after ~12 h of cold exposure. This was accompanied by a rise in serum FA levels during cold exposure, likely owing to the inability of activated BAT to efficiently take up exogenous FAs. Peroxisome proliferator-activated receptor alpha (PPAR-α) and gamma (PPAR-γ), two transcriptional regulators of cell differentiation and lipid metabolism, were shown to control expression of FATP1 in 3T3-L1 adipocytes (Frohnert et al., 1999). Treatment with activators of PPAR-γ, such as linoleic acid, induced expression of FATP1 and FA uptake. This is consistent with the induction of both PPAR-γ and FATP1 in BAT during cold exposure, although a direct path of regulation has not been demonstrated in cold-activated BAT. Similarly, BAT-specific KO of FATP1 has yet to be explored to elucidate the function of its FA transport role in thermogenesis. Future studies are required to contextualize the role of FATPs in exogenous FA uptake by BAT.

FABPs are a family of intracellular protein chaperones that bind long chain fatty acids (LCFAs) to expedite movement through membranes. There are currently 9 FABPs named for the tissue of discovery, but they are not exclusively expressed in the tissue for which they are named. The cytoplasmic FABP family differs in sequence, structure, and function from the plasma membrane associated FABP (FABP_PM). First described in 1985, the putative fatty acid transporter FAPB_PM was shown to have a high affinity for fatty acids (Stremmel et al., 1985). Later, FABP_PM was found to be identical to mitochondrial aspartate aminotransferase, a member of the malate-aspartate shuttle, although expression at the plasma membrane was associated with fatty acid uptake (Berk et al., 1990; Bradbury and Berk, 2000). Inhibition of FABP_PM by antibody treatment in multiple tissues yielded a decrease in LCFA uptake while overexpression in skeletal muscle yielded an increase, thus supporting FABP_PM as a lipid transporter (Berk et al., 1990; Clarke et al., 2004). More work is needed to understand the importance of FABP_PM expression in brown adipocytes and how it changes with cold exposure. The remainder of this section will focus on other FABP family members in the cytoplasm.

Cytoplasmic FABPs solubilize hydrophobic molecules in the aqueous cellular environment, acting as chaperones and delivery systems for lipids between organelles. Structural characterization has revealed an indirect role in LCFA uptake; that is, FABPs do not facilitate transport across the plasma membrane but are essential for delivery into the cell (Storch and McDermott, 2009). All FABPs share a common tertiary structure comprised of antiparallel beta sheets forming a small beta barrel within which a hydrophobic molecule may bind (Sacchettini et al., 1988). There are subtle structural differences between isoforms that dictate ligand specificity. For example, adipocyte FABP (FABP4), also known as adipocyte protein 2 (AP2), exclusively binds LCFA while liver FABP (FABP1) can additionally bind eicosanoids, lysophospholipids, and acyl-CoA (Rolf et al., 1995; Thompson et al., 1997; Furuhashi and Hotamisligil, 2008). These cytoplasmic FABPs have also been shown to be secreted into the bloodstream with several physiological stresses including high fat diet, insulin resistance, and cold exposure (Hotamisligil and Bernlohr, 2015; Shu et al., 2017).

The adipocyte associated FABP (FABP4) is the best characterized in its family. FABP4 was first isolated from mouse embryonic fibroblast (3T3)-derived adipocytes, although it is highly expressed in macrophages as well (Hunt et al., 1986; Hotamisligil et al., 1996). FABP4 KO mice were viable, developed normally, and were indistinguishable from control mice in appearance and metabolic health (Hotamisligil et al., 1996). Like control mice, they were sensitive to dietary and genetic obesity, but were protected from insulin resistance and diabetes. This implicates FABP4 in suppression of insulin signaling, perhaps through increasing the intracellular lipid pool. Additionally, these mice exhibited a compensatory increase in epidermal FABP (E-FABP or FABP5) expression in adipose tissue, despite basal FABP5 expression being almost 100-fold lower than FABP4. Besides binding FAs, FABP4 has been shown to suppress PPAR-γ, a master adipocyte transcription factor that activates genes involved in lipid synthesis and uptake to enhance adipogenesis (Garin-Shkolnik et al., 2014). FABP4 also stimulates intracellular lipolysis through direct binding of HSL as suggested by a reduction in lipolysis in FABP4 KO mice and demonstrated by fluorescence resonance energy transfer of FABP4 mutagenized within the putative HSL binding site (Smith et al., 2008). At a chemical level, FABP4 binds FAs entering the cell. However, substantial evidence illustrates its function as an intracellular signal for proper adipocyte function and systemic fuel utilization.

FABP4 is increased with cold exposure in BAT and in the blood plasma and was shown to be necessary for FFA uptake into BAT (Shu et al., 2017). Knockout of FABP4 and HFD led to cold sensitivity and lower UCP1 expression that could be rescued with recombinant FABP4. FABP4 was shown to be sufficient to drive increased oxygen consumption in mice and elevated UCP1 expression. This increase required FA binding to FABP4, as a mutant R126Q did not have the same rescue capacity. Mechanistically, FABP4 in the BAT was shown to drive the intracellular conversion of T4 to T3 which is necessary for the thermogenic program (Ahmadian et al., 2011). FABP4 has also been shown to be increased in the BAT of hibernating mammals (Hittel and Storey, 2001; Eddy and Storey, 2004). Other studies have shown that FABP4 and 5 are necessary for proper metabolic adaptation in cold-activated BAT and that fasting greatly impacts this contribution to BAT function as well. When fasted for 20 h prior to a 4-h cold exposure, FABP4/5 KO mice were cold intolerant but indistinguishable from control mice in a fed state (Syamsunarno et al., 2014). This suggests a requirement of FABP4/5 for proper adjustment of systemic fuel utilization during physiological stress, such as long-term fasting and cold.

Curiously, heart-type FABP (FABP3) has been shown to be the dominant isoform in BAT upon cold activation. Expression of FABP3 was significantly increased in BAT after 4 h of cold despite a basal level of expression 6-fold lower than FABP4. Further, FABP3 KO mice were cold intolerant but displayed similar changes in thermogenic transcriptional programs to WT mice, such as induction of the coactivator PGC1α, which regulates mitochondrial biogenesis. Expected physiological changes were observed in FABP3 KO mice as well, such as depletion of TGs in BAT and increased FFAs in plasma (Vergnes et al., 2011). FABP3 was also shown to increase in hibernating ground squirrels (Hittel and Storey, 2001). Together, these data highlight an essential role for FABP3 in the uptake of exogenous lipids into BAT, the dysfunction of which results in an inability to maintain body temperature. Although FABP3 is not the dominant isoform in BAT, its function in cold exposure suggests a dynamic response network of FABPs in certain cell types depending on the environmental stimulus (Daikoku et al., 1997; Yamashita et al., 2008). This also underscores the gaps in current knowledge of FABPs in BAT function.

CD36, also referred to as fatty acid translocase (FAT), is an integral membrane protein belonging to the scavenger receptor superfamily. SRs bind a host of ligands including LDL, oxidized LDL, phospholipids, cholesterol esters, collagen, thrombospondin, carbohydrates, and microbial pathogens. There are 11 classes of scavenger receptors categorized based on primary sequence (A-L). CD36 belongs to the scavenger receptor B class (SR-B), the members of which contain two transmembrane domains and a single extracellular loop (Pepino et al., 2014). Diversity in the extracellular loop allows SR-Bs to regulate a variety of signal transduction pathways. Loss of function mutations or genetic deficiencies of SR-B members causes dysregulation of apoptosis, inflammatory response, and intracellular metabolism resulting in metabolic disease (Bonen et al., 2007; Glatz et al., 2010; McFarlan et al., 2012).

Although originally identified in human platelets, CD36 was characterized as a high affinity long-chain fatty acid (LCFA) receptor and transporter in rat adipocytes (Abumrad et al., 1993; Harmon and Abumrad, 1993). Treatment with highly charged FA analogs (N-sulfosuccinimidyl LCFA esters) inhibited oleate transport into adipocytes, and covalent protein labeling identified the 88 kDa CD36 as the receptor candidate (Harmon and Abumrad, 1993). Northern blotting revealed transcript abundance across tissues including adipose, heart, intestine, and skeletal muscle. Oleate uptake was limited to cell lines expressing CD36, further implicating the membrane protein as a lipid transporter (Abumrad et al., 1993). Modern proteomics and transport kinetics studies in transgenic Chinese hamster ovary cells identified a definitive FA binding pocket in CD36 that supports a direct function in lipid binding (Kuda et al., 2013). This pocket contains lysine 164 which binds the CD36 inhibitor N-hydroxysuccinimidyl ester of oleate (SSO) and, when mutagenized to alanine, diminished FA uptake. Despite a host of evidence supporting its lipid binding function, its role as a transporter has long been disputed.

Transmembrane LCFA transport has been at the center of a contentious debate for decades. The two opposing sides are comprised of “diffusionists,” who argue for rapid diffusion-mediated LCFA transport, and “translocationists,” who suggest transport is facilitated by proteins (Pownall and Moore, 2014; Glatz and Luiken, 2020). There is a myriad of evidence supporting both sides. The diffusionists argue that in biological membranes, lipids constantly diffuse laterally and often exchange between membrane leaflets without proteins. The FFAs would transfer from albumin to the outer leaflet of the plasma membrane and then flip-flop to the inner leaflet (Hamilton, 1998). This hypothesis is supported by demonstration of diffusion through fluorescent LCFA uptake in lipid vesicles pre-treated with inhibitors of FA transporters, two of which were direct competitive inhibitors for CD36 (Jay et al., 2020). The rate of uptake in rat adipocytes was unaffected by inhibitor treatment, suggesting that proteins are not required for transport (Abumrad et al., 1993; Harmon and Abumrad, 1993; Coburn et al., 2000; Kuda et al., 2013). However, these are imperfect systems to study lipid transport. Artificial membranes do not reflect true biological lipid or protein composition, and cultured cells do not reflect membrane dynamics. For example, fatty acid chain composition is different in culture, having lower levels of poly-unsaturated fatty acids compared to cells in organisms (Else, 2020).

The translocationists have been supported by several key findings including the observation that FFA uptake can be saturated in multiple cell types at levels within physiological range and that the flip-flop between the outer and inner leaflet does not occur at high enough levels to support metabolic demand (Kleinfeld and Storch, 1993; Kleinfeld et al., 1997). Moreover, an antibody to CD36 blocked FFA uptake into mouse adipocytes, and mutation of Lys164 in the FA binding pocket demonstrates site specific functional regulation (Abumrad et al., 1993; Pepino et al., 2014). Recent work has found that palmitoylation of CD36 leads to caveolae-dependent internalization, and inhibition of palmitoylation prevents this uptake (Hao et al., 2020). While the mechanism of lipid transport remains controversial, the in vivo role of CD36 in regulating FA uptake and metabolism has been demonstrated across many models. Moreover, in humans, common SNPs in the promotor of CD36 cause protein deficiency and have been associated with high levels of FAs and LDL in serum, a phenotype recapitulated in Cd36^−/− mice (Miyaoka et al., 2001; Ma et al., 2004; Goudriaan et al., 2005; Yamashita et al., 2007; Love-Gregory et al., 2011). Newer models favor both diffusion and translocation that are important in different ranges and energy demands of the cell (Abumrad et al., 1999).

CD36 expression is known to be highest in adipose tissue, and it is required for body temperature regulation in mice (Putri et al., 2015). Additionally, its expression is significantly increased in BAT following cold exposure (Bartelt et al., 2011). CD36 is required for lipoprotein uptake in BAT, and global CD36 KO in mice results in TG buildup in BAT and cold intolerance (Yamashita et al., 2007; Bartelt et al., 2011). CD36 is necessary for transport of lipids via lipoproteins and FFA-bound albumin in BAT as well. Clearance of both lipoproteins and FFAs from plasma and uptake into BAT were diminished in cold-exposed Cd36 KO mice (Bartelt et al., 2011). CD36 is also required for uptake of lipid-related molecules into BAT, such as coenzyme Q (CoQ) (Anderson et al., 2015). CoQ contains an isoprenoid tail and is a necessary electron transporter between complexes in the electron transport chain. Therefore, it is required for both ATP synthesis and heat production in BAT. High levels of CoQ are present in BAT despite low levels of endogenous synthesis, suggesting a requirement for uptake of CoQ from the circulation. Whole-body Cd36^−/−mice displayed a 2-fold increase in serum CoQ accompanied by CoQ deficiency, TG accumulation, and diminished mitochondrial size and metabolic capacity in BAT. Although CD36 is present in other tissues, these phenotypes were not reflected throughout the mice. Cold intolerance in both whole-body and BAT-specific Cd36^−/−was driven by CoQ insufficiency causing reduced FA β-oxidation in BAT. These studies highlight the requirement of CD36 for proper uptake of lipids and hydrophobic molecules into cold-activated BAT.

Lipoproteins are means of transporting lipids of varying polarity from centers of lipid processing, such as the intestine and liver, to peripheral tissues through the circulation. They are comprised of a phospholipid monolayer interlaced with apolipoproteins that serve as structural reinforcements and as LDL receptor (LDLR) recognition sites to aid in endocytosis. Lipoproteins fall into four main classes depending on their relative composition of proteins and lipids: high density (HDL), low-density (LDL), very low-density (VLDL), and chylomicrons. In addition to the characteristic structure of each class, different lipoproteins also serve different functions. For example, LDL carries circulatory cholesterol and easily enters arterial walls, while chylomicrons are essential for dietary lipid transport to the liver and peripheral tissues. Nonetheless, each class has a vital role in lipid transport, endocytosis, and hydrolysis (Zanoni et al., 2018).

LDL endocytosis begins with recognition at the cell membrane by a low-density lipoprotein receptor (LDLR). Following recognition, a clathrin-coated pit forms a vesicle around the endocytosed lipoprotein. The vesicle is directed to endosomes where a drop in pH causes release of LDL from LDLR. Acidic lipases in the newly formed lysosome hydrolyze the contents of LDL to release FAs from TGs and unesterified cholesterol from cholesteryl esters (Figure 3). Cholesterol is incorporated into cell membranes and negatively regulates its own synthesis by preventing translocation of sterol regulatory binding protein (SREBP) from the ER to the Golgi, thereby blocking transcriptional activation of HMG-CoA reductase (Brown and Goldstein, 1997). FAs released from LDL are shuttled to the ER for synthesis into membrane lipids, the Golgi for protein acylation, and the mitochondria for FA β-oxidation.

In brown adipocytes, LDLR endocytosis fuels the electron pool needed for proton gradient uncoupling and heat production (Figure 4). LDLR is recycled back to the plasma membrane for addition rounds of endocytosis (Ikonen, 2008). LDLRs are also necessary in systemic cholesterol and triglyceride balance during cold exposure. Ldlr^−/− mice displayed a reduction in plasma TG upon activation of BAT by cold or with a β3 adrenergic receptor agonist but did not show a reduction in plasma cholesterol normally observed after BAT activation (Dong et al., 2013; Berbée et al., 2015). In this model, the liver is unable to clear lipoprotein remnants produced from FA uptake in BAT following LPL-mediated TG hydrolysis and thus cholesterol remains in the circulation. Coordinated plasma lipid clearance between the liver and BAT during cold exposure requires proper protein-mediated lipid uptake, such as LDLR endocytosis.

TGs are the neutral storage form of FAs and travel between tissues through the bloodstream via chylomicrons and VLDL. Chylomicrons are the largest form of lipoprotein and are assembled by the small intestine following emulsification of dietary fats. TGs packaged into chylomicrons mainly enter adipose tissue and skeletal muscle due to high lipoprotein lipase (LPL) activity. The remaining TG is taken up by the liver and can be repackaged into VLDL for use throughout the body. Alternatively, FAs are synthesized by the liver through de novo lipogenesis, assembled into TG, and mobilized in VLDL. As TGs are liberated from VLDL for use by peripheral tissues, smaller remnants of decreasing TG content are formed such as intermediate-density lipoproteins (IDL) and LDL. Like chylomicrons, VLDL is utilized by tissues after lipolysis by LPL.

LPL is the predominant lipase in tissues with high levels of exogenous lipid uptake, such as adipose tissue, heart, and skeletal muscle. LPL is a dimeric enzyme localized to the vascular lumen where it hydrolyzes plasma TGs into glycerol and FAs for tissue uptake. TG-derived FA uptake into BAT is reliant on localized LPL activity, as injection of an LPL inhibitor (tetrahydrolipstatin) or treatment with heparin to release LPL from the vascular wall in mice prior to cold exposure almost completely abolished labeled TRL uptake into BAT.²⁰ LPL is ferried and anchored to the endothelium via glycophosphatidylinositol-anchored HDL-binding protein (GPIHBP1). Mice lacking GPIHBP1 have increased plasma TGs due to reduced tissue uptake (Cushing et al., 2018). As circulating lipoprotein levels are highly modulated in response to metabolic state, such as fasting and cold exposure, tight regulation of LPL activity is required. Besides localization, LPL is primarily regulated post-translationally by several extracellular proteins. Angiopoietin-like proteins (ANGPTL) are the main class of LPL regulators. ANGPTLs are secreted into the lumen and directly interact with LPL, preventing its dimerization and inhibiting lipolysis (Hegele, 2016).

ANGPTL4 is the predominant isoform in brown adipose tissue. In BAT, ANGPTL4 expression is induced during periods of nutrient deprivation (fasting) and suppressed when substantial nutrients are available or needed, such as postprandially or during cold exposure (Singh et al., 2018). In cold-activated WAT, ANGPTL4 is upregulated to shift away from fat storage and toward output of FAs to support thermogenesis. BAT-specific ANGPTL4 KO mice exhibit reduced plasma TGs in the fed and fasted state, accumulation of ³H from labeled triolein in BAT, and increased expression of CD36 in BAT (Cushing et al., 2018). LPL expression was unchanged, but activity was significantly increased in BAT from ANGPTL4 KO mice. ANGPTL4 KO mice had higher rectal temperature over time during a 4-h cold exposure, likely due to enhanced LPL activity in BAT fueling FA uptake and β-oxidation. In contrast, Gpihbp1^−/− mice show no change in labeled triolein uptake into BAT. Double GPIHBP1/ANGPTL4 KO mice show a partial correction in plasma TG levels compared to GPIHBP1 KO mice (Dijk et al., 2015). However, this correction is lost after a 2-week HFD, with both GPIHBP1 KO and GPIHBP1/ANGPTL4 KO showing a similar increase in plasma TGs compared to mice of the same genotype on a normal chow diet. In sum, these studies highlight the necessity of tight LPL regulation in BAT to modulate TG and FA uptake during metabolic stress, including cold exposure. It has been shown that BAT relies more heavily on LPL-based uptake of TG-derived FAs rather than particle endocytosis, but cold exposure significantly enhances the uptake of TGs through both methods (Khedoe et al., 2015). This was observed through tracing of ³H-triolein and ¹⁴C cholesteryl esters which allow measurement of TG uptake by lipolysis and lipoprotein uptake by endocytosis, respectively.

While VLDL is overwhelmingly shuttled to BAT when activated by exposure to cold, transport of HDL to the liver is drastically increased (Schaltenberg et al., 2021). HDL balances cholesterol flux between peripheral tissues and is both synthesized and excreted by the liver. This hepatic processing is dependent on endothelial lipase (EL). Following extended cold exposure (1 week), expression of the gene encoding EL, Lipg, was increased in murine BAT. Cold exposure was also shown to enhance HDL clearance from plasma. Moreover, Lipg^−/− mice displayed higher plasma levels of HDL, indicating poor cholesterol clearance. The HDL particles from Lipg^−/− mice were enriched in phospholipids but lacking in cholesterol and TG compared to WT mice both at room temperature and after prolonged cold exposure. EL is known to promote TRL uptake, much like LPL. While EL expression is induced in BAT after cold exposure, it was not required for proper thermogenic activation, nor did its loss affect thermogenic transcriptional programs such as those controlled by PPAR-γ.

Acylcarnitines are fatty acids conjugated to a carnitine through esterification. At the cellular level, acylcarnitines function as intermediaries facilitating transport of FFAs into the mitochondria for β-oxidation. Carnitine palmitoyltransferase 1 (CPT1) is embedded on the outer-surface of the mitochondrial membrane and esterifies the fatty acid from an acyl-CoA to carnitine. This acylcarnitine can then diffuse into the porous outer mitochondrial member. Acylcarnitine is then brought into the inner mitochondria by carnitine acylcarnitine transferase (CACT) and de-carnitylated by CPT2. Besides their cellular role for fatty acid transport, acylcarnitines are also found in the blood plasma. Plasma acylcarnitines increase with chronic diseases such as type 2 diabetes, cardiovascular disease, and inborn errors of metabolism as well as in acute metabolic stresses such as fasting, exercise, and cold exposure (Muoio et al., 2012; Schooneman et al., 2013; McCoin et al., 2015; Simcox et al., 2017). The functional role of acylcarnitines in the plasma has been proposed to range from protection from toxicity to a distinct storage pool that can be pulled from during energy demanding conditions (Muoio et al., 2012).

In cold exposure, short chain, medium chain, and long chain acylcarnitines are increased in the plasma while carnitine levels decrease (Simcox et al., 2017; Pernes et al., 2021). This cold induction of increased plasma acylcarnitines is mediated by β3-adrenergic receptor induced WAT lipolysis, since adipose tissue-specific ATGL knockout mice had no changes in acylcarnitine levels with β3-adrenergic receptor agonist treatment. Once FFAs are released from the WAT, they are taken up into the liver, where they transcriptionally activate CPT1, CACT, and CPT2 through an HNF4α-mediated mechanism as well as serve as substrate for acylcarnitine production (Simcox et al., 2017; Jain et al., 2021). These liver-produced acylcarnitines are then taken up into the BAT, skeletal muscle, and heart. In the BAT, the acylcarnitines are catabolized as a fuel source for thermogenesis. Beyond the liver, there are other potential sources for cold induced plasma acylcarnitines; while ablation of acylcarnitine production in the liver causes cold intolerance, it is not sufficient to completely block the rise in plasma acylcarnitines with cold exposure. Recently it has been shown that the kidney may also contribute to the plasma acylcarnitine pool (Jain et al., 2021). These studies collectively demonstrate that plasma acylcarnitines are produced through a multi-tissue processing, and that they function as a fuel source for cold-activated BAT.

Several questions remain in understanding the regulation and transport of plasma acylcarnitines including how they are transported through the plasma membrane in the liver and in the brown adipose tissue. Studies in Xenopus oocytes have demonstrated that acylcarnitines require a transporter to cross the plasma membrane, and cDNA libraries from mouse liver have demonstrated that these unknown transporters are present in the liver (Berardi et al., 1998; Nakanishi et al., 2001). SLC22a1 was recently identified as an acylcarnitine exporter in the liver, and knockout of SLC22a1 led to decreased short and medium chain acylcarnitines in plasma but had no impact on long chain acylcarnitine levels (Kim et al., 2017). Moreover, there has been no identified BAT acylcarnitine transporter and SLC22a1 has low expression in brown adipocytes. Plasma long chain acylcarnitines have been shown to travel bound to albumin, while short and medium chain acylcarnitines are unbound. Whether albumin bound acylcarnitines are the dominant form of acylcarnitine in the circulation during cold exposure is unknown. Future work will be needed to understand their entry into brown adipocytes and the kinetics of their uptake compared to FFAs of the same acyl chain.

Ceramides are a long chain sphingoid base conjugated to a fatty acid through an amide bond and are the precursor to all sphingolipids. Ceramides are known to circulate during tissue dysfunction and metabolic disease in both mice and humans. Plasma ceramide levels have been shown to correlate with risk of diabetes and coronary artery disease in a species-specific manner across human cohorts (Tippetts et al., 2021). Reduction of plasma and WAT ceramides in mice via increased degradation (ceramidase overexpression) or inhibition of ceramide synthesis (SPTLC2 KO) ameliorated HFD-induced obesity, insulin resistance, and hepatic steatosis (Xia et al., 2015; Chaurasia et al., 2016). SPTLC2 KO in WAT also enhanced adipocyte browning and resulted in an increase in beige adipocyte differentiation (Chaurasia et al., 2016). This suggests that ceramides act as signals to increase lipid storage in WAT and inhibit the beige program. Moreover, liver SPTLC2 expression is upregulated in response to SPTLC2 KO in WAT, suggesting a means of communication to balance tissue ceramide levels.

Despite the wealth of literature on ceramide function in metabolic disease and its regulation of the adipocyte differentiation program, little is known about how ceramides control brown and beige adipocyte maintenance or their direct role in thermogenic metabolism. We have observed significant increases in plasma ceramides following acute cold exposure, with computational assessment revealing that these plasma levels are regulated by the BAT and the kidney (Jain et al., 2021). More work is needed to characterize how these plasma ceramides are regulated in acute cold exposure, what their functional role may be, and what complexes facilitate their transport in the plasma during cold exposure. At ambient temperature, ceramides are known to be associated with lipoproteins (primarily LDL) and have been shown to transfer between cells via extracellular vesicles (EVs) (Hammad et al., 2010; Crewe et al., 2018). These vesicles act as carriers for intercellular signaling molecules during metabolic stress, and many ceramide species act as second messengers for key metabolic pathways including insulin sensing and cell growth. For example, during fasting, white adipose tissue traffics EVs containing signaling molecules such as caveolin 1 and very long chain ceramides to neighboring endothelial cells (and vice versa) (Crewe et al., 2018). EVs are also produced by adipose-derived stem cells during beige adipocyte differentiation and were shown to be sufficient to differentiate these stem cells into beige adipocytes (Jung et al., 2020). Additionally, BAT has been shown to be a significant contributor of exosomes into the circulation (Thomou et al., 2017). It is unknown whether export of these vesicles is upregulated or if ceramides are enriched in these vesicles during cold exposure. Moreover, there are no known plasma membrane transporters of ceramides. Ceramides in cold exposure remain an exciting area of research with many outstanding questions, including the role of ceramides in thermogenic metabolism as well as the function of plasma ceramides compared to ceramides produced in brown and beige adipocytes.

12,13-diHOME is produced in brown adipose tissue and can function as an autocrine or paracrine signal to increase mitochondrial oxidation rates (Lynes et al., 2017). Upon β3-adrenergic receptor activation, linoleic acid is oxidized by cytochrome P450 and soluble epoxide hydrolase to produce 12,13-diHOME. Beyond BAT, other tissues are known to produce 12,13-diHOME, including the skeletal muscle, and contribute to the circulating pool to regulate body weight, energy expenditure, insulin sensitivity, and plasma lipid levels (Vasan et al., 2019). The uptake of 12,13-diHOME into brown adipocytes is regulated by CD36 and FATP1, and treatment of brown adipocytes with 12,13-diHOME is sufficient to increase translocation of CD36 and FATP1 (Lynes et al., 2017, 2019). Further studies are needed to understand how 12,13-diHOME regulates mitochondrial oxidation, determine the mechanism of secretion from cells, and understand how this lipid circulates in the plasma.

FAHFAs are a recently discovered class of signaling lipids that regulate brown and beige adipocyte differentiation and maintenance. Structurally, FAHFAs are fatty acids complexed to a hydroxy fatty acid through an ester bond (Yore et al., 2014). There are numerous types of FAHFAs named for the acyl chains and the location of the hydroxylation including stearic-acid-9-hydroxy stearic acid (9-SAHSA), oleic-acid-9-hydroxy stearic acid (9-OAHSA), and palmitic-acid-9-hydroxy stearic acids (9-PAHSA). Both 5- and 9- PAHSA have been shown to increase brown adipocyte differentiation, insulin sensitivity, decrease inflammation in adipose tissue, and improve whole body glucose tolerance. Treatment of 3T3-L1 adipocytes or leptin deficient mouse models with 9-PAHSA led to increased expression of thermogenic genes including UCP1 (Wang et al., 2018). Part of this signaling is mediated through binding and activating G-protein coupled receptor 120 (GRP120), and knockdown of GPR120 in 3T3-L1 cells abrogated the effect of 9-PASHA treatment (Oh et al., 2014; Wang et al., 2018). Cold exposure induced the production of 5- and 9-PAHSA from WAT, with this production being mediated by lipolysis from triglycerides since knockout of ATGL led to ablated the cold-induced production (Paluchova et al., 2020). Many outstanding questions remain on how various species of FAHFAs impact brown and beige adipocytes and how they are transported into cells.

The numerous plasma lipids that act upon BAT is still an open area of study. For the purpose of this review, we have chosen to focus on plasma lipids that are transported into brown adipocytes, however there are a number of other lipids that are altered in brown adipocytes themselves that regulate thermogenesis. These include ether lipids (such as plasmalogens) and cardiolipins (Lynes et al., 2018; Park et al., 2019; Von Bank et al., 2021b). Although FFAs produced in the BAT are not necessary for thermogenesis (Schreiber et al., 2017), they have been shown to be sufficient to drive the thermogenic program through mediation of GPCR signaling (Sveidahl Johansen et al., 2021). Interestingly, ether lipids have recently been observed to increase with cold exposure in studies where mice were acclimated to thermoneutrality then placed for 24 h in thermoneutrality, room temperature (22°C), and cold exposure (5°C) as well as fasted for the final 5 h of temperature stress (Pernes et al., 2021). More work is needed to understand these various lipids in the plasma and how their transport is regulated.