Palmitic acid induces UCP1-independent mitochondrial depolarization specifically in brown adipose tissue

TL;DR

This study shows that palmitic acid can cause mitochondrial depolarization in brown fat without UCP1, suggesting a new way brown fat generates heat.

Contribution

The study identifies palmitic acid as a direct inducer of UCP1-independent mitochondrial depolarization in brown adipose tissue.

Findings

Cold-mimicking stimulation causes mitochondrial depolarization in UCP1-deficient brown adipocytes.

Palmitic acid levels increase during cold-mimicking stimulation in a lipolysis-dependent manner.

Palmitic acid reduces mitochondrial membrane potential specifically in brown adipose tissue mitochondria.

Abstract

Brown adipose tissue (BAT) is a major site of nonshivering thermogenesis, where mitochondria generate heat instead of ATP. The thermogenesis occurs through the activity of uncoupling protein 1 (UCP1), which specifically resides in the mitochondrial inner membrane and dissipates the mitochondrial proton gradient upon activation by long-chain free fatty acids. Although UCP1-independent proton leak has been reported, the mechanism underlying UCP1-independent mitochondrial membrane depolarization remains largely unknown. Here, using primary brown adipocytes, we found that cold-mimicking stimulation induces mitochondrial membrane depolarization even under UCP1 KO and knockdown conditions. Furthermore, during cold-mimicking stimulation, palmitic acid shows the most prominent increase in a lipolysis-dependent manner. Notably, palmitic acid directly decreases mitochondrial membrane potential specifically in mitochondria isolated from BAT but not in those isolated from liver or brain. These findings suggest that palmitic acid contributes to mitochondrial depolarization in BAT, thereby contributing to UCP1-independent depolarization.

Figures

Full text

Brown adipose tissue (BAT) generates heat to adapt to cold environments, a process that predominantly occurs in mitochondria (1, 2). Mitochondria in BAT uniquely express uncoupling protein 1 (UCP1), which induces mitochondrial depolarization to generate heat (3, 4). Furthermore, free fatty acid (FFA) oxidation increases during UCP1-mediated mitochondrial depolarization induced by cold exposure or cold-mimicking stimulation. Therefore, BAT has emerged as a therapeutic target for obesity and related diseases (1, 5).

During thermogenesis, norepinephrine released from sympathetic neurons binds to β-adrenergic receptors located on the plasma membrane of BAT, thereby activating the cAMP–PKA cascade (1). PKA activates adipose triglyceride lipase (ATGL) and hormone-sensitive lipase (HSL), thereby promoting lipolysis and leading to the release of FFAs from lipid droplets (1, 6). In particular, the intrinsic ATGL- and HSL-mediated increase of FFAs in BAT is essential for maintaining body temperature in vivo (6, 7). FFAs released by lipase activity are utilized for oxidation and phospholipid remodeling and also bind to UCP1 to promote proton conductance (8, 9, 10, 11, 12). Notably, excessive FFAs promote uncoupled proton leak respiration regardless of whether UCP1 is present in primary brown adipocytes (13). Among the proposed mechanisms of UCP1-independent proton leaks, the ATP/ADP carrier has also been identified as a primary candidate (4). ATP/ADP carrier has been reported to mediate proton leak in other tissues, but its proton conductance is much smaller than the UCP1-mediated one (14, 15).

Of note, the composition of certain fatty acyl chains in phospholipids is upregulated in response to cold environments (9, 12), suggesting that the composition of FFAs also changes during thermogenesis. However, it remains unclear which FFA species are upregulated upon β-adrenergic stimulation and whether these FFAs directly contribute to mitochondrial membrane depolarization. In this study, using primary brown adipocytes, we found that mitochondrial membrane depolarization under UCP1 knockdown (KD) conditions is strongly induced by ATGL-mediated lipolysis-dependent manner. Furthermore, analyses using isolated mitochondria from mouse brain, liver, and BAT revealed that palmitic acid decreases mitochondrial membrane potential, specifically in BAT. These findings provide new insight into the lipolysis-mediated, UCP1-independent activation mechanism in BAT.

Results

β-adrenergic stimulation induces mitochondrial depolarization in a UCP1-independent manner

To examine whether mitochondrial depolarization depends on UCP1 in primary brown adipocytes, we compared the mitochondrial membrane potential between control cells and UCP1 KO cells treated with isoproterenol, which is a β-adrenergic receptor agonist that promotes BAT activity (13), using live-cell imaging (Fig. 1A). We costained with tetramethylrhodamine methyl ester perchlorate (TMRM), a dye labeling the inner membrane in a membrane potential–dependent manner, and MitoTracker Green, a dye labeling the cristae membrane in a membrane potential–independent manner (16). The mitochondria of UCP1 KO primary brown adipocytes depolarized upon stimulation, similar to control cells (Fig. 1, A and B). We also confirmed efficient UCP1 KD, in which UCP1 expression was suppressed to less than half of control levels, and found that these UCP1 KD cells also exhibited mitochondrial depolarization (Figs. S1A, 1, C and D). In addition, GDP, a purine nucleotide that binds to and inhibits UCP1 at high concentrations (8, 17, 18), did not suppress mitochondrial depolarization (Fig.S1, B and C). To further assess mitochondrial depolarization, we observed OPA1 cleavage. OPA1 shows five bands, including two long forms (L1 and L2) and three short forms (S3, S4, and S5), on immunoblots (Fig. 1E). When mitochondria depolarize, OPA1 L2 is cleaved to S5 by OMA1, which is activated upon depolarization, and this processing is used as an indicator of OMA1 activity (19, 20, 21). OPA1 was also cleaved from L2 to S5 in primary brown adipocytes treated with isoproterenol, similar to cells treated with carbonyl cyanide m-chlorophenylhydrazone (CCCP) or valinomycin, which are mitochondrial depolarization inducers (Fig. 1, E and F). Under isoproterenol treatment, stress-dependent markers, such as Parkin and phosphorylated DRP1 (Ser 616), did not change their levels (Fig. S1D). Thus, OPA1 cleavage can be used as an indicator of isoproterenol-induced mitochondrial depolarization. In UCP1 KO primary brown adipocytes, OPA1 was cleaved upon stimulation, similar to control cells (Fig. 1, G and H). This cleavage was also observed in UCP1 KD cells and in the presence of GDP (Figs. 1, I and J, S1, E and F). These observations suggested that primary brown adipocytes can depolarize through a UCP1-independent pathway.Figure 1**Isoproterenol induces mitochondrial depolarization in UCP1 KO and UCP1 KD primary brown adipocytes.A and B, mitochondria in primary brown adipocytes exhibit depolarization in response to isoproterenol treatment under UCP1 KO conditions. The cells were treated with 10 μM isoproterenol for 5 h and then stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. A, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. B, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in A from n = 3 independent experiments. For statistical analysis, two-way ANOVA followed by Fisher’s least significant difference (LSD) post hoc test was performed. p Values are denoted as ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. C and D, mitochondria in *primarybrown* adipocytes exhibit depolarization in response to isoproterenol treatment under UCP1 KD conditions. The cells were treated with 10 μM isoproterenol for 5 h, and then stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. C, representative images of mitochondrial depolarization in primary brown adipocytes are shown. Scale bars represent 50 μm. D, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in C from n = 3 independent experiments. For statistical analysis, two-way ANOVA followed by Fisher’s LSD post hoc test was performed. p Values are denoted as ∗∗p < 0.01; ∗∗∗p < 0.001. E and F, OPA1 was cleaved from L2 to S5 in response to isoproterenol, CCCP, or valinomycin. The cells were treated with 10 μM isoproterenol for 5 h, or with 10 μM CCCP or 10 μM valinomycin for 3 h. E, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. F, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in E from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey's test. p Values are denoted as ∗∗∗p < 0.001. G and H, OPA1 was cleaved from L2 to S5 in primary brown adipocytes in response to isoproterenol treatment under UCP1 KO conditions. The cells were treated with 10 μM isoproterenol for 5 h. G, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. H, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in G from n = 3 independent experiments. For statistical analysis, two-way ANOVA followed by Fisher’s LSD post hoc test was performed. p Values are denoted as ∗∗∗∗p < 0.0001. I and J, OPA1 was cleaved from L2 to S5 in primary brown adipocytes in response to isoproterenol treatment under UCP1 KD conditions. The cells were treated with 10 μM isoproterenol for 5 h. I, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. J, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in I from n = 3 independent experiments. For statistical analysis, two-way ANOVA followed by Fisher’s LSD post hoc test was performed. p Values are denoted as ∗∗∗∗p < 0.0001. CCCP, carbonyl cyanide m-chlorophenylhydrazone; KD, knockdown; TMRM, tetramethylrhodamine methyl ester perchlorate; UCP1, uncoupling protein 1.

Lipolysis induces mitochondrial depolarization in UCP1 KD primary brown adipocytes

Importantly, uncoupled respiration in UCP1 KO primary brown adipocytes requires lipolysis downstream of PKA activation (13). First, we examined whether UCP1-independent mitochondrial depolarization depends on PKA, which is activated following β-adrenergic receptor stimulation. Forskolin, which stimulates PKA activity, induced mitochondrial depolarization in UCP1 KD primary brown adipocytes (Fig. 2, A and B), and OPA1 was cleaved from L2 to S5 (Fig. 2, C and D). In contrast, H89, a PKA inhibitor, suppressed mitochondrial depolarization and OPA1 cleavage in UCP1 KD primary brown adipocytes (Fig. 2, E–H). Next, to further clarify whether lipolysis is involved, we focused on ATGL, HSL, and monoacylglycerol lipase, which are the key enzymes involved in this process (1). The cells were pretreated with atglistatin, CAY10499, and JZL184, which respectively inhibit ATGL, HSL, and monoacylglycerol lipase activity, before isoproterenol treatment. Atglistatin and CAY10499 suppressed mitochondrial depolarization (Fig. 3, A and B) and OPA1 cleavage (Fig. 3, C and D). These results are consistent with previous reports that ATGL and HSL are required for UCP1-mediated uncoupled respiration in primary brown adipocytes and for cold-induced BAT activity in vivo (6, 13). FFAs, especially long-chain FFAs, released from lipid droplets are converted to fatty-acyl-CoA by ACSL1 and then transported into mitochondria via CPT1 and CPT2. Through β-oxidation, these FFAs promote UCP1-mediated uncoupled respiration (11, 22). Notably, FFAs also accelerate UCP1-independent uncoupled respiration in primary brown adipocytes (13). Based on these observations, we hypothesized that enhanced β-oxidation might contribute to UCP1-independent mitochondrial depolarization. However, inhibition of β-oxidation-related factors, such as triacsin C, an inhibitor of acyl-CoA synthetase, etomoxir and perhexiline, inhibitors of CPT1 and CPT2, did not suppress isoproterenol-induced mitochondrial depolarization or OPA1 cleavage in UCP1 KD primary brown adipocytes (Fig. 3, E–H). Taken together, these data indicate that FFAs released in response to β-adrenergic stimulation are primary candidates responsible for inducing mitochondrial depolarization.Figure 2**Mitochondrial depolarization in UCP1 KD primary brown adipocytes in a PKA-dependent manner.A and B, mitochondria in primary brown adipocytes exhibited membrane depolarization in response to forskolin treatment even under UCP1 KD conditions. The cells were treated with 10 μM isoproterenol or 10 μM forskolin for 5 h and then stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. A, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. B, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in A from n = 3 independent experiments. For statistical analysis, two-way ANOVA followed by Fisher’s least significant difference (LSD) post hoc test was performed. p Values are denoted as ∗∗∗∗p < 0.0001. C and D, OPA1 was cleaved from L2 to S5 in primary brown adipocytes in response to forskolin treatment. The cells were treated with 10 μM isoproterenol or 10 μM forskolin for 5 h. C, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. D, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in C from n = 3 independent experiments. For statistical analysis, two-way ANOVA followed by Fisher’s LSD post hoc test was performed. p Values are denoted as ∗∗∗∗p < 0.0001. E and F, isoproterenol-induced mitochondrial depolarization was inhibited by H89 pretreatment in primary brown adipocytes with UCP1 KD. The cells were pretreated with 10 μM H89 for 1 h and then treated with 10 μM isoproterenol for 5 h. Before analysis, the cells were stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. E, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. F, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in E from n = 5 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗∗p < 0.01. G and H, H89 inhibited isoproterenol-induced OPA1 cleavage (L2 to S5) in primary brown adipocytes with UCP1 knockdown. The cells were pretreated with 10 μM H89 for 1 h and then treated with 10 μM isoproterenol for 5 h. G, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. H, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in G from n = 5 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗p < 0.05; ∗∗p < 0.01. KD, knockdown; TMRM, tetramethylrhodamine methyl ester perchlorate; UCP1, uncoupling protein 1.Figure 3ATGL or HSL inhibition blocks mitochondrial depolarization in UCP1 KD primary brown adipocytes.**A and B, lipase inhibitors suppressed isoproterenol-induced mitochondrial depolarization in primary brown adipocytes with UCP1 KD. The cells were pretreated with 50 μM JZL184, 50 μM CAY10499, or 50 μM atglistatin for 1 h and then treated with 10 μM isoproterenol for 5 h. Before analysis, the cells were stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. A, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. B, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in A from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗p < 0.05; ∗∗∗p < 0.001. C and D, lipase inhibitors suppressed OPA1 cleavage from the long (L2) to the short (S5) form in primary brown adipocytes with UCP1 KD. The cells were pretreated with 50 μM JZL184, 50 μM CAY10499, or 50 μM atglistatin for 1 h and then treated with 10 μM isoproterenol for 5 h. C, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. D, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in C from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. E and F, β-oxidation inhibitors did not suppress isoproterenol-induced mitochondrial depolarization in primary brown adipocytes with UCP1 KD. The cells were pretreated with 50 μM triacsin C, 50 μM etomoxir, or 10 μM perhexiline for 1 h and then treated with 10 μM isoproterenol for 5 h. Before analysis, the cells were stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. E, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. F, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in E from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ns, not significant. G and H, β-oxidation inhibitors did not suppress OPA1 cleavage from the long (L2) to the short (S5) form in primary brown adipocytes with UCP1 KD. The cells were pretreated with 50 μM triacsin C, 50 μM etomoxir, or 10 μM perhexiline for 1 h and then treated with 10 μM isoproterenol for 5 h. G, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. H, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in G from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ns, not significant. ATGL, adipose triglyceride lipase; HSL, hormone-sensitive lipase; KD, knockdown; TMRM, tetramethylrhodamine methyl ester perchlorate; UCP1, uncoupling protein 1.

Palmitic acid predominantly induces mitochondrial depolarization in BAT

Since fatty-acyl-CoA is not required for mitochondrial depolarization in UCP1 KD primary brown adipocytes, we focused on the role of ATGL, which hydrolyzes triacylglycerol into FFA and diacylglycerol. While diacylglycerol is hydrolyzed by HSL, it can act as a second messenger for PKC (23). To test whether PKC signaling is required, we pretreated the cells with sotrastaurin, an inhibitor of PKC activity, before isoproterenol treatment, but mitochondrial depolarization and OPA1 cleavage were not suppressed in UCP1 KD primary brown adipocytes (Fig. 4, A–D). We therefore speculated that FFA induces mitochondrial depolarization. Indeed, FFA has been reported to act as protonophores (24, 25, 26). Thus, we quantified the amounts of specific FFAs—myristic acid (C14:0), palmitic acid (C16:0), stearic acid (C18:0), oleic acid (C18:1), linoleic acid (C18:2), arachidonic acid (C20:4), eicosapentaenoic acid (C20:5), and docosahexaenoic acid (C22:6)—in primary brown adipocytes (Fig. 5A). C16:0 significantly increased upon isoproterenol treatment, and this increase was suppressed by atglistatin pretreatment. Remarkably, the increase in C16:0 upon isoproterenol treatment in our study is consistent with previous analyses of BAT tissue (7). From this perspective, we expected that C16:0 was a primary candidate to directly depolarize mitochondria. To directly assess this hypothesis, we incubated mitochondria from BAT, brain, or liver with several FFAs (Fig. 5, B–D). When C16:0 or C18:0 was added, isolated mitochondria from BAT depolarized, whereas those from brain and liver did not (Fig. 5, B–D). We also assessed the effects of bovine serum albumin (BSA)–conjugated FFAs and found that treatment of isolated mitochondria with BSA-conjugated C16:0 induced mitochondrial depolarization (Fig. S2A). Furthermore, to confirm whether C16:0 acts independently of β-oxidation and FFAs transporters, such as CPT1 and CPT2, we coincubated mitochondria from BAT with etomoxir. Etomoxir did not suppress mitochondrial depolarization (Fig. 5E), and these data are consistent with the experiments in primary brown adipocytes (Fig. 3, E–H). Moreover, to check whether this phenomenon is independent of UCP1 activity, we coincubated mitochondria from BAT with GDP. When mitochondria from BAT were incubated with high concentrations of GDP, they were still depolarized (Fig. 5F). Collectively, our data suggest that C16:0 predominantly induces UCP1-independent mitochondrial depolarization in BAT. It is known that palmitic acid induces endoplasmic reticulum stress, resulting in altered Ca^2+^ flux between the endoplasmic reticulum and mitochondria and subsequent mitochondrial depolarization (27). To confirm the possibility of mitochondrial Ca^2+^ involvement, we overexpressed a mitochondrial-targeted CEPIA2mt, which is a calcium indicator. The fluorescence of TMRM was completely decreased, whereas that of CEPIA2mt did not show a substantial change (Fig. S2B). Moreover, to suppress mitochondrial Ca^2+^ overload, we pretreated primary brown adipocytes with BAPTA-AM and Ruthenium Red, which, respectively, chelate Ca^2+^ and inhibit mitochondrial Ca^2+^ uniporter activity. However, neither treatment suppressed isoproterenol-induced mitochondrial depolarization or OPA1 cleavage in UCP1 KD primary brown adipocytes (Fig. S2, C–F). These results suggest that mitochondrial depolarization is unlikely to be driven by mitochondrial Ca^2+^ uptake under our experimental conditions. Previous reports showed that FFA-mediated depolarization is partially a result of the opening of the mitochondrial permeability transition pore (mPTP) (28, 29). In addition, cyclosporine A, an inhibitor of the mPTP, suppresses UCP1-independent uncoupled respiration in primary brown adipocytes (13). Thus, we pretreated the cells with cyclosporine A before isoproterenol treatment. Mitochondrial depolarization was not restored (Figs. 6, A and B, S3, A and B), whereas OPA1 cleavage was partially rescued (Figs. 6, C and D, S3, C and D). To directly confirm the involvement of mPTP, we coincubated mitochondria from BAT with cyclosporine A, but the mitochondrial depolarization was not suppressed (Fig. 6E). Therefore, these results indicate that mPTP is unlikely to be a major contributor to palmitic acid–mediated mitochondrial depolarization.Figure 4**PKC signaling inhibition does not block mitochondrial depolarization in UCP1 KD primary brown adipocytes.A and B, PKC inhibitor did not suppress isoproterenol-induced mitochondrial depolarization in primary brown adipocytes with UCP1 KD. The cells were pretreated with 10 μM sotrastaurin for 1 h and then treated with 10 μM isoproterenol for 5 h. Before analysis, the cells were stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. A, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. B, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in A from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ns, not significant. C and D, PKC inhibitor did not block OPA1 cleavage from the long (L2) to the short (S5) form in primary brown adipocytes with UCP1 KD. The cells were pretreated with 10 μM sotrastaurin for 1 h and then treated with 10 μM isoproterenol for 5 h. C, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. D, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in C from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ns, not significant. KD, knockdown; TMRM, tetramethylrhodamine methyl ester perchlorate; UCP1, uncoupling protein 1.Figure 5Palmitic acid predominantly induces mitochondrial depolarization in BAT.A, palmitic acid levels especially increased among free fatty acids in response to ATGL-mediated lipolysis upon isoproterenol treatment. The cells were pretreated with 50 μM atglistatin for 1 h and then treated with 10 μM isoproterenol for 5 h. The lipid compositions of the cell lysates were analyzed using LC–MS/MS. Data are from n = 3 independent experiments. Only C16:0 showed a significant change when compared with the overall lipid composition (left panel). Statistical analysis was performed using one-way ANOVA with Turkey's test. p Values are denoted as ∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001. The inset (upper right) shows magnified views of C18:0, C18:1, C18:2, C18:3, and C20:4; no statistical analyses were performed for the upper panel. B–D, palmitic acid and stearic acid induced mitochondrial depolarization in isolated mitochondria from BAT. The mitochondria were isolated from (B) BAT, (C) brain, or (D) liver and then incubated with 250 μM CCCP, 50 μM C16:0 (palmitic acid), 50 μM C18:0 (stearic acid), 50 μM C18:1 (oleic acid), 50 μM C18:2 (linoleic acid), 50 μM C18:3 (α-linolenic acid), or 50 μM C20:4 (arachidonic acid) for 30 min at 37 °C. The lysates were further incubated with 100 nM TMRM for 30 min at 37 °C, and then fluorescence intensity was measured. Data are from n = 4 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗∗∗∗p < 0.0001; ns, not significant. E, palmitic acid–mediated mitochondrial depolarization in isolated mitochondria from BAT was not suppressed by etomoxir. The mitochondria were isolated from BAT and then unincubated or incubated with 250 μM etomoxir and incubated with 50 μM C16:0. Data are from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗∗∗∗p < 0.0001; ns, not significant. F, palmitic acid–mediated mitochondrial depolarization in isolated mitochondria from BAT was not suppressed by GDP. The mitochondria were isolated from BAT and then incubated with 1 mM GDP and 50 μM C16:0. Data are from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗∗∗∗p < 0.0001; ns, not significant. ATGL, adipose triglyceride lipase; BAT, brown adipose tissue; CCCP, carbonyl cyanide m-chlorophenylhydrazone; TMRM, tetramethylrhodamine methyl ester perchlorate.Figure 6mPTP inhibition does not block mitochondrial depolarization in UCP1 KD primary brown adipocytes.**A and B, the mPTP inhibitor did not suppress isoproterenol-induced mitochondrial depolarization in primary brown adipocytes with UCP1 KD. The cells were pretreated with 5 μM cyclosporine A for 1 h and then treated with 10 μM isoproterenol for 5 h. Before analysis, the cells were stained with 1 μM TMRM and 500 nM MitoTracker Green for 30 min. A, representative images of mitochondrial depolarization in primary brown adipocytes are shown. The scale bars represent 50 μm. B, quantification of TMRM fluorescence intensity normalized to MitoTracker Green from live-cell imaging as shown in A from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ns, not significant. C and D, the mPTP inhibitor did not fully block OPA1 cleavage from the long (L2) to the short (S5) form in primary brown adipocytes with UCP1 KD. The cells were pretreated with 5 μM cyclosporine A for 1 h and then treated with 10 μM isoproterenol for 5 h. C, representative immunoblots of lysates from primary brown adipocytes. Lysates of the cells were subjected to immunoblot analysis with the indicated antibodies. D, quantification of the L-OPA1 (L2) band intensity normalized to S-OPA1 (S5) from immunoblots in C from n = 3 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗p < 0.05; ns, not significant. E, palmitic acid–mediated mitochondrial depolarization in isolated mitochondria from BAT was not suppressed by mPTP inhibitor. The mitochondria were isolated from BAT and then unincubated or incubated with 10 μM cyclosporine A and incubated with 50 μM C16:0. Data are from n = 4 independent experiments. Statistical analysis was performed using one-way ANOVA with Tukey’s test. p Values are denoted as ∗∗∗∗p < 0.0001; ns, not significant. BAT, brown adipose tissue; KD, knockdown; mPTP, mitochondrial permeability transition pore; MRM, tetramethylrhodamine methyl ester perchlorate; UCP1, uncoupling protein 1.

Discussion

FFAs have been reported to play important roles of thermogenesis in BAT. Among them, FFA-mediated UCP1-independent uncoupled respiration in primary brown adipocytes has been reported (13). However, which kinds of lipids are utilized for each function, especially in the UCP1-independent pathway, is not fully understood. This study provides the first findings about the amount of FFAs and identifies that C16:0 directly induces mitochondrial depolarization in BAT.

It is generally considered that the roles of FFAs in BAT are directly interacting with UCP1 to depolarize mitochondrial membrane potential or increasing β-oxidation (8, 11, 22). Here, we demonstrate that FFAs initiate mitochondrial depolarization in BAT (Fig. 5B). The UCP1-independent mitochondrial depolarization pathway is often indirectly assessed by measuring mitochondrial respiration because the respiration increases upon treatment with uncouplers such as CCCP or valinomycin. Indeed, a UCP1-independent uncoupled respiration has been reported, and this finding is consistent with our results (13). Of note, mitochondrial respiration is generally increased by treatment with FFAs (30, 31), so that it is important to use mitochondrial membrane potential–sensitive dyes such as TMRM in cell or isolated mitochondria experiments. In addition, our data initially indicate that C16:0 dissipates mitochondrial membrane potential under the experimental condition analyzed (Fig. 5B). C18:0 also induced mitochondrial depolarization. However, its levels did not significantly change in response to isoproterenol treatment when compared with overall lipid changes. The levels of C16:0 were significantly increased, suggesting that C16:0 predominantly induces mitochondrial depolarization under isoproterenol treatment. On the other hand, C14:0, not C16:0, dissipates proton gradient in liver mitochondria (24). These observations imply that the FFAs that induce mitochondrial depolarization differ across organs or tissues. Differences in mitochondrial characteristics, such as phospholipid composition and protein expression, may underlie tissue-specific FFA–induced mitochondrial depolarization. Future studies comparing these features will help to clarify the tissue-specific mechanism.

We show that C16:0 directly induces mitochondrial depolarization in vitro (Fig. 5B); however, whether FFAs are directly imported into mitochondria from lipid droplets in primary brown adipocytes remains uncertain. It is known that mitochondria are bound to lipid droplets in BAT (32); therefore, the contact site may function as a site for direct FFA transport. On the other hand, FFAs are generally bound proteins in the cytosol; thus, there is also a possibility that they are transported from lipid droplets to mitochondria by such proteins. Furthermore, the mechanism underlying FFA-induced mitochondrial depolarization remains to be clarified. Accordingly, although our results suggest that β-oxidation may not contribute to mitochondrial depolarization in UCP1-KD primary brown adipocytes, off-target effects of pharmacological inhibition cannot be completely ruled out. To exclude these possibilities, genetic approaches, such as CPT1 or CPT2 KD experiments, will be required. Another possible mechanism is the flip–flop, which is a well-known property of FFAs and has also been discussed as a mechanism underlying proton transport activity in BAT (8, 33). Indeed, C16:0 also has protonophore activity through flip–flop (26). To assess this possibility, patch-clamp experiments are one of the best approaches. However, this analysis was not performed in the present study. Further studies are required to compare the proton conductance between C16:0-mediated and UCP1-mediated pathways using patch-clamp analysis.

However, several reports have suggested that the proton conductance of C16:0 is smaller than that of longer-chain FFAs (26, 34). In addition, in UCP1-mediated proton leak, the conductance of C20:4, C18:1, and C22:6 is more efficient than C16:1 (8). Taken together, we hypothesize that each FFA plays a distinct role at different stages of thermogenesis; first, C16:0 induces UCP1-independent mitochondrial depolarization, and then longer-chain FFAs promote UCP1-dependent mitochondrial depolarization. Because reactive oxygen species are generated after mitochondrial depolarization and promote UCP1-mediated proton conductance (35, 36), C16:0-induced mitochondrial depolarization may initiate UCP1 activity through reactive oxygen species production. To confirm these ideas, it is necessary to comprehensively measure FFAs and then trace each FFA to delineate the specific roles of individual FFAs in thermogenesis. Moreover, we were unable to determine the physiological importance of C16:0-mediated mitochondrial depolarization for thermogenesis in BAT. To address the physiological relevance of C16:0-mediated mitochondrial depolarization, it should be examined whether dietary intake of C16:0-rich foods promotes UCP1-independent thermogenesis.

In summary, we identified that C16:0 induces mitochondrial depolarization following lipolysis under cold-mimicking conditions in BAT mitochondria, and this effect may represent an initial step toward UCP1 activation. Importantly, BAT is considered a therapeutic target for metabolic diseases, such as obesity and type 2 diabetes. Therefore, future animal studies may help establish new therapeutic strategies based on this mechanism.

Experimental procedures

Cell culture

The brown preadipocytes were isolated and cultured based on the described methods with modifications (37). The stromal vascular fraction (SVF) was dissected from interscapular brown adipose tissue of newborn ICR mice or C57BL/6 mice. UCP1-KO (Ucp1^−/−^) mice were kindly provided by Dr L. Kozak (Pennington Biomedical Research Center) (38). The SVF was minced and incubated in isolation buffer (0.125 M NaCl, 5 mM KCl, 1.3 mM CaCl_2_, 5 mM glucose, 100 mM Hepes–KOH, pH 7.4, 4% BSA) with 1 mg/ml collagenase (Wako, 034-22363) at 37 °C for 30 min. The digested SVF was filtered through a 100 μm nylon cell strainer and isolated by centrifugation at 1000g for 5 min. The pellet was washed with isolation buffer and cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Nacalai-Tesque, 08458-16) supplemented with 20% fetal bovine serum (FBS) and penicillin–streptomycin. The cells were cultured in DMEM supplemented with 20% FBS and penicillin–streptomycin for 2 to 4 days. The culture medium was replaced with differentiation induction medium (DMEM supplemented with 20% FBS, 20 nM insulin [Sigma–Aldrich, I1882], 1 nM triiodo-l-thyronine [T3; Nacalai-Tesque, 35006-44], 5 μM dexamethasone [Nacalai-Tesque, 11107-64], 0.125 mM indomethacin [Tokyo Kasei Kogyo, I0655], 0.5 mM IBMX [Nacalai-Tesque, 19624-86], and 1 μM rosiglitazone [Tokyo Chemical Industries, R0106]) for 2 days and then replaced with differentiation enhancement (DE) medium (DMEM supplemented with 20% FBS, 20 nM insulin, and 1 nM triiodo-L-thyronine) for an additional 4 days.

siRNA transfection

The cells were transfected with 20 nM scrambled siRNA or UCP1 siRNAs using Lipofectamine RNAiMAX (Invitrogen, 13778-150) following the manufacturer’s protocol. Transfection was performed at the time of both differentiation induction and DE medium replacement. Scrambled siRNA was purchased from QIAGEN (#1027281). The following UCP1 siRNAs were used: UCP1 #2: 5′-TACAATGCTTACAGAGTTATA-3′, UCP1 #5: 5′-AAGCTTGTCAACACTTTGGAA-3′. Both siRNAs were purchased from Japan Bio Services.

Live-cell imaging

The cells were plated in 9.5 mm glass-bottomed dishes (MATSUNAMI, D141400). After stimulation, the cells were incubated with fresh DE medium containing 1 μM TMRM (Invitrogen, T668), 500 nM MitoTracker Green (Invitrogen, M7514), and 100 nM Lipi-Blue (Dojindo, LD01) for 30 min at 37 °C in a 5% CO_2_ incubator. The culture medium was replaced with FluoroBrite-DMEM (Gibco) supplemented with 20% FBS, 100 μM pyruvate (Agilent, 103578-100), and 2 mM GlutaMAX (Gibco, 35050-061) prior to imaging. The samples were analyzed using an FV3000 confocal laser scanning microscope (Olympus) equipped with a 60× oil immersion objective (numerical aperture = 1.42) and then quantified using Fiji/ImageJ software (39). To minimize bias, we designed macros that allowed the TMRM-labeled channel to be hidden before selecting the region of interest (ROI). ROIs were drawn around 30 Lipi-Blue–positive cells using the “Freehand selection” tool on a single-slice image in each experiment. The ratio of TMRM fluorescence to MitoTracker Green was calculated using the “Divide” operation in the “Image Calculator.” The integrated density within each ROI was measured, and the quantification represents the average integrated density of the 30 cells in each experiment.

For calcium imaging, the cells were incubated with fresh DE medium containing 1 μM TMRM (Invitrogen, T668) for 30 min at 37 °C in a 5% CO_2_ incubator. The culture medium was replaced with FluoroBrite-DMEM (Gibco) supplemented with 20% FBS, 100 μM pyruvate (Agilent, 103578-100), and 2 mM GlutaMAX (Gibco, 35050-061) prior to imaging. The samples were also analyzed using an FV3000 confocal laser scanning microscope (Olympus) equipped with a 60× oil immersion objective (numerical aperture = 1.42) and a temperature-controlled incubator set to 37 °C.

Western blot analysis

Cells were lysed in 2× Laemmli sample buffer (125 mM Tris–HCl, pH 6.8, 10% 2-mercaptoethanol, 4% SDS, 10% sucrose, and 0.01% bromophenol blue) for 5 min at 95 °C. Protein concentration was determined using the Protein Quantification Assay (Takara Bio, 74967.50/250). Proteins were resolved by SDS-PAGE and transferred to Immobilon polyvinylidene fluoride membranes (Millipore, IPVH00010). Membranes were incubated with the indicated antibodies. Protein bands were detected using Immobilon Western Chemiluminescent Horseradish Peroxidase Substrate (Millipore, WBKLS0500) and imaged using a LuminoGraph I imager (ATTO). The relative band intensities were quantified using Fiji/ImageJ software.

The following antibodies were used: UCP1 (abcam, ab10983, 1:2000 dilution), OPA1 (BD Biosciences, 612606, 1:1000 dilution), ATP5a (abcam, ab14748, 1:2000 dilution), and Vinculin (Sigma, V9131, 1:2000 dilution).

Lipid LC–MS/MS analysis

Differentiated cells were pretreated with or without 50 μM atglistatin for 1 h and then treated with 10 μM isoproterenol for 5 h. After stimulation, the culture medium was removed, and cells were washed with PBS. The cells were collected using a cell scraper, followed by centrifugation at 15,000 rpm for 10 min at 4 °C to collect the cell pellet. The pellet was resuspended in 300 μl of cold methanol containing 1 μM arachidonic acid-d8 (Cayman Chemical Company, 390010) and 100 μM butylated hydroxytoluene and then subjected to sonication for 10 s. After centrifugation at 15,000 rpm for 10 min at 4°C, the supernatant was collected for LC–MS/MS analysis.

LC–MS/MS was performed using an LCMS-8060 system (Shimadzu) equipped with an electrospray ionization source. FFAs were quantified in multiple reaction monitoring mode. The multiple reaction monitoring transitions for each FFA are summarized in the supporting information table. Quantification was performed based on the peak area ratios of each analyte to the internal standard (arachidonic acid-d8).

The LC conditions were as follows: injection volume, 5 μl; autosampler temperature, 4 °C; column, InertSustain C18 column (2.1 mm × 150 mm, 3-μm particle size, GL Sciences); column temperature, 40 °C; mobile phase, 5 mM ammonium formate (Sigma–Aldrich, 516961) in acetonitrile (Wako, 016-19854):H_2_O (Wako, 210-01303) = 2:1 (A) and isopropanol (164-25533):methanol (Wako, 132-14524) = 95:5 (B); flow rate, 0.4 ml/min; gradient elution, 0 to 22.5 min (0%–100% B) and 22.5 to 27.5 min (100% B). The LC–MS/MS analysis was performed using Multi-ChromatoAnalysT, version 1.3.3.0 (Beforce).

Mitochondria isolation

Differentiated cells were resuspended in homogenization buffer (250 mM sucrose, 1 mM EDTA [pH 8.0], 10 mM Hepes–KOH [pH 7.4]). Cells were homogenized using a syringe with a 27G needle for 10 strokes. The cell homogenate was centrifuged at 800g for 10 min, and then the supernatant was centrifuged at 8000g for 10 min. The pellet was collected as crude mitochondria.

Organs were harvested from C57BL/6JmsSlc mice. Organs were homogenized using a Potter homogenizer at 300 rpm for 20 strokes in homogenization buffer. The organ homogenate was centrifuged twice at 800g for 10 min, and then the supernatant was centrifuged at 8000g for 10 min. The pellet was collected as crude mitochondria. All animals were maintained under university guidelines for the care and use of laboratory animals. The experiments were performed after securing protocol approval from the Gakushuin University Animal Use Committee.

Preparation of fatty acid–free BSA-conjugated fatty acid

Fatty acid–free BSA (pH 7.0; Nacalai-Tesque, 08587-26) was dissolved at 30% (w/v) in homogenization buffer and warmed to 37 °C. Then, 10 mM fatty acids were added to the warmed fatty acid–free BSA-containing homogenization buffer and incubated at 37 °C for 10 min. During incubation, the solution was mixed by vortexing. The final BSA:FFA molar ratio was 2:1.

Fluorescence measurement

The mitochondria were incubated in homogenization buffer supplemented with 2 mM ADP (Sigma, A2754), 1 mM pyruvate (Agilent, 103578-100), malic acid–NaOH (pH 7.0) (Nacalai-Tesque, 21030-44), and either 250 μM CCCP or 50 μM fatty acids, either fatty acid free BSA–conjugated or unconjugated, for 30 min at 37 °C. Then, 100 nM TMRM was added, and the mitochondria were further incubated for 30 min at 37 °C. Fluorescence was measured using a Synergy LX multimode reader (BioTek).

DNA constructs

To express mitochondrial calcium indicator, pCMV CEPIA2mt (Addgene, #58218) was subcloned into pLKO.1 vector (40, 41).

Lentivirus production and infection

Lenti-X 293T cells were plated in 6-well plates. Cells were transfected with lentiviral plasmid, psPAX2, which is the packaging plasmid, and VSV-G, which is the envelope plasmid, using Lipofectamine 3000 (L3000015, Invitrogen). Transfections were performed in Opti-MEM, and cells were maintained in DMEM containing 10% FBS. After 3 days, viral supernatants were collected and filtered through a 0.22-μm filter. The viral supernatants were diluted in culture medium at the time of plating SVF dissected from interscapular brown adipose tissue. After 24 h, the medium was changed to fresh culture medium.

List of reagents

Isoproterenol (I0620), forskolin (F0855), and stearic acid (S0163) were purchased from Tokyo Chemical Industries. SR 59230A (21407), H-89 (10010556), JZL184 (13158), CAY10499 (10007875), perhexiline (16982), sotrastaurin (16726), α-linolenic acid (90210), arachidonic acid (90010), valinomycin (10009125), and Ruthenium Red (14339) were purchased from Cayman Chemical Company. Atglistatin (10-4598) and triacsin C (10-2190) were purchased from FOCUS Biomolecules. Etomoxir (ab254445) was purchased from Abcam. Octanoic acid (07112-72), lauric acid (20108-32), myristic acid (23517-82), palmitic acid (25918-72), oleic acid (25630-51), linolic acid (20513-41), cyclosporine A (59865-13-3), and BAPTA-AM (03731-24) were purchased from Nacalai-Tesque. CCCP (038-16991) was purchased from Wako.

Statistical analysis

The data are presented as the mean ± SEM and expressed as percentages or multiples relative to the values in control cells. Statistical significance was analyzed using Prism 10 software (GraphPad Software). Statistical significance was analyzed either using one-way or two-way analysis of variance followed by Bonferroni’s test as appropriate. p Values are denoted as ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; and ns, not significant.

Data availability

All data are contained within the article.

Supporting information

This article contains supporting information.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.