*2.5. Effects of FO Supplementation on Gene Expression in subWAT*

FO dramatically increased *Ucp1* mRNA expression by 6.3-fold in the subWAT (Figure 4a). UCP1 protein levels were also analyzed (n = 7 in each group), and representative data (n=2 in each group) indicating a 2.7-fold increase compared with those in the Con-fed mice are shown in Figure 4b. FO supplementation also caused higher expressions of *Pparα* and its target genes of *Cpt I*, *Aco* and *Mcad* in comparison to those in the Con-fed mice (Figure 4a). Fgf21 expression in the FO-fed mice was also increased by 2.3-fold compared with that in the Con-fed mice (Figure 4a). β3-AR mRNA expression was higher in subWAT from the FO-fed mice than that in the Con-fed mice (Con, 100.0 ± 17.9%; FO, 246.1 ± 40.4%, *p* < 0.001).

Among the brown fat-selective genes, expression of cell death-inducing DNA fragmentation factor α-like effector a (Cidea) was significantly increased by 3.9-fold in the FO-fed mice compared with that in the Con-fed mice, whereas that of PR domain containing 16 (Prdm16) mRNA was not different (Figure 4a).

**Figure 4.** Effect of fish oil (FO) supplementation on gene expression and UCP1 protein levels in subcutaneous white adipose tissue. The mRNA levels of *Ucp1*, *Pparα* and its target genes and beige adipocyte-specific gene (**a**) and UCP1 protein (**b**) were assessed by quantitative RT-PCR or western blotting. β-actin was used as the normalization control. The percent of mRNA and protein levels relative to those of control fat (Con)-fed mice are indicated. White and gray columns represent data from the Con- and the FO-fed mice, respectively. Values are mean ± SEM (n = 7). \* *p* < 0.05, \*\* *p* < 0.01 vs. Con-fed mice. Significant differences between two groups were tested by Student *t*-test.

### *2.6. Effects of FO Supplementation on Gene Expression in Liver*

As the effects of supplementation could also be caused by increased metabolism in the liver, we analyzed gene expressions in the liver related to fatty acid β-oxidation and fatty acid synthesis. As shown in Table S1, fatty acid β-oxidation was induced the FO-fed mice, and fatty acid synthesis was decreased.

#### **3. Discussion**

We found that FO increased DIT in mice by 1.2-fold along with the activation of BAT caused by the increased expression of UCP1 and the browning of subWAT. As females are reported to produce less heat than males, we used male mice for our experiment [40]. We observed DIT in both the light period and dark period, although it was higher in the latter because mice eat principally in the dark period. Actually, the Con and FO groups of mice took about 70–80% and 20–30% of their total food intake in the dark and light periods, respectively. It appeared that maintenance of DIT in the light period was caused by feeding in the light period (Figure 1a).

In this study, FO increased the expression of the *Ucp1* gene in BAT by 1.5-fold. Other researchers also reported that FO administration both in vitro and in vivo induced the increased expression of *Ucp1* in BAT. *Ucp1* mRNA expression and UCP1 protein levels were both significantly increased in brown progenitor cells isolated from interscapular BAT supplemented with EPA [35]. EPA also increased mitochondrial content in a dose-dependent manner in HIB 1B brown adipose cells [41]. EPA administration to C57BL/6J mice for 11 weeks, and DHA-enriched FO (DHA 25%, EPA 8%) or EPA-enriched FO (DHA 12%, EPA 28%) administration to mice for 10 weeks, significantly increased UCP1 protein levels and *Ucp1* mRNA expression in BAT [34,42]. UCP1 activity in BAT was significantly increased in rats fed with EPA or a mixture of EPA and DHA for 4 weeks by GDP binding [43]. These reports support our results that UCP1 expression was significantly increased by FO administration in BAT (Figure 3a,b). The nuclear receptor PPARα regulates the expression of *Cpt I*, *Mcad* and *Aco*, which are involved in the fatty acid β-oxidation [41,44,45]. FO administration increased the expression of *Pparα* mRNA by 1.3-fold (*p* < 0.05), but expressions of its target genes *Cpt I*, *Mcad* and *Aco* were not affected in BAT, although that of the other Pparα target gene, *Ucp1*, increased (Figure 3a). Kim et al. reported that the expression of *Cpt I* mRNA in BAT of mice fed EPA-enriched FO increased significantly compared with that of control mice. In contrast, the expression of *Cpt I* mRNA did not increase in mice fed DHA-enriched FO, which has a similar fatty acid ratio as in the present study [34]. The reason why EPA-enriched FO could, but DHA-enriched FO could not, induce *Cpt I* expression in BAT is currently not clear. Further study will be required to reveal the mechanism.

FO also increased the expressions of the *Ucp1* gene and other genes related to the fatty acid β-oxidation in subWAT (Figure 4a). In terms of the marker of browning, *Cidea* mRNA was increased by 3.9-fold, but the expression of *Prdm16* was not increased (Figure 4a). FO enhances fatty acid oxidation through PPARα activation in WAT and causes browning of subWAT [34,46]. When cells derived from subcutaneous adipocytes from overweight females were treated with 200 μM EPA, expressions of *UCP1* and *CIDEA* mRNA increased significantly. The mRNA expression of *PRMD16* increased significantly with 100 μM EPA treatment but not with 200 μM EPA treatment [31]. When the stromal vascular cells isolated from subWAT of C57BL/6J mice were treated with 200 μM EPA during a differentiated process, the expressions of fatty acid β-oxidation-related genes *Ucp1*, *2*, *3* and *Cpt I*, and *Cidea*, increased significantly, but that of *Prdm16* was still not increased as in our results [32]. The reasons for FO causing different expressions of *Cidea* and *Prdm16* are currently unknown. Due to the increased expressions of *Ucp1* and *Cpt I* mRNA in FO-fed mice, the brown adipocyte-like phenotype was induced in subWAT [33]. The PPARα agonist is known to promote browning in subWAT [47,48] and increase the body temperature [48]. Contrary to these reports, UCP1 protein is reported to be very low or undetectable in subWAT even though mice were fed FO [42]. Our results supported the findings that FO administration markedly increased UCP1 expression in subWAT and induced subWAT browning. Beige adipocytes were shown to have potent thermogenic ability comparable to classical BAT [30], and the thermogenic density and total quantitative contribution in subWAT were maximally one-fifth and one-third of all BAT mitochondria, respectively [49]. Thus, the classical BAT depots would still be predominate in thermogenesis, but the browning of WAT would also contribute to thermogenesis. Sato et al. recently showed that phospholipase A2 group IID, which is expressed in M2-type macrophages in WAT, released n-3 fatty acid and increased energy expenditure and rectal temperature by facilitating subWAT browning, which ameliorated diet-induced obesity [50]. Thus, FO-caused browning of WAT might also contribute to inducing DIT.

FGF21 is reported to have an endocrinological role in BAT and WAT [51,52]. The expression of *Fgf21* mRNA in subWAT increased dramatically in mice after exposure to cold [52]. Moreover, the differentiated primary subWAT treated with β-agonist synthesized and secreted FGF21, suggesting that adipose FGF21 may act mainly in a paracrine/autocrine manner [52]. However, similar to the previous research concerning FO [34], the expression

of *Fgf21* did not increase with FO administration in BAT (Figure 3a). However, contrary to that report, *Fgf21* expression in the present study was significantly increased by FO administration in subWAT (Figure 4a). This result leads us to the hypothesis that increased *Fgf21* of subWAT might induce the browning of WAT observed in the present study.

FO is reported to induce UCP1 expression in BAT and WAT via the sympathetic nervous system and transient receptor potential vanilloid 1 [34]. In the present study, β3-AR mRNA expression was higher in subWAT from the FO-fed mice than that in the Con-fed mice. However, no difference in β3-AR mRNA was observed in BAT. Although we did not determine the direct influence of fish oil on sympathetic flow, over the short term of 10 days, FO intake might induce UCP1 expression in subWAT via the sympathetic nervous system at least in part.

G-protein-coupled receptor 120 (GPR120), a receptor for n-3 polyunsaturated fatty acids, was also suggested to contribute to thermogenic activation in BAT and WAT by n-3 fatty acids by suppressing tissue inflammation induced by macrophages, especially in obese mice [53–56]. We used non-obese mice, and the expression of *Gpr120* was not affected in BAT and subWAT by FO supplementation (data not shown).

A systematic review indicated that EPA and DHA lowered serum lipid levels such as TG concentration [57]. Some mechanisms of serum lipid lowering by FO have been reported. EPA increased lipid oxidation in rat liver and reduced serum lipids [58]. FO also lowered serum lipids in adult human subjects [59]. We previously reported that FO administration at the same dose as in the present study decreased fatty acid synthesis genes such as acetyl-CoA carboxylase and increased fatty acid oxidation genes such as *Cpt I*, *Mcad* and *Aco* in mouse liver [60]. The rate limiting step in mitochondrial fatty acid oxidation is mediated by CPT I [61]. Even though CPT I activity in WAT was still low compared with that in liver and BAT in rat [62], activation of CPT I by overexpression of CPT I in 3T3-L1 adipocytes reduced NEFA release [63]. These FO-induced mechanisms in liver and WAT may contribute to lowering of the serum lipid levels. In general, enhanced fatty acid oxidation in the whole body is related to decreased RER. However, in human, RER correlated negatively with plasma palmitate concentrations [64]. In the present study, FO administration caused decreased serum concentrations of NEFA and TG (Table 2). We showed here that the RER of the FO-fed mice was slightly higher than that of the Con-fed mice (Figure 1c), although not significantly so. This was probably due to the reduced serum lipid levels in the FO-fed mice.

The short period of FO administration of 10 days in the present study did not result in weight loss, but weight gain and the weights of epididymal and mesenteric WAT were significantly reduced. Mice fed 21.42 or 42.84 energy% (en%) FO for 6 weeks significantly reduced BW by about 1.5 g or 4 g, respectively [36]. It is likely that mice need to be fed FO for a longer period of time to reduce their weight. BAT-positive subjects would undergo higher DIT than BAT-negative subjects [65]. Thus, BAT activation is expected to have an anti-obesity effect. Interestingly, BAT-positive subjects (young healthy men) showed an increase in EE after oral ingestion of capsinoids (9 mg) [15]. Moreover, capsinoids 6 mg/day taken orally for 12 weeks promoted loss of human abdominal fat [66]. FO supplementation in the present study resulted in a 1.2-fold increase in DIT/intake (Figure 2e). In human, DIT uses 10% of the daily energy intake [67]. The estimated energy requirement for adult men is about 2500 kcal/day [68], and the energy consumed by DIT was calculated to be about 250 kcal/day, and 300 kcal/day if multiplied by 1.2. Thus, a 1.2-fold increase in DIT was estimated to maximally increase energy consumption by 50 kcal/day. Adult human adipose tissue contains 71.6% crude fat [69]. Therefore, an increase in DIT by 1.2-fold was estimated to indicate fat burning of 500 g of adipocytes over about 2 months.

In conclusion, we first showed that FO supplementation significantly increased DIT by 1.2-fold. DIT/intake and DIT/TEE for the FO-fed mice were 11.2% and 22.3%, respectively. The FO-increased DIT was complemented by the increased expression of UCP1, activation of BAT and subWAT browning. FO may be a promising dietary fat for the prevention of overweight and obesity.
