Male but not female lFgf21-/- mice on regular chow diet (defined as low fat diet, LFD) feeding show impaired fat tolerance
We have generated lFgf21-/- mice and the littermate controls (Fgf21fl/fl) for current study by mating Alb-Cre with Fgf21fl/fl, as illustrated in Fig. S1A. lFgf21-/- mouse liver showed barely detectable FGF21 (Fig. S1B-D), while FGF21 expression in both brown adipose tissue and hypothalamus were virtually unaffected (Fig. S1E-F).
Six-week-old male and female lFgf21-/- mice and the control Fgf21fl/fl mice were fed with regular chow diet (low fat diet, LFD) for 12 weeks, while three metabolic tolerance tests (glucose tolerance test, GTT; insulin tolerance test, ITT; and pyruvate tolerance test, PTT) were conducted at indicated time for all mice, as indicated in Fig. 1A. Apparently, on LFD feeding, male or female lFgf21-/- mice exhibited comparable glucose, pyruvate, and insulin tolerance when compared with sex-matched control littermates Fgf21fl/fl mice (Fig. 1B-G). There were no appreciable differences on body weight with hepatic FGF21 knockout in both male and female mice (Fig. 1H-I).
The above mice were then fed with LFD for an additional 3-week period (Fig. 1A), followed by collecting random blood samples and conducting fat tolerance tests (FTT). Mice were then fasted overnight before they were sacrificed for tissue sample collections. As shown, male lFgf21-/- mice exhibited elevated random and fasting plasma TG levels, as well as impaired fat tolerance when compared with correspondent littermate controls (Fig. 2A-B). Such defects were not observed in female lFgf21-/- mice (Fig. 2C-D). We then assessed the expression of a battery of hepatic genes that are related to the FGF21 cellular signaling. Expression levels of genes that encode FGFR1 (Fgfr1) and KLB (Klb), as well as the two FGF21 downstream mediators (Ehhadh, which encodes enoyl-CoA hydratase and 3-hydroxyacyl CoA dehydrogenase; and Ppargc1a, which encodes PPARG coactivator 1 alpha) were reduced in male lFgf21-/- mice (Fig. 2E). In female lFgf21-/- mice, only Ppargc1a level was reduced (Fig. 2F). We have also assessed expression of liver Fgf1 and gut (ileum) Fgf15, asking whether hepatic FGF21 deficiency results in elevated Fgf1 or Fgf15 expression for compensation. No such compensatory elevation was observed in lFgf21-/- mice (Fig. 2G-2H). Together, we conclude that male but not female lFgf21-/- mice showed impaired lipid homeostasis in the absence of obesogenic dietary challenge, associated with reduced hepatic expression of Fgfr1, Klb, and Ehhadh. Our following investigations were then performed on male mice only.
Dietary curcumin intervention improves lipid homeostasis in the control Fgf21fl/fl mice but not in lFgf21-/- mice with the obesogenic dietary challenge
We have reported previously that in HFD-challenged male mice, curcumin or anthocyanin intervention regulated hepatic FGF21 production and improved FGF21 sensitivity in hepatocytes (6, 23). Curcumin intervention stimulated hepatic FGF21 expression in mice on LFD feeding and attenuated HFD-induced hepatic FGF21 over-expression and FGF21 resistance (6). Here we aimed to determine whether hepatic FGF21 is required for curcumin in exerting its major metabolic beneficial effect, especially lipid homeostatic effect in mice on obesogenic dietary challenge. As shown, male lFgf21-/- mice or the control littermate Fgf21fl/fl mice were fed with high fat high fructose (HFHF) diet without or with curcumin intervention for 15 weeks (Fig. 3A). In the control littermates, curcumin intervention moderately attenuated HFHF-diet induced body weight gain (Fig. 3B-C), reduced fasting serum and hepatic TG contents (Fig. 3D-E), and improved lipid tolerance (Fig. 3F). Curcumin dietary intervention also reduced hepatic expression of ChREBP, as well as expression of genes that encode ChREBP (Mlxipl) and fatty acid synthase (Fasn) (Fig. S2A-B). In lFgf21-/- mice, none of the above regulatory effects of dietary curcumin intervention were observable (Fig. 3G-K and Fig. S2C-D). We hence conclude that liver FGF21 is required for curcumin intervention in exerting its metabolic beneficial effect, especially the improvement of lipid homeostasis.
In obesogenic diet challenged male mice, dietary resveratrol intervention also regulates hepatic FGF21
In 2014, Li and colleagues have reported that resveratrol treatment increased the transcriptional activity of the FGF21 gene promoter (7). We hence ask whether in vivo resveratrol intervention in wild type mice with an obesogenic dietary challenge affects hepatic FGF21 expression, FGF21 sensitivity, or FGF21 mediated cellular signaling events. Here we conducted such assessments in two different sets of mice. In the first set, wild type C57BL/6J mice were fed with LFD, HFD, or HFD with resveratrol (HFD+Res) intervention for 8 weeks (Fig. 4A). In such experimental settings, hepatic Fgf21 expression was reduced by HFD feeding, and the reduction was attenuated by resveratrol intervention (Fig. 4B). Hepatic FGF21 protein expression was not significantly affected by HFD while concomitant resveratrol intervention exhibited a stimulatory effect on hepatic FGF21 expression (Fig. 4C). Importantly, resveratrol intervention increased expression of Fgfr1, which was reduced by HFD challenge (Fig. 4D). Klb level was not significantly affected by 8-week HFD challenge, while resveratrol intervention exhibited a stimulatory effect on hepatic Klb expression (Fig. 4E). Among the four downstream effectors of FGF21 we have assessed, expression of Ehhadh was inhibited by HFD and the inhibition was effectively reversed by concomitant resveratrol intervention. Acox1 (encodes Acyl-CoA Oxidase 1) expression was not affected by HFD challenge while resveratrol intervention elevated its expression level. HFD challenge significantly reduced expression level of Ppargc1a, while resveratrol intervention generated no appreciable reversing effect in the current experimental settings. Finally, hepatic Pdk4 (which encodes pyruvate dehydrogenase lipoamide kinase isozyme 4) level was not affected by HFD challenge or resveratrol intervention in our current experimental settings (Fig. 4F).
The above observations suggest that like curcumin intervention previously reported by our team and by others (6, 24, 25), resveratrol intervention can also target hepatic FGF21 or its downstream signaling events. Fructose consumption is known to stimulate hepatic FGF21 expression, associated with the development of insulin resistance (26). In the second set of mice, we challenged Fgf21fl/fl mice with HFHF-diet without and with resveratrol intervention for an extended period, as indicated in Fig. 5A. As shown, 16-week HFHF-diet challenge significantly increased hepatic Fgf21 levels, while dietary resveratrol intervention attenuated the elevation effectively (Fig. 5B). At FGF21 protein level, elevation was observed in mice with HFHF challenge, without or with 16-week resveratrol intervention (Fig. 5C).
We then collected liver tissues from those mice for RNAseq analysis. As we have anticipated (Fig. 5D, Fig. S3, and Table S4), HFHF-diet challenge increased hepatic expression of Fgf21 but reduced the expression of Klb, while those effects were reciprocally reversed by 16-week dietary resveratrol intervention. The attenuation effect of HFHF-diet and reversible effect of resveratrol intervention was also observed on certain downstream effectors of FGF21 signaling, which were then further verified by our qRT-PCR experiment (Fig. 5E-G). As shown in Fig. S3, HFHF diet feeding most significantly repressed expression of genes including Eif4ebp3 (which encodes eukaryotic translation initiation factor 4E binding protein 3) and Zbtb16 (which encodes zinc finger and BTB domain-containing protein 16), and those genes were also significantly restored by resveratrol intervention. Exact metabolic functions of these two genes remain to be further explored. Additional information on the effect of HFHF diet challenge and resveratrol intervention on hepatic gene expression are presented in Fig. S4 and S5.
Dietary resveratrol intervention improves glucose tolerance and reduces serum and hepatic TG levels in HFHF challenged Fgf21fl/fl mice but not in HFHF challenged lFgf21-/- mice
We then directly compared the effect of dietary resveratrol intervention in Fgf21fl/fl mice and lFgf21-/- mice with HFHF-diet challenge (Fig. 6A). As shown, glucose tolerance was improved by dietary resveratrol intervention in the control littermate Fgf21fl/fl mice but not in lFgf21-/- mice (Fig. 6B). Dietary resveratrol intervention also attenuated HFHF-diet induced fasting hyperglycemia and hyperinsulinemia in Fgf21fl/fl mice but not in lFgf21-/- mice (Fig. 6C-D). Interestingly, in both Fgf21fl/fl and lFgf21-/- mice, 16-week dietary resveratrol intervention attenuated HFHF-diet induced body weight gain (Fig. 6E-F) and fat accumulation (Fig. S6), although the degree of the attenuation in Fgf21fl/fl mice appeared much stronger than that in lFgf21-/- mice (Figure 6E-F and Fig. S6). Nevertheless, in Fgf21fl/fl mice but not in lFgf21-/- mice, dietary resveratrol intervention attenuated HFHF diet induced elevation on serum as well as hepatic TG levels (Figure 6G-H). Thus, although hepatic FGF21 is required for resveratrol intervention in exerting its metabolic beneficial effect on improving energy homeostasis, the body weight lowering effect of dietary resveratrol intervention does not completely rely on hepatic FGF21. Whether this involves FGF21 expressed in adipose tissue, or the brain deserves further investigations.