FGF21 KO mice developed cardiac dysfunction before impaired metabolic state onset
To understand the mechanism of how FGF21 possesses beneficial effects on cardiovascular health, we conducted FGF21 KO mice (Fig.S1 A-D), and we closely monitored the metabolic baseline data and cardiac function of FGF21 KO mice. Unexpectedly, though FGF21 is a key regulator of metabolism, there are significant differences in cardiac diastolic function between FGF21 KO mice and their littermate in young age (16 weeks old) but not in blood sugar and body weight, as evidenced by comparable global longitudinal strain, longitudinal and radial strain time to peak (Fig. 1A-C, and Fig.S1E-1H). In young age FGF21 KO heart also found mild hypertrophic features and increased cardiac fibrosis (Fig. 1D, 1E). Meanwhile, the cardiac impaired marker BNP, hypertrophy marker MYH7, and fibrosis marker was also increased (Fig. 1F). Although eject fraction has no significant impairment in young age FGF21 KO mice, severe systolic dysfunction developed after 22 weeks old (Fig. S2A, B). These results suggest that FGF21 play a key role in maintain cardiac function.
FGF21 KO mice OPA1 degradation increased, and mitochondrial structure was severely damaged
To investigate how FGF21 deficiency mediates the cardiac dysfunction, we next examined the morphological and function of heart tissue in FGF21 KO mice and their littermates. TEM analysis demonstrated that sarcomere arrangement and mitochondrial morphological were disrupted in FGF21 KO mice (Fig. 2A). The average cross-sectional area of individual mitochondrial, mitochondrial perimeter, and mitochondrial area were significantly reduced in FGF21 KO mice (Fig. 2B). A 5-grade scoring system analysis showed that FGF21 deficiency induced severe damage in mitochondrial morphology with inflation, warped membranes, irregularities, and absence of cristae (Fig. 2C,2D, and Fig.S3A).
To elucidate the mechanism of mitochondrial damage, we further investigated mitochondrial dynamic related protein, mitochondrial genesis associated protein, and mitophagy indicator. There was no difference of mitochondrial genesis protein but in mitochondrial dynamic related protein OPA1 was significantly decreased, and mitophagy related protein PINK and PARKIN were significantly increased in FGF21 KO mice (Fig. 2E, 2F). We further investigated the mRNA levels of related protein, there are no differences between FGF21 KO mice and their littermates (Fig. 2G). To further investigate whether the OPA1 decreased impact the mitochondrial dynamics in cardiomyocytes, we use Mitotracker staining to evaluate the mitochondrial morphology in isolated adult cardiomyocytes. As expected, the FGF21 KO isolated adult cardiomyocytes mitochondrial is characterized by increased splitting and swelling (Fig. 2H-J). These results suggest that FGF21 deficiency may cause increased OPA1 degradation and mitochondrial damage.
Restoration of cardiac FGF21 reverses impaired cardiac function in FGF21 KO mice
To confirm FGF21 has cardiac-specific effect on protecting mitochondrial damage, an adeno-associated virus carrying a TNT-GFP promoter (AAV-TNT-GFP) was used to monitor FGF21 cardiac-specific overexpression in FGF21 KO mice (Fig. 3A, Fig.S4A). FGF21 cardiac-specific overexpression had no effect on metabolic status (Fig.S4B, C), but obviously improved the cardiac dysfunction caused by FGF21 deficiency 16 weeks after injection, as shown by the EF%, GLS% (Fig. 3B, C). Consistently, FGF21 cardiac-specific overexpression reversed cardiac hypertrophy and fibrosis caused by FGF21 deficiency (Fig. 3D, 3E). Cardiac impaired marker BNP, hypertrophy marker MYH7, and fibrosis marker was also decreased in FGF21 cardiac-specific overexpression mice (Fig. 3F). Western blotting identified that OPA1 expression was increased and mitophagy is inhibited in FGF21 cardiac-specific overexpression mice (Fig. 3G, H).
Furthermore, Mitotracker staining of the FGF21 KO mice mitochondrial in isolated adult cardiomyocytes reveal that FGF21 cardiac-specific overexpression reversed increased mitochondrial splitting and swelling (Fig. 3I-3K). These results demonstrated that FGF21 cardiac-specific overexpression could reverse cardiac dysfunction caused by FGF21 deficiency.
FGF21 deficiency caused OPA1 degradation and abnormal mitochondrial structure and function in vitro
To further explore the effect of FGF21 in maintaining mitochondrial homeostasis, we transfected cardiomyocytes with siFGF21. As expected FGF21 deficiency caused OPA1 degradation (Fig. 4A). Mitotracker staining showed increased fission with decreased mitochondrial network branches and branch length in cardiomyocytes with siFGF21 transfection (Fig. 4B, C). To validate whether FGF21 deficiency could affect mitochondrial oxidative phosphorylation, we used an extracellular flux analyzer to assess mitochondrial function in cardiomyocytes. FGF21 deficiency cardiomyocytes had similar levels of basal respiration and proton leak compared with normal cardiomyocytes, but the maximal respiratory capacity and spare capacity were markedly reduced (Fig. 4E-G).
Overexpression of FGF21 has no effect on physiological mitochondrial function
As the role of FGF21 in maintaining mitochondrial homeostasis was confirmed, we explored whether FGF21 could regulates mitochondrial function under physiological condition. We further evaluated the mitochondrial dynamics and function after given recombination FGF21 or adFGF21. Under physiological status recombination FGF21 or AdFGF21 has no effect on mitochondrial dynamics and function (FigS5A-J).
In vitro both autocrine and exogenous FGF21 could defenses mitochondrial damage under stressing
To further explored whether FGF21 could defenses mitochondrial damage under stressing conditions, we use different stimulators to detect the expression of FGF21 and OPA1 in cardiomyocytes. As shown in (Fig.S6 A-F), CoCl2 can active Hif1-a expression, and inhibited FGF21 and OPA1 expression in dose- and time-dependent manner. Moreover, CoCl2 stimulation also cause increased mitochondrial fission and impairment of mitochondrial function (Fig.S6 G-K).
Under CoCl2 stimulation, recombination FGF21 and adFGF21 could both restore OPA1 levels (Fig. 5A-D). Then we further investigated the mitochondrial morphology, given recombination FGF21or adFGF21 both can restore mitochondrial dynamic, decrease fission, also could increase fusion and branches forming, under CoCl2 stimulation (Fig. 5E-H). As expected, FGF21 overexpression dramatically improved both spare and maximal respiratory capacity but did not affect basal respiration, H+ leak, or respiration attributed to basal F1/F0 ATP synthase activity (Fig. 5I-K).
FGF21 is an endocrine protein, mainly secreted by liver but cardiomyocyte could also synthesis and secrete FGF21. As shown in (Fig.S7A), the adFGF21 cardiomyocytes culture supernatant was with significantly high FGF21 levels. When given Monensin, which is an intracellular protein transport blocker, the high FGF21 levels in the adFGF21 cardiomyocytes culture supernatant was decreased. In contrast, Monensin administration ignored the rescue of adFGF21 on mitochondrial damage induced by CoCl2 (Fig.S7B-7G).
Inhibition of FGFR1 impairs the mitochondrial protected effect of FGF21 under CoCl 2 stimulation in vitro
As previous study(22, 23), FGFR1 and FGFR3 are the mainly expressed FGF21 receptors on heart (Fig. S8A). To determine FGF21 to exert its effect on heart through FGFR1 or FGFR3. First, we transfected cardiomyocytes with siFGFR1 or siFGFR3 respectively. Only FGFR1 deficiency could inhibitor OPA1 restoration caused both by recombination FGF21 and endogenous FGF21 under CoCl2 stimulation (Fig. S8B, C, Fig. 6A-6D). We next sought to validate whether FGF21 exert its effect on mitochondrial through activating FGFR1. As expected, both endogenous or exogenous FGF21 couldn’t prevent OPA1 degradation in FGFR1 absence cardiomyocytes (Fig. 6A-D) and Mitotracker staining shows in FGFR1 absent cardiomyocytes FGF21 couldn’t restore decreased mitochondrial fusion caused by CoCl2 (Fig. 6E-H). We also test the mitochondrial oxidative phosphorylation in these FGFR1 absent cardiomyocytes. Upregulation of spare and maximal respiratory capacity by FGF21 in CoCl2 stimulated cardiomyocytes was absent (Fig. 6I-N). These results suggest that activation of FGFR1 is required for FGF21-mediated mitochondrial protected effect.
OPA1 mediated mitochondrial damage caused by FGF21 deficiency or under CoCl2 stimulation
To determine whether the regulating of OPA1 plays an important role in FGF21-induced the mitochondrial homeostasis. We overexpressed the OPA1 by transfected with OPA1 plasmid in FGF21 knock down cardiomyocytes or cardiomyocytes under CoCl2 stimulation and investigated the mitochondrial morphological change. As shown in (Fig. 7A-D), overexpression of OPA1 could restore mitochondrial dynamics with increased mitochondrial network branches and branch length in FGF21 knock down cardiomyocytes or cardiomyocytes under CoCl2 stimulation. Overexpression of OPA1 could also restore the mitochondrial function and reduce the mitophagy which cause by FGF21 deficiency or CoCl2 stimulation (Fig. 7E-I). These results suggest that FGF21 maintain mitochondrial homeostasis through regulating OPA1.
FGFR1 regulate OPA1 degradation through E3 ligase FBXW11
Due to FGF21 deficiency and overexpression of FGF21 has no effect on mRNA levels of OPA1, the possible OPA1 degradation modulators were identified in cardiomyocytes. Firstly, we found that MG132 could inhibitor OPA1 degradation in time-dependent manner (Fig. 8A, B). In contrast, FGF21 administration remarkedly reduced the OPA1 degradation in cardiomyocytes under CHX stimulation (Fig. 8C, D). These results indicate that FGF21 regulated the degradation of OPA1.
To determine the E3 ligase binding to OPA1, we use UbiBrowser to predict potential OPA1 E3 ligase (Fig.S9A). We evaluated several E3 ligases expression, FBXW11 was detected to obvious increase in FGF21 KO mice heart compared to that in littermate mice in protein level but not in mRNA level (Fig. 8E, F, and Fig. S9B). Further with CoCl2 stimulation, FBXW11 remarkedly increased in cardiomyocytes associated with the reduction of OPA1 and FGF21 (Fig. 8G, H). As expected, silence of FGF21 increased the FBXW11 expression, and overexpression of FGF21 decreased the FBXW11 expression (Fig. 8I-L). Moreover, the OPA1 degradation increased after overexpression of FBXW11, and transfection of siRNAs targeting FBXW11 reduced OPA1 degradation in cardiomyocytes caused by CoCl2 stimulation (Fig. 8M-P), further suggesting that OPA1 is a possible physiologic substrate of FBXW11. To confirm the result, we detected the interaction between FBXW11 and OPA1 by IP assay. As shown in (Fig. 8Q, R) overexpression of FBXW11 increased the ubiquitin of OPA1.
Because FGF21 exert its effect mainly through activation of FGFR1 which is tyrosine kinase receptor, we continued to investigate whether FGFR1 could interact with FBXW11. In transfected siFGFR1 cardiomyocytes we found increase FBXW11 (Fig.S10A, B). Next, we further constructed FGFR1 plasmid then we transfected FGFR1 after evaluated the phosphorylation levels of FBXW11, we found out increased FBXW11 degradation (Fig.S10C, D). Collectively, these results support the notion that OPA1 is a degradation substrate of the FBXW11 and FGFR1 activation could increase FBXW11 degradation.