FGF19 increases mitochondrial biogenesis and fusion in chondrocytes via the AMPKα

To explore the influence of FGF19 on the biological behaviors of mitochondria, we first used TEM to observe the morphological change of mitochondria in chondrocytes induced by FGF19 with the help of β Klotho (KLB), a vital accessory transmembrane glycoprotein for assisting the binding of FGF19 to its receptor [20]. We found that FGF19 at 200 ng/ml could significantly increase the mitochondrial biogenesis as indicated in Fig. 1a. Quantification confirmed that the number of mitochondria in chondrocytes in the FGF19 + KLB group was significantly enhanced relative to that of the single FGF19 group or the KLB control group (Fig. 1b). To further confirm the number change of mitochondria in living chondrocytes, we then used mitochondria staining kit for living cells (Cell Navigator™ Mitochondrion Staining) and performed immunofluorescence. The results revealed that the mitochondria number was significantly enhanced in FGF19-treated living chondrocytes in the presence of KLB (Fig. 1c). Linear quantification of fluorescence intensity (red) also showed that the number of mitochondria was increased and the distribution of mitochondria was broader in the cytoplasm region of living chondrocytes by FGF19 (Fig. 1d). The increase of mitochondrial biogenesis is usually accompanied with the generation of ATP products [30]. Thus, intracellular ATP products were tested by enhanced ATP assay kit in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml) at 48 h. Results confirmed that the intracellular ATP products in chondrocytes were considerably increased by FGF19 (Fig. 1e). By using western blotting, we detected the expression of citrate synthase (CS), one of the key enzymes of aerobic respiration in mitochondria, was up-regulated in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) (Fig. 1f, g). Taking together, these results indicated that FGF19 could increase the mitochondrial biogenesis and thus promote energy generation.

FGF19 induces a transient increase in mitochondrial number and an enhanced generation of ATP products. a Representative TEM images showing the changes of mitochondrial number in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). Orange arrows indicated individual mitochondrion. b Quantification of mitochondrial number (per cell) in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). Quantitative analyses of the mitochondrial number were based on nine cells (per group) from three independent experiments (n = 3). c Representative immunofluorescent staining showing the number changes of mitochondria in living chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) for 72 h. The images were chosen based on three independent experiments (n = 3). Red, individual mitochondrion; Green, F-actin; Blue, nucleus. d Linear quantification of fluorescence intensity of mitochondrion number in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml) by Image Pro Plus 6.0. e ATP assay showing the increase of intracellular ATP products in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The results were based on three independent experiments (n = 3). f Representative western blotting showing the expression change of CS in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). g Quantification of CS by western blotting in (f). The data in b are shown as box (from 25, 50 to 75%) and whisker (minimum to maximum values) plots. The significant difference analysis in b, e and g was based on Student T-test

We performed RNA sequencing to precisely explore the associated gene changes in mitochondrial metabolism of chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). Genes shown in red are upregulated and genes in green are downregulated in the pheatmap (Fig. 2a and gene information in Additional file 1: Table S1). We analyzed the expression of all changed genes and screened 22 changed mitochondrion-related genes in chondrocytes induced by FGF19 in the presence of KLB. Among them, mitochondrial fusion genes, Mfn1, Mfn2 and Opa1, were substantially upregulated, which indicates that FGF19 enhances the expression of mitochondrial fusion genes in chondrocytes. By using western blotting, we then confirmed the protein changes of Mfn1, Mfn2 and Opa1 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml) at 72 h. As shown in Fig. 2b, FGF19 significantly upregulated the expression of Mfn1, Mfn2 and Opa1 in chondrocytes. Quantitative analysis confirmed the significant increase in mitochondrial-fusion proteins in chondrocytes induced by FGF19 (Fig. 2c). Since mitochondrial fission–fusion is a dynamic process [4], we also detected the protein expression of mitochondrial fission-related proteins, Drp1 and Fis1, by western blotting in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml) (Additional file 1: Figure S1). Results showed that FGF19 did not significantly change the expressions of Drp1 and Fis1 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of β-Klotho (200 ng/ml). We next used TEM to explore the fission–fusion change of mitochondrial morphology in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The results showed that FGF19 could elongate the individual mitochondrial morphology in chondrocytes (Fig. 2d), especially, the boxed images in yellow showing the elongation of mitochondrial morphology in chondrocytes induced by FGF19. The schematic diagram showed the morphological changes of individual mitochondria transferred from regularly shaped and circular to irregular and elongated. Further, we analyzed the mitochondrial morphology with Image J (Fig. 2e). Quantitative results confirmed a significant increase in spreading area (in nm2), perimeter in 2D (in nm), aspect ratio (major to minor axis) and Feret’s diameter (longest distance in one single mitochondrion) of individual mitochondria and a significant decrease of circularity (rated by 4π × area/perimeter2) and roundness (rated by 4 × area/π × major axis2) of individual mitochondria in chondrocytes induced by FGF19. Together, FGF19 could also mediate fission–fusion process of mitochondria by characterizing the enhancement of fusion proteins and elongation of mitochondrial morphology.

FGF19 promotes the elongation of mitochondrial morphology by up-regulating the expression of mitochondrial fusion proteins. a RNA sequencing showing the change of mitochondrial metabolism-related genes in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). Three pairs of samples were obtained from three independent cell isolates (n = 3), namely, samples 1, 1′, 1′′and 1′′′, samples 2, 2′, 2′′and 2′′′, and samples 3, 3′, 3′′and 3′′′. The data were present as log2(FPKM + 1). FPKM, Fragments per kilobase of exon model per million mapped fragments. b Representative western blotting showing the expression changes of Opa1, Mfn1 and Mfn2 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). c Quantification of Opa1, Mfn1 and Mfn2 by western blotting in b was performed to confirm these protein changes (n = 3). d Representative TEM images showing the changes of mitochondrial network’s morphology in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). Cyan arrows indicated the elongation of mitochondrial morphology. Schematic diagram illustrated that elongation was correlated with mitochondrial fusion. e Measurements of mitochondrial network’s morphology in d by Image J. Quantitative analyses of mitochondrial network’s morphology were based on three independent experiments (n = 3). The data in e were shown as box (from 25, 50 to 75%) and whisker (minimum to maximum values) plots. The significant difference analysis in c and e was based on Student T-test

It is widely recognized that FGF19 can bind to FGFR1, 2, 3 and 4 receptors but has a high affinity for FGFR4 with the help of KLB [20]. In order to explore the gene expressions of FGFRs in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml), we analyzed RNA sequencing and the results were shown in the form of pheatmap (Fig. 3a and gene information in Additional file 1: Table S2). The gene expression of FGFR1 and FGFR4 were significantly increased by FGF19 in the presence of KLB, and moreover, the expression of FGFR4 was much higher than that of FGFR1 in the chondrocytes. Then, we performed qPCR and western blotting to affirm the change of FGFR4 expression in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The results in Fig. 3b showed that FGF19 could significantly increase the gene expression of FGFR4 in chondrocytes by qPCR and the up-regulation of FGFR4 gene in the FGF19 + KLB group was remarkably enhanced relative to that without the help of KLB (the single FGF19 group). The protein expression of FGFR4 was also increased in chondrocytes induced by FGF19 (Fig. 3c). With the help of KLB, FGFR4 in the FGF19 + KLB group showed a higher expression than that in the single FGF19 group.

FGF19 increases the mitochondrial biogenesis by up-regulating the expression of AMPKα signalling related proteins in chondrocytes. a RNA sequencing showing the change of FGFRs genes in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). Three pairs of samples were obtained from three independent cell isolates (n = 3), namely, samples 1, 1′, 1′′and 1′′′, samples 2, 2′, 2′′and 2′′′, and samples 3, 3′, 3′′and 3′′′. The data were present as log2(FPKM + 1). b q-PCR showing the gene changes of FGFR4 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The results were based on three independent experiments (n = 3). c Representative western blotting showing the expression change of FGFR4 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). d Representative western blotting showing the expression changes of AMPKα, p-AMPKα, PGC-1α and SIRT1 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). e Quantifications of AMPKα, p-AMPKα, PGC-1α and SIRT1 by western blotting in (d). f Representative immunofluorescent staining showing the change in the distribution of p-AMPKα in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) for 72 h. The images were chosen based on three independent experiments (n = 3). Red, p-AMPKα; Green, F-actin; Blue, nucleus. g Quantification of fluorescence intensity of p-AMPKα in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The data were based on at least eight cells from three independent experiments. h Representative immunofluorescent staining showing the change in the distribution of PGC-1α in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) for 72 h. The images were chosen based on three independent experiments (n = 3). Red, PGC-1α; Green, F-actin; Blue, nucleus. i Quantification of fluorescence intensity of PGC-1α in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The data were based on at least eight cells from three independent experiments. The data in g and i were shown as box (from 25, 50 to 75%) and whisker (minimum to maximum values) plots. The significant difference analysis in b, e, g and i was based on Student T-test

AMPKα signalling directly regulates the biogenesis of mitochondria through the AMPKα-PGC-1α-SIRT1 axis, a putative mitochondrial biogenesis relevant signalling [30]. From western blotting, we found that the expression of AMPKα, p-AMPKα, PGC-1α and SIRT1 was up-regulated in chondrocytes by FGF19 (Fig. 3d). Quantitative analysis of these proteins further confirmed the increase in AMPKα-PGC-1α-SIRT1 signalling in chondrocytes induced by FGF19 in the presence of KLB (Fig. 3e). As the phosphorylation of AMPKα and activation of PGC-1α play a vital role in mitochondrial biogenesis. We further performed immunofluorescent staining to explore the expression and distribution of p-AMPKα and PGC-1α in chondrocytes induced by FGF19 (Fig. 3f–i). The results showed that FGF19 could increase the expression of p-AMPKα and PGC-1α. The expression of p-AMPKα was notably accumulated in the nuclear region (Fig. 3f) while the expression of PGC-1α was increased in whole cytoplasm of chondrocytes (Fig. 3h). Quantification of fluorescent intensity (per cell) confirmed the increase of p-AMPKα and PGC-1α in chondrocytes induced by FGF19 in the presence of KLB (Fig. 3g, i). Taking together, these results indicated that FGF19 could enhance the biogenesis and fusion by up-regulation of AMPKα signalling.

To determine the key cytoplasmic pathways related to FGF19-mediated mitochondrial biogenesis and fusion, we analyzed the RNA sequencing data and screened out all changed kinases involving classical pathways. These kinases were clustered by pheatmap (Fig. 4a and gene information in Additional file 1: Table S3). It showed that most of the kinases were related to MAPK signaling. In particular, MAP kinases such as Dusp4 and Dusp2 were shown to be significantly enhanced in chondrocytes. We then performed western blotting to confirm the changes of ERK/p-ERK, p38/p-p38 and JNK/p-JNK in chondrocytes induced by FGF19 (Fig. 4b). Among them, we found that the enhancement of total p38 and p-p38 were higher than the other two. Quantitative analysis confirmed a significant increase in total p38 and p-p38 but the increase of ERK/p-ERK and JNK/p-JNK was not as obvious as p38/p-p38 in chondrocytes induced by FGF19 in the presence of KLB (Fig. 4c and Additional file 1: S2). We further used immunofluorescent staining to explore the expression and distribution of p-p38 in chondrocytes induced by FGF19 in the presence of KLB (Fig. 4d, e). From the CLSM images, we found that FGF19 could increase the expression of p-p38 in the cytoplasm of chondrocytes, especially in the nuclear region (Fig. 4d). Quantification of total fluorescent intensity confirmed the increased expression of p-p38 in chondrocytes induced by FGF19 in the presence of KLB (Fig. 4e).

FGF19 activates p38/MAPK signalling in chondrocytes. a RNA sequencing showing the changes in the expression of MAPK-related mediators in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). Three pairs of samples were obtained from three independent cell isolates (n = 3), namely, samples 1, 1′, 1′′and 1′′′, samples 2, 2′, 2′′and 2′′′, and samples 3, 3′, 3′′and 3′′′. The data were present as log2(FPKM + 1). b Representative western blotting showing the expression change of ERK, p-ERK, p38, p-p38, JNK and p-JNK in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). c Quantification of p38 and p-p38 by western blotting in (b). d Representative immunofluorescent staining showing the change in the expression and distribution of p-p38 in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) for 72 h. The images were chosen based on three independent experiments (n = 3). Red, p-p38; Green, F-actin; Blue, nucleus. e Quantification of fluorescence intensity of p-p38 in chondrocytes induced by FGF19 at 200 ng/ml in the presence of KLB (200 ng/ml). The data were based on nine cells from three independent experiments. The data in e were shown as box (from 25, 50 to 75%) and whisker (minimum to maximum values) plots. The significant difference analysis in c and e was based on Student T-test

To further determine the importance of p38/MAPK in regulating the expression of AMPKα signalling, SB203580, a specific inhibitor of p38/p-p38 signalling, was utilized [31]. We detected the expressions of p38/p-p38, AMPKα/p-AMPKα, PGC-1α and Sirt1 in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) after pretreatment of SB203580 (10 µM) for 2 h (Fig. 5a). From western blotting, we revealed that SB203580 could effectively impair the up-regulation of p38/p-p38 and also attenuated the mitochondrial biogenesis proteins including AMPKα/p-AMPKα, PGC-1α and Sirt1. Quantitative analysis further confirmed a significant decrease in the expression of p38 pathway and AMPKα signalling in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) after pretreatment of SB203580 (10 µM) (Fig. 5b). We then used immunofluorescent staining to show the expression and distribution of p-AMPKα and PGC-1α (Fig. 5c–e). The results showed that the expressions of p-AMPKα (Fig. 5c) and PGC-1α (Fig. 5d) were largely reduced in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) after pretreatment of SB203580 (10 µM). Fluorescence quantification further confirmed the changes of p-AMPKα and PGC-1α (Fig. 5e).

Inhibition of p38 attenuated FGF19-enhanced AMPKα activity. a Representative western blotting showing the expression change of p38, p-p38, AMPKα, p-AMPKα, PGC-1α and SIRT1 in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). b Quantification of p38, p-p38, AMPKα, p-AMPKα, PGC-1α and SIRT1 by western blotting in (a). c Representative immunofluorescent staining showing the change in the distribution of p-AMPKα in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). Red, p-AMPKα; Green, F-actin; Blue, nucleus. d Representative immunofluorescent staining showing the change in the expression and distribution of PGC-1α in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). Red, PGC-1α; Green, F-actin; Blue, nucleus. e Quantification of fluorescence intensity of p-AMPKα and PGC-1α in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The data in e were shown as box (from 25, 50 to 75%) and whisker (minimum to maximum values) plots. The significant difference analysis in b and e was based on Student T-test

Next, we explored the role of p38/MAPK in regulating biogenesis and fusion of mitochondria. We detected the protein mediators in the fission–fusion process (Additional file 1: Figure S3 and 6a). The results showed that inhibition of p38 did not significantly change the expression of FGF19-induced mitochondrial fission proteins, i.e., Drp1 and Fis1 (Additional file 1: Figure S3), but indeed decreased the expression of FGF19-induced mitochondrial fusion proteins, i.e., Opa1, Mfn1 and Mfn2 (Fig. 6a, b). Further, we performed immunofluorescence and found the impairment of Opa1 (Fig. 6c) and Mfn2 (Fig. 6d) in chondrocytes induced by FGF19 (200 ng/ml) in the presence of KLB (200 ng/ml) after pretreatment of SB203580 (10 µM). Fluorescence quantification further confirmed the changes of Opa1 and Mfn2 (Fig. 6e).

Inhibition of p38 decreases the expressions of mitochondrial fusion proteins induced by FGF19 in chondrocytes. a Representative western blotting showing the expression change of Opa1, Mfn1 and Mfn2 in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). b Quantification of Opa1, Mfn1 and Mfn2 by western blotting in (a). c Representative immunofluorescent staining showing the change in the expression and distribution of Opa1 in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). Red, Opa1; Green, F-actin; Blue, nucleus. d Representative immunofluorescent staining showing the change in the distribution of Mfn2 in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The images were chosen based on three independent experiments (n = 3). Red, Mfn2; Green, F-actin; Blue, nucleus. e Quantification of fluorescence intensity of Opa1 and Mfn2 in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml). The data were based on at least eight cells from three independent experiments. f Representative immunofluorescent staining showing the changes of morphology mitochondrial network in living chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) and KLB (200 ng/ml) for 72 h. Image J shows the change of mitochondrial network morphology analysis in cyan boxes. The images were chosen based on three independent experiments (n = 3). Red, mitochondrial network; Blue, nucleus. g Quantification of mitochondrial number (per cell) and mitochondrial elongated number (per cell) in chondrocytes induced by SB203580 (10 µM) in the presence of FGF19 (200 ng/ml) by Image J. Quantitative analyses were based on three independent experiments (n = 3). The data in e and g were shown as box (from 25, 50 to 75%) and whisker (minimum to maximum values) plots. The significant difference analysis in b, e and g were based on Student T-test

Finally, to confirm the role of p38/MAPK in controlling mitochondrial network morphology, we applied mitochondrial living cell staining (Fig. 6f, g). From CLSM, we observed that SB203580 could significantly decrease the FGF19-enhanced mitochondrial number, and moreover, it could sharply reduce the mitochondrial network morphology formed by FGF19 (cyan boxes). Quantitative analysis confirmed that about a 50% decrease in the total change of mitochondrial number (per cell) and a 60% decrease in the number of mitochondrial elongation (per cell) in chondrocytes induced by SB203580 (Fig. 6g).

Together, the inhibition of p38/MAPK could decrease the expression of AMPKα signalling, and thus impair the biogenesis and fusion of mitochondria by characterizing mitochondrial fusion proteins, and mitochondrial network morphology in living chondrocytes.