Loss of Mfn1 but not Mfn2 enhances adipogenesis

Abstract

Objective: A biallelic missense mutation in mitofusin 2 (MFN2) causes multiple symmetric lipomatosis and partial lipodystrophy, implicating disruption of mitochondrial fusion or interaction with other organelles in adipocyte differentiation, growth and/or survival. In this study, we aimed to document the impact of loss of mitofusin 1 (Mfn1) or 2 (Mfn2) on adipogenesis in cultured cells.

Methods: We characterised adipocyte differentiation of wildtype (WT), Mfn1-/- and Mfn2-/- mouse embryonic fibroblasts (MEFs) and 3T3-L1 preadipocytes in which Mfn1 or 2 levels were reduced using siRNA.

Results: Mfn1-/- MEFs displayed striking fragmentation of the mitochondrial network, with surprisingly enhanced propensity to differentiate into adipocytes, as assessed by lipid accumulation, expression of adipocyte markers (Plin1, Fabp4, Glut4, Adipoq), and insulin-stimulated glucose uptake. RNA sequencing revealed a corresponding pro-adipogenic transcriptional profile including Pparg upregulation. Mfn2-/- MEFs also had a disrupted mitochondrial morphology, but in contrast to Mfn1-/- MEFs they showed reduced expression of adipocyte markers. Mfn1 and Mfn2 siRNA mediated knockdown studies in 3T3-L1 adipocytes generally replicated these findings.

Conclusions: Loss of Mfn1 but not Mfn2 in cultured pre-adipocyte models is pro-adipogenic. This suggests distinct, non-redundant roles for the two mitofusin orthologues in adipocyte differentiation.

Fig 1. Loss of mitofusins induces mitochondrial network fragmentation in mouse embryonic fibroblasts (MEFs).

MEFs deficient for the mitochondrial fusion proteins Mfn1, Mfn2, both Mfn1 and Mfn2 (Mfn1^-/-2^-/-), or Opa1^-/- were studied in their undifferentiated state. (A) Western blot confirming loss of protein from knock-out MEFs with calnexin (Canx) as loading control. (B) Mitotracker Orange imaging of MEFs on confocal microscopy demonstrated mitochondrial network fragmentation in knockout cell lines. (C) Transmission electron microscope (TEM) images of MEFs with zoomed-in images of mitochondria (highlighted in yellow). (D) Quantification of mitochondrial circularity (length/width) from TEM. Each red/blue dot represents one mitochondrion from two biological replicates. (E) Quantification of mitochondrial perimeter from TEM. Each red/blue dot represents one mitochondrion from two biological replicates. (F) Western blot of mitochondrial Oxphos complex subunits. (G) Mitochondrial DNA content expressed as mtRnr2 / nuclear Hk2 DNA from quantitative PCR. Each data point represents a separate biological replicate. All p-values represent pairwise comparisons between knock-outs and wild-type using t-tests, adjusted for multiple testing (p.adj). Data is representative of at least 3 independent replicates.

Fig 2. Mfn1^-/- MEFs demonstrate enhanced adipogenesis.

MEFs underwent adipogenic differentiation using standard protocols and were assessed at day +4 and +8 of differentiation. (A) Oil Red O staining of differentiated MEFs compared to undifferentiated wild-type MEFs. (B) Fluorometric quantification of AdipoRed neutral lipid dye at day +4 & day +8. Each data point represents a separate biological experiment. (C) Glucose uptake assay (2-deoxy-D-glucose) at day +8 under basal and 100nM insulin conditions. Each data point represents a separate biological experiment. FC = Fold increase from basal to insulin-stimulated glucose uptake. (D) Western blot for components of the insulin signalling cascade and markers of adipocyte differentiation at day +10 with and without insulin stimulation. Calnexin (Canx) used as loading control. All p-values represent pairwise comparisons between knock-outs and wild-type using T-tests, adjusted for multiple testing (p.adj). Data is representative of at least 3 independent replicates.

Fig 3. Mfn1^-/- MEFs demonstrate a pro-adipogenic transcriptional signature even in an undifferentiated state.

Bulk RNA sequencing was performed on wild-type (WT), Mfn1^-/-, and Mfn2^-/- at day -2, 0, +3, and +8 of differentiation. (A) Volcano plot demonstrating significant differential gene expression (DGE) in Mfn1^-/- MEFs vs. WT at day-2. Data from n = 3 biological replicates. All genes in orange show significant DGE (adjusted p-value <0.001 and log₂ fold change (log₂FC) greater or less than 1.5. Adipogenic transcription factors are shown in red. (B) Volcano plot for Mfn2^-/- MEFs vs. WT at day -2. Genes in green show significant DGE. (C) Venn diagram illustrating the significantly upregulated and downregulated Hallmark gene sets on transcriptomic analysis of Mfn1^-/- and Mfn2^-/- MEFs versus wild-type MEFs at day -2. (D) Time course from RNAseq demonstrating the change in seven key adipogenic transcription factors in WT and Mfn1^-/- MEFs during differentiation. Data represents normalised counts per million. Log₂FC and adjusted p-values (p.adj) are derived from calculations of DGE at day -2, as described above. (E) Volcano plot for Mfn1^-/- MEFs vs. WT at day+8. Data from n = 3 biological replicates. All genes in red show significant DGE. (F) Volcano plot for Mfn2^-/- MEFs vs. WT at day +8. Genes in blue show significant DGE. (G) Significantly upregulated and downregulated Hallmark gene sets from Mfn1^-/- and Mfn2^-/- MEFs versus wild-type MEFs at day +8. (H) Hallmark adipogenesis gene set enrichment score from pathway analysis following DGE comparison of day -2 versus day +8 for each cell line. (I) Heatmap illustrating log2fold change for all the 200 genes included in the Hallmark adipogenesis gene set (data from (H)). Pathway analysis showing up-regulated (J-K) and down-regulated (L-M) pathways for Mfn1^-/- and Mfn2^-/- MEFs versus WT at day +8.

Fig 4. siRNA knockdown of Mfn1 in 3T3-L1s enhances adipogenesis.

3T3-L1s were treated with scrambled, Mfn1-targeted, or Mfn2-targeted siRNA on alternate days from day -2 to day +12 of differentiation and assessed on days +6 and +12. (A) Western blot from 3T3-L1 at day +12 showing efficacy of Mfn1/Mfn2 knock-down and expression of selected adipogenic genes. Data represents three biological replicates. (B) Light microscope images of AdipoRed fluorescence (green = lipid) and Oil Red O stained 3T3-L1s at day +6 and +12 of differentiation. (C) Oil Red O staining of 3T3-L1s at day +6 and +12 of differentiation. (D) Effect of siRNA knock-down on mitochondrial morphology assessed by confocal microscopy in 3T3-L1s cells at day +12 of differentiation under high-glucose conditions. Mitochondria were labelled using an anti-TOMM20 antibody (purple) and neutral lipids were stained with LipidTox (blue). Scale bars: 10 μm. (E) Quantification of mitochondrial morphology, and different mitochondrial parameters including number, area, and length in 225 μm² region of interests (ROI) from (D). (F) Effect of siRNA knock-down on mitochondrial morphology assessed by confocal microscopy in 3T3-L1s cells at day +12 of differentiation under low glucose conditions. Mitochondria were labelled using an anti-TOMM20 antibody (purple) and neutral lipids were stained with LipidTox (blue). Scale bars: 10 μm. (G) Quantification of mitochondrial morphology, and different mitochondrial parameters including number, area, and length in 225 μm² ROIs from (F). All p-values represent pairwise comparisons between knock-outs and wild-type using t-tests, adjusted for multiple testing (p.adj). Data is representative of at least 3 independent replicates.

Fig 5. Reactive oxygen species scavenger inhibits differentiation in wild-type MEFs.

Wild-type (WT), Mfn1^-/-, and Mfn2^-/- MEFs were treated with N-acetylcysteine (NAC) at 1mM, 2.5mM, and 5mM on alternate days from day -2 to day +8 of differentiation. (A) Oil Red O staining of MEFs at day +8 of differentiation. (B) Western blot from MEFs at day +8 showing expression of selected adipocyte proteins. Data is representative of at least 3 independent replicates.