Culture and adipogenic differentiation of MEFs

Mouse embryonic fibroblasts (MEFs) null for Mfn1^-/-, Mfn2^-/-, Mfn1^-/-2^-/-, and Opa1^-/-, and wild-type MEFs, originally derived by Chen et al. [14], were obtained for use in this study from American Type Culture Collection in February 2019 (see S1 Table for a list of reagents).

MEFs were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% (vol/vol) foetal bovine serum (FBS), 1% (1mM) sodium pyruvate (from stock 100mM), 1% non-essential amino acids 100x, 1% penicillin/streptomycin (10,000 units penicillin and 10 mg streptomycin per mL), 2mM L-glutamine, and 50mM 2-mercaptoethanol. All cells were grown at 37°C in 20% O₂ and 5% CO₂.

Cells were maintained and passaged whilst sub-confluent prior to adipogenic differentiation. All experiments were performed with cells at passage <10. To initiate differentiation, cells were grown to confluence (day -2) in 6- or 12-well plates and then, two days later, treated with 8μg/mL D-pantothenic acid, 8μg/mL biotin, 1μM rosiglitazone, 0.5mM 3-isobutyl-1-methylxanthine (IBMX), 1μM dexamethasone, and 1μM insulin. Treatment with D-pantothenic acid, biotin, rosiglitazone, and insulin was repeated on day +2, +4, and +6. Differentiation continued until day +8.

For fluorometric quantification of intracellular lipid accumulation, cells were incubated with AdipoRed at 1:25 dilution and 37°C for 20 minutes in the dark. Cells were washed twice in PBS then fluorescence was measured on Tecan M1000 Pro plate reader with excitation at 485nm and emission at 572nm.

For studies using the reactive oxygen species (ROS) scavenger N-acetylcysteine (NAC), MEFs were differentiated according to the above protocol in the presence of pH7 adjusted NAC (Sigma) at either 1mM, 2.5mM, or 5mM. NAC treatment began at day -2 and was repeated with each media change.

Over-expression of Pparg2 in MEFs

To improve the adipogenic potential of MEFs, cells were retrovirally transduced with murine Peroxisome proliferator-activated receptor gamma 2 (Pparg2). Phoenix-AMPHO (ATCC) cells were used for packaging. 70% confluent Phoenix cells were transfected with 12μg pBABE (puromycin-resistant)-mPparg2 vector or pBABE-EGFP (control) plasmid DNA using Lipofectamine 3000 transfection reagent. 72-hours later the supernatant media containing the virus was collected and filtered through a 0.45μm membrane. Retroviral stocks were used to transduce 50–60% confluent MEFs with addition of 12μg/mL polybrene. From 24-hours after transduction, puromycin selection began using 4μg/mL, and MEFs were subsequently cultured in media containing this concentration of puromycin. Adipogenic differentiation was then performed according to the above protocol.

Confocal microscopy of mitochondrial network

Cells were stained with Mitotracker Orange CMTMRos (ThermoFisher) for imaging of the mitochondrial network using confocal microscopy. MEFs were grown on cover slips to 80–90% confluence (pre-adipocytes at day -2 for MEFs and day 0 for 3T3-L1s) and treated with 400nM Mitotracker for 20 minutes at 37°C in the dark. Cells were washed with fresh media four times, each for 30 minutes at 37°C in the dark, followed by three washes in phosphate buffered saline (PBS). Cells were then fixed in 4% formaldehyde at room temperature for 10 minutes and mounted onto slides using ProLong Gold or Diamond Antifade Mountant with 4′,6-diamidino-2-phenylindole (DAPI) (ThermoFisher). Images were obtained using a Leica SP8 confocal microscope using excitation/emission spectra: Mitotracker Orange CMTMRos (550/560-580), BODIPY (490/515-535), and DAPI (405/450-470).

Culture and adipogenic differentiation of 3T3-L1 fibroblasts

Mouse fibroblast 3T3-L1 fibroblasts (3T3-L1s) [33] were used as an in vitro model of adipocyte differentiation and were obtained as a gift from D.J. Fazakerley and D.E. James (University of Sydney), originally obtained from Dr. Howard Green (Harvard Medical School, Boston, MA). 3T3-L1s were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% (vol/vol) foetal bovine serum and 1% GlutaMax. Cells were grown at 37°C in 20% O₂ and 10% CO₂. Cells were maintained and passaged sub-confluent prior to adipogenic differentiation.

To initiate differentiation, cells were grown to confluence (day -2) in 12-well plates and then two days later (day 0) treated with 0.5mM IBMX, 410nM biotin, 220nM dexamethasone, and 350nM insulin. After 72-hours (day +3), media was replaced with addition of 350nM insulin. A further 72-hours later (day +6) media was changed (without any additional chemicals) and subsequently replaced on day +8 and day +10 with experiments ending on day +12. On day +12, lipid accumulation was assessed using Oil Red O staining and AdipoRed quantification. RNA and protein were also extracted at day +12.

siRNA knock-down experiments in 3T3-L1s

For knock-down, confluent pre-adipocyte (day -2) 3T3-L1s grown in 12-well plates were transfected with a pool of two anti-Mfn1 siRNAs (ThermoFisher Scientific, CatIDs: s85002 & s85004), each at 40nM to achieve a final concentration of 80nM, with 2.5μl RNAiMax (ThermoFisher Scientific) in OptiMEM (ThermoFisher Scientific). In addition, we tested the knock-down efficacy of a third anti-Mfn1 siRNA (ThermoFisher Scientific, CatID s85003), which had limited efficacy and was not taken forward in further experiments. For knock-down of Mfn2, an anti-Mfn2 SmartPool (Dharmacon) was used at a final concentration of 40nM. In all experiments, a scrambled Silencer Negative Control siRNA (ThermoFisher Scientific) was used at 80nM. Knock-down was performed with each media change, i.e. day -2, 0, +3, +6, +8, and +10.

Immunofluorescence in 3T3-L1s

Mature (day +8) differentiated 3T3-L1 adipocytes were reseeded onto matrigel-coated (Corning) coverslips at 50% density. siRNA knockdown was performed during reseeding. Reseeded adipocytes (day 9) were cultured under standard “high glucose” conditions (Dulbecco’s modified eagle’s medium containing 25mM glucose supplemented with 10% (vol/vol) foetal bovine serum and 1% GlutaMax) or “low glucose” (serum-free Dulbecco’s modified eagle’s medium containing 5mM glucose supplemented with 0.2% (w/v) bovine serum albumin and 1% GlutaMax) conditions for 16 hours. Cells were then fixed in 4% paraformaldehyde (PFA) in PBS at room temperature for 20 minutes, quenched with 100mM glycine in PBS, washed three times with PBS, followed by incubation with 50mM ammonium chloride in PBS to quench the unspecific fluorescence signal from aldehyde groups. Cells were washed again three times with PBS and permeabilized in 0.1% Triton X-100 in PBS for 10 minutes. Then, cells were blocked with 10% FBS in PBS, followed by incubation with the appropriate primary antibodies in 5% FBS in PBS, for two hours at room temperature. After three washes with 5% FBS in PBS, cells were incubated with specific secondary antibodies (1:1,000) for one hour at RT. After three washes in PBS, coverslips were mounted onto slides using Dako fluorescence mounting medium (Dako).

Mitochondria were stained using a rabbit anti-TOMM20 antibody (11802-1-AP/Proteintech) (1:1,000). Alexa Fluor 488 (anti-rabbit) was used as a secondary antibody (1:1,000) (Invitrogen). Lipid droplets were stained by incubating coverslips with LipidTOX deep red (Thermo-fisher) for 2 hours (1:1,000).

For confocal image acquisition, cells were visualized using a 100X objective lenses (NA1.4) on a Nikon Eclipse TiE inverted microscope and 7 stacks of 0.2μm each were acquired with appropriate lasers using an Andor Dragonfly 500 spinning disk system, equipped with a Zyla 4.2 PLUS sCMOS camera (Andor), coupled with Fusion software (Andor). Images from the same experiment were acquired following the same parameters including exposure time and laser intensities.

For image analysis, mitochondrial morphology was manually analysed and classified as intermediate, hyperfused or fragmented as already described [34]. The different mitochondrial parameters (area, length and number) were quantified by randomly selecting region of interests (ROIs) of 225μm² at the cell periphery and analysed using MitoMapr (Fiji) [35]. Error bars displayed on graphs represent the mean ± S.D from 3 independent experiments (at least 18 cells per experiment and condition). Statistical significance was analysed using Nested t-test or Two‐way ANOVA test using GraphPad Prism software. *p < 0.05, **p < 0.01, *** p < 0.001 and **** p < 0.0001 were considered significant.

Electron microscopic imaging of mitochondria

Transmission electron microscope (TEM) studies were performed at the Cambridge Advanced Imaging Centre, University of Cambridge. Undifferentiated MEFs (day -2) were grown to sub-confluence in 6-well plates. Cells were washed twice in 0.9% NaCl then fixed in (2% glutaraldehyde/2% formaldehyde in 0.05M sodium cacodylate buffer pH 7.4 containing 2mM calcium chloride) for 4 hours at 4°C. Cells were then scraped from plates and, after washing five times with 0.05M sodium cacodylate buffer pH 7.4, samples were osmicated (1% osmium tetroxide, 1.5% potassium ferricyanide, 0.05M sodium cacodylate buffer pH 7.4) for three days at 4°C. After washing five times in DIW (deionised water), samples were treated with 0.1% (w/v) thiocarbohydrazide/DIW for 20 minutes at room temperature in the dark. After washing five times in DIW, samples were osmicated a second time for 1 hour at room temperature (2% osmium tetroxide/DIW). After washing five times in DIW, samples were block stained with uranyl acetate (2% uranyl acetate in 0.05M maleate buffer pH 5.5) for three days at 4°C. Samples were washed five times in DIW and then dehydrated in a graded series of ethanol (50%/70%/95%/100%/100% dry) 100% dry acetone and 100% dry acetonitrile, three times in each for at least 5 minutes. Samples were infiltrated with a 50/50 mixture of 100% dry acetonitrile/Quetol resin (without benzyldimethylamine (BDMA)) overnight, followed by three days in 100% Quetol (without BDMA). Then, samples were infiltrated for 5 days in 100% Quetol resin with BDMA, exchanging the resin each day. The Quetol resin mixture is: 12g Quetol 651, 15.7g NSA, 5.7g MNA and 0.5g BDMA (all from TAAB). Samples were placed in embedding moulds and cured at 60°C for 3 days.

Thin sections were cut using an ultramicrotome (Leica Ultracut E) and placed on bare 300 mesh copper TEM grids. Samples were imaged using a Tecnai G2 TEM (FEI/Thermo Fisher Scientific) run at 200 keV using a 20μm objective aperture to improve contrast. Images were acquired using an ORCA HR high resolution CCD camera (Advanced Microscopy Techniques Corp, Danvers USA).

Analysis was performed by manual measurement of individual mitochondria from all obtained images using ImageJ. Despite multiple mitochondria measured per sample, statistical tests (pairwise comparisons relative to wild-type) were based on the mean from two independent biological replicates and false-discovery rate adjustment for the number of tests [36].

mtDNA content assay

Relative mitochondrial DNA (mtDNA) content was assayed using real-time quantitative polymerase chain reaction (RT-qPCR) for a mitochondrial DNA gene (Rnr2) and a nuclear DNA gene (Hk2). DNA was extracted from undifferentiated MEFs at 80% confluence using DNeasy kit (Qiagen). DNA was quantified on a Nanodrop and diluted to 4ng/μl. RT-qPCR was performed in triplicate for each sample using 8ng DNA with primers for Hk2 (Forward: GCCAGCCTCTCCTGATTTTAGTGT, Reverse: GGGAACACAAAAGACCTCTTCTGG) and Rnr2 (Forward: AACTCGGCAAACAAGAACCC, Reverse: CCCTCGTTTAGCCGTTCATG). The ratio mtRnr2/nHk2 was calculated using the standard curve method and expressed relative to wild-type.

Oil Red O staining

Mature adipocytes (day +8 MEFS or day +12 3T3-L1s) in six-well plates were gently washed in PBS before fixing in 10% formalin for 30 minutes at room temperature. Cells were then washed twice in PBS before dehydration in 60% isopropanol for 5 minutes (twice). Oil Red O (ORO) working solution was then added for 20 minutes. ORO working solution was made from 1g ORO (Sigma) dissolved in 400ml isopropanol, diluted 3:2 with milli-Q water and filtered through a 0.45μm membrane. Cells were washed with water then imaged on a flat-bed scanner and with light microscopy. Lipid droplet width was measured in day +12 3T3-L1s using ImageJ [37].

2-deoxy-glucose uptake assay

2-deoxy-glucose (2DOG) uptake assays were performed on mature adipocytes (day +8) following differentiation of MEFs in 24-well plates, according to standard protocols [38]. Briefly, MEFs were serum starved in Dulbecco’s Modified Eagle’s Medium (DMEM)/0.2% bovine serum albumin (BSA)/1% GlutaMax for 2 hours then washed three times in KRP buffer with 0.2% BSA. KRP buffer was prepared as: 0.6mM Na₂HPO₄, 0.4mM NaH₂PO₄, 120mM NaCl, 6 mM KCl, 1mM CaCl₂, 1.2mM MgSO₄ and 12.5mM HEPES adjusted to pH 7.4. At least one well on each plate served as a negative control by addition of 25μM cytochalasin B (Sigma Aldrich) to block all transporter-mediated glucose uptake. Cells were incubated with KRP/0.2% BSA ±100nM insulin for 20 minutes. After 15 minutes, 2-deoxyglucose (Sigma/Merck; radiolabelled 2DOG from PerkinElmer) was added to each well to a final concentration of 50μM and 0.25μCi/well. 5 minutes after addition of 2DOG, cells were quickly washed three times in ice-cold PBS then lysed in 1% (v/v) Triton X-100. Uptake of 2DOG (³H) was quantified using liquid scintillation counting and normalised to protein content.

For immunoblots of the insulin signalling cascade, MEFs were grown to day +10 of adipogenic differentiation, serum starved for 2 hours, then stimulated with 1nM insulin for 30 minutes. Cells were then washed in cold PBS and lysed in RIPA buffer (Sigma) with PhosSTOP Phosphatase Inhibitor (Roche) and Complete-Mini Protease Inhibitor (Roche), as described below. Lysates for immunoblotting of Glut1 and Glut4 were sonicated prior to quantification and were not denatured.

RNAseq in knock-out MEFs

Bulk RNA sequencing (RNAseq) was performed on wild-type (WT) and Mfn1^-/- MEFs at four time points: day -2 (pre-adipocyte), day 0 (initiation of adipogenic differentiation), day +3 (early adipocyte), and day +8 (mature adipocyte). RNAseq was performed on Mfn2^-/- MEFs at two time points: day -2 (pre-adipocyte) and day +8 (mature adipocyte). Three independent biological replicates of each sample were used for analysis.

Cells were cultured and differentiated as described above, then briefly washed in PBS and lysed in RLT Buffer (Qiagen) with 1% 2-mercaptoethanol then passed through Qiashredder tubes (Qiagen). RNA was extracted using the RNeasy Isolation Kit (Qiagen) according to the manufacturers’ protocol. RNA was quantified using Agilent 2100 Bioanalyzer (Agilent Technologies Inc) and only samples with RNA Integrity Number ≥8 were used for library preparation. cDNA libraries were made using Illumina TruSeq RNA sample kits and sequencing was performed on Illumina NovaSeq 6000 with paired-end 150bp reads (Novogene, Cambridge, UK). Raw reads all passed quality control for Qscore, error rate distribution, and AT/GC distribution.

Adapter sequences were removed from raw FASTQ files using cutadapt [39] and aligned to Mus musculus reference genome (GRCm38) using STAR [40]. Binary alignment/map (BAM) files were sorted using samtools [41] and counts were performed using featureCounts [42]. Differential gene expression (DGE) was performed using DESeq2 [43], where significance was considered as a Benjamini-Hochberg false-discovery rate (FDR) corrected p-value < .01.

Pathway analysis was performed with the EnrichR package for R [44–46] using significantly differentially expressed genes to determine enriched Hallmark gene sets [47]. Gene sets with FDR-corrected p-value < .05 were considered enriched. Figures were generated in R 4.0.2 [48] using packages pheatmap, ggplot2, and dplyr. RNAseq data is available in S1 Appendix and at Gene Expression Omnibus Study ID: GSE232502. All code used in analysis is available from: https://doi.org/10.5281/zenodo.5770057.

Western blot studies

Cells were washed in cold PBS and lysed in RIPA buffer (Sigma) with PhosSTOP Phosphatase Inhibitor (Roche) and Complete-Mini Protease Inhibitor (Roche). Lysates were spun at 13,000rpm at for 15 minutes at 4°C then protein concentration quantified using BioRad DC Protein Assay Kit. 30–45μg protein lysates were mixed with NuPAGE 4x LDS buffer (ThermoFisher Scientific), containing 0.05% 2-mercaptoethanol, and denatured for 10 minutes at 95°C. Samples were run on 4–12% Bis-Tris gels (Invitrogen) and transferred onto a nitrocellulose membrane using iBlot-2 (ThermoFisher Scientific). Membranes were washed in Tris-buffered saline with 0.1% (vol/vol) Tween 20 (TBST, Sigma) before blocking in 5% (wt/vol) skimmed milk powder dissolved in TBST. Membranes were incubated with primary antibodies (S2 Table) at 4°C for 16 hours. Membranes were then washed with TBST five times for 5 minutes, followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 hour at room temperature. Blots were developed using Immobilon Western Chemiluminescent HRP Substrate (Millipore) with images acquired on BioRad ChemiDoc Imaging system or GE ImageQuant LAS 4000.

Statistical analysis

Continuous data were expressed as mean ± standard deviation. Normally distributed data were analysed by t-test (for two group pairwise comparison) and one-way ANOVA (for three or more groups) with post-hoc Bonferroni multiple comparisons test, where FDR-corrected p-value <0.05 was considered significant. Details of specific analyses and the number of replicates performed are reported in figure legends. All experiments were conducted at least three times using, where possible, randomisation of sample order and blinding of experimenters handling samples. Data were analysed using R 4.0.2 [48] and GraphPad Prism version 9 (GraphPad, San Diego).

Characterization of undifferentiated knock-out MEFs

To assess the role of mitofusins in adipogenesis we utilised MEF lines deficient for Mfn1 (Mfn1^-/-), Mfn2 (Mfn2^-/-), or both (Mfn1^-/-2^-/-) (Fig 1A). We initially also characterised a line deficient for Opa1 (Opa1^-/-), the GTPase that mediates fusion of the inner mitochondrial membrane downstream of mitofusin engagement [49, 50]. Opa1^-/- MEFs are expected to have a specific defect in mitochondrial fusion.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

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.

https://doi.org/10.1371/journal.pone.0306243.g001

We first assessed expression of core proteins involved in mitochondrial fusion and fission in undifferentiated MEFs. Loss of Mfn1 increased expression of Mfn2, whilst loss of Mfn2 or Opa1 reduced expression of Mfn1 (Fig 1A & S1 Fig). Loss of each of the mitofusins was associated with no change in Opa1 expression whereas loss of both mitofusins led to a substantial reduction in Opa1 expression. Drp1, the main regulator of mitochondrial division [51], was expressed at similar levels in wild-type (WT) and Mfn1^-/-, but increased in Mfn2^-/- MEFs and reduced in Mfn1^-/-2^-/- and Opa1^-/- MEFs. Fis1, a protein involved in, but not required for, mitochondrial fission [52, 53], was expressed at higher levels in Mfn2^-/- MEFs but at similar levels in the other MEF lines.

As previously reported [14], Mfn1^-/-, Mfn2^-/- and Mfn1^-/-2^-/- MEFs exhibited fragmented mitochondria when compared to WT controls using either confocal- (Fig 1B) or transmission electron- microscopy (TEM) (Fig 1C). Analysis of the TEM images revealed more circular mitochondria (reduced length/width) in Mfn1^-/- and Mfn2^-/- MEFs (Fig 1D) without any difference in the mean mitochondrial perimeter (Fig 1E). Mitochondrial network disruption was most striking in Mfn1^-/-2^-/- MEFs, which also consistently multiplied very slowly and failed to show any signs of adipocyte differentiation. Growth of Opa1^-/- MEFs, which are also known to manifest severely disrupted mitochondrial morphology [54, 55], was similarly delayed and they too failed to show any signs of adipocyte differentiation, so these cells were not studied further in terms of adipogenesis. Loss of Opa1 or both mitofusins has previously been found to prevent cells from responding to changes in metabolic demand [56] as well as manifesting low growth rates [49], which would limit the proliferation necessary for in vitro adipogenesis [57].

In the Mfn1^-/-2^-/- and Opa1^-/- MEFs we observed substantially reduced expression of mitochondrial oxidative phosphorylation complexes II, III, and IV (Fig 1F & S1 Fig). In contrast, Oxphos subunit expression was similar to WT cells in Mfn1^-/- MEFs, except for slightly increased complex III expression. TOMM20 expression was increased in Mfn1^-/- MEFs (Fig 1F & S1 Fig), suggesting increased mitochondrial mass. In Mfn2^-/- cells expression of complexes II-V appeared normal whereas complex I expression was reduced (Fig 1F).

Mitochondrial DNA (mtDNA) is prone to depletion in cells with impaired mitochondrial fusion [58]. We found that double knock-out Mfn1^-/-2^-/- and Opa1^-/- MEFs had reduced mtDNA levels whereas loss of Mfn1 or Mfn2 in isolation did not significantly decrease mtDNA content (Fig 1G).

Adipogenic differentiation of knock-out MEFs

We next assessed the ability of mitofusin KO MEFs to differentiate into adipocytes using a well-established adipogenic protocol. Surprisingly, we observed significantly increased lipid accumulation in Mfn1^-/- MEFs, but not in Mfn2^-/- MEFs, from day 4 of differentiation (Fig 2A), with increased neutral lipid content on both days 4 and 8 (Fig 2). A panel of mature adipocyte proteins, namely Plin1, Pparg, Fabp4, adiponectin and Glut4 (Fig 2D & S2 Fig) showed concomitant increases in expression. Expression of the insulin receptor (Insr) was also increased in these cells (Fig 2D & S2 Fig). Loss of Mfn1 was associated with a greater increase in Pparg1 than Pparg2 (Fig 2D & S2 Fig). These expression changes were associated with increased insulin-stimulated Akt phosphorylation (Fig 2D & S2 Fig) and elevated insulin-stimulated 2-deoxy-glucose uptake compared to WT MEFs (p = 1.1x10^-12, Fig 2C). These findings suggest bona fide enhanced adipogenic differentiation rather than simple lipid accumulation in Mfn1^-/- MEFs. In Mfn2^-/- MEFs, adipogenic markers were generally reduced (Fig 2D & S2 Fig) and the impact of insulin on glucose uptake was decreased (Fig 2C). However, these cells did manifest high basal glucose uptake (WT 2.8 ± .6 vs. Mfn2^-/- 11.7 ± 0.9 nmol/mg/min, p = 4.8x10^-13). We hypothesized that this might reflect increased Glut1 expression, but this was not the case (Fig 2D & S2 Fig), so at this stage we do not have a definitive explanation for this observation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

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.

https://doi.org/10.1371/journal.pone.0306243.g002

Differentiation of knock-out MEFs over-expressing Pparg

MEFs are not fully committed preadipocytes, and show a relatively low rate of adipocyte differentiation in response to hormonal stimuli alone [59]. We thus next assessed whether over-expression of Pparg2, the master transcriptional driver of adipogenesis, would modify the impact of mitofusins deficiency on adipogenesis. Pparg2 was retrovirally overexpressed in WT and different mitofusins KO MEFs, with overexpression higher in Mfn2^-/- MEFs than in WT and Mfn1^-/- MEFs (S3A, S3B Fig). Upon adipogenic differentiation, we observed results consistent with those in untransduced cells, namely enhanced lipid accumulation in Mfn1^-/- MEFs but similar lipid accumulation in the WT and Mfn2^-/- MEFs (S3C & S3D Fig). There was increased expression of Pparg1, Fabp4 and Glut4 in Mfn1^-/- MEFs with reduced expression of Glut4 in Mfn2^-/- MEFs despite higher baseline expression of Pparg2 (S3E & S4 Figs). Expression of Plin1 was higher in both Mfn1^-/- and Mfn2^-/- MEFs compared to differentiated WT MEFs, demonstrating that loss of Mfn1, but not Mfn2, was associated with enhanced differentiation.

Transcriptomic profiling of knock-out MEFs

To investigate how Mfn1 deficiency enhances adipogenic differentiation we next performed bulk RNA sequencing of WT, Mfn1^-/-, and Mfn2^-/- MEFs in undifferentiated (day -2) and differentiated (day +8) states. Mfn1 and Mfn2 null pre-adipocytes (day -2) both demonstrated significant differences to WT cells, and differences from each other (Fig 3A & 3B). Most strikingly, gene set enrichment analysis (GSEA) showed that the most highly enriched pathway in Mfn1^-/- MEFs was ‘hypoxia-related genes’ (Fig 3C), whereas this gene set was downregulated in Mfn2^-/- MEFs (Fig 3C). This gene set includes Vegfa as well as genes involved in the response to oxidative stress, such as Selenbp1 and Ankzf1. Both knock-out lines showed enrichment of growth-related pathways (e.g. Estrogen Response Early), which was most evident in the Mfn1^-/- MEFs (Fig 3C). Genes implicated in the G2-M checkpoint gene set were upregulated in Mfn2^-/- MEFs but downregulated in Mfn1^-/- MEFs. Mfn2^-/- MEFs also showed upregulation of other pathways related to growth, including ‘Apoptosis’, ‘Mitotic spindle’, and ‘E2F targets’ (Fig 3C), consistent with the rapid growth of this cell line [60].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

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.

https://doi.org/10.1371/journal.pone.0306243.g003

Although this analysis did not highlight the ‘adipogenesis pathway’, it did show that mRNA expression of some pro-adipogenic transcription factors, including Pparg, Cebpa, and Zfp467 [61] was increased and that expression of Zfp521 (coloured red in Fig 3A), an anti-adipogenic factor [62, 63], was reduced in Mfn1^-/- cells compared to WT(Fig 3D). These differences were sustained throughout the differentiation time-course. Despite the higher basal glucose uptake in Mfn2^-/- MEFs compared to WT (Fig 2D), there was no significant difference in Slc2a1 (gene encoding Glut1): 0.2 log2FC, q-value = .03. Unlike models of Mfn2 deletion in vivo [11], there was no evidence of activation of the integrated stress response in Mfn2^-/- MEFs in the undifferentiated state (Fig 3C).

RNA sequencing of mature (day +8) adipocytes (Fig 3E, 3F) demonstrated expression profiles consistent with enhanced differentiation in Mfn1^-/- MEFs and reduced differentiation in Mfn2^-/- MEFs. For example, relative to WT, Mfn1^-/- MEFs showed increased expression of Plin1, Pparg, Slc2a4 (encoding Glut4), Fabp4, and Cd36, whilst each of these were reduced in Mfn2^-/- MEFs. This was also observed on gene set enrichment analysis using Hallmark gene sets [47]: when comparing day -2 pre-adipocytes and day +8 mature adipocytes within each cell line, Mfn1^-/- had markedly greater enrichment of the adipogenesis gene set (Fig 3H–3I). Similarly, pathway analysis comparing day +8 WT and KO cell lines revealed upregulation of the adipogenesis gene set for Mfn1^-/- MEFs and downregulation of the same gene set in Mfn2^-/- MEFs (Fig 3G–3M). Mfn1^-/- MEFs also manifested upregulation of oxidative phosphorylation, hypoxia, and reactive oxygen species pathway gene sets (Fig 3J). Thus, the transcriptomic data supports a pro-adipogenic phenotype in Mfn1^-/-, but not Mfn2^-/-, MEFs.

siRNA knockdown and differentiation of 3T3-L1 preadipocytes

Given the surprisingly divergent effects of Mfn1 and Mfn2 deficiencies on MEF adipogenesis, we next sought to corroborate findings in an independent cellular model. We thus evaluated the impact of siRNA mediated Mfn1 and Mfn2 knockdown in 3T3-L1 pre-adipocytes. The 3T3-L1 line is a MEF subclone first established in the 1970s based on its high propensity for adipogenic differentiation, and has since been widely used as a model of in vitro adipogenesis [64, 65]. Treating 3T3-L1s with siRNAs targeting Mfn1 or Mfn2 effectively reduced target protein expression by more than 95% (Fig 4A and S5A and S5B Fig). Like the KO cell lines, knock-down of Mfn2 reduced Mfn1 expression, whereas knock-down of Mfn1 modestly increased Mfn2 expression (Fig 4A & S5 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

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.

https://doi.org/10.1371/journal.pone.0306243.g004

Throughout differentiation of wild-type 3T3-L1s there was an increase in expression of both markers of mitochondrial fusion (Mfn1 & Mfn2) and fission (Fis1, S6A Fig). In keeping with the MEF data, 3T3-L1s treated with Mfn1-targeted siRNA demonstrated increased lipid accumulation (Fig 4B, 4C). This observation was broadly replicated when the two anti-Mfn1 siRNAs were used independently, such that Mfn1 knock-down efficiency (S7A Fig) correlated with Pparg expression (S7B Fig) and degree of lipid accumulation (S7C and S7D Fig). Whilst there was no difference in total lipid accumulation between scrambled and Mfn2 knockdown cells, lipid droplet size was increased in the Mfn2 knockdown cells (Fig 4B & S6B Fig), as has been previously reported in knock-out MEFs [66]. There was no difference in lipid droplet size between scrambled and Mfn1 knock-down cells (Fig 4B & S6B Fig).

Expression of Pparg1 and Glut4 protein was increased in Mfn1 knockdown cells (Fig 4A & S5 Fig), whereas expression of Plin1 and Fabp4 were similar to that of the control cells. Mfn2 knock-down reduced lipid accumulation and expression of Plin1, Fabp4, Pparg1, Pparg2, and Glut4 proteins (Fig 4 & S5 Fig).

Mitochondrial morphology analysis revealed that loss of Mfn1 and Mfn2 both led to mitochondrial fragmentation, with Mfn1 silencing inducing more drastic changes of the mitochondrial network in mature (day +12) differentiated adipocytes (Fig 4D, 4E), even though, as previously reported [25], the mitochondrial network of wild-type 3T3-L1s manifests increased fragmentation in late differentiation (Fig 4D, 4E). Nutritional deprivation is known to induce mitochondrial fusion in several different cell lines [67]. In keeping with this, incubating the 3T3-L1 adipocytes in low glucose media or without serum increased fusion of the mitochondrial network in siNT treated control cells characterized by mitochondrial elongation (S6 Fig), and accentuating the impact of Mfn1 and Mfn2 deficiencies on the mitochondrial network (Fig 4F, 4G). These findings in 3T3-L1 adipocytes support the observation that loss of mitofusins 1 and 2 have specific effects on adipogenic potential despite both causing mitochondrial network fragmentation.

Effect of ROS scavenging on MEF differentiation

Prior work has suggested that reactive oxygen species (ROS) promote adipogenesis [68, 69]. To test the hypothesis that altered ROS production might be mediating the enhanced adipogenesis observed in Mfn1^-/- MEFs, we added a ROS-scavenger (N-acetylcysteine (NAC)) during adipogenic differentiation of the knock-out cell lines. Consistent with previous reports [68, 69], NAC inhibited lipid accumulation and expression of adipogenic markers in wild-type MEFs (Fig 5). However, i n Mfn1^-/- MEFs, addition of NAC had very little impact on either lipid accumulation (Fig 5A) or adipocyte protein expression (Fig 5B) expression. Similarly, NAC had no effect on adipogenic differentiation of Mfn2^-/- MEFs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

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.

https://doi.org/10.1371/journal.pone.0306243.g005