Loss of lipin 1–mediated phosphatidic acid phosphohydrolase activity in muscle leads to skeletal myopathy in mice

Abstract

Lipin 1 regulates glycerolipid homeostasis by acting as a phosphatidic acid phosphohydrolase (PAP) enzyme in the triglyceride-synthesis pathway and by regulating transcription factor activity. Mutations in human lipin 1 are a common cause of recurrent rhabdomyolysis in children. Mice with constitutive whole-body lipin 1 deficiency have been used to examine mechanisms connecting lipin 1 deficiency to myocyte injury. However, that mouse model is confounded by lipodystrophy not phenocopied in people. Herein, 2 muscle-specific mouse models were studied: 1) Lpin1 exon 3 and 4 deletion, resulting in a hypomorphic protein without PAP activity, but which preserved transcriptional coregulatory function; and 2) Lpin1 exon 7 deletion, resulting in total protein loss. In both models, skeletal muscles exhibited a chronic myopathy with ongoing muscle fiber necrosis and regeneration and accumulation of phosphatidic acid and, paradoxically, diacylglycerol. Additionally, lipin 1–deficient mice had abundant, but abnormal, mitochondria likely because of impaired autophagy. Finally, these mice exhibited increased plasma creatine kinase following exhaustive exercise when unfed. These data suggest that mice lacking lipin 1–mediated PAP activity in skeletal muscle may serve as a model for determining the mechanisms by which lipin 1 deficiency leads to myocyte injury and for testing potential therapeutic approaches.—Schweitzer, G. G., Collier, S. L., Chen, Z., McCommis, K. S., Pittman, S. K., Yoshino, J., Matkovich, S. J., Hsu, F.-F., Chrast, R., Eaton, J. M., Harris, T. E., Weihl, C. C., Finck, B. N. Loss of lipin 1–mediated phosphatidic acid phosphohydrolase activity in muscle leads to skeletal myopathy in mice.

Lipin 1 is an intracellular protein that controls metabolism by acting at multiple regulatory levels (1). This soluble protein acts at the endoplasmic reticulum to dephosphorylate phosphatidic acid (PA) to form diacylglycerol (DAG), the penultimate step in triacylglycerol (TAG) synthesis (2). Lipin 1 also acts in the nucleus to directly interact with DNA-bound transcription factors to regulate gene expression (3–6). Mutations in the gene encoding lipin 1 (LPIN1) are a common cause of early onset, recurrent pediatric rhabdomyolysis (7–15), a condition in which damage to skeletal myocytes results in substantial buildup of myocellular proteins in the blood. Severe rhabdomyolysis can cause acute kidney failure and may lead to death.

Most mutations in LPIN1 that have been associated with rhabdomyolysis are frameshift mutations leading to complete loss of protein expression or generation of severely truncated proteins lacking the carboxyl terminus, which contains its catalytic site and transcription factor–interacting domains. However, other mutations are single amino acid substitutions that we have recently shown (11) to cause a loss of intrinsic phosphatidic acid phosphohydrolase (PAP) activity, but which does not affect transcriptional regulatory function. These in vitro findings suggest that defects in lipin 1–mediated PAP activity underscore rhabdomyolysis in people with LPIN1 mutations, but this has not been tested in vivo. Zhang et al. (16) recently identified previously uncharacterized skeletal muscle myocyte damage in the fatty-liver dystrophic (fld) mouse model caused by a spontaneous mutation in the mouse Lpin1 gene. Mechanistic studies suggested that myopathy in fld mice, and potentially rhabdomyolysis in people with LPIN1 mutations, may be due to impaired autophagy and lysosomal function secondary to PA accumulation and reduced abundance of DAG in skeletal muscle (16). However, fld mice also exhibit progressive lipodystrophy (17) and neuropathy (18), which complicates the interpretation of data obtained with that model. In addition, the adipocyte and neuropathic abnormalities associated with lipin 1 deficiency in fld mice are not phenocopied in humans with LPIN1 mutations (10).

To separate the myocyte-autonomous effects of lipin 1 from the neuropathic and lipodystrophic phenotype of the fld mice, we generated and studied 2 models of muscle-specific Lpin1 deficiency dependent on muscle creatine kinase (MCK)-promoter driven Cre expression. First, we crossed MCK-Cre mice with a previously described Lpin1 flox mouse that serendipitously expresses a truncated lipin 1 protein missing the N-terminal 115 aa after Cre-mediated recombination. The truncated protein lacks intrinsic PAP activity but has preserved transcriptional regulatory function (19). We also crossed MCK-Cre mice with mice harboring a novel Lpin1 flox allele that leads to complete lipin 1 protein knockout. We demonstrate that either selective loss of PAP activity or complete deletion of lipin 1 protein in skeletal muscle leads to active myopathy and myofibrillar regeneration, increased muscle PA levels, evidence for mitochondrial dysfunction, and blunted myocellular autophagy. The phenotype of these mice seems somewhat more severe but mechanistically similar to the reported phenotype of fld mice (16, 20). However, compared to the fld model, the muscle-specific modulation of lipin 1 activity shows differences in the muscle abundance of other key lipids, such as DAG and TAG, suggesting that low levels of these lipids in fld mice are due to lipodystrophy and are not copied in humans (7–9, 11). Lastly, we demonstrate that skeletal muscle lipin 1 deficiency leads to increased exercise endurance in younger mice, but that capacity to exercise declines with age and is associated with elevated plasma creatine kinase after exhaustive exercise, which parallels some reports of people with LPIN1 mutations. These data suggest that myocyte-autonomous deletion of lipin 1-mediated PAP activity leads to active myopathy and eventual decline in exercise capacity in mice. These mice may have utility for better defining mechanisms and as a model for potential therapies to prevent rhabdomyolysis in patients with LPIN1 mutations.

MATERIALS AND METHODS

Animal studies

All mouse studies were approved by the Institutional Animal Care and Use Committee of Washington University and fulfilled U.S. National Institutes of Health (NIH) requirements for humane care. All mice were in the C57BL/6J background. The generation of mice harboring an Lpin1 allele with LoxP sites flanking exons 3 and 4 (Lpin1^Δ115) has been previously described (21). MCK-Lpin1^Δ115 mice were generated by crossing mice harboring the Lpin1^Δ115 allele with mice expressing Cre recombinase under control of the muscle creatine kinase (MCK) promoter (Fig. 1A). To generate mice with complete, but conditional, loss of lipin 1, embryonic stem cells containing a knockout allele of Lpin1 were obtained from the Knockout Mouse Project (IKMC project ID 89403; International Knockout Mouse Consortium). This construct contains a targeted allele of Lpin1 containing inserted cassettes for promoter-driven LacZ and Neo cDNAs flanked by flippase recognition target sites upstream of exon 7 of the Lpin1 gene, which are flanked by LoxP sites. Propagated embryonic stem cells were injected into developing embryos and then implanted into pseudopregnant females. Chimeric offspring were mated to establish germline transmission. Resulting heterozygous offspring were mated with C57BL/6J mice expressing Flp recombinase in a global manner (chicken α actin promoter-driven transgenic) to remove LacZ and Neo cassettes and generate mice harboring the conditional floxed allele. Lpin1 floxed mice were then crossed with hemizygous C57BL/6J mice expressing Cre recombinase under the MCK promoter to create MCK-Lpin1^−/− mice (Fig. 1D). Littermates not expressing Cre [wild-type (WT) mice] were used as control mice in all experiments. All mice were group housed and maintained on standard laboratory chow diet (PicoLab Rodent Diet 5053; LabDiet, St. Louis, MO, USA) and on a 12/12-h light/dark cycle.

Figure 1.

Generation of mouse models with muscle-specific lipin 1 deficiency. A) Schematic showing Lpin1 gene-targeting strategy for generation of truncated lipin 1. MCK-Lpin1^Δ115 animals were generated using LoxP sites flanking exons 3 and 4 of lipin 1, which encode the principal start codon. MCK-driven Cre recombination led to enforcement of an alternative start codon in exon 5. B) Western blot analysis of gastrocnemius, heart, liver, and eWAT from WT and MCK-Lpin1^Δ115 mice. C) PAP activity in vastus, heart, liver, and eWAT of WT and MCK-Lpin1^Δ115 mice. Data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^Δ115 compared with WT mice. n = 5–6 and animals were 3–4 mo old. D) Schematic showing Lpin1 gene-targeting strategy for generation of full-protein knockout of lipin 1. MCK-Lpin1^−/− animals were generated with LoxP sites flanking exon 7 of lipin 1, and MCK-driven Cre recombination led to complete ablation of the lipin 1 protein. Western blot analysis of gastrocnemius, heart, liver, and eWAT from WT and MCK-Lpin1^−/− mice 3–4 mo old; n = 5.

Protein isolation and abundance assessment

Protein from frozen tissue was homogenized in 0.3–1 ml ice-cold homogenization buffer [25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, pH 8.0], supplemented with 1 mM activated Na₃VO₄, 1 mM phenylmethanesulfonyl fluoride, 5 mM sodium fluoride, and 1× complete protease inhibitor cocktail tablet (04693116001; F. Hoffmann-La Roche, Basel, Switzerland) using high-speed tissue disruption with the TissueLyser II (Qiagen, Hilden, Germany). For tissue used for LC3 and p62 protein assessment, tissue was homogenized in RIPA buffer [50 mM Tris, ph 8.0, 150 mM NaCl, 1% nonidet-p40 (octylphenoxypolyethoxyethanol; MilliporeSigma, Burlington, MA, USA), 0.5% sodium deoxycholate, 0.1% SDS, and 1× σ protease inhibitor tablet (S8820; MilliporeSigma)] and disrupted via glass Dounce homogenization. Tissue homogenates were subsequently solubilized by rotating at 4°C at 50 rpm for 1 h before being centrifuged (15,000 g for 15 min at 4°C) and the supernatant collected. Aliquots of the lysate from tissues were used to determine protein concentration by the bicinchoninic acid assay (BCA), according to the manufacturer’s instructions (23227; Thermo Fisher Scientific, Waltham, MA, USA). The remaining lysates were stored at −80°C until further analyzed. Mitochondrial protein was isolated by fractionation as per previous studies (22, 23). In brief, skeletal muscles (gastrocnemius and vastus tissue combined) from mice were Dounce-homogenized in Tris-sucrose buffer. Cellular debris was spun in 4°C at 1000 g for 5 min, and the supernatant containing mitochondria was spun in 4°C at 10,000 g for 10 min and washed in Tris-sucrose buffer 3 times. The final pellet was resuspended in mito buffer (14 mM NaCl, 10 mM Tris, 1 mM EDTA, 0.2% NP-40, 10% glycerol supplemented with 1 mM activated Na₃VO₄, 1 mM phenylmethanesulfonyl fluoride, 5 mM sodium fluoride, and 1× complete protease inhibitor cocktail tablet). The total protein content of this pellet, measured by BCA, was greater in MCK Lpin1^Δ115 vs. WT littermates.

Lysates were subjected to SDS-PAGE and transferred to PVDF membranes. Blots were then rinsed with Tris-buffered saline plus Tween 20 (TBST; 0.14 µm NaCl, 0.02 µm Tris base, pH 7.6, and 0.1% Tween), blocked with 5% bovine serum albumin (BSA) in TBST for 1 h at room temperature, washed 3 times for 10 min at room temperature, and incubated with the relevant primary antibody (1:1000 for all antibodies; 1:500 for LC3) in 5% BSA overnight at 4°C. Blots were then washed 3 times for 5 min with TBST, incubated with relevant secondary antibodies for 1 h at room temperature, washed again 3 times for 10 min with TBST, and washed 2 times for 10 min with Tris-buffered saline. For lipin 1 and α-tubulin, protein bands were visualized with the Odyssey Imaging System (Li-Cor Biosciences, Lincoln, NE, USA). For p62 and LC3, protein bands were visualized by administering ECL detection reagent (RPN2209; GE Healthcare Life Sciences, Little Chalfont St. Giles, United Kingdom) and quantifying chemiluminescence with G:Box Chemi XT4 (Syngene International, Bangalore, India). Lipin 1 (sc-98450) antibody was obtained from Santa Cruz Biotechnology (Dallas, TX, USA). Anti–α-tubulin clone B-5-1-2 (T5168) and LC3 (L7543) were purchased from MilliporeSigma. Goat anti-mouse IRDye 680 (926-32220) and goat anti-rabbit IRDye 680 (926-68021) secondary antibodies were obtained from Li-Cor Biosciences. p62 (18420-1-AP) was obtained from the Proteintech Group (Rosemont, IL, USA). Anti-rabbit secondary antibody (7074) was obtained from Cell Signaling Technology (Danvers, MA, USA).

PAP assay

The method to assess PAP activity was modified from Martin et al. (24) and has also been previously described (25). Approximately 50 mg of tissue were homogenized in 1 ml of lysis buffer [0.01 M Tris-HCl, pH 7.3, 0.25 M sucrose, 0.5% Tween-20, 1 mM DTT, 1× EDTA-free protease inhibitor (04693132001; F. Hoffmann-La Roche)] using Tissue Master-125 (Omni International, Kennesaw, GA, USA). After homogenization, aliquots of the supernatant were used to determine protein concentration by BCA. The ¹⁴C-phosphatidic acid used was prepared as follows: 15 µl 0.1 mCi/ml of ¹⁴C-phosphatidic acid l-α-dioleoyl (oleoyl-1-¹⁴C) Na salt (¹⁴C -PA) (1303; American Radiolabeled Chemicals, St. Louis, MO, USA) was added to 3 mM 3-sn-phosphatidic acid sodium salt (P9511; MilliporeSigma) and 2 mM l-α-phosphatidylcholine (P3556; MilliporeSigma) in 1 ml of chloroform, evaporated under a stream of N₂, and resuspended in 1 ml of ice-cold double-distilled H₂O. This mixture was sonicated 3 times for 10 s and kept on ice between sonications. To run the assay reaction, 5 µl of each sample was incubated with 40 µl of assay buffer [0.1 M Tris/maleate, pH 6.9, 10 mM MgCl₂, 0.2% fatty acid-free BSA, 1 mM DTT, and 1× EDTA-free protease inhibitor with or without 12.5 mM N-ethylmaleimide (E3876; MilliporeSigma)] and 5 µl of ¹⁴C-phosphatidic acid, which was incubated at 37°C for 20 min. The reaction was stopped by adding 1 ml of a 19:1 (v/v) chloroform/methanol mixture containing 0.08% olive oil to each reaction and vortexing briefly; 500 mg of activated aluminum oxide was added to the tubes, and they were capped securely. Then, samples were vortexed for 3 cycles of 30 s, followed by being left undisturbed for 10 min. Samples were then spun at 10,000 g for 5 min, and 250 µl of sample was added to the scintillation fluid to determine the ¹⁴C radioactivity.

Microarray for transcriptional profile and pathway analysis

Microarray gene expression analysis of mouse vastus tissue was performed by the Washington University Genome Technology Access Center (St. Louis, MO, USA) using the Illumina (San Diego, CA, USA) 8v2 platform. Total RNA was isolated with RNeasy Mini Kit (Qiagen).

The background-subtracted raw microarray data were subjected to quantile normalization. Parametric analysis of gene set enrichment was performed with the normalized data to determine the effect of Lpin1 deletion on skeletal muscle biologic pathways, as previously described (26, 27). Gene sets were obtained from Molecular Signatures Database (gene ontology gene sets, C5 collection; Broad Institute of Massachusetts Institute of Technology and Harvard, Cambridge, MA, USA), and z scores and significance values were calculated for each gene set. A value of P < 0.05 was considered significantly changed. All data were obtained using an R-based statistical software package (R Foundation for Statistical Computing, Vienna, Austria), Parametric Gene Set Enrichment Analysis, available at the Bioconductor (Seattle, WA, USA) website. The heat map was generated from microarray data using Partek software (St. Louis, MO, USA), and data are shown for all probes with a false-discovery rate (FDR) < 0.05. The FDR was determined using the step-up method between MCK-Lpin1^Δ115 and WT littermate control animals.

Muscle histology

Isolated muscle was mounted with tragacanth gum (G1128; MilliporeSigma) and quick frozen in liquid nitrogen–cooled 2-methylbutane. Samples were stored at −80°C until sectioning into 12-µm sections. Cryostat sections were processed and fixed in acetone. Hematoxylin and eosin (H&E) staining was conducted, and muscle fiber cross-sections were visualized using EVOS FL Color Imaging system (Thermo Fisher Scientific). Succinate dehydrogenase staining of muscle fiber cross sections were visualized with an Eclipse 80i microscope (Nikon, Tokyo, Japan) with NIS-Elements AR 3.0 software (Nikon). For immunohistochemistry, the sections were fixed with 10 min incubation in ice-cold acetone, then blocked in TSA blocking reagent (FP1012; PerkinElmer, Waltham, MA, USA) for 2 h, and incubated with primary antibodies: TOM20 (sc-11415; Santa Cruz Biotechnology) 1:500, SQSTM1/p62 (H00008878; Novus Biologicals, Littleton, CO, USA) 1:100, LC3 (0231-100; NanoTools Antikoerpertechnik, Teningen, Germany) 1:250, and BNIP3 (ab10433; Abcam, Cambridge, United Kingdom) 1:200; overnight at 4°C, followed by the appropriate secondary antibody. Coverslips were mounted with Mowiol 4–88 (MilliporeSigma) + DAPI and examined with a fluorescence microscope [Nikon 80i upright+ and Scientific EZ monochrome charge-coupled device (CCD) camera (Roper Technologies, Sarasota, FL, USA) with deconvolution software analysis (NIS Elements; Nikon)]. Nonfluorescent images were taken with a 5-megapixel color CCD camera (Nikon).

mRNA isolation and gene expression quantification

For all analyses, total RNA from skeletal muscle was isolated in RNAzol (Thermo Fisher Scientific) reagent. RNA (500 ng) was subjected to high-capacity cDNA reverse transcription (Thermo Fisher Scientific) and SYBR Green RT-PCR (Thermo Fisher Scientific), following the manufacturer’s instructions. Results were corrected to the expression of Rplp0. Sequence of gene-specific primers are listed in Supplemental Table 1.

Mitochondrial DNA determination

Tissue was homogenized in RNAzol (1 ml of RNAzol/50 mg tissue), chloroform added, and the mixture was vortexed vigorously. Samples were spun at 10,000 g for 15 min at 4°C. The top aqueous phase was removed and used for RNA isolation, and 500 μl of back-extraction buffer (stock: 4 M guanidine thiocyanate, 50 mM sodium citrate, 1 M Tris-free base) was added per 1 ml remaining interphase and organic phase. Sample was mixed well and stood for 10 min at room temperature before being centrifuged at 3000 g for 30 min at 4°C. The upper phase containing DNA was transferred, and 4 μl of polyacryl carrier was added/ml RNAzol used; DNA was precipitated by adding 400 μl of isopropanol/ml of RNAzol used, mixed, left standing for 5 min at room temperature, and centrifuged at 12,000 g for 5 min at 4°C. The pellet was washed in ethanol 3 times, air dried, dissolved in 8 mM NaOH, vortexed, and heated, if needed (55°C) to dissolve; 0.34 μl of 1 M HEPES and 0.3 μl of 100 mM EDTA were added. ND1 and LPL genes were quantified for mitochondrial DNA (mtDNA) and nuclear DNA.

Mitochondrial respiration

High-resolution respirometry was conducted using a 2-chamber Oxygraph O2k (Oroboros Instruments, Innsbruck, Austria). Skeletal muscle respiration protocol was derived from Ryan et al. (28) with modification for mouse tissue. Soleus muscles were isolated and placed in muscle biopsy buffer [50 mM K-2-(N-morpholino)ethanesulfonic acid, 7.23 mM K₂EGTA, 2.77 mM CaK₂EGTA, 20 mM imidazole, 20 mM taurine, 5.7 mM Na2ATP, 14.3 mM phosphocreatine, 0.5 mM DTT, and 6.56 mM MgCl₂0 · 6H₂O, pH 7.1] before further processing. Fiber strips from soleus were separated along their longitudinal axis with needle-tipped forceps under magnification (SZX7; Olympus, Tokyo, Japan) and permeabilized in muscle biopsy buffer with 0.05 mg/ml of saponin for 20 min at 4°C with gentle rocking. Fibers were transferred to MiRO5 buffer (20 mM HEPES, 0.5 mM EGTA, 10 mM K₂HPO₄, 3 mM MgCl₂0 · 6H₂O, 60 mM lactobionic acid, 20 mM taurine, 110 mM d-sucrose, 1.0 mg/ml BSA, pH 7.4) and gently rocked for 15 min at 4°C. Fibers were blotted dry and weighed. Approximately 3 mg of fibers were placed in an Oxygraph chamber with MiRO5 buffer containing 3 mg/ml creatine and 10 µM blebbistatin. The respiration protocol was conducted with sequential additions of the following: 10 mM glutamate, 2 mM malate, 5 mM pyruvate (complex I substrates), 5 mM ADP, 10 mM succinate (complex II substrate), 3 titrations of 0.5 mM carbonyl cyanide‑p‑trifluoromethoxyphenylhydrazone (FCCP; uncoupling agent), 10 μM cytochrome c to test for mitochondrial membrane integrity, and 0.5 μM rotenone (complex I inhibitor). State 2 (i.e., leak) respiration was measured when oxygen flux reached a steady state after the addition of pyruvate. State 3 (i.e., OxPhos) respiration was measured when the oxygen flux reached steady state after the addition of succinate. Uncoupled respiration (i.e., electron transport system) was measured as the highest oxygen flux achieved after FCCP was added.

Autophagy assessment

The method to assess autophagy in skeletal muscle has been previously described (29). MCK-Lpin1^Δ115 mice were i.p. injected with 0.4 mg/kg colchicine (1364; Bio-Techne, Minneapolis, MN, USA) on d −2 and −1. Mice were euthanized on d 0, and skeletal muscles were isolated and prepared for protein isolation, as described above.

Electron microscopy

Tissue was fixed in a modified Karnovsky’s fixative of 3% glutaraldehyde and 1% paraformaldehyde in 0.1 M sodium cacodylate buffer, postfixed in 2% osmium tetroxide in 0.1 M sodium cacodylate buffer for 1 h, en bloc stained with 3% aqueous uranyl acetate for 30 min, dehydrated in graded ethanols, and embedded in PolyBed 812 (08792-1; Polysciences, Warrington, PA, USA). Tissue blocks were sectioned at 90 nm thick, after staining with Venable’s lead citrate, and being viewed with a JEOL model 1400EX electron microscope (Jeol, Tokyo, Japan). Digital images were acquired with the AMT Advantage HR (Advanced Microscopy Technology, Woburn, MA, USA) high-definition CCD, 11-megapixel transmission electron microscopy camera. Lipid droplet area was quantified from 6 scans/mouse and analyzed with ImageJ (National Institutes of Health, Bethesda, MD, USA) software.

Lipid analysis

Electrospray ionization liquid chromatography–mass spectrometry (LC-MS) analysis of lysophosphatidic acid (LPA), PA, DAG, and TAG, phosphatidylglycerol (PG), and cardiolipin (CL) in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− vastus and gastrocnemius muscles, was determined as described with modification (30). The LC-MS analysis was performed with either a 10A HPLC system (Shimadzu, Kyoto, Japan) or a SIL-20AC HT auto-sampler (Shimadzu) coupled to a Thermo Fisher Scientific TSQ Quantum Ultra Triple Quadrupole Mass Spectrometer operated in selected reaction monitoring mode or with a Thermo Fisher Scientific Vantage TSQ Mass Spectrometer with Thermo Fisher Scientific Accela Ultra Performance Liquid Chromatography operated by Xcalibur software (Thermo Fisher Scientific) using selected ion monitoring mode.

LPA-(17:0), PA-(14:0)₂, DAG-(15:0)₂, TG-(17:0)₃, PG-(15:0)₂, and CL-(14:0–14:0)₂ were used as internal standards for LPA, PA, DG, TG, PG, and CL, respectively. Quantitation of lipids was based on the ratio of the peak area of the analyte to the internal standard. For example, the ratio of CL-(18:2–18:2)₂ and CL-(14:0–14:0)₂ is used for measurement of CL-(18:2–18:2)₂.

Endurance exercise assessment and plasma creatine kinase measures

MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice with their respective WT littermate controls (3–4 mo or 1 yr old) performed a single bout of treadmill endurance exercise to exhaustion. The mice were unfed overnight (∼12 h) before the exercise session. After a 5 min familiarization period on the treadmill, the exercise protocol consisted of graduated intensity running at 5 m/min for 5 min, 10 m/min for 5 min, 15 m/min for 5 min, 20 m/min for 5 min, and 25 m/min until exhaustion. The mice were encouraged to run with a mild electric stimulus (20 V) placed at the end of the treadmill. Mice were judged to be exhausted by refusal to remain on treadmill belt for >5 s. In MCK-Lpin1^−/− mice, blood was collected 3 h after completing exercise via mandibular bleed, which was spun at 8000 g for 8 min to isolate plasma. Plasma was also isolated from blood collected from those mice during a sedentary, fed state as a basal measurement. Plasma creatine kinase was determined with the Hoffmann-La Roche Creatine Kinase Kit and analyzed on the Hoffmann-La Roche Cobas c501 analyzer.

Statistical analysis

Data were analyzed via SPSS software (IBM, Armonk, NY, USA) and Excel (Microsoft, Redmond, WA, USA). Statistical comparisons were made with independent-sample Student’s t tests or 2-way ANOVA with Tukey’s post hoc testing to identify specific differences. Before analysis, normality and equal variance of samples were determined by Kolmogorov-Smirnov testing and Levene testing, respectively. If those conditions were violated, a square root transformation, intended to produce data to satisfy normality and equal variance assumptions, was used. Otherwise, nonparametric methods (e.g., Kruskal-Wallis for 1-way ANOVA with Mann-Whitney U post hoc analysis and Friedman’s test for 2-way ANOVA) were used as an alternative to the more standard analyses. All data are presented as the means ± sem, with a statistically significant difference defined as a P < 0.05.

RESULTS

Characterization of lipin 1 protein in mouse models

We generated mice lacking lipin 1–mediated PAP activity in a muscle-specific manner by crossing mice harboring a Lpin1 allele flanked by LoxP sites around exons 3 and 4 (21) with mice expressing Cre recombinase driven by the MCK gene promoter (MCK-Lpin1^Δ115 mice) (Fig. 1A). Consistent with previous reports using this lipin 1 floxed allele (19, 25), Western blot analyses revealed that a truncated lipin 1 protein of diminished abundance was expressed in a skeletal muscle– and heart-specific manner in MCK-Lpin1^Δ115 mice (Fig. 1B). That protein lacks 115 aa of the N terminus of lipin 1, and the resulting protein is deficient in PAP activity but retains transcriptional regulatory function (19, 25). Accordingly, compared with WT littermate controls, PAP activity in the tissues of MCK-Lpin1^Δ115 mice was greatly diminished in skeletal muscle and heart (Fig. 1C). PAP activity was not affected in liver and epididymal white adipose tissue (eWAT). MCK-Lpin1^Δ115 mice were viable and outwardly normal. Compared with WT littermate control mice, MCK-Lpin1^Δ115 mice exhibited similar body weights and tissue masses (Supplemental Fig. 1).

Using a novel line of mice with a floxed Lpin1 exon 7 (Fig. 1D), which creates a frameshift mutation, we also generated mice completely lacking lipin 1 protein in a skeletal muscle– and heart-specific manner that retained protein in liver and eWAT (MCK-Lpin1^−/− mice; Fig. 1D). Those mice were also outwardly normal, viable, and exhibited similar body weights and tissue masses (Supplemental Fig. 1). In studies described below, MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice are often used interchangeably, but which strain was used is always indicated.

Myocyte damage in mouse models of lipin 1 deficiency

Analysis of H&E-stained sections of MCK-Lpin1^Δ115 gastrocnemius muscles and MCK-Lpin1^−/− vastus muscles showed an active myopathy, with muscle fiber necrosis and regeneration that was more prominent with increasing age (Fig. 2A). Nearly all muscle fibers in 1-yr-old mice were centrally nucleated, suggesting almost complete turnover of muscle fibers. The similarity of the phenotypes of the 2 models suggests that loss of lipin 1–mediated PAP activity is sufficient to induce myopathy in vivo, whereas the transcriptional regulatory function of the protein is dispensable.

Figure 2.

Mice with muscle-specific Lpin1 deficiency exhibit active and progressive myopathy. A) H&E stains of skeletal muscle cross sections from MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice with respective age-matched WT controls at 3–4 mo and 1 yr old. Field width is ∼200 µm. B) Quantitative RT-PCR performed to quantify expression of the indicated genes from MCK-Lpin1^Δ115 mice 3–4 mo and 1 yr old. Various markers of macrophage, proinflammatory cytokines, necrosis, and muscle regeneration were measured. Values were normalized (1.0) to age-matched WT mice, and data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^Δ115 compared with age-matched WT mice; n = 7–9.

Consistent with those observations, quantitative RT-PCR detected increased expression of genes encoding macrophage markers (Emr1, Cd68, Cd11c), proinflammatory cytokines (Tnfa, Il6, Il1b) as well as markers of necrosis (Ripk3 and Mlkl) in MCK-Lpin1^Δ115 muscles compared with WT littermate muscle at 3–4 mo and 1 yr old (Fig. 2B). Expression of regeneration markers (Fsp1, Col12a, Adam12) was observed in MCK-Lpin1^Δ115 mice, compared with age-matched WT littermates. These histologic and molecular biologic findings are consistent with active and progressive myopathy with regeneration because of loss of lipin 1–mediated PAP activity in muscle.

Alterations in transcript profiles with muscle deficiency in PAP activity

To characterize the global transcriptional differences secondary to loss of lipin 1–mediated PAP activity in muscle without the confounding effects of the loss of transcriptional regulatory function, microarray analyses were conducted with RNA from vastus muscle tissue of MCK-Lpin1^Δ115 and WT mice. Unsupervised, hierarchical clustering of the data from differentially expressed mRNAs (fold change 1.8, FDR <0.05) revealed low levels of intragroup variance between samples and demonstrated that these mRNAs were sufficient to distinguish MCK-Lpin1^Δ115 mice from WT control (Fig. 3A). To identify the nature of these unique genotypic signatures, analyses of the microarray data revealed an up-regulation of several cellular pathways involved in inflammatory responses, cytokine activity, and cellular defense in the MCK-Lpin1^Δ115 compared with WT controls, indicating myocellular inflammation or damage (Fig. 3B, red highlighted pathways). Additionally, down-regulated pathways of MCK-Lpin1^Δ115 include those involved in mitochondrial partitioning and function (Fig. 3B, blue highlighted pathways). Collectively, the transcriptomic profile suggests that loss of lipin 1–mediated PAP activity provokes an inflammatory response characterized by heightened cytokine expression and infiltration of immune cells.

Figure 3.

Microarray analysis of MCK-Lpin1^Δ115. A) Standardized heat map of differentially expressed mRNAs from microarray of vastus from mice 3–4 mo old, using unsupervised hierarchical clustering selected at FDR <0.05. B) Pathway analyses of vastus muscle gene-expression microarray studies. The top 15 significantly changed cellular pathways up- and down-regulated are listed to highlight significant transcriptional alterations of MCK-Lpin1^Δ115 (vs. WT) studied; n = 4.

Alterations in lipid intermediates in muscle specific lipin 1 deficiency

Electron microscopy revealed a 2-fold increase in lipid droplet area (WT: 4895 ± 717 pixels², MCK-Lpin1^Δ115: 10197.2 ± 2020 pixels²) in skeletal muscle of MCK-Lpin1^Δ115 mice compared with normal littermates in both the intermyofibrillar and subsarcolemmal space (Fig. 4A). To characterize the nature of that lipid biochemically, we quantified triglyceride and several glycerol-3-phosphate (G3P) pathway lipid intermediates (Fig. 4B) in mouse skeletal muscle. As would be predicted, PA, the substrate of lipin 1, was significantly elevated in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice vs. respective WT littermates (Fig. 4C) with contributions from many individual PA species of variable fatty acid chain composition (Supplemental Fig. 2). Curiously, DAG, the product of lipin 1–mediated PAP activity, was also significantly elevated in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice compared with their respective normal littermates (Fig. 4C, D) with contributions from many individual DAG species (Supplemental Fig. 2). TAG levels were not different in either model compared with their respective WT littermate (Fig. 4C, D), and LPA levels in MCK-Lpin1^Δ115 mice were also no different from WT control (Supplemental Fig. 2). These results suggest that skeletal muscle deficiency lipin 1–mediated PAP activity leads to accumulation of PA and, unexpectedly, DAG, which contradicts previous studies conducted in lipodystrophic fld mice.

Figure 4.

Mice with muscle-specific Lpin1 deficiency exhibit muscle lipid accumulation of G3P lipid and exhibit muscle accumulation of cardiolipin. A) Electron micrographs of lipid droplets in soleus muscle. A greater area of lipid droplets is present in both the contractile and sarcolemmal regions of the muscle fiber in 3–4-mo-old MCK-Lpin1^Δ115 mice. Field width is ∼6 µm. B) The pathway schematic depicts the synthesis of TAG from G3P. Circled is lipin 1, the primary skeletal muscle PAP enzyme, and DAG. Also depicted is the PA branch with downstream phospholipids, including PG and CL. C, D) PA, DAG, and TAG from vastus and gastrocnemius muscles of MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice, respectively, were assessed via LC-MS analysis. E, F) PG and tetra linoleoyl CL (18:2/18:2)₂ from vastus and gastrocnemius muscles of MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice with respective WT littermate controls were assessed via LC-MS analysis. Values were normalized (1.0) to age-matched WT mice (3–4 mo), and data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^Δ115 or MCK-Lpin1^−/− compared with respective WT littermate mice; n = 5–7.

Abnormal mitochondria in lipin 1 deficiency

Because lipin 1 deficiency might preferentially shunt PA away from the classic G3P pathway to PG and CL (Fig. 4E, F), those lipids were also measured. In both MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice, increased CL abundance was detected. In MCK-Lpin1^Δ115 mice, PG was significantly increased and tended to be increased in MCK-Lpin1^−/− mice. Because PG and CL are phospholipids found almost exclusively in mitochondrial membranes and CL is an important and abundant membrane lipid in the inner mitochondrial membrane (31, 32), these results could also suggest an accumulation of mitochondria in skeletal muscle in these mouse models of Lpin1 deficiency.

Supporting the observation of increased mitochondria in skeletal muscle-specific lipin 1–deficient animals, MCK-Lpin1^−/− mice had increased staining of Tom20, a mitochondrial protein, in vastus skeletal muscle compared with WT littermate controls both in the young 3–4 mo-old and the 1-yr-old animals. Additionally, MCK-Lpin1^Δ115 mice had increased mtDNA content compared with WT controls (Fig. 5B) and increased total mitochondrial protein normalized to skeletal muscle tissue weight (Fig. 5B). Further, in MCK-Lpin1^−/− mice, succinate dehydrogenase staining revealed abnormal internal architecture with some muscle fibers having increased accumulation of mitochondria around the periphery compared with WT controls (Fig. 5C). Because succinate dehydrogenase is a marker of oxidative muscle fibers, we also assessed the expression of myosin isoforms associated with type 1, 2A, 2X, and 2B fibers in gastrocnemius tissue. We observed no increases in the expression of the myosin isoforms associated with the more-oxidative fiber types, MyHC-1 (Myh7), in knockout mice of any age, whereas expression of MyHC-2A (Myh2) and MyHC-2B (Myh1) were reduced in 3–4-mo-old MCK-Lpin1^−/− mice compared with littermate controls by quantitative PCR (33).

Figure 5.

Mice with muscle-specific Lpin1 deficiency exhibit mitochondrial accumulation and abnormalities with noneffectual alterations in muscle fiber–type composition. A) Tom20 staining (bright punctate pattern) of vastus muscle in MCK-Lpin1^−/− mice 3–4 mo and 1 yr old. B) Protein quantity per tissue weight of mitochondrial fraction and mtDNA content in MCK-Lpin1^Δ115 mice and age-matched WT controls. C) Succinate dehydrogenase staining of skeletal muscle cross sections from MCK-Lpin1^−/− mice with respective age-matched WT controls at 3–4 mo and 1 yr of age. Red arrows indicate examples of “ragged blue fibers,” which show intense, irregular staining of succinate dehydrogenase in the periphery of muscle fibers. Field width is ∼400 µm. D) Quantitative RT-PCR to quantify expression of muscle fiber types (type 1, 2A, 2X, and 2B) via assessing myosin heavy-chain genes in MCK-Lpin1^−/− mice 3–4 mo and 1 yr old. Values were normalized (1.0) to age-matched WT mice (3–4 mo), and data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^Δ115 or MCK-Lpin1^−/− compared with respective WT littermate mice; n = 6–7 for MCK-Lpin1^Δ115 and WT controls; n = 10–11 for MCK-Lpin1^−/− and WT controls.

The increase in mitochondrial content in MCK-Lpin1^−/− muscle could be due to enhanced mitochondrial biogenesis. However, consistent with the pathway analysis conducted with microarray data (Fig. 3B), the expression of Cox2 and the master regulator of mitochondrial biogenesis PGC-1α (Ppargc1a) was reduced in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice compared with littermate controls (Fig. 6A). In contrast, Cytc expression was not affected, suggesting a partial pattern of reduced mitochondrial gene expression and mitochondrial biogenesis not underlying the observed accumulation of mitochondria. Both models of lipin 1–deficient mice had increases in swollen mitochondria with rarefied cristae in the intermyofibrillar and subsarcolemmal space (Fig. 6B). Finally, skeletal muscle respiration was altered in 1-yr-old MCK-Lpin1^−/− mice compared with WT littermates with a statistically significant increase in oxygen flux in the leak state (state 2 respiration without addition of ADP), a trend toward a decrease in oxygen flux with uncoupled respiration, and overall, a significantly reduced respiratory control ratio (state 3:2). These data collectively suggest that skeletal muscle lipin 1 deficiency is associated with accumulation of mitochondria with abnormal ultrastructure and overall dysfunctional respiratory capability.

Figure 6.

Mice with muscle-specific Lpin1 deficiency display mitochondrial abnormalities and noneffectual alterations in muscle fiber–type composition. A) Quantitative RT-PCR to quantify expression of muscle fiber types (type 1, type 2A, type 2X, and type 2B) via assessing myosin heavy-chain genes in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice at 3–4 mo old. Values were normalized (1.0) to age-matched WT mice, and data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^Δ115 or MCK-Lpin1^−/− compared with respective WT littermate mice. n = 10–11. B) Electron micrographs of lipid droplets in soleus muscle showing swollen and rarefied mitochondria in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice in the intermyofibrillar and subsarcolemmal space. Field width is ∼3 µm. C) Oxygen flux indicating mitochondrial respiration for Leak (state 2 respiration), oxphos (state 3 respiration), electron transport system (ETS; uncoupled respiration), and respiratory control ratio (RCR), the oxphos:Leak. Values are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^−/− compared with WT littermate mice; n = 8.

Skeletal muscle autophagy

Increased abundance of abnormal mitochondria in lipodystrophic fld mice was associated with an inability to clear damaged mitochondria through the process of autophagy (16). Autophagy is an intracellular degradation system that brings targeted cellular contents to the lysosome, a membrane-bound organelle capable of degrading many biomolecules and maintaining a homeostatic environment through regulation of metabolism, cell signaling, and membrane repair (34). To determine whether autophagy-lysosomal activity might be altered in muscle-specific lipin 1-deficient mice, we assessed protein abundance of p62, LC3-I, and LC3-II in WT or MCK-Lpin1^Δ115 mice after treating with vehicle or colchicine, a microtubule depolarizing agent that blocks the trafficking of autophagosomes to lysosomes resulting in autophagosome accumulation. p62 transports targeted proteins to autophagosomes for lysosomal degradation, and the p62 also becomes degraded. In whole-cell lysates, we observed increased p62 protein abundance in tibialis anterior muscle of MCK-Lpin1^Δ115 mice compared with WT controls (Fig. 7A), suggesting impaired lysosomal function (35). The ratio of inactive cytosolic LC3-I to its activated form LC3-II under basal conditions was also decreased in MCK-Lpin1^Δ115 mice compared with WT controls, suggesting abnormal autophagic flux (Fig. 7A). Treatment with colchicine resulted in increased LC3-II accumulation in WT mice but had no effect in MCK-Lpin1^Δ115 mice. These data suggest a reduction in autophagic activity of mice with skeletal muscle-specific lipin 1 deficiency (Fig. 7A).

Figure 7.

MCK-Lpin1^Δ115 mice display autophagy in skeletal muscle. A) Increased p62 protein abundance in tibialis anterior muscles of vehicle-treated MCK-Lpin1^Δ115 mice compared with WT. With 2 d of intraperitoneal injections of 0.4 mg/kg colchicine, although p62 abundance is increased, there was no increase in LC3-II protein abundance detected in MCK-Lpin1^Δ115 despite increased LC3-II in WT. Values were normalized (1.0) to age-matched WT mice (3–4 mo) treated with vehicle, and data are shown as means ± sem. *P < 0.05 (ANOVA post hoc) for differences in MCK-Lpin1^Δ115 compared with WT littermate mice. N.s., nonspecific band. n = 5/condition. B) Increased p62, LC3-I, and LC3-II protein abundance in mitochondrial fraction from vastus muscles of MCK-Lpin1^−/− of 1-yr-old mice compared with WT controls. VDAC is a mitochondrial-specific loading control. Values were normalized (1.0) to age-matched WT mice (3–4 mo) treated with vehicle, and data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^−/− compared with WT littermate mice; n = 5–6. C) Increased abundance of Tom20, p62, LC3, and BNIP3 in vastus muscle fiber from 1-yr-old WT or MCK-Lpin1^−/− mice. The third column overlay is Tom20 (green), and p62, LC3, or BNIP3 (red).

To assess the interaction of mitochondria with autophagic mediators in muscle of these mice, mitochondria were isolated from WT and MCK-Lpin1^−/− mice, normalized for protein content, and then blotted for p62 and LC3. The mitochondrial fraction from gastrocnemius and vastus tissues from MCK-Lpin1^−/− mice contained increased p62, LC3-I, and -II protein compared with WT littermates (Fig. 7B). Finally, skeletal muscle histology from gastrocnemius tissue revealed qualitatively increased p62, LC3, and BNIP3, another protein involved in autophagy, in the mitochondrial fraction (Fig. 7C). These findings are also consistent with impaired autophagic flux.

Reduced exercise tolerance and increased plasma creatine kinase in muscle-specific lipin 1 deficiency

We next evaluated whether the MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice exhibited signs of rhabdomyolysis paralleling the phenotype of humans with LPIN1 mutations. Previous work has linked febrile illness, prolonged fasting, and intensive exercise to acute rhabdomyolysis in patients with LPIN1 mutations (11, 12). Therefore, WT and lipin 1–deficient mice were kept unfed overnight, and then, an exhaustive exercise test was conducted. Although the young 3–4-mo-old MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice had increased endurance during an exercise bout to exhaustion (Fig. 8A, B), the older 1-yr-old mice had significantly reduced endurance (Fig. 8C, D) compared with their littermate WT controls. Plasma creatine kinase concentration was increased in both young and older MCK-Lpin1^−/− mice compared with WT controls after being unfed overnight and 3 h after the endurance exercise bout (Fig. 8B, D). During sedentary, fed conditions, those mice also had higher levels of plasma creatine kinase (Fig. 8D). Elevations of plasma creatine kinase in those basal conditions suggest a baseline level of myocyte damage with skeletal muscle–specific Lpin1 deficiency that is exacerbated with environmental stressors of being unfed and performing considerable physical activity. These findings parallel observations that fasting and/or exercise can provoke rhabdomyolysis in human patients with LPIN1 mutations.

Figure 8.

MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice have altered exercise endurance and increased exercise plasma creatine kinase when unfed. A, C) Exercise endurance time to exhaustion in 3–4-mo-old and 1-yr-old MCK-Lpin1^Δ115, respectively. B, D) Exercise endurance time to exhaustion in 3–4-mo-old and 1-yr-old MCK-Lpin1^−/−, respectively. Additionally, plasma creatine kinase is elevated in MCK-Lpin1^−/− mice compared with their WT controls. Sedentary indicates that plasma creatine kinase was measured under fed and rested conditions; 3-h post-ex indicates that plasma creatine kinase was measured after being unfed overnight, and 3 h after the time to exhaustion in an acute bout of treadmill exercise. Data are shown as means ± sem. *P < 0.05 (Student’s t test) for differences in MCK-Lpin1^Δ115 or MCK-Lpin1^−/− compared with respective WT littermate mice; n = 7–9.

DISCUSSION

Inborn errors in intermediary metabolism are among the most common causes of unexplained episodic rhabdomyolysis in children. Among those genetic mutations, LPIN1 deficiency has garnered attention with continued reports of otherwise healthy children being afflicted with acute and serious febrile bouts with muscle pain, hyperkalemia, and ventricular tachycardia (7–9, 11, 15). Efforts to develop mouse models of Lpin1 deficiency to understand the mechanism and to test potential treatments are imperative to unraveling the pathophysiology. Because lipin 1 is bifunctional and can affect metabolism by multiple modes of action, dissecting the pathogenic mechanisms of lipin 1 deficiency is more difficult. The present study used 2 muscle-specific models of Lpin1 deficiency in mice. The first model employed a mouse with hypomorphic lipin 1 lacking PAP activity, but with retained transcriptional coregulatory function (19, 25). The second model used a novel gene-deletion strategy that resulted in complete ablation of lipin 1 protein in muscle. Herein, we report that both lines of mice with muscle lipin 1 deficiency exhibited an active and progressive myopathic phenotype and that muscle-specific loss of lipin 1–mediated PAP activity is sufficient to induce muscle injury, which may provide new clues about pathogenic mechanisms.

Recent work by Zhang et al. (16) to dissect the mechanistic aspects of myocyte injury with lipin 1 deficiency used fld mice, which are deficient in lipin 1 in all tissues because of a spontaneous mutation in Lpin1. Those studies suggested that loss of lipin 1 led to muscle injury by alterations in signaling cascades that regulate autophagy, accumulation of dysfunctional and damaged mitochondria, and regenerating muscle fibers and suggested that accumulation of PA and depletion of DAG in fld muscle were important mediators of the phenotype (16, 20). Studies were also conducted in fld mice, complemented by transgenic reexpression of lipin 1 in skeletal muscle, to show a myocyte-specific rescue could be achieved. Nevertheless, the interpretation of the data was complicated by the lipodystrophic phenotype of fld mice. Indeed, although use of MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice in the present study confirmed the aberrant autophagy, accumulation of abnormal mitochondria, and accumulation of PA in muscle, the steady state DAG content of muscle was actually increased compared to WT littermate controls. Previous studies have shown that human LPIN1-deficient myocytes accumulate TAG normally in vitro (36), and patients with LPIN1 mutations have lipid accumulation in muscle and heart (10); a key phenotypic difference from the fld mice. Taken together, the present data and previous studies suggest that PA accumulation, rather than depletion of downstream lipids, is linked to myocyte injury. Interestingly, PA accumulation has been linked to potentially deleterious signaling events and cytotoxicity (21, 37–40), and it is possible that multiple intracellular signaling cascades are activated.

The unexpected finding that the product of lipin 1–mediated enzymatic activity, DAG, was increased in both models of skeletal muscle lipin 1 deficiency, whereas muscle TAG abundance showed no change, remains to be explained. Whether the increase in DAG is due to reduced DAG hydrolysis by DAG lipases, such as hormone-sensitive lipase (HSL), decreased phosphorylation by DAG kinase, reduced diacylglycerol acyltransferase activity, increased monoacylglycerol acyltransferase activity, or synthesized from myriad other phospholipids is not known (41–44). We also recently found that DAG was increased and TAG was unchanged in hepatic tissue of mice with liver-specific lipin 1 deficiency using the Δ115 model (25). In addition, increased DAG in skeletal muscle has been demonstrated to result in impaired insulin sensitivity. However, compared with WT littermate controls, MCK-Lpin1^Δ115 mice had normal glucose tolerance and isolated soleus and extensor digitorum longus muscles from MCK-Lpin1^Δ115 animals exhibited normal insulin-stimulated glucose uptake (unpublished results). Consistent with that, previous work by Michot et al. (7) has suggested that patients with LPIN1 mutations also exhibit normal glucose tolerance and hemoglobin A1c concentrations.

The current model, based on this study and previous work by Zhang et al. (16), suggests that accumulation of mitochondria with abnormal morphology in lipin 1–deficient mice is not a consequence of increased biogenesis but, rather, is explained by impairments in autophagic degradation of mitochondria (mitophagy). Indeed, pathway analyses revealed reduced expression of genes encoding enzymes involved in mitochondrial metabolism as well as the master regulator of mitochondrial biogenesis, PGC-1α, but also accumulation of mitochondrial protein, mitochondrial DNA, and cardiolipin and increased mitochondrial staining of tissue. Furthermore, increased abundance of autophagic proteins (p62, LC3, and BNIP3) in the total cellular protein and mitochondrial fractions of lipin 1–deficient mice are consistent with impaired autophagic and mitophagic flux. It is possible that accumulation of dysfunctional mitochondria, as demonstrated by ultrastructural abnormalities and reduced respiratory control ratio, have a causative role in the myocyte dropout and myopathic phenotype observed in the mice. Strategies to enhance autophagic flux and overcome that deficit may have utility in treating this and other related disorders.

Previous work has suggested that lysosomal membrane PA content may be an important regulator of lysosome permeability, function, and dysfunction (37–39). Our work in a study published very recently (45), and another publication (16) detected a general defect in skeletal muscle autophagy in mice lacking lipin 1. In Alshudukhi et al. (45), LC3-eGFP transgenic mice were used to confirm impairments in autophagic flux and to show that loss of lipin 1 impaired Bnip3-regulated mitophagy by affecting its interaction with LC3. This is a critical step in recruiting mitochondria destined for autophagy to nascent autophagosomes. Abnormalities in lysosomal function and autophagy have been linked to several forms of muscle injury and myopathy, including muscular dystrophy (46–49). How deficits in mitophagy or macroautophagy in general lead to muscle injury and rhabdomyolysis remains to be determined. Some patients with LPIN1 deficiency exhibit mitochondrial aggregation (7), and thus, a mitochondrial phenotype may also be present in humans as well.

Rhabdomyolysis associated with lipin 1 deficiency is often induced by trauma, illness, or environmental stressors and then accompanied by extreme elevations of creatine kinase in the blood (12). The myopathy in MCK-Lpin1^Δ115 and MCK-Lpin1^−/− mice, although dramatic, appears to differ from severe, episodic rhabdomyolysis and the predominant lack of muscle pathology in children with LPIN1 mutations (7). The dissimilarities could be due to basic differences in muscle metabolism, physiology, or lipin function across species. However, other data suggest that some patients with LPIN1 mutations exhibit a chronic myopathy that seems to affect the oldest patients, suggesting that this phenotype may manifest with age (7). This parallels our observations that the exercise intolerance of muscle-specific lipin 1–deficient mice was apparent in 1-yr-old mice and suggests that careful follow-up examinations of older patients with LPIN1 mutations could reveal a higher incidence of myalgia, myopathy, or exercise intolerance.

In a recent study of children and young adults with LPIN1 mutations undergoing a maximal exercise tests, maximal oxygen consumption tended to be lower in subjects with LPIN1 mutations compared with healthy control subjects (12). In most of the subjects in that study, creatine kinase levels at 8, 16, and 24 h after exercise were within range of their resting, pre-exercise levels, the same resting levels as that found in healthy children. However, 1 subject who exhibited a fever had creatine kinase levels increasing 2-fold, 5-fold, and 10-fold at 8, 16, and 24 h after exercise. Additionally, other reports describe patients to be exercise intolerant between rhabdomyolytic episodes (15). It has also been reported that a high percentage of LPIN1 heterozygous carriers experience exercise-induced myalgia (7). This observation resembles the present mouse study after mice were unfed and engaged in exhaustive exercise intervention, which appears to trigger rhabdomyolysis. Future work will be needed to confirm those observations and to examine the efficacy of preventative measures to alleviate muscle injury in people affected by partial or complete LPIN1 deficiency.

In summary, our data are consistent with a role for skeletal myocyte-autonomous defects in lipin 1–mediated PAP activity in the development of skeletal muscle injury, which is likely due to autophagic and mitochondrial perturbations leading to myocyte death and development of myopathy. The muscle-specific knockout mouse models reported herein will be useful in further dissection of the pathogenic mechanisms and for testing potential therapeutic approaches.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

BCA

bicinchoninic acid assay

BSA

bovine serum albumin

CCD

charge-coupled device

CL

cardiolipin

DAG

diacylglycerol

eWAT

epididymal white adipose tissue

FCCP

carbonyl cyanide‑p‑trifluoromethoxyphenylhydrazone

FDR

false-discovery rate

fld

fatty liver dystrophic

G3P

glycerol-3-phosphate

H&E

hematoxylin and eosin

LC-MS

liquid chromatography–mass spectrometry

LPA

lysophosphatidic acid

LPIN1

lipin 1

MCK

muscle creatine kinase

mtDNA

mitochondrial DNA

PA

phosphatidic acid

PAP

phosphatidic acid phosphohydrolase

PG

phosphatidylglycerol

TAG

triacylglycerol

TBST

Tris-buffered saline plus Tween 20

WT

wild type

AUTHOR CONTRIBUTIONS

G. G. Schweitzer, S. L. Collier, Z. Chen, K. S. McCommis, S. K. Pittman, J. Yoshino, S. J. Matkovich, F.-F. Hsu, J. M. Eaton, T. E. Harris, C. C. Weihl., and B. N. Finck designed and performed the experiments; G. G. Schweitzer, J. Yoshino, S. J. Matkovich, F.-F. Hsu, T. E. Harris, C. C. Weihl, and B. N. Finck analyzed the data; R. Chrast and C. C. Weihl provided mice and reagent; and G. G. Schweitzer and B. N. Finck wrote and edited the manuscript.