1 Introduction
Obesity and its complications including coronary heart disease (CHD), non-alcoholic fatty liver disorder (NAFLD), diabetes mellitus, and certain types of cancers are crucial risk factors threatening human health [
1,
2]. According to a recent survey study, more than 0.21 billion children and 1.31 billion adults worldwide have an underlying overweight condition [
3]. The increasing prevalence of obesity and related diseases has become one of the most serious global health problems [
4]. At present, nearly 400 genes represented by
FTO genes that may be related to obesity have been identified [
5,
6]. Among these genes, triglyceride (TG) and its metabolism-related genes such as
RTN3,
Pex11a, and
CTRP2 are linked to the occurrence of obesity and related complication [
7–
9]. Anabolism and/or catabolism in TG play a critical role in hypertriglyceridemia (HTG), obesity, and related complications. Previous studies have shown that sterol regulatory element binding proteins (SREBPs) and reactive oxygen species (ROS) are responsible for the anabolism of TG [
10,
11]. Conversely, peroxisome proliferator-activated receptor alpha (PPARα) and mitochondrial damage are related to TG catabolism [
12,
13].
The lipin protein family consists of Lipin1, Lipin2, and Lipin3 in mammalian systems and has two highly conserved domains, namely, amino-terminal (N-LIP) and carboxy-terminal (C-LIP) lipin domains [
14]. The C-LIP domain is responsible for regulating phosphatidate phosphatase enzyme activity and transcriptional co-activator function, whereas the N-LIP domain is important for its catalytic activity, nuclear localization, and binding to protein phosphatase-1cγ [
15,
16]. Previous studies have shown that the lipin protein family plays a crucial role in lipid metabolism, inflammatory responses, and cell differentiation [
17–
20]. Many studies have focused on Lipin1, whereas the roles of Lipin2 and Lipin3 are less understood.
In the present study, we identified a novel heterozygous mutation (NM_001301860: p.1835A>T/p.D612V) of Lipin3 (NM_001301860) in a large family with HTG and obesity. Functional studies revealed that the mutation (p.D612V) of Lipin3 altered the half-life and stability of Lipin3 protein. We then generated Lipin3 knockout (KO) mice to explore the pathophysiological roles of Lipin3 in TG metabolism. Interestingly, Lipin3 heterozygous KO (Lipin3-heKO) mice showed obesity phenotypes and high levels of TG in the liver, fat tissue, and plasma. Lipin3-heKO mice fed a high-fat diet (HFD) for 3 months also presented more serious obesity and HTG. Mechanistic studies further revealed that the haploinsufficiency of Lipin3 in primary hepatocytes can promote the expression of Lipin1, an important gene that can regulate balance between TG anabolism and catabolism. In Lipin3-heKO primary hepatocytes, the expression of Lipin1 in cytoplasm was found to increase, which may further contribute to TG anabolism. Simultaneously, the expression of Lipin1 in nucleoplasm was reduced, which may decrease the expression of downstream targets, such as PPARα and PPARγ coactivator 1 alpha (PGC1α). Subsequently, mitochondrial dysfunction increased and TG catabolism decreased. These phenomena may have finally led to HTG, obesity, and NAFLD.
2 Materials and methods
2.1 Pedigrees and participators
A 17-person, five-generation Han Chinese pedigree with HTG and/or CHD was recruited at the Third Xiangya Hospital of Central South University, Changsha, China (Fig.1). Clinical data and peripheral blood samples were obtained from 12 members, including six affected (II-2, II-7, II-8, III-3, III-4, and IV-2) and six unaffected members. Two hundred unrelated local healthy people were also enrolled to serve as normal controls.
2.2 Whole-exome sequencing and Sanger sequencing
Genomic DNA was extracted from the peripheral blood lymphocytes of all family members by using a DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instruction. We then selected II-8, III-3, and III-2 to perform whole-exome sequencing. The central part of the whole-exome sequence was provided by the Novogene Bioinformatics Institute (Beijing, China). The exomes were captured using Agilent SureSelect Human All Exon V6 kits, and high-throughput sequencing was performed using Illumina HiSeq X-10. The necessary bioinformatics analyses including reads, mapping, variant detection, filtering, and annotation were also provided by Novogene Bioinformatics Institute as previously described [
21]. The strategies of data filtering are shown in Fig.1 and 1C.
All filtered mutations of the family members were validated by Sanger sequencing. Primer pairs (the primer sequences can be provided upon request) were designed using Primer 5. Sequences of the polymerase chain reaction products were determined using an ABI 3100 Genetic Analyzer (ABI, Foster City, CA, USA).
2.3 Bioinformatic analysis
Polyphen-2, SIFT, and MutationTaster programs were used to predict the effect of mutation on protein function. Swiss-Model software was used to identify the function of the mutation. Local hydrophobicity was predicted by ProtScale. Conservation analysis was performed with ConSurf Server software.
2.4 Mouse strains, cell lines, and key reagents
Lipin3-heKO mice, with exon 3–7 deleted in the C57BL/6J background, were generated by Cyagen company (Suzhou, China). Genotyping was performed as described previously [
7]. Wild-type (WT) control mice were bred from female and male
Lipin3-heKO mice and raised in the Department of Zoology, Central South University. HFD mice were treated with feedstuff comprising 60% fat, 20% protein, and 20% carbohydrate, which was purchased from TROPHIC Animal Feed High-Tech Co., Ltd., China.
Primary hepatocytes were isolated from mouse liver tissues as follows. Fresh liver tissues were obtained from newborn mice. Primary mice hepatocytes were isolated from the respective liver tissues through collagenase/hyaluronidase digestion (STEMCELL Technologies, Vancouver, Canada). Cells were cultured in F-Y medium comprising complete DMEM, F12 nutrient mixture, hydrocortisone/epidermal growth factor mixture, insulin, amphotericin B, gentamicin, cholera toxin, ROCK inhibitor Y-27632, and HEPES in a 37 °C incubator with 5% CO2.
L02 cell line was purchased from the Cell Bank of the Shanghai Institutes for Biological Sciences (Shanghai, China). It was maintained at 37 °C in a humidified, 5% CO
2-controlled atmosphere in DMEM supplemented with 10% fetal bovine serum, 50 IU/mL penicillin, 50 mmol/L streptomycin, and glutamine [
22].
Lipin3 antibody was generated in our laboratory. The antibodies of Flag (Cat No. 80010-1-RR) and β-actin (Cat No. 20536-1-AP) were purchased from Proteintech Group, Inc. The antibodies of Mfn2 (#11925S) and OPA1 (#67589) were purchased from Cell Signaling Technology. The antibodies of Fis1 (sc-376447), Lipin1 (sc-376874), and PGC1α (sc-518025) were purchased from Santa Cruz Biotechnology. PPARα (ab24509) was purchased from Abcam. Hematoxylin–eosin (HE) Staining Kit (G1120), Red O Staining Kit (G1260), ATP (BC0300), Reactive Oxygen Species (ROS) Assay Kit (CA1410), Mitochondrial Membrane Potential Assay Kit with JC-1 (M8650), and Nuclear Protein Extraction Kit (R0050) were purchased from Beijing Solarbio Science & Technology Co., Ltd. BCA Protein Assays and Analysis kit (23227) and OxPhos Rodent WB Antibody Cocktail (Cat # 45-8099) were purchased from Thermo Fisher Scientific.
2.5 Plasmid construction and transfection
WT Lipin3 CDS with a C-terminal Flag-tag in pEnter was designed. The D612V-Lipin3 missense mutation was engineered into the vector above by using a Takara MutanBEST Kit (Takara Bio, Otsu, Shiga, Japan). L02 cells were transiently transfected with Lipin3-Flag-pEnter (WT and/or mutation) by using a Lipofectamine™ 2000 CD Transfection Reagent (Thermo Fisher Scientific) following the manufacturer’s instructions.
2.6 HE staining and oil red O staining
Paraformaldehyde-fixed tissue was embedded in paraffin and sliced into 6 μm sections. The sections were stained with a HE staining kit and examined by routine light microscopy. Paraformaldehyde-fixed tissue was embedded in OCT and sliced into 12 μm sections. The sections were stained with a Red O Staining Kit and examined by routine light microscopy. Staining was performed according to established protocols. The liver sections were scored on the basis of the criteria of the NAFLD activity score (NAS).
2.7 Western blot (WB) analyses
For WB analysis, tissues or cells were homogenized on ice in 1% 3-((3-cholamidopropyl) dimethylammonio)-1-propanesulfonate extraction buffer containing complete protease inhibitors (Roche Bioscience, No. 04693159001) and 0.1 mmol/L Na3VO4 to inhibit phosphatase. The homogenates were rotated for 30 min at 4 °C to ensure the extraction of membrane proteins. After centrifugation at 15 000× g for 120 min, the supernatants were collected, and protein concentrations were measured with bicinchoninic acid protein assay reagent. Nuclear proteins were isolated with a nuclear protein extraction kit according to the manufacturer’s instructions. Equal amounts of protein lysates were resolved by 4% to 12% Bis-Tris NuPAGE gel electrophoresis followed by standard WB with the antibodies specified above. The chemiluminescence signals were scanned, and the integrated density values were calculated with a chemiluminescence imaging system (Alpha Innotech).
2.8 Transmission electron microscopy (TEM)
TEM samples were prepared as previously described with modifications [
23]. After dissection and fixation, each sample was sectioned to generate 70 nm-thick sections with an ultramicrotome (EM UC7; Leica Microsystems) and then stained with uranyl acetate and lead citrate. Images were acquired by TEM (H-7650; Hitachi). Five samples were prepared for each experimental condition.
2.9 ATP assay, ROS assay, and JC-1 staining
Adenosine triphosphate (ATP) assays and ROS assays were performed using appropriate assay kits. Phosphomolybdic acid colorimetry was used to detect the ATP levels. ROS generation was determined by fluorometric analysis by using 2,7-dichlorofluorescein diacetate.
Mitochondrial membrane potential (ΔΨm) was detected using JC-1 according to the manufacturer’s instructions. Visualized images were acquired using a fluorescence microscope.
2.10 Statistical analysis
Data were statistically analyzed with Graph-Pad Prism 5 (GraphPad Software) and plotted with AI Illustrator (Adobe). The results represented the mean ± SEM from at least three independent experiments, as indicated in the figure legends. All data are presented as the mean ± standard error of the mean. Data were analyzed using paired Student’s t-test. Differences were considered statistically significant at P < 0.05, with significance indicated in the figures as *P < 0.05, **P < 0.01, and ***P < 0.001 (ns represents no significant difference).
3 Results
3.1 Whole-exome sequencing identified a novel mutation of Lipin3 in a family with obesity and HTG
The proband, a 49-year-old woman from Hunan province in central south China, had a history of nearly 10 years of atherosclerosis (AS) in her left common iliac artery and had high plasma levels of HTG (TG = 11.2 mmol/L). The proband’s body mass index was 24.7 kg/m2, with no other disease detected. Family-history investigation revealed that four members died from CHD, and another six family members also suffered from HTG and/or AS. The plasma lipid-testing data for the family members are presented in Tab.1. One of the proband’s uncles (II-7) also died from CHD two months after our investigation. The family was an isolated HTG and/or CHD group with an autosomal dominant pattern (Fig.1).
Whole-exome sequencing yielded 9.29 Gb of data with 99.4% coverage of target regions, and 98.6% of the target regions were covered over 10×. After alignment and single-nucleotide variant calling, 69 427 variants were identified in the proband. Data filtering was performed as shown in Fig.1 and 1C. After filtering the data, only 15 variants were included (Table S1). After co-segregation analysis and bioinformatics prediction, only the novel mutation (p.1835A>T/p.D612V) of Lipin3 (NM_001301860) was observed in all affected patients but not in healthy individuals (Fig.1). The novel mutation was also absent in our 200 local control cohorts, as well as public databases such as 1000G, ExAC, and genomAD. Hence, we may have identified a new HTG disease-causing gene in a large Chinese family via whole-exome sequencing and Sanger sequencing.
3.2 The novel mutation (p.D612V) may disrupt the stability of Lipin3 protein
The novel mutation, p.D612V, which resulted in the substitution of aspartic acid by valine at code 612 in exon 15 of Lipin3, was located in the highly evolutionarily conserved C-LIP (Fig.2). Swiss-Model online software revealed that the D612V mutation increased the hydrophobic surface area of Lipin3 protein and may induce changes in surface charge and polarity, as marked by arrows in the figure (Fig.2). We then constructed WT and mutant Lipin3 plasmids and transfected them into L02 cells. WB experiments demonstrated that the expression of both total Lipin3 and transfected Lipin3 (Flag) in the Lipin3 mutated group decreased more overtly than that in the WT group (Fig.2), whereas the mRNA levels of both groups were similar (data not shown). We then used cyclohexamide (CHX) to treat the cells after transfection with the plasmids for 36 h, and WB analysis revealed that the degradation rate of transfected Lipin3 (Flag) in the Lipin3 mutated group was faster than that in the WT group (Fig.2 and 2E). Site-mutation functional studies showed that the variant (p.1835A>T/p.D612V) of Lipin3 altered the half-life and stability of Lipin3 protein.
3.3 Haploinsufficiency of Lipin3 led to obesity and HTG in mice
We then generated the Lipin3 KO mice with the help of Cyagen (Suzhou, China) company (Fig.3). Interestingly, Lipin3-heKO mice exhibited an overtly larger body size than that of WT controls at 8 months of age with standard chow (Fig.3). The average body weight of Lipin3-heKO mice (n = 5) was approximately 27.2% greater than that of their WT control littermates (n = 5) (Fig.3). However, food intake showed no difference between these two groups of mice (Fig. S1A). Epididymal adipose tissues that accumulated in Lipin3-heKO mice were approximately 222.5% greater than those in WT littermates (Fig.3), and the liver weight of Lipin3-heKO mice was also approximately 26.1% greater than that in WT littermates (Fig.3). Plasma TG levels in Lipin3-heKO mice obviously increased compared with those in WT control littermates after 8 months of standard chow. The level of plasma TG in Lipin3-heKO mice increased to approximately 56.1% compared with that in WT controls (Fig.3), whereas the plasma total cholesterol, high-density lipoprotein, or low-density lipoprotein levels did not change (Fig. S1B–S1D). No difference existed in glucose, glucose tolerance test, and insulin tolerance test between WT and Lipin3-heKO mice (Fig. S1E–S1G). The Lipin3-hoKO mice in our study did not show any problems in body weights, lipids, and livers compared with WT mice (Fig. S2A–S2C). HE staining and Oil red O staining showed that the lipid droplets in adipocytes and hepatocytes of Lipin3-heKO mice were larger than those in WT littermates (Fig.3 and 3H). In primary hepatocytes, the TG levels were also much higher in the Lipin3-heKO group than in the WT (Fig.3). These observations in Lipin3-heKO mice suggested that haploinsufficiency of Lipin3 may lead to HTG and obesity in mice, consistent with our clinical genetic studies.
3.4 Lipin3-heKO mice exhibited a severe overweight phenotype underlying HFD feeding
To further confirm the phenotypes of HTG and obesity in Lipin3-heKO mice, we selected Lipin3-heKO mice and WT controls at 2 months of age, which did not show body-size differences, for HFD for 12 weeks. We found that HFD-Lipin3-heKO mice had a larger body size than that of HFD-WT mice (Fig.4). The body weight of HFD-Lipin3-heKO mice was approximately 21.9% greater than that of HFD-WT controls (Fig.4). The epididymal adipose tissues and liver weight of HFD-Lipin3-heKO mice were also greater than those of HFD-WT controls (Fig.4 and 4D). Plasma TG detection further revealed that the levels of plasma TG in HFD-Lipin3-heKO mice were greater than those in HFD-WT controls (Fig.4). HE staining revealed that the lipid droplet size in the adipocytes of HFD-Lipin3-heKO mice was larger than that of HFD-WT controls (Fig.4). Meanwhile, HE staining of liver tissues suggested that HFD-Lipin3-heKO mice had significantly aggravated histological parameters, including steatosis, ballooning, and inflammation, on the basis of the criteria of the NAS, compared with the HFD-WT mice group (Fig.4). These data suggested that under the pressure of HFD, Lipin3-heKO mice presented more severe overweight and HTG phenotypes than WT controls, which further confirmed that haploinsufficiency of Lipin3 may lead to HTG and obesity.
3.5 Haploinsufficiency of Lipin3 may disrupt the nucleocytoplasmic localization of Lipin1 and induce mitochondrial dysfunction
Previous studies on lipin proteins and lipid metabolism indicated that Lipin3 and Lipin1 may present a compensatory mechanism for TG metabolism in adipose tissues and liver tissues [
24]. Lipin1 may play the major phosphatidic acid phosphohydrolase enzyme activity [
24,
25]. Hence, we detected Lipin1 expression in primary hepatocytes and found that it was extremely increased in
Lipin3-heKO primary hepatocytes, whereas Lipin2 showed no difference (Fig.5 and S3). Some studies have also revealed that the Lipin1 can play dual roles in TG metabolism [
20,
26], i.e., Lipin1 in cytoplasm may regulate TG anabolism, and the Lipin1 in nucleoplasm may regulate mitochondrial structure and function via the PPARα-PGC1α pathway [
27–
29]. WB analysis of Lipin1 isolated from the cytoplasm and the nucleoplasm suggested that Lipin1 distribution in nucleoplasm overtly decreased in
Lipin3-heKO compared with that in WT primary hepatocytes, and Lipin1 distribution in cytoplasm obviously increased in
Lipin3-heKO compared with that in WT primary hepatocytes (Fig.5). According to previous studies, decreased Lipin1 in nucleoplasm may reduce the expression of the PPARα-PGC1α pathway, as confirmed by WB results in
Lipin3-heKO primary hepatocytes in this study (Fig.5). Subsequently, we detected the mitochondrial structure and function in
Lipin3-heKO primary hepatocytes, and found that the following results. First, the expression of mitochondrial electronic respiratory chain-related proteins decreased in
Lipin3-heKO primary-cultured hepatocytes (Fig.5). Second, JC-1 staining of primary hepatocytes also indicated that the mitochondrial membrane potential of
Lipin3-heKO cells was disrupted (Fig.5). Third, TEM results showed that the morphology of mitochondria was disrupted in
Lipin3-heKO primary-cultured hepatocytes (Fig.5). Fourth, proteins such as fission-1 (FIS1) and optic atrophy 1 (OPA1) [
30,
31], which could control mitochondrial morphology, were also altered in
Lipin3-heKO primary hepatocytes (Fig.5). Fifth, ATP levels were reduced in
Lipin3-heKO primary hepatocytes (Fig.5), and the levels of ROS in
Lipin3-heKO primary hepatocytes increased dramatically (Fig.5).
Collectively, our mechanistic studies revealed that the haploinsufficiency of Lipin3 in hepatocytes can promote Lipin1 expression. Lipin1 overexpression also showed high levels in cytoplasm and low levels in nucleoplasm. The high levels of Lipin1 in cytoplasm may further contribute to TG anabolism. The low levels of Lipin1 in nucleoplasm may also reduce the expression of the PPARα-PGC1α pathway, further leading to mitochondrial dysfunction and reducing the ability in TG catabolism. The effects of increased TG anabolism and reduced TG catabolism may finally lead to HTG, obesity, and NAFLD (Fig.6).
4 Discussion
We identified a novel mutation (p.1835A>T/p.D612V) of
Lipin3 in a Chinese family with HTG and obesity through whole-exome sequencing. According to ACMG guidelines, this mutation is likely pathogenic (PS3+PM2+PP1+PP3). Functional studies have revealed that this mutation may disrupt the stability of Lipin3 and increase its degradation. Previous studies have revealed that Lipin1 and Lipin3 together determine adiposity
in vivo. Another study has detected three
Lipin3 mutations in 46 patients with moderate rhabdomyolysis [
32]. Rhabdomyolysis is a muscle disease caused by direct or indirect injury to the muscle and is closely related to lipid metabolism [
33]. In the present study, we established for the first time a relationship between
Lipin3 mutations and HTG.
Gene-expression studies have found that each
Lipin has a unique tissue-expression pattern, but the overlapping expression of more than one
Lipin gene occurs in many tissues. This finding indicated that Lipin family proteins may play different roles in addition to the same biochemical reaction [
34]. For example, Lipin1 deficiency can cause sarcoplasmic reticulum stress and chaperone-responsive myopathy [
35]. Deficiency of Lipin2 can activate NF-κB signaling, which may be associated with inflammation and immune reactions [
36]. In lipid metabolism, fatty liver dystrophy mice carrying mutations within the
Lipin1 gene display life-long deficiency in adipogenesis, insulin resistance, neonatal hepatosteatosis, and HTG, as well as increased AS susceptibility through the regulation of SREBP1 and ROS-dependent SREBP2 activation [
17]. Lipin1 has also been regarded as a key integrator of hormonal signals in the liver in diabetic dyslipidemia [
37]. Lipin2–3 deficiency can lead to increased plasma TG levels and decreased weight by activating the mTORC1 pathway and the regulation of chylomicron synthesis [
38]. Interestingly, Lipin1 and Lipin3 together reportedly influence adiposity
in vivo; however, Lipin3-deficient mice exhibits only subtle metabolic abnormalities, which may be caused by the compensation of Lipin1 when Lipin3 is completely deficient [
24]. At present, no studies have focused on the effects of only Lipin3 deficiency.
Mitochondria plays a central role in obesity and its complications because they control cell metabolism and regulate important processes such as ATP production, lipid β-oxidation, oxidative stress, and inflammation. Mitochondrial dysfunction leads to decreased substrate oxidation, particularly fatty acids, resulting in lipid accumulation [
12,
39]. Decreased substrate oxidation can also affect electron flow through the electron-transport chain, triggering electron leakage toward oxygen and ROS formation [
40]. ROS overproduction can further contribute to obesity, NAFLD, and other complication [
41]. Our study showed that the haploinsufficiency of Lipin3 in hepatocytes may lead to reduced Lipin1 in nucleoplasm, which further decreased the expression of the PPARα-PGC1α pathway and led to mitochondrial dysfunction. This work was also the first to establish a relationship between Lipin3 and mitochondria.
Previous research on adipose tissue has indicated that Lipin3 and Lipin1 may present a compensatory mechanism for TG metabolism. Csaki et al. reported that
Lipin3-hoKO mice do not show any problems in body weights, lipids, and livers compared with WT mice. The same results were also found in our
Lipin3-hoKO mice. However, in our study, the
Lipin3 heterozygous mutation carriers and the
Lipin3-heKO mice showed obvious HTG and obesity. The reason may be the balance disruption among Lipin family members. We further found that the expression of Lipin1 increased in
Lipin3-heKO hepatocyte [
24]. We detected as well that the distribution of Lipin1 in
Lipin3-heKO hepatocyte decreased in nucleoplasm and increased cytoplasm compared with that in WT hepatocyte. Previous studies on alcoholic fatty liver disease have shown that increased Lipin1 in cytoplasm may play the crucial role of phosphatidic acid phosphohydrolase enzyme activity and contribute to TG anabolism. Meanwhile, decreased Lipin1 in nucleoplasm may directly reduce the function of co-activating with PGC-1α and PPAR, further decreasing the expression of mitochondrial genes involved in fatty acid oxidation [
16,
26,
42]. The discoveries of Lipin1 in alcoholic fatty liver disease were similar to our studies between Lipin3 and TG accumulation. Hence, our study further confirmed the function of Lipin1 in cytoplasm and nucleoplasm.
Lipin1 can regulate the expression of PPARγ during adipogenesis [
43]. In the current work, we also detected the levels of activated PPARγ in the liver tissues of WT and
Lipin3-heKO mice, but no changes were observed. However, we further detected PPARα and found that PPARα expression was reduced in liver tissues of
Lipin3-heKO mice. Previous studies have suggested that PPARα activation can contribute to oxidative energy production [
44]. PPARα exhibits marked anti-inflammatory capacities and has become an interesting drug target in various metabolic disorders, including obesity and NAFLD [
45]. Here, we found that PPARα activation decreased with decreased Lipin3, indicating reduced Lipin3-induced mitochondrial damage and obesity via the PPARα-PGC1α pathway. We also found for the first time that Lipin3 can regulate PPARα activation.
In summary, our study suggested that the haploinsufficiency of Lipin3 in hepatocytes may disrupt the expression and nucleocytoplasmic localization of Lipin1. Subsequently, the balance between TG anabolism and catabolism may be broken by regulating Lipin1 phosphatidic acid phosphohydrolase enzyme activity and the Lipin1-PPARα-PGC1α pathway, further leading to HTG, NAFLD, and obesity. Our findings identified a new disease-causing gene in TG accumulation and revealed for the first time the function of Lipin3 in regulating mitochondrial biological homeostasis and PPARα.
PDF
(4480KB)
