Seipin expression in hepatocytes impairs the assembly of VLDLs and exacerbates steatohepatitis
Qianqian Dong, Yidan Ma, Xin Chen, Xiaowei Wang, Ziwei Liu, Chenxi Liang, Liwen Qiu, Jinye Tang, Jin Wu, Yang Liu, Xiaoqin Wu, Yaru Zhou, Mingming Gao, Hongyuan Yang
TL;DR
Overexpression of seipin in liver cells disrupts VLDL assembly and worsens liver disease, suggesting a link between lipid droplet formation and lipid transport.
Contribution
This study reveals a novel role of seipin in impairing VLDL lipidation and exacerbating steatohepatitis through altered lipid droplet dynamics.
Findings
Liver-specific seipin overexpression increases cytoplasmic lipid droplets and reduces plasma lipid levels.
Seipin overexpression impairs VLDL lipidation and exacerbates liver inflammation and fibrosis on a high-fat diet.
Seipin redirects lipid storage to the cytoplasm, disrupting endoplasmic reticulum lipid droplet biogenesis and VLDL assembly.
Abstract
VLDLs are crucial for maintaining liver and whole-body lipid homeostasis. Limited knowledge exists regarding the lipidation process of VLDL. Endoplasmic reticulum (ER) luminal lipid droplets (LLDs) have been suggested to provide lipids for VLDL lipidation and maturation. Seipin, an integral membrane protein of the ER, plays key roles in the formation of cytoplasmic LDs (CLDs) and adipocyte differentiation. Surprisingly, seipin is hardly detectable in hepatocytes. Given the critical contribution of seipin in forming CLDs, we hypothesize that the absence of seipin in hepatocytes might ensure the proper formation of LLDs and the lipidation and assembly of VLDLs. To explore the functional interactions between CLDs, LLDs, and VLDLs, we generated liver-specific human seipin (hSeipin) overexpression (adeno-associated virus [AAV]-hSeipin) mice using AAV. We examined hepatic lipid accumulation,…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsLipid Membrane Structure and Behavior · Lipid metabolism and biosynthesis · Fatty Acid Research and Health
Lipid droplets (LDs) are dynamic organelles that play critical roles in cellular lipid storage and metabolism (1). LDs arise from the endoplasmic reticulum (ER) in eukaryotic cells, and each LD comprises a neutral lipid core wrapped by a monolayer of amphipathic lipids. Based on their subcellular localization, three types of LDs have been described: cytoplasmic LD (CLD), nuclear LD, and ER luminal LD (LLD) (1). Among these, CLDs are best characterized, and a growing number of recent studies have also shed light on the biogenesis of nuclear LDs (2, 3). By contrast, little is known about how LLDs are generated and how they grow in the lumen of the ER (1). Notably, LLDs are believed to interact with nascent VLDL particles to promote their maturation (1, 4).
Seipin and its orthologs are well-established regulators of the biogenesis and expansion of CLDs in eukaryotic cells (5, 6, 7, 8, 9, 10, 11, 12, 13, 14). Null mutations of seipin also cause congenital generalized lipodystrophy (also known as Berardinelli-Seip congenital lipodystrophy [BSCL]), an autosomal recessive disorder characterized by a near-total loss of adipose tissue, severe insulin resistance, and fatty liver (15, 16, 17). Therefore, seipin is a unique protein that can regulate lipid storage at both systemic (adipogenesis) and cellular (LD biogenesis and expansion) levels (18). As an integral membrane protein of the ER, seipin is conserved from yeast to man (6, 7, 19). Structural studies have demonstrated that yeast, fly, and human seipin (hSeipin) exist as homo-oligomers (undecamer for hSeipin) and that the evolutionarily conserved luminal domain forms an eight-stranded β sandwich fold, which resembles lipid-binding domains, such as that of Niemann-Pick C2 (20, 21). The oligomerization state of hSeipin is essential to its function (20). Recent biochemical evidence suggests that seipin and its interacting proteins may directly facilitate phase separation of neutral lipids to initiate LD biogenesis and may also help define the ER domains/sites where LDs originate (12, 13, 14). Notably, adipogenin was very recently identified to interact with seipin and disrupt seipin's role in the initiation of nascent LDs (22, 23).
Alternatively, seipin may regulate LD biogenesis and growth through controlling the synthesis and/or partitioning of phospholipids (18). Seipin deficiency has been shown to accumulate certain phospholipid species, such as phosphatidic acid (PA) (24, 25, 26, 27), which are strong PPARγ antagonists (28, 29). The accumulation of PA, a negatively charged conical lipid, may impact LD formation by altering the tension and curvature of the ER membrane. PA may also block PPARγ activity and cause lipodystrophy (19, 28, 29, 30, 31, 32). Our previous results suggested that seipin may function to inhibit the activity of ER-localized glycerol-3 phosphate acyl transferases (GPATs) (30), whose activation in seipin deficiency may lead to accumulation of PA, thereby impacting LD growth and adipogenesis. Consistently, depleting GPAT3 partially rescued the differentiation defects of seipin-deficient preadipocytes in vitro and in vivo (30, 33). Seipin can also bind 1-palmitoyl-2-oleoyl phosphatidic acid in in vitro lipid binding assays (20). Moreover, adipogenin binding to seipin freed GPAT3/4 from seipin inhibition (22, 23). Recent studies have also implicated seipin in autophagy, the biogenesis of peroxisomes, the regulation of calcium homeostasis, and sphingolipid synthesis (34, 35, 36, 37, 38, 39). Together, these results highlight seipin's functions beyond LD biogenesis.
Distinct from other cell types where CLDs dominate, hepatocytes and enterocytes can generate both CLDs in the cytoplasm and LLDs in the ER lumen (1). LLDs are believed to interact with nascent lipoprotein particles to facilitate the maturation and secretion of lipoproteins (4). Exactly how LLDs are formed remains largely unknown. Moreover, no mechanistic connection between the biogenesis of CLDs and LLDs has been uncovered (1). As one of the most relevant proteins in LD biogenesis, seipin expression is surprisingly extremely low in murine hepatocytes (40). Given the presence of both CLDs and LLDs in hepatocytes, we hypothesize that the presence of seipin in hepatocytes may interfere with the generation of LLDs, thereby disrupting the assembly of lipoprotein particles. To examine this hypothesis, we overexpressed hSeipin in the mouse liver. Our results indicate that enhanced seipin function in the mouse liver accumulated CLDs and disrupted VLDL assembly. Our results thus provided the first molecular link between the biogenesis of CLDs and that of LLDs, as well as VLDL secretion.
Materials and Methods
Animals
C57BL/6J mice were purchased from Hebei INVIVO Biotech Co, Ltd (Shijiazhuang, China). All animals were housed under barrier system conditions in the Animal Center of Hebei Medical University, with controlled environmental parameters: room temperature maintained between 23 and 29°C, relative humidity between 40% and 70%, and a 12-h light/dark cycle. Mice had free access to food and water. The experimental protocols were approved by the Laboratory Animal Ethical and Welfare Committee of Hebei Medical University (IACUC-Hebmu-P 2023031).
The high-fat diet (HFD) consisted of 0.5% cholesterol, 20% fat, with the following composition: 5 g of cholesterol, 200 g of lard, and 795 g of powdered diet. The sugar water provided contained 23.1 g of fructose and 18.9 g of glucose per liter (41, 42).
Adeno-associated virus 8 construction and injection
The adeno-associated virus 8 (AAV8) delivery system was used to overexpress seipin in mouse livers. Recombinant AAV8 vectors carrying hSeipin or an empty vector with a thyroxine-binding globulin (TBG) promoter (Serpin A7, a liver-specific protein) were constructed by Obio Technology Co, Ltd (Shanghai, China). The vectors used were AAV-hSeipin: pAAV-TBG-sfGFP-P2A-BSCL2-3×FLAG-WPRE and AAV-Ctrl: pAAV-TBG-sfGFP-3×FLAG-WPRE. Mice were administered 150 μl of virus containing 2 × 10^11^ AAV8 vector genomes via tail vein injection.
Blood biochemistry
Blood samples were collected from mice fasted for 4 or 12 h via retro-orbital venipuncture. Triglycerides (TAGs) in blood samples were determined using a GPO-PAP method kit (Bio Sino, Beijing, China). Total cholesterol and glucose concentrations were measured following the same procedure. Plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were determined by colorimetric assays and standard curve analysis using ALT and AST assay kits (Jiancheng Bioengineering Research Institute, Jiangsu, China), following the manufacturer's instructions. Plasma FFA levels were determined using an enzymatic kit (Jiancheng Bioengineering Research Institute, Jiangsu, China), following the manufacturer's instructions.
Fast-protein liquid chromatography assay
Lipoprotein distribution analysis was performed using pooled 4 h fasting plasma from 4 mice per group, which was applied to a fast-protein liquid chromatography column (GE AKTA Purifier 100 Fast-Protein Liquid Chromatography system). Fractions were eluted at 0.5 ml per minute using PBS and automatically collected. A total of 35 fractions (0.5 ml each) were analyzed for TAG and total cholesterol content.
VLDL secretion assay
After a 12-h fast, mice were injected with Triton WR1339 (500 mg/kg body weight) via the tail vein. Blood samples were collected at 0, 0.5, 1, 2, and 4 h after injection. Plasma TAG concentrations were determined using the method described in the blood biochemical analysis section.
Liver lipid extraction and assay
Frozen liver tissue (approximately 100 μg) was homogenized in 1 ml of cold PBS. Lipids were extracted using chloroform/methanol (2:1, v/v), dried under nitrogen, and dissolved in 0.5 ml 3% Triton X-100. Liver TAG (Bio Sino, Beijing, China) and total cholesterol content (Bio Sino, Beijing, China) were determined following the manufacturer's instructions. Liver cholesteryl ester (CE) content was calculated as the difference between total cholesterol and free cholesterol. Free cholesterol and total cholesterol contents in liver tissue were measured using assay kits from Applygen Technologies (Beijing, China).
Electron microscopy
Mice were anesthetized with an intraperitoneal injection of sodium pentobarbital. Following anesthesia, the liver was perfused with PBS (0.01 M, pH 7.4) to remove blood. Liver tissue was then excised and immediately fixed in a prefixative solution containing 2.5% glutaraldehyde (v/v). The tissue was subsequently trimmed into small pieces (approximately 0.1 × 0.2 × 0.2 mm^3^). The tissue pieces were further fixed in the same glutaraldehyde solution overnight at 4°C after an initial fixation for 2 h at room temperature. After fixation, liver tissues were processed for staining, embedding, and sectioning. The prepared ultrathin sections were placed on copper grids and examined using a transmission electron microscope.
For VLDL particle analysis, plasma samples were collected from mice after an overnight fasting period. The plasma density was adjusted to 1.006 g/ml, and lipoproteins were separated by ultracentrifugation at 42,000 rpm for 6 h using a P42AT rotor (Hitachi, Japan). The white lipoprotein-containing supernatant was carefully collected. Negative-stain images of VLDL particles were obtained using a transmission electron microscope to analyze their size and morphology.
Quantitative real-time PCR
Total RNA was extracted from liver tissue using Trizol reagent (TIANGAN, Beijing, China), and complementary DNA was synthesized using SweScript All-in-One RT SuperMix (Servicebio, Wuhan, China). The primers for quantitative PCR experiments were synthesized by Sangon Biotech Co, Ltd (Shanghai, China). Gene expression in liver tissue was quantified using quantitative real-time PCR with Applied Biosystems and 2× SYBR Green Quantitative PCR Master Mix (Servicebio, Wuhan, China). Gapdh was used as the reference gene.
Western blotting
Frozen liver tissues were homogenized in cold RIPA buffer containing protease inhibitors (Servicebio, Beijing, China). Protein concentration was determined using the BCA protein assay kit (Seven Biotech, Beijing, China). Equal amounts of protein (30 μg) were separated by 10% SDS-PAGE, followed by wet transfer to PVDF membranes (Applygen Technologies, Beijing, China). The membranes were blocked in Tris-buffered saline with Tween-20 containing 5% milk at room temperature for 1 h, then incubated overnight with primary antibodies at 4°C. Afterward, the membranes were incubated with HRP-conjugated secondary antibodies at room temperature for 1.5 h. Band intensities were analyzed using ImageJ software ( National Institutes of Health, Bethesda, MD).
Histological analysis and confocal microscopy
Liver tissues were perfused with cold PBS and fixed in 4% paraformaldehyde. Paraffin-embedded liver sections (4–5 μm) were stained with H&E for general histological analysis. Immunohistochemistry was performed using F4/80 (Servicebio, Wuhan, China) antibody to assess macrophage infiltration in liver tissue and α-smooth muscle actin (Servicebio, Wuhan, China) antibody to examine hepatic stellate cell activation. Sirius red staining was employed to evaluate the degree of liver fibrosis.
For frozen section preparation, fixed liver tissues were incubated in 20% sucrose solution for 48 h, embedded in OCT compound, and stored at −20°C. Liver frozen sections (5–7 μm) were stained with Oil Red O at 37°C for 30 min, followed by hematoxylin counterstaining to assess neutral lipid deposition. Polarized light microscopy (SOPTO CX40; NINGBO SUNNY Instruments Co., Ltd, Ningbo, China) was used to visualize cholesterol crystals in liver frozen sections.
For free cholesterol detection, liver frozen sections (5–7 μm) were incubated with filipin for 30 min at room temperature and mounted without 4',6-diamidino-2-phenylindole. Filipin fluorescent staining was used to visualize free cholesterol in liver tissue by confocal microscopy (Leica TCS SP5; Leica Microsystems, Germany). Filipin fluorescence intensity was quantified using ImageJ software.
Statistical analysis
Statistical analyses were performed using GraphPad Prism 9.0 software （GraphPad Software, Inc., Boston, MA）. Data were presented as mean ± SD (parametric) or median (interquartile range) (nonparametric). Normality was tested by the Shapiro-Wilk test, and homogeneity of variance was tested by Fisher's F-test. Comparisons between two groups were performed using a two-tailed Student's *t-*test/Welch's *t-*test (parametric) or a two-tailed Mann-Whitney U test (nonparametric). One-way ANOVA followed by Tukey post hoc test was used for comparisons between multiple groups. For comparisons with multiple variables, two-way ANOVA was conducted, followed by Sidak's multiple comparison test as appropriate. P < 0.05 was considered statistically significant.
Results
Liver-specific overexpression of hSeipin in C57BL/6J mice reduced plasma lipids
To examine the gain-of-function effects of seipin on hepatocytes in vivo, we have successfully expressed hSeipin in mouse liver by using AAV8 containing the human TBG promoter (Fig. 1A–C; Supplemental Fig. S1A, B). These mice are referred to as AAV-hSeipin mice hereafter. Endogenous seipin can be detected in the testis (the antibody used can recognize both human and mouse seipin), where seipin is known to be highly expressed (Fig. 1C) (43). By contrast, mouse liver had little, if any, endogenous seipin (Fig. 1C), consistent with a previous report (40). We further examined the expression of endogenous seipin in WT and SKO liver only, as using overexpressed seipin or testes, where seipin is highly expressed, may mask low seipin expression levels present in the liver. We were not able to detect any corresponding band of seipin (Supplemental Fig. S1C), confirming the low level of seipin expression in mouse liver (40). No significant changes in body weight and tissue weight were detected between AAV-Ctrl and AAV-hSeipin mice fed a normal chow diet (Supplemental Fig. S1D, E). Upon fasting for 4 h, the AAV-hSeipin mice on a chow diet showed a dramatic reduction of plasma TAG and total cholesterol, with little change in plasma glucose (Fig. 1D–F). Similar changes were observed under fed conditions (Supplemental Fig. S1F). The above observations are consistent with the results of our previous study, where the AAV-hSeipin mice were examined 7 days after AAV injection in the context of partial hepatectomy and liver regeneration (44). When the plasma is fractionated, the VLDL fraction from fasted AAV-hSeipin mice contained much less TAG than that from AAV-Ctrl mice (Fig. 1G). Likewise, the HDL fraction from AAV-hSeipin mice contained much less cholesterol (Fig. 1H). Furthermore, plasma FFA levels were significantly reduced in AAV-hSeipin mice compared with controls following both 4 h and 12 h fasting periods (Fig. 1I).Fig. 1. Liver-specific overexpression of seipin in C57BL/6J mice reduces plasma lipid levels. Male C57BL/6J mice (6–8 weeks old) were injected via the tail vein with 2 × 10^11^ AAV particles of either AAV-Ctrl or AAV-hSeipin and fed a normal rodent diet (ND) for 6 weeks. Mice were sacrificed under fed conditions. A: Western blot analysis of seipin protein levels in the livers of AAV-Ctrl and AAV-hSeipin mice after 6 weeks of AAV infection (n = 6). B: RT-quantitative PCR analysis of relative Seipin mRNA expression in the livers of AAV-Ctrl and AAV-hSeipin mice after 6 weeks of AAV infection (n = 10–11). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Welch's *t-*test. ∗∗∗∗P < 0.0001. C: Western blot (left) and quantitative analysis (right) of seipin protein levels in the livers and testes of AAV-Ctrl and AAV-hSeipin mice and seipin knockout (SKO) mice (n = 3). Data are presented as mean ± SD. D–F: Plasma TAG (D), total cholesterol (E), and glucose (F) levels in AAV-Ctrl and AAV-hSeipin mice fasted for 4 h after AAV injection (n = 8). Data are presented as mean ± SD. Statistical significance was determined by two-way ANOVA with Sidak's multiple comparison test. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. G and H: TAG (G) and total cholesterol (H) levels in each fraction of plasma from AAV-Ctrl and AAV-hSeipin mice fasted for 4 h. Pooled plasma from AAV-Ctrl and AAV-hSeipin mice was fractionated into VLDL, LDL, and HDL by fast-protein liquid chromatography (FPLC) (n = 4). Mice blood was collected at the fourth week after AAV injection. Different species of lipoproteins are indicated. I: Plasma FFA levels in AAV-Ctrl and AAV-hSeipin mice fasted for 4 and 12 h (n = 8–9). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. ∗P < 0.05, ∗∗P < 0.01.
Reduced VLDL size and VLDL-TAG secretion in AAV-hSeipin mice
To understand the basis of reduced plasma TAG in the AAV-hSeipin mice, we measured plasma TAG and ApoB after 12 h of fasting. The levels of plasma ApoB100 and ApoB48, as well as TAG, were significantly reduced in the AAV-hSeipin mice (Fig. 2A, B). Upon examination by transmission electron microscopy, the sizes of VLDL particles were reduced (Fig. 2C). The diameter of VLDL particles isolated from AAV-hSeipin mice was 42.83 ± 12.74 nm, whereas the diameter of VLDL particles from AAV-Ctrl mice was 63.68 ± 32.43 nm, suggesting a clear reduction in VLDL lipid content.Fig. 2. Liver-specific overexpression of seipin in C57BL/6J mice reduces plasma VLDL particle size. A: Plasma TAG levels in AAV-Ctrl and AAV-hSeipin mice fasted for 12 h (n = 8). Blood samples were collected at the fourth week after AAV injection. Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. ∗∗P < 0.01. B: Plasma ApoB protein levels in AAV-Ctrl and AAV-hSeipin mice fasted for 12 h. Western blot analysis (left panel) and quantitative analysis (right panel) of ApoB100/48 were performed (n = 4). Coomassie Brilliant Blue R250 staining was used to determine the reference protein. Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. ∗∗P < 0.01. C: Representative negative-stain images of plasma VLDL particles from AAV-Ctrl and AAV-hSeipin mice (n = 4 independent experiments). Isolated VLDL particles from AAV-Ctrl or AAV-hSeipin mice fasted for 4 h were visualized by transmission electron microscopy (left). Particle sizes were quantified (right). The scale bar represents 100 nm.
The reduced levels of plasma VLDL-TAG and ApoB of AAV-hSeipin mice suggest that the synthesis and/or secretion of ApoB and VLDL particles might be impaired. To evaluate this possibility, we first examined the hepatic expression of key genes involved in VLDL assembly. The mRNA levels of Apob, Mttp, and Pdi remained unchanged upon seipin overexpression (Fig. 3A), and protein levels of ApoB, MTP (microsomal triglyceride transfer protein), and PDI were similarly unaffected (Fig. 3B), indicating that seipin does not alter their expression. We next assessed hepatic VLDL secretion by intravenously injecting Triton WR-1339—a lipoprotein lipase inhibitor that blocks peripheral lipoprotein clearance—into mice fasted for 12 h. Under these conditions, the accumulation of plasma TAG and ApoB primarily reflects the hepatic output of TAG-rich VLDL particles. Plasma TAG levels in AAV-hSeipin mice injected with Triton WR-1339 were significantly lower than those in AAV-Ctrl mice at all time points tested (Fig. 3C), and the calculated rate of VLDL-TAG secretion was markedly reduced (Fig. 3C), consistent with defective lipidation of VLDL particles. Interestingly, while the steady-state plasma levels of ApoB100/48—which represent VLDL particle number, as each VLDL particle contains a single ApoB molecule—were significantly lower in AAV-hSeipin mice, the rates of ApoB secretion following Triton injection were largely comparable between AAV-Ctrl and AAV-hSeipin mice (Fig. 3D, E), although the rate of ApoB48 secretion appeared to be slowed by the expression of hSeipin. These results suggest that the number of VLDL particles secreted from the liver is not significantly decreased but that the particles are poorly lipidated. The reduced plasma levels of ApoB100/48 are likely because of increased clearance of these lipid-poor VLDL particles in AAV-hSeipin mice, akin to the CideB^−/−^ mice (45). This is further supported by our finding that the plasma lipid clearance rate is elevated in AAV-hSeipin mice following intravenous injection of intralipid (Supplemental Fig. S2). Together, these findings suggest that hepatic seipin overexpression impairs VLDL lipidation without reducing the number of secreted particles, resulting in smaller, lipid-poor VLDL particles that are rapidly cleared from circulation.Fig. 3. Liver-specific overexpression of seipin reduced VLDL-TAG secretion in C57BL/6J mice. A: RT-quantitative PCR analysis of relative Apob, Mttp, and Pdi mRNA expression in the liver of AAV-Ctrl and AAV-hSeipin mice (n = 11). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. B: Western blot (left) and quantitative analysis (right) of ApoB, MTP, and PDI protein expression levels in the liver of AAV-Ctrl and AAV-hSeipin mice (n = 6). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. C: VLDL-TAG secretion and secretion rate in AAV-Ctrl and AAV-hSeipin mice. AAV-Ctrl and AAV-hSeipin mice fasted for 12 h were injected with 500 mg/kg body weight of tyloxapol (Triton WR-1339) through the tail vein at the fourth week after AAV injection. Blood samples were collected from the retroorbital plexus at 0, 0.5, 1, 2, and 4 h after injection (n = 4). These are three independent experiments. Data are presented as mean ± SD or median (interquartile range). Statistical significance was determined using two-way ANOVA followed by Sidak's multiple comparison test (left panel) and a Mann-Whitney U test (right panel). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. D: Western blot of plasma ApoB protein levels at 0, 0.5, 1, 2, and 4 h after injection. Coomassie Brilliant Blue R250 staining was used to determine the reference protein (n = 4 independent experiments). E: Quantitative analysis of ApoB100 (left), ApoB48 (middle), and secretion rate (right) from D (n = 4). Data are presented as mean ± SD. Statistical significance was determined using two-way ANOVA followed by Sidak's multiple comparison test (left and middle panels) and a two-tailed Student's *t-*test (right panel). TC, total cholesterol.
Accumulation of LDs and neutral lipids in AAV-hSeipin liver
The reduced VLDL-TAG secretion of AAV-hSeipin mice may cause hepatic lipid accumulation. Indeed, livers from AAV-hSeipin mice on a normal chow diet accumulated lipids as examined by Oil Red O staining (Fig. 4A). Transmission electron microscopy revealed significantly more LDs in AAV-hSeipin liver (Fig. 4B). Moreover, there appears to be more extensive ER membrane expansion as well as more abundant mitochondria and autophagosomes in the AAV-hSeipin liver. These changes are consistent with seipin's expanding roles in organelle dynamics such as autophagy (39). AAV-hSeipin livers have significantly more TAG and total cholesterol content (Fig. 4C). The results described in Fig. 4A, C are consistent with the results of our previous study (44). Despite the accumulation of lipids, the expression of key genes involved in lipogenesis and lipid uptake was not increased (Fig. 4D, E). In fact, the expression of SREBP1/2 and especially their target genes Scd1 and Fasn was significantly reduced (Fig. 4D), suggesting that defective VLDL secretion may cause lipids to accumulate in the ER, triggering CLD formation and suppressing lipid synthesis. The expression of genes involved in lipolysis and fatty acid oxidation remained unchanged after expressing hSeipin (Fig. 4F). Western blots further confirmed the reduction of stearoyl-CoA desaturase 1 (SCD1) with little change in lipid uptake proteins in AAV-hSeipin liver (Fig. 4G, H). We further carried out lipidomic analyses of livers from AAV-Ctrl and AAV-hSeipin mice (Supplemental Fig. S3). Total PA was significantly decreased in AAV-hSeipin liver, whereas total TAG, CE, ceramide, and SM were increased (Supplemental Fig. S3C–E). Although there is no difference in total phospholipids, the amount of monounsaturated fatty acids was reduced in the AAV-hSeipin liver (Supplemental Fig. S3F, G), consistent with reduced SCD1 expression. Interestingly, the protein level of PLIN2 increased but not its mRNA level, suggesting that PLIN2 may be stabilized by the presence of more LDs in the liver of AAV-hSeipin mice (Fig. 4G).Fig. 4. Liver-specific overexpression of seipin increases hepatic lipid accumulation in C57BL/6J mice. A: Representative images of liver, H&E, and Oil Red O staining of liver sections from AAV-Ctrl and AAV-hSeipin mice under normal rodent diet conditions. The scale bar represents 50 μm. B: Representative images of the hepatocyte ultrastructure in AAV-Ctrl (upper) and AAV-hSeipin (lower) mice revealed by transmission electron microscopy. The scale bars represent 2 μm (top) and 500 nm (bottom). C: Liver TAG and total cholesterol contents of AAV-Ctrl and AAV-hSeipin mice (n = 7–8). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test for liver TAG content and a two-tailed Welch's *t-*test for liver total cholesterol contents. ∗∗P < 0.01. D–F: RT-quantitative PCR analysis of relative mRNA expression of lipogenic genes (D), lipid uptake and apoproteins (E), and lipolysis and fatty acid oxidation genes (F) in the liver of AAV-Ctrl and AAV-hSeipin mice, 6 weeks post-AAV infection (n = 10–11). Data are presented as mean ± SD or median (interquartile range). Statistical significance was determined using a two-tailed Student's *t-*test for normally distributed data with equal variances. Statistical significance was determined using a two-tailed Welch's *t-*test (genes: Srebf1, Fasn, and Scd1) and a two-tailed Mann-Whitney U test (genes: Srebf2, Plin2, Lrp1, Atgl, Lpl, and Cpt1α). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. G: Representative Western blot (left) and quantitative analysis (right) of protein expression of PPARγ, SCD1, PLIN2, and ATGL in the liver of AAV-Ctrl and AAV-hSeipin mice after AAV infection for 6 weeks (n = 6). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test (protein: PPARγ, PLIN2, and ATGL) and a two-tailed Welch's *t-*test (protein: SCD1). ∗∗P < 0.01. H: Representative Western blot (left) and quantitative analysis (right) of protein expression of CD36, LDLR, LRP1, and SRB1 in the liver of AAV-Ctrl and AAV-hSeipin mice after AAV infection for 6 weeks (n = 6). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. TC, total cholesterol.
Liver-specific overexpression of hSeipin aggravates diet-induced liver inflammation and fibrosis
To elucidate the role of hepatic seipin in metabolic dysfunction-associated steatohepatitis (MASH, formerly known as nonalcoholic steatohepatitis or NASH), we first analyzed hSeipin*/BSCL2* expression patterns. Examination of three independent human liver Gene Expression Omnibus datasets (GSE37031, GSE33814, and GSE135251) revealed moderate but significant elevation of seipin mRNA levels in MASH/NASH liver tissues—particularly in those with advanced fibrosis—compared with normal controls, whereas no such increase was observed in simple steatosis tissues (Fig. 5A–C). This clinical observation was further corroborated in our experimental model, where mice fed an HFD (high fat, high cholesterol) with sugar water showed markedly increased hepatic seipin mRNA levels at 20 weeks but not at 12 weeks (Fig. 5D). When AAV-Ctrl and AAV-hSeipin mice were fed on the same HFD with sugar water, plasma lipids of the AAV-hSeipin mice remained lower than those of AAV-Ctrl mice at both 12 and 20 weeks (Fig. 5E, F). No significant differences in body weight were observed between AAV-Ctrl and AAV-hSeipin mice on the HFD (Supplemental Fig. S4A), and the HFD had no impact on the expression of hSeipin (Supplemental Fig. S4B). Liver weight increased more in the AAV-hSeipin mice (Fig. 5G, H). While there appears to be more neutral lipid accumulation based on Oil Red O staining (Fig. 5I, J), there is no apparent difference in the amount of TAG and cholesteryl ester between AAV-Ctrl and AAV-hSeipin livers (Fig. 5K–M). Moreover, the expression of lipogenic genes (Srebf1, Srebf2, Fasn, and Scd1) and plasma glucose remains largely unchanged under these conditions (Supplemental Fig. S4C, D). Interestingly, there is increased free cholesterol in AAV-hSeipin liver (Fig. 5N–P), a common feature of MASH/NASH (46).Fig. 5. Liver-specific overexpression of seipin aggravates diet-induced hepatic lipid accumulation. C57BL/6J male mice were infected with 2 × 10^11^ AAV particles of AAV-Ctrl or AAV-hSeipin through single tail-vein injection and fed with HFD (20% lard and 0.5% cholesterol) with sugar water (18.9% glucose and 23.1% fructose) for 12 or 20 weeks. Mice were sacrificed under feeding conditions. A: The fragments per kilobase of transcript per million mapped reads (FPKMs) of Seipin/BSCL2 mRNA expression in the human liver tissue from healthy controls (normal, n = 7) and MASH patients (n = 8) from the Gene Expression Omnibus (GEO) database GSE37031. Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. ∗∗P < 0.01. B: The FPKM of Seipin/BSCL2 mRNA expression in the human liver tissue surgical samples (normal, n = 13; steatosis, n = 19; and steatohepatitis, n = 12) from the GEO database GSE33814. Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test. ∗∗P < 0.01. C: The FPKM of Seipin/BSCL2 mRNA expression in the human liver tissue from healthy control (normal, n = 10); steatosis (n = 51), and MASH with different fibrosis stages (MASH_F0, n = 5; MASH_F1, n = 29; MASH_F2, n = 53; MASH_F3, n = 54; and MASH_F4, n = 14) from GEO database GSE135251. Data are presented as mean ± SD. Statistical significance was determined using one-way ANOVA. ∗P < 0.05, ∗∗P < 0.01. D: RT-quantitative PCR analysis of relative seipin mRNA expression in the liver from C57BL/6J male mice fed a chow diet, an HFD with sugar water for 12 weeks and 20 weeks (n = 4–8). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's t-test. ∗P < 0.05. E and F: Fasting 12-h plasma TAG (E) and total cholesterol (F) levels of AAV-Ctrl and AAV-hSeipin mice fed the HFD with sugar water for 12 or 20 weeks (n = 4). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's t-test. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. G and H: Liver weight (G) and the ratio of liver weight to body weight (H) of AAV-Ctrl and AAV-hSeipin mice fed the HFD with sugar water for 12 or 20 weeks (n = 4). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's t-test. ∗P < 0.05, ∗∗∗P < 0.001. I and J: Representative images of liver, H&E, and Oil Red O staining of liver section from AAV-Ctrl and AAV-hSeipin mice fed the HFD with sugar water for 12 (I) or 20 (J) weeks (n = 4). The scale bar represents 50 μm. K–N: Liver TAG (K), total cholesterol (L), cholesterol ester (M), and free cholesterol (N) levels in AAV-Ctrl and AAV-hSeipin mice fed the HFD with sugar water for 12 or 20 weeks (n = 4). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test for all indicators except liver free cholesterol (N), whose statistical analysis was performed with a two-tailed Welch's *t-*test. ∗P < 0.05, ∗∗P < 0.01. O: Representative polarization microscopy images (left) and quantification (right) of polarization positive area (%) in liver frozen section from AAV-Ctrl and AAV-hSeipin mice fed the HFD with sugar water for 12 or 20 weeks (n = 4). The scale bar represents 100 μm. Data are presented as median (interquartile range). Statistical significance was determined using a two-tailed Mann-Whitney U test. ∗∗P < 0.01. P: Representative confocal microscopy images (left) and quantification (right) of filipin fluorescence in liver sections from AAV-Ctrl and AAV-hSeipin mice fed the HFD with sugar water for 12 or 20 weeks (n = 4). The scale bar represents 10 μm. Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Welch's *t-*test. ∗P < 0.05, ∗∗∗P < 0.001. CE, cholesteryl ester; FC, free cholesterol; TC, total cholesterol.
We also examined the impact of seipin expression on liver inflammation/damage. No change in plasma ALT and AST levels was detected between AAV-Ctrl and AAV-hSeipin mice before HFD feeding (Fig. 6A, B). However, after 12 weeks on an HFD, both plasma ALT and AST increased in AAV-hSeipin mice (Fig. 6A, B). Plasma ALT, but not AST, remained increased in AAV-hSeipin mice after 20 weeks of HFD (Fig. 6A, B). Although the area of F4/80 staining remains similar between AAV-Ctrl and AAV-hSeipin mice (Fig. 6C, D), more crown-like structures were detected in AAV-hSeipin liver after 20 weeks of HFD (Fig. 6E). The mRNA expression levels of many inflammatory genes examined showed a significant increase in AAV-hSeipin mice fed with HFD for 20 weeks (Fig. 6F). Sirius red and α-smooth muscle actin staining revealed much worsened liver fibrosis in AAV-hSeipin mice fed with HFD for 20 weeks (Fig. 6G–J). Consistently, most fibrosis genes examined were upregulated in AAV-hSeipin mice (Fig. 6K).Fig. 6. Liver-specific seipin overexpression aggravates liver inflammation and fibrosis. A and B: Plasma ALT (A) and AST (B) levels in AAV-Ctrl and AAV-hSeipin mice at baseline (prediet) and after 12 or 20 weeks of HFD with sugar water feeding (n = 4–8). Data are presented as median (interquartile range). Statistical significance was determined using a two-tailed Mann-Whitney U test. ∗P < 0.05. C: Representative immunohistochemical staining images of F4/80 in liver sections from AAV-Ctrl and AAV-hSeipin mice fed an HFD with sugar water for 12 or 20 weeks. The scale bar represents 50 or 25 μm. D and E: Quantification of F4/80 positive area (D) and crown-like structures (E) from C (n = 4). Data are presented as mean ± SD or median (interquartile range). Statistical significance was determined using a two-tailed Student's *t-*test (D) and a two-tailed Mann-Whitney U test (E). ∗∗P < 0.01. F: RT-quantitative PCR analysis of relative mRNA expression of inflammatory genes in the liver from AAV-Ctrl and AAV-hSeipin mice fed an HFD with sugar water for 20 weeks (n = 4). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test (genes: Adgre1, Cxcl1) and a two-tailed Welch's t-test (genes: Cd68, Tnfa, Il1b, and Ccl2). ∗∗P < 0.01. G and H: Representative Sirius red staining images (G) and quantification (H) of liver sections from AAV-Ctrl and AAV-hSeipin mice fed an HFD with sugar water for 12 or 20 weeks (n = 4). The scale bar represents 50 or 100 μm. Data are presented as median (interquartile range). Statistical significance was determined using a two-tailed Mann-Whitney U test. ∗∗P < 0.01. I and J: Representative immunohistochemical staining images (I) and quantification (J) of α-smooth muscle actin in liver sections from AAV-Ctrl and AAV-hSeipin mice fed an HFD with sugar water for 12 or 20 weeks (n = 4). The scale bar represents 50 or 25 μm. Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Welch's *t-*test. ∗∗P < 0.01. K: RT-quantitative PCR analysis of relative mRNA expression of fibrosis-related genes in the liver from AAV-Ctrl and AAV-hSeipin mice fed an HFD with sugar water for 20 weeks (n = 4). Data are presented as mean ± SD. Statistical significance was determined using a two-tailed Student's *t-*test for all indicators except Trimp1 and Col1a1, whose statistical analysis was performed with a two-tailed Welch's *t-*test. ∗P < 0.05, ∗∗P < 0.01.
Discussion
Distinct from most other cell types, hepatocytes can produce CLDs in the cytoplasm as well as ER LLDs and lipoproteins in the lumen of the ER. Although hepatocytes can make and accumulate a large amount of CLDs during hepatic steatosis, the primary function of hepatocytes under normal conditions is not to store lipids. Instead, a major task of hepatocytes is to produce lipoproteins. Notably, seipin, one of the most critical proteins for CLD biogenesis, is almost absent from hepatocytes (Fig. 1) (40). We thus hypothesize that seipin might favor the biogenesis of CLDs and interfere with the production of LLDs and the assembly of VLDLs if it is present in hepatocytes. Our results in this study support such a hypothesis because we demonstrate here that elevated expression of hSeipin in chow-fed mouse liver severely disrupted VLDL maturation but increased hepatic lipid accumulation and CLD biogenesis. Results from the current study thus unveil an important molecular link between the biogenesis of CLDs and that of LLDs and VLDL assembly.
Although much progress has been made on the molecular mechanisms governing the biogenesis of CLDs, almost no knowledge exists on how LLDs are formed and on if and how the biogenesis of CLDs and LLDs is connected (1). While direct evidence is still lacking, our results do suggest, for the first time, that enhanced biogenesis of CLDs in hepatocytes can impede LLD formation and VLDL assembly. Based on its well-established function, seipin may preferably channel neutral lipids toward the synthesis of CLDs, thereby impairing LLD formation and VLDL lipidation (Fig. 7). Specifically, seipin may facilitate the nucleation of newly synthesized TAGs and push the nucleated TAGs toward the cytoplasm. As a consequence, more CLDs are formed, and less TAGs are available for delivery to pre-VLDL particles by MTP. Moreover, many fewer and smaller LLDs may be formed, further reducing the lipidation of pre-VLDLs (Fig. 7). As mentioned above, the key results of this study, that is, disrupted VLDL lipidation and hepatic steatosis of AAV-hSeipin liver, are consistent with this hypothesis. The reduced expression of SREBP and its target genes, for example, SCD1, in AAV-hSeipin mice may arise from the diversion of lipid flow from secretion to cellular storage, thereby suppressing lipogenesis. Enhanced seipin function in the liver exacerbated abnormal LD enlargement, hepatic inflammation, and fibrosis when mice were fed an HFD with sugar water, further suggesting that seipin promotes the cytoplasmic storage of hepatic lipids but disrupts their secretion. The hepatic lipid storage capacity of AAV-hSeipin mice may be exceeded after prolonged HFD with sugar water feeding, leading to increased hydrolysis of neutral lipids. Thus, the level of free cholesterol, but not TAG, is significantly elevated in the liver of AAV-hSeipin mice fed HFD, as we observed. Notably, accumulation of free cholesterol in hepatocytes is well known to induce liver inflammation and fibrosis (46, 47, 48).Fig. 7. Hypothetical model. Under normal conditions, nucleating and nucleated TAG are preferably delivered to pre-VLDL particles by MTP. Moreover, nucleated TAG can bud into the ER lumen, generating LLDs, which may also contribute to VLDL lipidation. Few CLDs are formed in normal hepatocytes. When seipin is present and active in hepatocytes, it may shield nucleating and nucleated TAG from access by MTP, thereby reducing lipidation of pre-VLDLs. Moreover, seipin may directly (physically) or indirectly (changing the partition of conical lipids) drive the formation of CLDs, thereby channeling TAG toward the cytoplasm and causing steatosis. Consequently, fewer and smaller LLDs are formed, further diminishing the lipidation of pre-VLDLs. DAG, diacylglycerol.
Exactly how seipin may promote CLD formation and disrupt VLDL lipidation in hepatocytes remains to be determined. Seipin may facilitate the phase separation of neutral lipids within the two leaflets of the ER and then physically drive the budding of nascent LDs toward the cytoplasm, as discussed above. This function may likely require LDAF1 (14), which is also lowly expressed in hepatocytes (40). It would be informative to examine the effects of overexpressing both seipin and LDAF1 in the mouse liver in the future. Alternatively, by changing local phospholipid concentration, seipin may reduce the tension of the cytoplasmic leaflet of the ER and increase that of the luminal leaflet. This can be achieved by increasing the amount of conical lipids, such as PA and diacylglycerol, in the luminal leaflet of the ER (Fig. 7). Such changes are known to promote the budding of LDs toward the cytoplasm (11, 49). Finally, seipin contains a large luminal domain, which may directly disrupt the functions of key proteins for VLDL lipidation, such as MTP, ApoE, ApoB, etc. Future studies will examine these possibilities.
In addition to mechanistic insights, our findings may also carry translational implications. The observation that forced hepatic expression of seipin leads to impaired VLDL lipidation and worsened hepatic inflammation and fibrosis highlights the importance of optimizing CLD biogenesis in maintaining liver lipid homeostasis. These results raise the possibility that limiting excessive CLD formation or redirecting lipid flux toward lipoprotein secretion could represent a novel therapeutic strategy for metabolic dysfunction-associated steatotic liver disease and its progressive form, MASH.
Previously published work from us and others demonstrated that seipin overexpression reduced the level of TAG (5, 50), which appears to contradict our current results in the mouse liver. However, those previously published works used cultured cell lines (HeLa, 3T3L1), where GPAT3/4 play a major role in TAG synthesis. A key function of seipin is to limit GPAT3/4 function; therefore, its overexpression would reduce TAG accumulation (23, 30). In the current study, we focus on the effect of seipin on hepatocytes in vivo. GPAT1, but not GPAT3/4, dominates in hepatocytes, and seipin does not inhibit GPAT1 (our unpublished data). Most importantly, only hepatocytes and enterocytes, but not any other cultured cell lines, produce LLDs and lipoproteins in the ER lumen, and our current data suggest that seipin diverts the lipids from LLDs and lipoproteins to CLDs. The different results of seipin overexpression between cell types likely result from the two key characteristics of hepatocytes: the presence of GPAT1 and the production of LLDs/lipoproteins. Our results also differ from a previous study where seipin overexpression in mouse liver alleviated HFD-induced hepatic steatosis in part by promoting mitochondrial activity (51). Notably, a significant amount of endogenous seipin was detected in mouse liver (51), contrary to results from our current study and those of others (40). Future studies are needed to further examine the role of seipin in mouse liver.
Despite the important mechanistic and therapeutic implications of our results, this study has several limitations. First, our findings are based solely on the AAV-mediated overexpression of hSeipin in mouse hepatocytes, which may not fully reflect physiological conditions and could potentially introduce supraphysiological or off-target effects. Second, although the data clearly demonstrate an inhibitory role of seipin in VLDL lipidation, we were unable to directly visualize or quantify ER LLDs because of current technical limitations in imaging lipid structures within the ER lumen. Third, while the observed defects in VLDL maturation are consistent with impaired lipidation, the precise molecular mechanisms by which seipin disrupts this process—whether through altered lipid partitioning, changes in ER membrane dynamics, or interference with key proteins, such as MTP or ApoB—remain to be fully elucidated. Fourth, while the expression of seipin was moderately increased in human patients with steatohepatitis or in mice after HFD, the exact mechanism governing seipin expression under such conditions remains unknown. Fifth, upon seipin expression, free cholesterol accumulated under HFD with sugar water conditions. We speculate that the free cholesterol accumulated specifically on the LD surface, as recently reported, but not the ER (46), which may explain why there is little change in the expression of low-density lipoprotein receptor. Sixth, while AAV-hSeipin livers accumulated more neutral lipids than control under a normal chow diet, the levels of neutral lipids between control and AAV-hSeipin livers became similar after HFD with sugar water feeding, despite significantly enlarged LDs in AAV-hSeipin livers. It is possible that high-fat feeding may have overwhelmed the lipid storage capacity of both AAV-hSeipin and control livers. Alternatively, although significantly more fat is believed to be diverted to the cytoplasm of AAV-hSeipin hepatocytes, it may also cause stronger inhibition of lipogenesis and TAG synthesis. Finally, Oil Red O staining may not capture small LDs and therefore cannot accurately reflect the quantity of neutral lipids. Future studies using complementary genetic models, high-resolution imaging techniques, and dynamic lipid flux analyses will be needed to clarify these mechanistic questions.
In summary, we studied, for the first time, the impact of enhanced biogenesis of CLDs on lipoprotein assembly in mouse liver. Our results suggest that for the proper secretion/production of lipoproteins, the biogenesis of CLDs should be restricted. Our results thus provide important new insights into how lipid balance/flux may be maintained in hepatocytes and highlight the intricate relationship between CLDs and LLDs in hepatocytes.
Data Availability
All data reported in this article will be shared by the lead contact upon request.
Supplemental Data
This article contains supplemental data.
Conflict of Interest
The authors declare that they have no conflicts of interest with the contents of this article.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Zadoorian A.Du X.Yang H.Lipid droplet biogenesis and functions in health and disease Nat. Rev. Endocrinol.1920234434593722140210.1038/s 41574-023-00845-0PMC 10204695 · doi ↗ · pubmed ↗
- 2Soltysik K.Ohsaki Y.Tatematsu T.Cheng J.Maeda A.Morita S.Y.Nuclear lipid droplets form in the inner nuclear membrane in a seipin-independent manner J. Cell Biol.2202021 e 20200502610.1083/jcb.202005026 PMC 773770333315072 · doi ↗ · pubmed ↗
- 3Romanauska A.Kohler A.The inner nuclear membrane is a metabolically active territory that generates nuclear lipid droplets Cell 1742018700715.e 182993722710.1016/j.cell.2018.05.047PMC 6371920 · doi ↗ · pubmed ↗
- 4Wang H.Gilham D.Lehner R.Proteomic and lipid characterization of apolipoprotein B-free luminal lipid droplets from mouse liver microsomes: implications for very low density lipoprotein assembly J. Biol. Chem.282200733218332261784854610.1074/jbc.M 706841200 · doi ↗ · pubmed ↗
- 5Fei W.Li H.Shui G.Kapterian T.S.Bielby C.Du X.Molecular characterization of seipin and its mutants: implications for seipin in triacylglycerol synthesis J. Lipid Res.522011213621472195719610.1194/jlr.M 017566 PMC 3220282 · doi ↗ · pubmed ↗
- 6Fei W.Shui G.Gaeta B.Du X.Kuerschner L.Li P.Fld 1p, a functional homologue of human seipin, regulates the size of lipid droplets in yeast J. Cell Biol.18020084734821825020110.1083/jcb.200711136 PMC 2234226 · doi ↗ · pubmed ↗
- 7Szymanski K.M.Binns D.Bartz R.Grishin N.V.Li W.P.Agarwal A.K.The lipodystrophy protein seipin is found at endoplasmic reticulum lipid droplet junctions and is important for droplet morphology Proc. Natl. Acad. Sci. U. S. A.104200720890208951809393710.1073/pnas.0704154104 PMC 2409237 · doi ↗ · pubmed ↗
- 8Tian Y.Bi J.Shui G.Liu Z.Xiang Y.Liu Y.Tissue-autonomous function of Drosophila seipin in preventing ectopic lipid droplet formation P Lo S Genet.72011 e 10013642153322710.1371/journal.pgen.1001364 PMC 3077376 · doi ↗ · pubmed ↗