Expression of Plin5 was decreased in NASH models

To identify molecules important for NASH, we first obtained 4 RNA-seq data for livers with NASH from the GEO database: 1 human (GSE185051), 2 mus musculus (GSE200409, GSE215225) and 1guinea pig (GSE192497). To identify specific genes during NASH, we focused on oscillating genes that featured significant changes (> twofold change and p < 0.05) in the NASH group compared with the control group in the four datasets. Results revealed only two genes including one down-regulated gene Plin5 and one up-regulated gene Slc35f2 after overlapping (Fig. 1A). Of note, Plin5 has been shown to be highly abundant in the liver and important for lipid metabolism [17].

Fig. 1

Expression of Plin5 was decreased in fatty liver models. A Venn diagram of overlapped genes in NASH group versus control group in the 4 GEO datasets. B Representation of western blot and qualification of Plin5 protein level in mice after 24 weeks of ND or HFHC feeding. C Quantification of Plin5 mRNA level in mice after 24 weeks of ND or HFHC feeding. D qPCR analysis of Plin5 mRNA level in mice under ND or CDA-HFD feeding conditions. A two-tailed Student t-test was used for statistical analysis. *< 0.05, **< 0.01. All data are shown as mean ± SEM. (n = 6/group)

To determine whether Plin5 participates in the development of NASH, we identified Plin5 expression changes in two murine models of fatty liver disease. Mice developed serious steatohepatitis after being fed an HFHC diet for 24 weeks or choline-deficient amino acid-defined high fat diet (CDA-HFD) for 12 weeks [18, 19]. Compared with the normal diet (ND) controls, hepatic Plin5 level was dramatically reduced in the liver after feeding an HFHC diet for 24 weeks (Fig. 1B, C) or CDA-HFD for 12 weeks (Fig. 1D). Our data as a whole showed that Plin5 expression was reduced in NASH models.

Deficiency of Plin5 had no effect on the liver of mice fed a normal chow diet

In order to better understand Plin5’s function in the development of NASH, mouse strains with knockout of Plin5 (Plin5 KO) were constructed by CRISPR/Cas-mediated genome engineering. Western blotting and qPCR analysis performed on liver tissue revealed that Plin5 was knocked out in Plin5 KO compared with WT mice (Fig. 2A, B). Plin5 deficiency in mice fed with ND caused a slight increase in body weight but had no effect on liver mass (Fig. 2C–E) or histological appearance (Fig. 2F, G). In addition, serum level of ALT, AST and lipids level were unchanged in Plin5 KO mice fed ND (Fig. 2H–K). Meanwhile, Plin5 deficiency did not influence the expression of inflammatory or collagen related genes in the liver (Fig. 2L, M). These results suggested that Plin5 deficiency had no effect on the liver in mice fed a normal chow diet.

Fig. 2

Plin5 deficiency had no effect on liver of mice fed a normal chow diet. A Representative western blot and quantification of Plin5 measured in livers of WT and Plin5 KO mice. B qPCR analysis of mRNA level of Plin5. C–E Body weight (C), liver mass (D) liver mass/body weight (E). F, G H&E and Masson staining images of liver sections from WT and Plin5 KO mice. Scale bar, 50 um. H, I Serum ALT and AST levels of mice in the indicated groups. J, K Serum TG (Triglyceride) and TC (total cholesterol) concentration in WT and Plin5 KO mice. L qPCR analysis of mRNA level of inflammation-related genes in the liver of WT and Plin5 KO mice fed a ND diet. M Relative mRNA level of profibrotic genes in liver tissue from WT and Plin5KO mice. A two-tailed student t test was used for statistical analysis. *< 0.05, **< 0.01, n.s. not significant. All data are shown as mean ± SEM. Abbreviations: MCP-1: Monocyte Chemoattractant Protein-1; TNF-α: Tumor necrosis factor-α; Collagen α-I: collagen type I alpha1 chain; TIMP-1: TIMP metallopeptidase inhibitor 1. (n = 6/group)

Plin5 deficiency aggravated HFHC diet-induced steatohepatitis

To evaluate the potential effects of Plin5 in NASH progression, we examined Plin5 KO mice in an HFHC diet-induced NASH model with induced hepatic steatosis, inflammation and fibrosis, features similar to human NASH. As shown in Fig. 3A, mice fed an HFHC diet or ND were subjected to GTT and ITT, then serum and liver tissue were collected for further analysis. We found that Plin5KO mice had insulin resistance and impaired glucose tolerance (Additional file 1: Fig. S1A, B).

Fig. 3

Systemic knockout of Plin5 aggravated HFHC diet-induced steatohepatitis. A Schematic diagram of the experimental procedure used to examine the harmful role of Plin5 in mice fed an HFHC for 24 weeks. B–D Body weight (B), Liver mass (C), and liver mass/body weight (D) values in WT and Plin5 KO mice after feeding an HFHC diet for 24 weeks. E, F Serum TG and TC level. G Representative images of H&E and oil red O staining of liver sections from mice fed chow diet or HFHC diet. The statistics of oil red O-positive areas are shown. Scale bar, 50 um. H Relative images and quantitative data of PSR staining in liver sections from WT and Plin5 KO mice (scale bars 50 um. I, J Serum ALT (I) and AST (J) were measured in WT and Plin5 KO mice after 24 weeks of HFHC diet feeding (n = 6/group). K NAFLD activity score. L qPCR analysis of mRNA level of inflammation-related genes in the liver of WT and Plin5 KO mice fed the HFHC diet. M Relative mRNA level of profibrotic genes in liver tissue from WT and Plin5 KO mice. A two-tailed Student t test was used for statistical analysis. The data are presented as the mean ± SEM. A significant difference between the WT-ND group and the Plin5 KO-HFHC. group. *< 0.05, **< 0.01; Abbreviations: NAS, NAFLD activity score. (n = 6/group)

Weight of Plin5 KO mice was significantly higher than that of the WT mice (Fig. 3B) with liver weight and liver weight-to-body weight ratio remarkably increased (Fig. 3C, D). We also observed that Plin5 KO increased TG and TC level in the liver, indicating aggravation of hypertriglyceridemia and hypercholesterolemia by Plin5 deficiency in an HFHC diet-induced NASH mouse model (Fig. 3E, F). Compared with WT mice, Plin5 KO mice showed more significant hepatic fibrosis, as demonstrated by Masson staining and Sirius red staining (Fig. 3H). Meanwhile, Plin5 KO mice fed an HFHC diet showed more severe liver damage, as evidenced by higher serum ALT and AST level (Fig. 3I, J). In addition, NAS score revealed that Plin5 KO aggravated the grade of liver steatohepatitis in the presence of an HFHC-diet (Fig. 3K). NASH is strongly associated with severe inflammation and progressive liver fibrosis [18]. We examined whether Plin5 affected inflammation and fibrosis accompanied by augmented fibrogenic and inflammatory gene expression after feeding an HFHC diet (Fig. 3L, M).

The findings revealed that the HFHC-induced NASH phenotype was worsened by Plin5 KO in hepatocytes.

Ferroptosis occurs in HFHC-diet fed mouse livers and knockout of Plin5 diminished HFHC diet-induced ferroptosis

Oxidative stress and its associated lipid peroxidation-mediated ferroptosis play key roles in NASH progression [13, 16, 20]. We established that lipid ROS was markedly increased in liver from mice fed an HFHC-diet for 24 weeks, as assessed by specific fluorescent probe C11-BODIPY (581/591) and compared with mice fed an NC diet (Additional file 1: Fig. S2 A). Furthermore, the mRNA levels of genes implicated in ferroptosis, such as Hmox1, Acsl4, Ptgs2, and Nox2, were increased, whereas the protein level of glutathione peroxidase 4 (GPX4), the essential regulator of ferroptosis [21], was lowered (Additional file 1: Fig. S2B, C). These results suggested the occurrence of ferroptosis during NASH progression.

We further investigated whether lipid droplet Plin5 was involved in the process of iron metabolism and ferroptosis. We first measured total iron and Fe2+ level in the liver and serum of WT and Plin5 KO mice. As shown in Fig. 4A, B, Plin5 KO mice showed an increased liver total iron and Fe2+ level. Prussian blue staining (Enhance with DAB) also showed increased iron deposits in Plin5 KO mouse liver (Fig. 4C). Meanwhile, Plin5 KO mice demonstrated increased serum total iron and Fe2+ level (Fig. 4D, E). Iron metabolism-related genes including Hamp1, Hamp2 and iron transport gene Tfr1 were significantly increased, while iron storage related genes Ferritin H heavy chain (Fth) and light chain (Ftl) were significantly decreased in Plin5 KO mouse liver compared with that of WT mice (Fig. 4F). Furthermore, increased MDA and diminished SOD content were evident in the liver of Plin5 KO mice (Fig. 4G, H).

Fig. 4

Knockout of Plin5 diminished HFHC diet-induced ferroptosis. A, B Total and ferrous iron levels in liver were measured by iron probe in WT and Plin5 KO mice fed an HFHC-diet. C Prussian blue iron staining (Enhanced with DAB) of liver sections from mice fed an HFHC-diet. Scale bar, 50 um. D, E Serum total and ferrous iron levels were measured by iron probe in WT and Plin5 KO mice fed an HFHC-diet. F Expression of hepatic iron metabolic genes, Tfr1, Hamp1, Hamp2, Fth, Ftl, was measured by qPCR from WT and Plin5 KO mice fed an HFHC-diet. G, H MDA and SOD content was measured using a malondialdehyde assay kit (TBA method) and total superoxide dismutase assay kit (hydroxylamine method) respectively. I Hepatic mRNA level of Hmox1, Ptgs2 and Nox2 was measured by qPCR in HFHC-diet fed mice. J Confocal images of WT and Plin5 KO mouse liver sections labeled with C11-BODIPY and DAPI from mice fed on HFHC-diet. Green and blue colors indicate lipid ROS (peroxidated lipids) and nucleus respectively. Scale bar, 50 um. K Western blot analysis of GPX4 in WT and Plin5 KO mouse livers. A two-tailed Student t-test was used for statistical analysis. The data are presented as the mean ± SEM. *< 0.05, **< 0.01. Abbreviations: MDA: malondialdehyde; SOD: total superoxide dismutase. (n = 6/group)

In addition, the positive regulators of ferroptosis Hmox1, Ptgs2 and NOX2 were significantly up-regulated in Plin5 KO mice (Fig. 4I). The occurrence of lipid peroxidation in the liver was further confirmed by markedly increased lipid ROS production (Fig. 4J). Nonetheless the expression of GPX4, which is the negative regulator of ferroptosis, was decreased in Plin5 KO mice (Fig. 4K). These results suggested that deletion of Plin5 aggravated the occurrence of ferroptosis in an HFHC-induced mouse model.

Overexpression of Plin5 ameliorated MCD diet-induced NASH and ferroptosis

To further validate the role of Plin5 in steatohepatitis, we administered AAV-Plin5 or AAV-Vector to mice and fed them an MCD for 3 weeks to induce steatohepatitis [22]. Western blot experiments showed that the expression of Plin5 was significantly increased in liver tissue from AAV-Plin5 mice compared with the control group (Fig. 5A). Compared with MCD fed WT mice, serum levels of hepatic enzymes (AST and ALT) were significantly reduced in AAV-Plin5 mice, indicating that Plin5 significantly ameliorated liver injury (Fig. 5B, C). Furthermore, histological staining and the grade of steatohepatitis determined by NAS score demonstrated a reduction in hepatic steatosis and fibrosis for MCD-diet fed AAV-Plin5 mice (Fig. 5D) accompanied by reduced inflammatory and fibrosis-related genes (Fig. 5E, F).

Fig. 5

Hepatocyte-specific overexpression of Plin5 alleviated MCD-induced NASH and ferroptosis. A Representative western blot of Plin5 measured in livers of AAV-vector and AAV-Plin5 mice. B, C Serum level of ALT and AST measured in WT and Plin5 KO mice after 3 weeks of MCD feeding. D Representative images of H&E, oil red O staining, Masson staining and PSR staining of liver sections from mice and NAFLD activity scores (Scale bar,100 um). E, F qPCR analysis of mRNA level of inflammation-related genes and profibrotic genes in the liver of AAV-vector and AAV-Plin5 mice fed an HFHC diet. G, H MDA and SOD content were measured by malondialdehyde assay kit (TBA method) and total superoxide dismutase assay kit (hydroxylamine method) respectively. I Expression of hepatic iron metabolic genes Tfr1, Hamp1, Hamp2, Fth and Ftl were measured by qPCR from AAV-vector and AAV-Plin5 mice under MCD-diet condition. J Hepatic mRNA level of Hmox1, Ptgs2, and Nox2 were measured by qPCR in MCD-diet mice. K The protein level of GPX4 and GAPDH in AAV-vector and AAV-Plin5 mice fed an MCD-diet were detected by western blot. A two-tailed Student t-test was used for statistical analysis. Data are presented as the mean ± SEM. *< 0.05, **< 0.01. (n = 6/group)

Meanwhile, the reduction of MDA and excess of SOD were significantly improved after Plin5 over-expression (Fig. 5G, H). Hepatic iron metabolism genes including Hamp1, Hamp2, and Tfr1 were significantly down-regulated, while Fth and Ftl were significantly up-regulated in the livers of the AAV-Plin5 group (Fig. 5I). We also found that the Hmox1, Ptgs2 and NOX2 were decreased and protein expression of GPX4 was increased in AAV-Plin5 mice, indicating that Plin5 over-expression effectively reduced ferroptosis in MCD-induced NASH (Fig. 5J, K). These data demonstrate that overexpression of Plin5 inhibited ferroptosis and alleviated NASH progression.

Plin5 KO aggravated RSL-3-induced ferroptosis and could be reversed by 11-Dodecenoic acid, a downstream metabolite of Plin5

Exogenous MUFAs have been shown to prevent ferroptosis in both transformed and non-transformed cells through a structure-specific mechanism [23]. Exogenous monounsaturated fatty acids inhibit ferroptosis, a non-apoptotic mechanism, via increasing ACSL3-dependent polyunsaturated fatty acid substitution from plasma membrane phospholipids [24].

Since Plin5 is involved in lipid metabolism, we explored the possible downstream metabolite by which Plin5 affected NASH and ferroptosis via targeted lipidomics with liver tissue from Plin5 KO mice and WT mice fed an HFHC-diet (Fig. 6A). A total of 40 metabolites were detected: 11 types of MUFAs, 3 types of TUFAs, 14 types of PUFAs and 12 types of SFAs. Interestingly, volcano plot revealed that compared with the WT group, 11-Dodecenoic acid (11-DA) in liver was significantly down-regulated (p < 0.05 and |log2FC|≥ 1) in the Plin5 KO group under HFHC diet conditions (Fig. 6B, C), suggesting that 11-DA may be an important metabolite downstream of Plin5 during NASH progression. In addition, the total amount of MUFAs and PUFAs was not significantly altered in Plin5KO group compared with WT group (Additional file 1: Fig. S3A, B).

Fig. 6

Plin5 KO aggravated RSL-3 induced ferroptosis that was reversed by 11-Dodecenoic acid, a downstream targeting molecular of Plin5. A Schematic of the experimental strategy used to identify the potential target metabolites of Plin5. in steatotic hepatocytes. B Volcano plot indicates different metabolites between the Plin5 KO group and the control group. C Concentration of 11-Dodecenoic acid in WT and KO groups. D Protein expression and quantification of Plin5 in SK-HEP1 cells infected with siPlin5. E Schematic diagram of cells experiment. F Cell viability was assayed by trypan blue staining. G SK-HEP1 cells with empty vector, Plin5 knockdown, and adding 11-DA were treated with RSL-3 for 12 h and cell states were captured by microscope. Dead cells were stained by propidium iodide (PI). Scale bar, 100 um

To examine the functional roles of Plin5 in hepatocytes, we created Plin5-knockdown hepatocytes by transducing the Plin5 gene with a short interfering RNA (siRNA). Western blot analysis revealed that siRNA (siPlin5) significantly inhibited Plin5 expression in the siPlin5 group relative to the negative control (NC) group (Fig. 6D). Hepatocyte cell line SK-HEP1 is frequently utilized for ferroptosis research because to its extreme sensitivity to ferroptosis-inducing chemicals such the GPX4 inhibitor RSL-3, which causes lipid peroxidation and ferroptosis when present in sufficient amounts [25]. We treated control or siPlin5 hepatocytes with or without 11-Dodecenoic acid following RSL-3 administration (Fig. 6E). Of note, siPlin5 hepatocytes demonstrated significantly increased RSL-3-induced ferroptosis and cell death while 11-DA (200 μM) rescued the effects (Fig. 6F, G).

Excessive accumulation of lipid peroxide (lipid ROS), generated by the family of lipoxygenases, is a critical cause leading to ferroptosis [26]. Lipid ROS can be detected by using fluorescent radio-probe C11-BODIPY 581/591. We examined lipid ROS expression levels in different groups. It was found that under RSL3 treatment conditions, the expression of lipid ROS was increased after knocking down Plin5, while the level of lipid ROS was decreased after adding 11-DA (Additional file 1: Fig. S3C).

We concluded that Plin5 knockdown in hepatocytes aggravated RSL-3-induced ferroptosis, while addition of 11-DA rescued the phenotype, suggesting that 11-DA may be a downstream metabolite effector of Plin5.