Celastrol ameliorates acute liver injury through modulation of PPARα

Abstract

Celastrol, derived from the roots of the Tripterygium Wilfordi, has attracted interest for its potential anti-inflammatory and lipid-lowering activities. In the present study, the protective effect of celastrol on carbon tetrachloride (CCl₄)-induced acute liver injury was investigated. Celastrol improved the increased transaminase activity, inflammation, and oxidative stress induced by CCl₄, resulting in improved metabolic disorders found in mice with liver injury. Dual-luciferase reporter assays and primary hepatocyte studies demonstrated that the peroxisome proliferator-activated receptor α (PPARα) signaling mediated the protective effect of celastrol, which was not observed in Ppara-null mice, and co-treatment of wild-type mice with the PPARα antagonist GW6471. Mechanistically, PPARα deficiency potentiated CCl₄-induced liver injury through a deoxycholic acid (DCA)-EGR1-inflammatory factor axis. These data demonstrate a novel role for celastrol in protection against acute liver injury through modulating PPARα signaling.

1. Introduction

Celastrol is a natural compound isolated from the root extracts of Tripterygium wilfordi (thunder god vine). Celastrol shows significant pharmacological activities, including anti-inflammatory, anti-cancer, anti-obesity and treating mesangioproliferative glomerulonephritis [1–3]. It was reported that celastrol could modulate various targets, such as NFκB, Nur77, and the HSF1-PGC1α axis [2,4,5]. Celastrol protects against experimental acetaminophen (APAP)-induced liver injury and α-naphthyl isothiocyanate (ANIT)-induced cholestasis [6,7]. The liver injury induced by APAP, ANIT, triptolide, and sunitinib, disrupts mitochondrial fatty acid β-oxidation and the accumulated serum long-chain acylcarnitines [8–10]. Peroxisome proliferator-activated receptor α (PPARα) regulates mitochondrial fatty acid transport and β-oxidation, bile acid synthesis, and inflammation [11]. PPARα agonist reverses the increase in acylcarnitines levels that result from mitochondrial damage in mouse models of hepatotoxicity [8]. Therefore, PPARα plays an important role in protecting against chemically-induced liver injury.

LC-MS-based metabolomics has been used to identify drug metabolites related to toxicity [12,13], and was applied to investigate the mechanism of liver injury, including cholestasis [8], liver dysfunction [14], and steatohepatitis [15]. Furthermore, metabolomics could determine the roles of nuclear receptors, such as pregnane X receptor (PXR) [16], farnesoid X receptor (FXR) [17], and PPARα [8] in physiology, metabolic diseases and chemically-induced liver toxicities. The results of the present study reveal that celastrol regulates PPARα signaling and significantly attenuates carbon tetrachloride (CCl₄)-induced acute liver injury, through modulation of inflammation, oxidative stress, bile acid metabolism and acylcarnitine utilization. The protective effect of celastrol on CCl₄-induced liver injury was attenuated in Ppara-null mice (Ppara^−/− mice) and co-treatment with PPARα antagonist GW6471 in wild-type (WT) mice. These findings provide a novel role for celastrol in protecting against acute liver injury through the activation of PPARα signaling.

2. Materials and methods

2.1. Chemicals and reagents

Celastrol was provided by Chengdu Mansite Bio-technology Co Ltd (Chengdu, China). CCl₄ and corn oil were obtained from the Shanghai Aladdin Bio-Chem Technology (Shanghai, China). Lauroylcarnitine (12:0-carnitine), myristoylcarnitine (14:0-carnitine), palmitoylcarnitine (16:0-carnitine), stearoylcarnitine (18:0-carnitine), deoxycholic acid (DCA), taurocholic acid (TCA), taurohyodeoxycholic acid (THDCA), and taurochenodeoxycholic acid (TCDCA) were ordered from Sigma-Aldrich (St. Louis, Missouri, USA). Tauro-β-muricholic acid (TβMCA) was provided by Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). Tauro-α-muricholic acid (TαMCA) was purchased from Steraloids (Newport, RI, USA). All other chemical reagents and solvents were of the highest grade commercially available.

2.2. Animals

Male WT mice and Ppara^−/− mice (6- to 8-weeks-old) on the 129/Sv genetic background were previously described [18]. All mice were maintained in a controlled environment with a standard 12 h light/12 h dark cycle and humidity 50%–60%. Animal experiments were approved by the institutional ethical committee of Kunming Institute of Botany.

Experiment 1: To determine the protective effect of celastrol on CCl₄-induced acute liver damage, the WT mice were randomly divided into four groups (n = 5): (1) control group; (2) CCl₄ group; (3) CCl₄ + celastrol group; (4) celastrol group (Fig. 1D). The Ppara^−/− mice were randomly divided into three groups (n = 5): (1) control group; (2) CCl₄ group; (3) CCl₄ + celastrol group; (4) celastrol group.CCl₄ + celastrol and celastrol groups were orally treated with celastrol (10 mg/kg dissolved in 1% DMSO + 2% Tween 80 + 97% water) for five consecutive days [7]. After the mice were treated with celastrol for three days, the mice of CCl₄ and CCl₄ + celastrol groups were given a single intraperitoneal dose of CCl₄ (20% CCl₄ solution in corn oil, 1 ml/kg body weight) [19,20].

Fig. 1.

Celastrol attenuated CCl₄-induced liver injury in WT mice. (A) Phenotype and H&E staining of liver. (B) Serum AST, ALT, and ALP enzyme activities in control, CCl₄, and CCl₄ + celastrol groups. (C) Celastrol does not cause hepatotoxicity at the therapeutic dose (10 mg/kg). (D) Experimental scheme for animal experiments. All data are expressed as mean ± SEM (n = 5). *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant.

Experiment 2: To determine the role of PPARα on CCl₄-induced acute liver damage, the WT and Ppara^−/− mice were randomly assigned into four groups (n = 5), respectively: (1) control group; (2) CCl₄ group; (3) Ppara^−/− control group; (4) Ppara^−/− CCl₄ group (Fig. 1D). CCl₄ and Ppara^−/− CCl₄ groups were given a single intraperitoneal dose of CCl₄.

Experiment 3: To investigate the role of PPARα inhibition in the protective effect of celastrol, the WT mice were randomly assigned into four groups (n = 5): (1) control group; (2) CCl₄ group; (3) CCl₄ + celastrol group; (4) CCl₄ + celastrol + GW6471 group (Fig. 1D). The CCl₄ + celastrol group was treated with celastrol (10 mg/kg) for 5 consecutive days. The CCl₄ + celastrol + GW6471 group was cotreated with GW6471 (10 mg/kg dissolved in 4% DMSO + 2% Tween 80 + 94% normal saline, intraperitoneal administration, 30 min before celastrol) and celastrol for 5 consecutive days. After celastrol treatment for three days, mice in CCl₄, CCl₄ + celastrol, and CCl₄ + celastrol + GW6471 groups were given a single intraperitoneal dose of CCl₄.

All mice were anesthetized by CO₂ and killed 48 h after CCl₄ treatment. Whole blood was collected in anti-coagulative tubes. Plasma was got by centrifugation of the whole blood at 2000 g for 5 min at 4 °C. Partial liver tissue was stored at −80 °C, and partial liver tissue was preserved in 10% buffered formalin for histological analysis.

2.3. Biochemical analysis and histological examination

Assay kits for aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), catalase (CAT), and malondialdehyde (MDA) were obtained from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). The procedure was performed following by the according protocols. The liver tissues were fixed in 10% buffer formalin, which was processed by soaking in different alcoholic concentration gradient, cleared in xylene, and embedded in paraffin. Four μm sections were stained with hematoxylin and eosin and examined by light microscopy.

2.4. Sample preparation and metabolomics analysis

Samples of plasma and liver were prepared using a method described previously [21]. The working conditions of the LC-MS system described in a previous report [8]. A 5 μl aliquot of extract was injected into the UPLC-ESI-QTOFMS system. The chromatographic and spectral data were extracted by MassHunter Workstation software (Agilent, Santa Clara, CA, USA). The data matrix was processed using Mass Profinder software (Agilent, Santa Clara, CA, USA) and analyzed by SIMCA-P + 13.0 (Umetrics, Kinnelon, New Jersey, USA) for principal component analysis (PCA) and orthogonal projection to latent structures-discriminant analysis (OPLS-DA). HMDB was assisted to determine the chemical structures of changed metabolites, which were confirmed by comparing retention time and MS/MS fragmentation with authentic standards (Table 1).

Table 1.

Biomarkers recovered by celastrol.

No.	Identity	Rt (min)		Observed	Error (ppm)	Formula	MS/MS
		PR	HILIC
1	C8:0-carnitine	6.3	-	288.2169	0.3	C15H29NO4[H+]	144;85;60
2	C10:0-carnitine	7.4	-	316.2483	−1.3	C17H33NO4[H+]	257;144;85
3a	C12:0-carnitine	8.4	-	344.2783	−3.2	C19H37NO4[H+]	285;183;144;85;60
4	C13:0-carnitine	8.8	-	358.2952	0.6	C20H39NO4[H+]	299;144;85;60
5a	C14:0-carnitine	9.4	-	372.3109	0.5	C21H41NO4[H+]	313;211;144;85
6	C15:0-carnitine	9.7	-	386.3231	−8.3	C22H43NO4[H+]	327;144;225;85;60
7a	C16:0-carnitine	10.3	-	400.3423	0.7	C23H45NO4[H+]	341;239;144;85
8	C17:0-carnitine	10.8	-	414.3573	−0.7	C24H47NO4[H+]	355;85;60
9a	C18:0-carnitine	11.3	-	428.3711	−5.1	C25H49NO4[H+]	369;267;144;85
10	C8:1-carnitine	4.8	-	286.2011	−0.6	C15H27NO4[H+]	85;60
11	C10:1-carnitine	6.9	-	314.2329	1.0	C17H31NO4[H+]	85
12	C11:1-carnitine	7.6	-	328.2483	0.4	C18H33NO4[H+]	85;60
13	C12:1-carnitine	7.9	-	342.2657	5.8	C19H35NO4[H+]	283;85;60
14	C14:1-carnitine	8.9	-	370.2943	−1.9	C21H39NO4[H+]	311;209;144;85;60
15	C16:1-carnitine	9.7	-	398.3261	−0.5	C23H43NO4[H+]	339;237;144;85;60
16	C17:1-carnitine	10.1	-	412.3422	0.2	C24H45NO4[H+]	353
17	C18:1-carnitine	10.6	-	426.3573	−0.7	C25H47NO4[H+]	367;265;144;85;60
18	C20:1-carnitine	11.4	-	454.3884	−1.5	C27H51NO4[H+]	395;293;144;85;60
19	C10:2-carnitine	6.6	-	312.2173	1.4	C17H29NO4[H+]	85;60
20	C12:2-carnitine	7.4	-	340.2481	−0.2	C19H33NO4[H+]	85;60
21	C14:2-carnitine	8.3	-	368.2801	1.9	C21H37NO4[H+]	207;144;85;60
22	C15:2-carnitine	8.7	-	382.2965	3.5	C22H39NO4[H+]	221;85;60
23	C16:2-carnitine	9.2	-	396.3109	0.3	C23H41NO4[H+]	85;60
24	C17:2-carnitine	9.5	-	410.3264	−0.2	C24H43NO4[H+]	85;60
25	C18:2-carnitine	10.0	-	424.3443	5.4	C25H45NO4[H+]	365;263;144;85;60
26	C20:2-carnitine	10.8	-	452.3720	−3.1	C27H49NO4[H+]	393;144;85;60
27	C8:0-OH-carnitine	4.1	-	304.2113	−1.3	C15H29NO5[H+]	85;60
28	C10:0-OH-carnitine	6.3	-	332.2430	0.0	C17H33NO5[H+]	85
29	C12:0-OH-carnitine	7.4	-	360.2743	−0.6	C19H37NO5[H+]	199;144;85
30	C13:0-OH-carnitine	8.2	-	374.2907	1.7	C20H39NO5[H+]	85;60
31	C14:0-OH-carnitine	8.4	-	388.3052	−1.0	C21H41NO5[H+]	329;85;60
32	C16:0-OH-carnitine	9.3	-	416.3367	−0.5	C23H45NO5[H+]	357;85
33	C18:0-OH-carnitine	10.3	-	444.3683	0.2	C25H49NO5[H+]	283;85;60
34	C12:1-OH-carnitine	6.9	-	358.2583	−1.4	C19H35NO5[H+]	85;60
35	C14:1-OH-carnitine	8.0	-	386.2896	−1.0	C21H39NO5[H+]	327;225;85
36	C16:1-OH-carnitine	8.7	-	414.3223	2.4	C23H43NO5[H+]	253;144;85;60
37	C18:1-OH-carnitine	9.6	-	442.3534	1.8	C25H47NO5[H+]	383;281;144;85;60
38	C20:1-OH-carnitine	10.5	-	470.3844	0.8	C27H51NO5[H+]	85;60
39a	DCA	10.2	-	391.2857	0.8	C24H40O4[H−]	373;355;345;327
40a	Tα/βMCA	6.3	-	514.2825	−3.5	C26H45NO7S[H−]	124;80
41a	TCA	7.2	-	514.2873	5.8	C26H45NO7S[H−]	124;80
42a	TCDCA	8.1	-	498.2890	−1.0	C26H48NO6S[H−]	124;80
43a	THDCA	7.1	-	498.2893	−0.4	C26H48NO6S[H−]	124;80
44a	Glutamine	0.9	5.3	147.0762	−1.2	C5H10N2O3[H+]	84;56
45a	Serine	0.8	5.3	106.0497	−1.9	C3H7NO3[H+]	70;60;42
46a	Methionine	1.0	5.2	150.0583	0.0	C5H11O2NS[H+]	133;104;102;56
47a	Phenylalanine	1.0	5.0	166.0865	1.2	C9H11NO2[H+]	149;131;120;103;93;77
48a	Proline	0.9	5.5	116.0701	−4.3	C5H9NO2[H+]	58;59;70;74
49a	Threonine	0.8	5.4	120.0653	−1.7	C4H9NO3[H+]	102;84;74;56

^aConfirmed by authentic standards.

2.5. Gene expression analysis

Total RNA was extracted from frozen liver tissues or primary hepatocytes using TRIzol reagent (Life technology, Carlsbad, CA, USA). QPCR was carried out using SYBR green PCR master mix (Takara, Dalian, China) in a CFX Connect Real-Time System (Bio-Rad Laboratories, Hercules, CA, USA). QPCR primer sequences were shown in Table 2. All results were normalized to 18S mRNA. Thermal cycling condition was carried according to a previous study [8].

Table 2.

Primer sequences for QPCR.

Gene	Abbreviation	Sequence
Choline kinase α	Chkα	AAAGTGCTCTTGCGGCTCTA GACCTCTCTGCAAGAATGGC
Chemokine (C-C motif) ligand 2	Ccl2	AGGTCCCTGTCATGCTTCTG GGGATCATCTTGCTGGTGAA
Carnitine palmitoyltransferase 1	Cptlb	CCTCTCATGGTGAACAGCAA GGTCCAGTTTACGGCGATAC
Carnitine palmitoyltransferase 2	Cpt2	CAGCACAGCATCGTACCCA TCCCAATGCCGTTCTCAAAAT
Chemokine (C-X-C motif) ligand 1	Cxcl1	AACCGAAGTCATAGCCACAC CAGACGGTGCCATCAGAG
Chemokine (C-X-C motif) ligand 10	CxcllO	TCAGCACCATGAACCCAAG CTATGGCCCTCATTCTCACTG GGGAATGCCATTTACTTGGA GTCCGGATATTCAAGGATGC
Cholesterol 7α-hydroxylase	Cyp7a1
Sterol 12α-hydroxylase	Cyp8b1	TCCTCAGGGTGGTACAGGAG GATAGGGGAAGAGAGCCACC
Early growth response 1	Egrl	ACGACAGCAGTCCCATCTACTCGG GGACTCGACAGGGCAAGCATATGG
Glutathione peroxidase 2	Gpx2	GGGCTGTGCTGATTGAGA CGGACATACTTGAGGCTGTT
Glutathione peroxidase 3	Gpx3	GGCTTCCCTTCCAACC AATTTCTGCTCTTTCTCCC
Glutathione peroxidase 4	Gpx4	ACGATGCCCACCCACT CCACGCAGCCGTTCTT
Glutathione S-transferase α2	Gsta2	TTATGTCCCCCAGACCAAAG CCTGTTGCCCACAAGGTAGT
Glutathione S-transferase α4	Gsta4	AGACCACGGAGAGGCT CCTGACCACCTCAACATAGGG
Hydroxyacyl-CoA dehydrogenase	Hadha	AAGGGGATGTGGCAGTTATT ACTCCTGATTTGGTCGTTGG
Interleukin 1	Mb	CCCTGCAGCTGGAGAGTGTGGA TGTGCTCTGCTTGTGAGGTGCTG
Interleukin 6	H6	TGATGCACTTGCAGAAAACA ACCAGAGGAAATTTTCAATAGGC
Lysophosphatidylcholine acyltransferase 1	Lpcatl	CACGAGCTGCGACTGAGC ATGAAAGCAGCGAACAGGAG
Lysophosphatidylcholine acyltransferase 4	Lpcat4	GAGTTACACCTCTCCGGCCT GGCCAGAGGAGAAAGAGGAC
Medium-chain acyl-CoA dehydrogenase	Mcad	GCGAGCAGAAATGAAACTCC AGCTCTAGACGAAGCCACGA
Sodium taurocholate contransporting polypeptide	Ntcp	AGGGGGACATGAACCTCAG TCCGTCGTAGATTCCTTTGC
Organic anion transporting protein 1	Oatpl	ACTCCCATAATGCCCTTGG TAATCGGGCCAACAATCTTC
Organic anion transporting protein 4	Oatp4	ACCAAACTCAGCATCCAAGC TAGCTGAATGAGAGGGCTGC
Phosphate cytidylyltransferase 1 α	Pcyt1α	AGCCCTATGTCAGGGTGACT GGCATGACCAGAGTGAAACA
Phosphate cytidylyltransferase 1 β	Pcytlb	ATAGAGCACACATGCCCACA GGCAACGGTCAGTTTTTCAT
Phospholipase D1	Pldl	CTGCATCCTCAAACGGAAAG GCTTGCTGTACTCGCTGTTG
Peroxisome proliferator-activated receptor α	Pparα	CCCAAGGGAGGAATAGCTTCT CTCTGCGATGCGGTTCCAA
Peroxisome proliferator-activated receptor γ	Pparg	CCACCAACTTCGGAATCAGCT TTTGTGGATCCGGCAGTTAAGA
Sphingomyelinphosphodiesterase 3	Smpd3	CCTGACCAGTGCCATTCTTT AGAAACCCGGTCCTCGTACT
Tumour necrosis factor α	Tnfa	CCACCACGCTCTTCTGTCTAC AGGGTCTGGGCCATAGAACT
18S ribosomal RNA	18S	ATTGGAGCTGGAATTACCGC CGGCTACCACATCCAAGGAA

2.6. Primary mouse hepatocytes cultures and luciferase reporter assays

Primary mouse hepatocytes were isolated from 6-week-old 129/Sv mice as described previously [8]. To evaluate the effect of celastrol on PPARα signaling, hepatocytes were harvested after incubation with celastrol (30, 60, and, 120 nM) for 24 h [7,22]. To evaluate the function of DCA, TCA, THDCA, and TCDCA, primary mouse hepatocytes were harvested after incubation with celastrol (120 nM), DCA (50, 100, and 200 μM), TCA (100 μM), THDCA (100 μM), and TCDCA (100 μM) for 24 h [7,22,23].

For luciferase reporter gene assays, HEK293 cells were transfected with PPARα, PPRE-luciferase, and renilla-luciferase [24]. The transfection procedure was detailed in the Lipofectamine 2000 instruction manual (Invitrogen, Grand Island, NY). Transfected cells were treated with celastrol (30, 60, and, 120 nM) and fenofibrate (50 μM) for 24 h [7,22]. Luciferase activity was assayed in a Dual-luciferase Reporter Assay System.

2.7. Data analysis

All data were expressed as mean ± SEM. Statistical analysis was performed using the one-way ANOVA followed by Dunnett’s test. P value < 0.05 was considered statistically significant.

3. Results

3.1. Celastrol protected against CCl₄-induced acute liver injury

Celastrol, a pentacyclic triterpene isolated from the roots of the Tripterygium Wilfordi, has anti-inflammatory effects against various inflammatory diseases [1]. Usually, liver injury is accompanied by inflammatory infiltration. Therefore, it was predicted that celastrol could protect mice from acute liver injury induced by CCl₄. The hepatic phenotype revealed that the histology of the CCl₄ + celastrol group was similar to the control group. Celastrol alleviated the periportal parenchymal necrosis induced by CCl₄ (Fig. 1A). The increased AST, ALT, and ALP by CCl₄ in mice were significantly decreased by celastrol treatment (Fig. 1B). These results showed that celastrol could protect against CCl₄-induced acute liver injury.

3.2. Metabolic disorder of serum and hepatic metabolites was improved by celastrol

PCA modeling was used to analyze the serum data sets from the control, CCl₄, and CCl₄ + celastrol groups. The CCl₄ group was separated from the control and CCl₄ + celastrol groups, indicating that celastrol treatment significantly normalized the metabolites changed by CCl₄ (Fig. 2A). Three ions m/z 344.2794⁺, 372.3107⁺, and 398.3263⁺ were found to deviate from the ions cloud in OPLS-DA S-plot compared the control group with the CCl₄ group (Fig. 2B). Chemical formula calculation showed these three ions corresponded to C₁₉H₃₇NO₄, C₂₁H₄₁NO₄, and C₂₃H₄₃NO₄. These ions m/z 344.2794⁺ (Rt = 8.426), 372.3107⁺ (Rt = 9.376), and 398.3263⁺ (Rt = 9.677) were identified as C12:0-carnitine, C14:0-carnitine, and C16:1-carnitine based on their MS/MS fragmentation, respectively (Fig. 2B and Table 1). Targeted metabolomic analysis showed that celastrol decreased the levels of 29 long-chain acylcarnitines that were increased by CCl₄ (Fig. 2C). A previous study found that the increased long-chain acylcarnitine resulted from mitochondrial dysfunction and the inability of efficiently metabolize fatty acids [8]. The mRNA levels of carnitine palmitoyltransferase 1b (Cpt1b) and Cpt2 were increased by CCl₄, however, the levels of Ppara, medium-chain acyl-CoA dehydrogenase (Mcad), and hydroxyacyl-CoA dehydrogenase (Hadha) mRNAs remained unchanged in the CCl₄ group compared with the control group (Fig. 2D). Using non-targeted metabolomics analysis in positive and negative modes, other metabolites, such as six amino acids and five bile acids, were improved after celastrol treatment (Fig. 2C). The expression level of mRNAs encoded by genes involved in bile acid synthesis (sterol 12α-hydroxylase (Cyp8b1) and cholesterol 7α-hydroxylase (Cyp7a1)) and transport (sodium taurocholate cotransporting polypeptide (Ntcp), organic anion transporting polypeptide 1 (Oatp1), and Oatp4) were measured [8], and also improved by celastrol treatment (Fig. 2E). These results showed that celastrol could improve serum metabolites influenced by CCl₄.

Fig. 2.

Celastrol decreased the accumulation of acylcarnitines and bile acids induced by CCl₄ in serum. PCA score plot (A) and OPLS-DA S-plot (B) derived from LCMS data of serum ions in positive mode. Each point represented an individual mouse serum sample (-left) and an ion in the samples (-right). Metabolites were labeled in the S-plot (●, control group; ■, CCl₄ group; ▲, CCl₄ + celastrol group). (C) Heat map analysis of the relative abundance of long-chain acylcarnitines, amino acids, and bile acids in serum of control, CCl₄, and CCl₄ + celastrol groups. QPCR analysis was performed to measure the mRNAs coded by acylcarnitine-related genes and Pparg (D) and bile acid-related genes in liver (E). All data were repressed as mean ± SEM (n = 5). Value represents fold change after normalization to control. *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant.

PCA and OPLS-DA models were then used to analyze metabolites from livers of control, CCl₄, and CCl₄ + celastrol groups. Significant differences in the hepatic metabolites between the CCl₄ and control groups were found, including increased acylcarnitines, lyso-phosphocholine 18:2 (LPC18:2), and lyso-phosphatidylethanolamine 22:6 (LPE22:6) that largely contributed to the separation (Fig. 3A,B). Further acylcarnitine targeted analysis indicated that the levels of 34 medium-and long-chain acylcarnitines that were increased in the CCl₄ group were significantly decreased after celastrol treatment (Fig. 3C). The levels of LPE22:6, LPC16:0, and LPC18:3 were recovered by celastrol (Fig. 3D). The mRNAs associated with LPC metabolism (lysophosphatidylcholine acyltransferase 1 (Lpcat1) and Lpcat4), PC synthesis (choline kinase α (Chka) and phosphate cytidylyltransferase 1 α (Pcyt1a)), PC metabolism (phospholipase D1 (Pld1)), and SM metabolism (sphingomyelin phosphodiesterase 3 (Smpd3)) also improved by celastrol (Fig. 3E). These results showed that celastrol improved hepatic metabolites influenced by CCl₄.

Fig. 3.

Celastrol decreased the accumulation of acylcarnitines and lipids induced by CCl₄ in liver. PCA score plot (A) and OPLS-DA S-plot (B) derived from LC-MS data of hepatic ions in positive mode. Each point represented an individual mouse hepatic sample (-left) and an ion in the samples (-right). Metabolites were labeled in the S-plot (●, control group; ■, CCl₄ group; ▲, CCl₄ + celastrol group). (C) Heap map analysis of the relative abundance of medium- and long-chain acylcarnitines in liver of control, CCl₄, and CCl₄ + celastrol groups. (D) Celastrol decreased LPEs and LPCs levels in liver. (E) Lipid-related mRNAs were attenuated after celastrol treatment in liver. All data are expressed as mean ± SEM (n = 5). Values represent fold change after normalization to control. **P < 0.01, ***P < 0.001, ns = not significant.

3.3. Inflammatory cytokine and oxidative stress in acute liver injury were decreased by celastrol

Increased serum acylcarnitines is an indication of mitochondrial dysfunction, which induces oxidative stress in vitro [8], suggesting that the increase of acylcarnitines in CCl₄-induced liver injury resulted in increased oxidative stress. Therefore, oxidative stress was evaluated. Hepatic CAT and MDA that were increased in the CCl₄ group, were decreased after celastrol treatment (Fig. 4A). The expression levels of several anti-oxidative gene mRNAs that were increased in the CCl₄ group, were lower after celastrol treatment, including glutathione peroxidases (glutathione peroxidase 2 (Gpx2), Gpx3, and Gpx4) and glutathione S-transferases (glutathione S-transferase α2 (Gsta2) and Gsta4) (Fig. 4B). Early growth response 1 (EGR1) pathway analysis indicated that the up-regulated Egr1 mRNA and its downstream inflammatory cytokines (chemokine (C-X-C motif) ligand 1 (Cxcl1), chemokine (C–C motif) ligand 2 (Ccl2), Cxcl10, tumor necrosis factor α (Tnfa), and interleukin 6 (Il6)) mRNAs in the CCl₄ group were improved by celastrol treatment (Fig. 4C). After celastrol treatment, the expression of Egr1, Cxcl1, Ccl2, Cxcl10, Tnfa, and Il6 mRNAs were reduced 43.5%, 65.2%, 76.2%, 92.8%, 52.4%, and 53.1%, respectively, compared with the CCl₄ group. Celastrol did not reverse the expression of Il1b mRNA (Fig. 4C), which was observed in ANIT-induced cholestasis [7]. Whether celastrol can directly combined with these inflammatory cytokines, needs further studies. These results showed that celastrol reduced inflammatory cytokine expression and oxidative stress induced by CCl₄.

Fig. 4.

Celastrol eliminated oxidative stress and activated the PPARα signaling pathway. (A) Hepatic MDA and CAT levels in control, CCl₄, and CCl₄ + celastrol groups. (B) QPCR analysis of the mRNA expression of hepatic Gpx and Gst isoforms. (C) QPCR analysis of the mRNA expression of Egr1 and its downstream genes in liver. **P < 0.01 and ***P < 0.001 verse control; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001, and ns means not significant verse CCl₄. (D) QPCR analysis of the gene expression of PPARα and its target genes in primary mouse hepatocyte after celastrol treatment for 24 h in vitro. (E) Luciferase assays of the activation of PPARα in HEK293 cells. All data are expressed as mean ± SEM (n = 5). Values represent fold change after normalization to control. *P < 0.05, **P < 0.01, ns = not significant.

3.4. Celastrol activates PPARα signaling pathway

Since the levels of acylcarnitine and lipids were modulated by celastrol, and the PPARα signaling pathway participates in the metabolism of acylcarnitines and lipids [11], the effect of celastrol on PPARα signaling was investigated. Low concentrations of celastrol (120 nM) could activate PPARα and increase its target gene mRNAs Cpt1b, Cpt2, Mcad, and Hadha in primary mouse hepatocytes after a 24 h exposure (Fig. 4D). Dual-luciferase reporter gene assays performed with HEK293 cells co-transfected with PPARα and PPRE-luciferase expression plasmid, demonstrated that 120 nM celastrol significantly increased the luciferase reporter gene activity (Fig. 4E). These results demonstrated a positive regulatory role of celastrol on PPARα signaling.

Furthermore, the role of PPARα in the protective effects of celastrol in the CCl₄-induced liver damage was explored using Ppara^−/− mice and a PPARα antagonist. H&E staining revealed that the CCl₄-induced liver injury was not attenuated by celastrol in Ppara^−/− mice (Fig. 5A). The decreased AST, ALT, and ALP levels after celastrol treatment in WT mice were not observed in the Ppara^−/− mice (Fig. 5B). Serum and hepatic metabolomics analysis in positive and negative modes showed that the levels of metabolites, such as acylcarnitines, bile acids, and amine acids, in the CCl₄ + celastrol group were similar to the CCl₄ group metabolites in Ppara^−/− mice (Figs. 5C–E and 6A–E). Although celastrol increased AST and ALT levels in Ppara^−/− mice (Fig. 5B), the ALP levels, histologic injury, and level of metabolites were not increased by celastrol (Figs. 5 and 6A–E). Hepatotoxicity was also not observed in WT mice treated for 5 days with celastrol (Fig. 1C). These data suggested that the levels of ALT and AST were more sensitive to celastrol exposure when PPARα signaling is low under pathological conditions. At the same time, the protective effect of celastrol was attenuated in WT mice after cotreatment with the PPARα antagonist GW6471 (Fig. 6F). These results showed that the protective effect of celastrol on liver injury was via activation of PPARα signaling.

Fig. 5.

Role of PPARα in celastrol protected against liver injury using Ppara^−/− mice. (A) H&E staining of liver. (B) Serum AST, ALT, and ALP enzyme activities. PCA score plot derived from LC-MS data of serum ions in both positive (C) and negative (D) modes. Each point represents an individual mouse serum sample (●, control group; ■, CCl₄ group; ▲, CCl₄ + celastrol group). (E) Serum acylcarnitine levels in Ppara^−/− mice. All data are expressed as mean ± SEM (n = 5). *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant.

Fig. 6.

Role of celastrol was dependent on PPARα using Ppara^−/− mice and PPARα inhibition GW6471. Serum bile acids (A) and amino acids (B) levels in Ppara^−/− mice. PCA score plot derived from LC-MS data of liver ions in both positive (C) and negative (D) modes. Each point represented an individual mouse serum sample (●, control group; ■, CCl₄ group; ▲, CCl₄ + celastrol group). (E)Hepatic acylcarnitine levels in Ppara^−/− mice. (F) The protective effect of celastrol was attenuated after GW6471 cotreatment. All data are expressed as mean ± SEM (n = 5). *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant.

3.5. Deficiency of PPARα increased CCl₄-induced liver injury

Ppara^−/− mice were used to evaluate the role of PPARα in CCl_4-induced acute liver injury. Histology analysis showed that CCl₄ induced obvious parenchymal necrosis in WT mice, and the parenchymal necrosis was more severe in Ppara^−/− mice (Fig. 7A). The levels of AST, ALT, and ALP in CCl₄-induced liver injury were higher in Ppara^−/− mice compared with WT mice (Fig. 7B). Serum metabolomics was used to determine the differences among control, CCl₄, Ppara^−/− control, and Ppara^−/− CCl₄ groups (Fig. 7C). The four top increased ions 391.2854⁻, 498.2895⁻, 514.2843⁻, and 514.2843⁻ were observed in control group compared with CCl₄ group (Fig. 7D). These ions were identified as DCA, THDCA, TCA, and Tα/βMCA, respectively. Bile acid analysis revealed that the increase of bile acids in the CCl₄ group was further increased in the Ppara^−/− CCl₄ group, especially DCA (Fig. 7E). Other bile acids such as Tα/βMCA, TCA, THDCA, and TCDCA were not significantly increased in the Ppara^−/− CCl₄ group compared with the CCl₄ group (Fig. 7E). The mRNA levels produced by bile acids synthesis and transport genes were further decreased in Ppara^−/− mice treated with CCl₄ (Fig. 7F). DCA, THDCA and TCDCA could increase the expression of Egr1 mRNA and its downstream inflammatory cytokine Cxcl10 mRNA in mouse primary hepatocytes (Fig. 8A,B). Celastrol reversed the down-regulation of cell viability induced by DCA (Fig. 8C), and inhibited the increase of Egr1 and Il6 mRNA expression (Fig. 8D). These results indicated that PPARα plays an important role in CCl_4-induced liver injury, and the potentiation of CCl₄-induced liver injury in Ppara^−/− mice might be due to DCA-EGR1-inflammatory factor pathway.

Fig. 7.

Role of PPARα in CCl₄-induced liver injury using Ppara^−/− mice. (A) H&E staining of liver. (B) Serum AST, ALT, and ALP enzyme activities. PCA score plot (C) and OPLS-DA S-plot (D) derived from LC-MS data of serum ions. Each point represented an individual mouse serum sample (-up) and an ion in the sample (-down). Metabolites were labeled in the OPLS-DA S-plot. (E) Serum bile acids levels in CCl₄-induced liver injury. (F) QPCR analysis of the hepatic mRNA expression of bile acids synthesis and basolateral uptake transporters. All data are expressed as mean ± SEM (n = 5). Values represent fold change after normalization to control. *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant.

Fig. 8.

Bile acids, especially DCA, activated EGR1-inflammatory factor axis. (A) DCA increased Egr1 mRNA and its downstream mRNAs Cxcl10 in primary mouse hepatocyte. (B) THDCA and TCDCA increased Egr1 mRNA in primary mouse hepatocyte. (C) Celastrol reversed the down-regulation of cell viability induced by DCA in primary mouse hepatocyte. (D) Celastrol reversed the increase of Egr1 and its downstream genes in primary mouse hepatocyte. (E) Proposed mechanism of hepatoprotective effect of celastrol against liver injury. All data are expressed as mean ± SEM (n = 5). Value represented fold change after normalization to control. *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant.

4. Discussion

The present study demonstrated a protective role for celastrol in CCl₄-induced acute liver injury. The CCl₄ model is frequently-used to study hepatotoxicity and liver fibrosis and to evaluate the hepatoprotective effects of drugs or natural products. One of the intriguing findings in this study was that celastrol could activate PPARα in primary hepatocytes and luciferase reporter gene assays. Further studies using Ppara^−/− mice revealed a protective role for celastrol dependent on PPARα activation (Fig. 8E).

Long-chain acylcarnitines are associated with the cellular stress response and activate some receptors, such as TLR2, NFκB, and JNK [25]. A previous study found that elevated long-chain acylcarnitines protected TP-induced liver injury through activation of the NOTCH-NRF2 pathway and induced a defense response against liver redundant line feed injury in vitro and in vivo [9]. More importantly, acylcarnitines that result from impaired mitochondrial function and fatty acid β-oxidation, were common biomarkers for liver injury. Acylcarnitines levels were increased and mitochondrial fatty acid β-oxidation was inhibited in triptolide-, ANIT-, APAP-, and sunitinib-induced liver injury [8–10,26]. Serum acylcarnitines levels were used as clinical biomarkers for screening inborn genetic defects in fatty acid β-oxidation [27]. The current study found that acylcarnitine levels were significantly increased after CCl₄ treatment, indicating that mitochondrial fatty acid β-oxidation was disrupted. It was reported that PPARα activation could reduce the accumulation of acylcarnitines and increased mitochondrial fatty acid β-oxidation in mice administered ANIT and cocaine [8,28]. Therefore, PPARα activation may be considered as a therapeutic target for the treatment of chemical-induced liver injury.

PPARα is expressed in metabolically-active tissues and regulates genes involved in mitochondrial and peroxisomal fatty acid β-oxidation, bile acid and amino acid metabolism, and inflammation [11]. PPARα also plays an important role in liver injury. PPARα deficiency potentiates chemically-induced hepatic injury. PPARα expression was lower in patients with hepatitis C virus (HCV) infection and steatohepatitis [29,30]. Ppara^−/− mice treated with cholic acid had disrupted bile acid and phospholipids homeostasis, indicating that PPARα was an essential regulator of bile acid synthesis and secretion [31]. Activation of PPARα may protect against ANIT-induced cholestasis [8], sunitinib-induced liver injury [10], concanavalin A- or diet-induced hepatitis [32,33], triptolide-induced liver injury [9], and pyrazinamide-induced hepatotoxicity [34]. Many traditional Chinese medicines protect against liver injury through activating PPARα signaling. For example, formononetin could activate PPARα, thereby inhibiting hepatic inflammation in cholestasis [35]. Nuciferine attenuated hepatic steatosis by activating the PPARα/PGC1α pathway [36]. In the present study, in vitro data demonstrated that celastrol was a PPARα agonist. The protective role of celastrol disappeared in CCl₄-induced liver injury when celastrol was co-treated with the PPARα antagonist GW6471. More importantly, celastrol could not protect against CCl₄-induced liver injury in Ppara^−/− mice, demonstrating the actual role of PPARα in the protective effect of celastrol. The activation of the Cpt1b and Cpt2 expression by celastrol would contribute to the elimination of acylcarnitines and improve the disruption of mitochondrial fatty acid β-oxidation by CCl₄. Therefore, this study provided evidences that the PPARα signaling has an important role in the protective effects of celastrol against chemically-induced liver injury.

Generally, bile acids facilitate the digestion and absorption of fat. But excessive accumulation of bile acids can result in apoptosis and inflammation in vivo and in vitro. It was reported that DCA could induce apoptosis in hepatocytes upon activation of death receptors [37]. DCA enhanced miR-34a/SIRT1/p53 proapoptotic signaling in a dose and time-dependent manner [23]. Furthermore, bile acids, such as DCA, chenodeoxycholic acid (CDCA), and TCA, up-regulated EGR1, which then regulated production of inflammatory mediators [38]. A previous study found that celastrol modulated the expression of pro-inflammatory cytokines [1]. Therefore, we hypothesized that celastrol could protect DCA-induced inflammatory. Indeed, DCA elevated the expression of Egr1 mRNA and the inflammatory factors Cxcl10 mRNAs. Celastrol improved the hepatic EGR1-inflammatory factor pathway that was elevated by CCl₄ treatment. Several bile acids, including DCA, THDCA, and TCDCA, could activate EGR1-infalmmaroty factor pathway in the present study. Compared with the WT mice, the increase of DCA was higher than other bile acid metabolites in Ppara^−/− mice after CCl₄ treatment, suggesting that DCA likely contributed to increased inflammation.

In this study, the protective role of celastrol in acute liver injury induced by CCl₄ exposure was demonstrated by histological and biochemical analysis, and the mechanism was deciphered using UPLC-ESI-QTOFMS-based metabolomics, primary hgepatocyte cultures, reporter gene assays, and Ppara^−/− mice. Finally, celastrol decreased proinflammatory cytokines and oxidative stress, and recovered the bile acid and acylcarnitine homeostasis from mice treated with CCl₄. This study revealed that PPARα activation by celastrol significantly attenuated CCl₄-induced liver injury.

Abbreviations:

ANIT

α-naphthyl isothiocyanate

ALP

alkaline phosphatase

ALT

alanine aminotransferase

APAP

acetaminophen

AST

aspartate aminotransferase

CAT

catalase

Ccl2

chemokine (C-C motif) ligand 2

CCl₄

carbon tetrachloride

Chka

choline kinase α

Cpt2

carnitine palmitoyltransferase 2

Cxcl1

chemokine (C-X-C motif) ligand 1

Cyp7a1

cholesterol 7α-hydroxylase

Cyp8b1

sterol 12α-hydroxylase

DCA

deoxycholic acid

Egr1

early growth response 1

Gpx2

glutathione peroxidase 2

Gsta2

glutathione S-transferase α2

Hadha

hydroxyacyl-CoA dehydrogenase

Il6

interleukin 6

Lpcat1

lysophosphatidylcholine acyltransferase 1

Mcad

medium-chain acyl-CoA dehydrogenase

MDA

malondialdehyde

Ntcp

sodium taurocholate cotransporting polypeptide

Oatp1

organic anion transporting polypeptide 1

OPLS-DA

orthogonal projection to latent structures-discriminant analysis

PCA

principal component analysis

Pcyt1a

phosphate cytidylyltransferase 1α

Pld1

phospholipase D1

Ppara-null mice

Ppara^−/− mice

PPARα

peroxisome proliferator-activated receptor α

Smpd3

sphingomyelin phosphodiesterase 3

TCA

taurocholic acid

TCDCA

taurochenodeoxycholic acid

THDCA

taurohyodeoxycholic acid

Tnfa

tumour necrosis factor α

WT

wild type

12:0-carnitine

lauroylcarnitine

14:0-carnitine

myristoylcarnitine

16:0-carnitine

palmitoylcarnitine

18:0-carnitine

stearoylcarnitine