Cyclophilin D deficiency attenuates mitochondrial perturbation and ameliorates hepatic steatosis

ANIMALS

Eight-week-old C57BL/6j male mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd., and housed in colony cages with a 12-hour light/dark cycle in a temperature-controlled environment. For high-fat diet (HFD) feeding experiments, mice were fed a diet containing 60 kcal% fat (Research Diets; D12492). All mice were killed from 10:00 pm to midnight and the liver tissues acquired. The protocol was approved by the Research Ethics Committee of Shandong Provincial Hospital.

Ppif^–/– mice and Srebp1c^–/– mice were kindly provided by Dr. Heng Du and Prof. Youfei Guan (Dalian Medical University, China), respectively. These mice were backcrossed to the C57BL/6j strain for 10 generations.

Liver-specific Ire1α knockout (Ire1α-LKO) mice were produced by intercrossing Ire1α^flox/flox mice, which were kindly provided by Prof. Yong Liu (Chinese Academy of Sciences, China), with Alb-Cre transgenic mice.

Mouse adenoviral (Ad) PPIF (CypD overexpression adenovirus) and adenoviral enhanced green fluorescent protein (AdEGFP; empty vector adenovirus) were purchased from Hanbio Biotechnology Co., Ltd.; dissolved in sterile phosphate-buffered saline (PBS); and injected through the caudal vein. The amount of viral particles used in these experiments was 2 × 10⁸ pfu per mouse, given three times with a 6-day interval.

Clinical-grade cyclosporine A (CsA; Novartis) was diluted fresh in PBS. PBS or CsA (10 mg • kg^–1 • day^–1 or 20 mg • kg^–1 • day^–1) was injected intraperitoneally daily to C57BL/6j mice for 6 weeks. Tunicamycin (Tm; Sigma; no. T7765) was soluble in dimethyl sulfoxide at 10 mg • mL^–1 and diluted fresh in PBS. PBS or Tm (1 mg • kg^–1 • day^–1) was injected intraperitoneally daily to Ppif^–/– (CypD knockout) mice and Ppif^+/+ mice for 2 days.

STATISTICAL ANALYSES

The analyses were performed using SPSS 19.0 software (SPSS, Chicago, IL). The results were reported as the mean ± SD. Comparison of different groups was performed using one-way analysis of variance. Two-tailed P < 0.05 was considered significant.

Further materials and methods are described in the Supporting Information.

HEPATIC STEATOSIS IS SECONDARY TO THE INCREASE OF CypD EXPRESSION

To assess the role of CypD in the etiology of hepatic steatosis, we fed mice a chow diet or the HFD containing 60 kcal% fat for different time periods, and the mice were examined at the end of the second, fourth, eighth, and twelfth weeks. The in vivo imaging analysis (Fig. 1A) showed that the luciferase activity of CypD was increased as early as the end of week 4 and gradually enhanced as time passed. Elevated CypD protein expression also started from the end of week 4, as shown by western blot analysis (Fig. 1B). However, hepatic TG deposition and steatosis, corroborated by gross morphology and hematoxylin and eosin staining, were observed until the end of the eighth week of HFD feeding and continued to increase (Fig. 1C,D). The hepatic total cholesterol (TC) content was not significantly changed (data not shown). The results showed that increased CypD expression occurred earlier than the TG accumulation in the liver, suggesting a potential relationship between CypD and the etiology of steatosis.

ABLATION OF CypD PREVENTS DIET-INDUCED HEPATIC MITOCHONDRIAL STRESS AND STEATOSIS

In view of the increased expression of CypD associated with hepatic TG contents, we explored whether CypD serves as a mitochondrial target potentiating diet-induced steatosis. Hepatic lipid accumulation induced by the HFD was ameliorated in Ppif^–/– (CypD deficiency) mice compared with Ppif^+/+ (littermate counterparts) mice (Fig. 2A). And the hepatic TG levels in Ppif^–/– mice were lower than those in Ppif^+/+ mice, while this difference was enlarged by HFD treatment (Fig. 2B). However, the hepatic TC level showed no significant change (Fig. 2C). In addition, the serum TG and fasting blood glucose levels were decreased in Ppif^–/– mice with the HFD, while the serum TC level was not significantly changed (Supporting Fig. S1A,B). This murine model also revealed increased phosphorylation of kinases associated with insulin signaling, including Akt and glycogen synthase kinase 3β, in the liver (Supporting Fig. S1C). On the other hand, no significant changes in serum TG and TC levels were observed between Ppif^–/– and Ppif^+/+ mice on the chow diet (Supporting Fig. S1D).

To confirm whether remission of steatosis in Ppif^–/– mice was related to the promotion of mitochondrial function, we detected hepatic mitochondrial stress in HFD-induced mice. Mitochondrial swelling is an index that reflects mPTP opening and mitochondrial function.14 Hepatic mitochondria of Ppif^–/– mice were more resistant to swelling and permeability transition in response to Ca²⁺ than those of Ppif^+/+ mice. The mitochondria in HFD-treated Ppif^+/+ mice showed greater swelling than those in the corresponding chow group, but this difference was eliminated between Ppif^–/– mice fed the chow diet and those fed the HFD diet, implying the indispensable role of CypD in HFD-induced swelling (Fig. 2D). Further, transmission electron microscopic analysis showed that mitochondrial swelling, vacuolation, and destruction of cristae in HFD-fed mouse livers were attenuated in the liver of Ppif^–/– mice (Fig. 2E). Mitochondrial swelling reflects excessive mPTP opening, which results in mitochondrial ROS generation, cytochrome c release and cell apoptosis.19, 20 To evaluate mitochondrial ROS generation, we used MitoSox red staining, a unique fluorogenic dye that allows for selective detection of superoxide production in the mitochondria. Ppif^–/– mice showed much less MitoSox staining compared to Ppif^+/+ mice (Fig. 2F), indicating that the absence of CypD reduced HFD-induced mitochondrial ROS generation.

It is well known that approximately 90% of mammalian oxygen consumption (VO₂) is mitochondrial and that an increase in the VO₂ level is a direct reflection of increased mitochondrial oxidative metabolism.15 We also checked the contribution of CypD to VO₂ and energy expenditure (H) at the whole-body level using metabolic cage analysis. Ppif^–/– mice fed the HFD have lower body weight and percentage of fat mass, which may be influenced by an increase in physical activity,15 while food intake was the same as that in Ppif^+/+ mice (Supporting Table S1). Higher VO₂ and energy expenditure (H) in Ppif^–/– mice fed the HFD indicated accelerated fatty acid oxidation (Supporting Fig. S1E). The higher body temperature found in Ppif^–/– mice on the HFD suggests that thermogenic mechanisms may be more active in these mice to maintain homeothermy (Supporting Table S1). Metabolic cage results showed no clear changes between Ppif^–/– and Ppif^+/+ mice on the chow diet (Supporting Fig. S1F and Table S1). Taken together, these results showed that CypD deficiency alleviated HFD-induced mitochondrial dysfunction and hepatic steatosis.

HEPATIC CypD OVEREXPRESSION INDUCED MITOCHONDRIAL STRESS AND LIPID ACCUMULATION

In order to directly determine the role of CypD in hepatic TG accumulation, we overexpressed CypD in mouse liver by infecting AdPPIF (adenoviral overexpression of CypD) through the caudal vein. Compared to the control group, the alanine aminotransferase and aspartate aminotransferase levels in the CypD overexpression group showed no significant changes, which means that CypD adenoviral infection did not result in obvious hepatic toxicity (Supporting Table S2). We verified the localization in the mitochondria and expression of CypD by double immunofluorescence staining (Supporting Fig. S2A) and western blot (Fig. 3A), respectively. Compared to AdEGFP (empty vector adenovirus), AdPPIF in the mice did not change the physical activity, body weight, percentage of fat mass, and VO₂ significantly (Supporting Fig. S2B and Table S1) but increased lipid accumulation (Fig. 3B-D). Serum free fatty acid, TG, and TC levels did not show a clear change with CypD overexpression adenovirus injection (Supporting Table S2). Furthermore, in vivo overexpression of CypD aggravated mitochondrial swelling (Fig. 3E) and increased mitochondrial ROS generation (Fig. 3F), suggesting more serious mitochondrial dysfunction. All of the above results showed that hepatic TG accumulation depended on mitochondrial CypD.

CypD MEDIATES HEPATIC TG ACCUMULATION BY UP-REGULATING SREBP-1c

To explore the mechanism of CypD-induced TG accumulation, we performed RT-PCR array to analyze the expression of fatty liver disease–focused genes involved in TG biosynthesis, fatty acid uptake, and oxidation (Supporting Table S3). The representative genes were validated by real-time quantitative PCR. It was demonstrated that the mRNA expression of synthesis and oxidation-related genes was significantly changed in Ppif^–/– (CypD deficiency) mice fed the HFD for 8 weeks (Fig. 4A). SREBP-1c is the key transcription factor of lipogenesis, and we further detected the protein expression of the precursor and mature forms. Mature SREBP-1c could translocate into the nucleus and bind to specific sterol regulatory element DNA sequences, leading to the up-regulation of genes involved in TG biosynthesis. CypD was increased as early as the end of the fourth week of HFD treatment (Fig. 1B). And mature SREBP-1c expression was increased at the end of the eighth week of HFD treatment (Fig. 4B), while CypD deficiency eliminated HFD-induced SREBP-1c up-regulation (Fig. 4C), indicating the potential regulation of SREBP-1c by CypD.

Further, the promoter activity and protein expression of SREBP-1c in AdPPIF (CypD overexpression adenovirus)–infected mice were increased compared with the AdEGFP (empty vector adenovirus) group (Fig. 4D,E). To determine whether SREBP-1c was indispensable for CypD-induced TG accumulation, Srebp1c^–/– and Srebp1c^+/+ (littermate counterparts) mice were injected with AdPPIF and AdEGFP, respectively. Unlike the littermate Srebp1c^+/+ mice, the lipid contents in Srebp1c^–/– mice were not increased even in those treated with AdPPIF (Fig. 4F), which supported the idea that CypD stimulates TG accumulation through SREBP-1c.

INOSITOL-REQUIRING ENZYME 1α IS INVOLVED IN THE PROCESS OF CypD-STIMULATED SREBP-1c EXPRESSION

Based on the above results, we wondered how CypD affected SREBP-1c. CypD-induced ROS accumulation could activate the downstream p38 MAPK signaling pathway in synaptic damage.21 In our study, we found that the phosphorylation of p38 MAPK was suppressed in Ppif^–/– (CypD deficiency) mice (Fig. 5A). However, phosphorylation of both extracellular signal–regulated kinase and c-Jun N-terminal kinase, the other two important members of the MAPK family, was not changed. These results supported the idea that CypD-induced mitochondrial stress might affect p38 MAPK. Previous studies showed that p38 MAPK might associate mitochondrial dysfunction with ER stress.22-24 We found that ER dilatation was attenuated in Ppif^–/– mice by transmission electron microscopy (Fig. 5B). The phosphorylation of inositol-requiring enzyme 1α (IRE1α), a representative ER stress–related protein, was down-regulated in Ppif^–/– mice, while phospho-PERK, phospho-eIF2α, and ATF6 were not changed (Fig. 5C). Tm, one of the most common ER stress inducers,25 was used as a positive control in our study to test the sensitivity of Ppif^–/– mice to the ER-active chemical. Tm could induce the activation of all three unfolded protein response arms of ER stress in Ppif^–/– mice; more importantly, SREBP-1c expression could also be activated by Tm in Ppif^–/– mice (Supporting Fig. S3). Our results indicated that CypD might cause ER stress through p38 MAPK activation.

ER stress can trigger SREBP-1c expression and/or activation.26-28 Next, to determine the role of IRE1α in CypD-induced SREBP-1c expression, we established an Ire1α-LKO mouse model. Ire1α-LKO mice and littermate counterpart (flox/flox) mice were injected with AdPPIF (adenoviral overexpression of CypD) and AdEGFP (empty vector adenovirus), respectively. In flox/flox littermates, both the hepatic mitochondria and ER were destroyed by AdPPIF injection relative to AdEGFP (Fig. 5D, b versus a), while AdPPIF-infected Ire1α-LKO mice had abnormal mitochondria compared to the AdEGFP-infected Ire1α-LKO group, but no significant difference was observed in the ER (Fig. 5D, d versus c). Moreover, the expression of hepatic SREBP-1c (Fig. 5E,F) and lipid accumulation (Fig. 5G) in Ire1α-LKO mice were not increased even after treatment with AdPPIF compared to littermate flox/flox mice. These results suggested the indispensable role of IRE1α in CypD-induced SREBP-1c expression.

Activated IRE1 can cleave the mRNA encoding a specific transcription factor called XBP1, giving rise to an active spliced mRNA (XBP1s) that regulates the transcription of genes involved in protein folding.25 Our results showed that IRE1α activation and Xbp1 mRNA splicing were increased when hepatic steatosis was aggravated (Supporting Fig. S4), which is in accordance with the study of Lee et al.29

CypD INCREASES SREBP-1c EXPRESSION THROUGH Ca²⁺/p38 MAPK/IRE1α SIGNALING

Considering the complexity of the in vivo experiments, we set up palmitate-treated HepG₂ cells and found that lipid accumulation was significant at the end of the 12 hours (Fig. 6A) but that corresponding expression of CypD was up-regulated as early as the end of the 6 hours before detectable lipid accumulation (Fig. 6B,C). Similarly, mitochondrial ROS generation was increased following the CypD change (Supporting Fig. S5A). As soon as CypD expression was increased, the phosphorylation level of p38 MAPK was elevated at the end of 6 hours, while IRE1α and SREBP-1c were changed at the end of 12 hours (Supporting Fig. S5B). When PPIF small interfering RNA (siRNA; the siRNA of CypD) was used, the phosphorylation levels of p38 MAPK and IRE1α and the expression level of SREBP-1c were suppressed with palmitate treatment (Fig. 6D). These results verified that CypD-induced mitochondrial stress occurred before detectable TG accumulation.

Next, pLV-PPIF (CypD overexpression plasmid) was used to confirm the effects of CypD on TG accumulation in HepG₂ cells and mouse primary hepatocytes, respectively (Fig. 6E; Supporting Fig. S6A). In HepG₂ cells, the pLV-PPIF plasmid increased mitochondrial ROS generation (Fig. 6F) and the intracellular concentration of Ca²⁺ (Fig. 6G), which was consistent with the result that CypD-induced ROS accumulation may cause Ca²⁺ balance disruption.21 CypD overexpression also aggravated mitochondrial dysfunction, including decreasing the oxygen consumption ratios of spare respiratory capacity and adenosine triphosphate production (Supporting Fig. S5C,D), which identified its role in mitochondrial respiratory function. Overexpression of CypD caused TG deposition (Fig. 6H) and increased mRNA and protein expression of SREBP-1c (Fig. 6I,J). Betulin, an inhibitor of SREBP maturation,30 impeded the effect of CypD on SREBP-1c maturation (Fig. 6J). At the same time, we found that CypD could specifically activate the IRE1α branch without changing the other two pathways in mouse primary hepatocytes from Ire1α-LKO and flox/flox mice (Supporting Fig. S7), so an IRE1α inhibitor (4μ8C) was used in our study. SB203580 (p38 MAPK inhibitor) or 4μ8C significantly decreased the expression of SREBP-1c promoted by CypD overexpression in HepG₂ cells (Fig. 6K,L).

Similar to the results for HepG₂ cells, in primary hepatocytes SB203580 or Ire1α-LKO significantly decreased the expression of SREBP-1c and TG accumulation promoted by CypD overexpression plasmid (Supporting Fig. S6B-G). These results indicated that Ca²⁺/p38 MAPK/IRE1α signaling might be the connection point between CypD and SREBP-1c expression. Given that activation of the ER stress response might parallel that of p38 MAPK, we tested the effect of p38 MAPK inhibitor on ER stress in primary hepatocytes (Supporting Fig. S6D). The results indicated that p38 MAPK inhibitor (SB203580) can suppress the IRE1α branch and decrease SREBP-1c expression promoted by CypD overexpression. On the other hand, in hepatocytes treated with SB203580, the effect of CypD on IRE1α phosphorylation also disappeared, while in Ire1α-deficient hepatocytes, CypD still could increase the expression of phospho-p38, which confirmed that Ire1α was the downstream target of the p38 signaling pathway (Supporting Fig. S6G).

PREVENTION AND THERAPEUTIC EFFECT OF CypD INHIBITOR IN HEPATIC STEATOSIS

Considering the effect of hepatic CypD on hepatic TG accumulation, we applied clinical-grade CsA (Sandimmune), an inhibitor of CypD, to determine the preventive and therapeutic effects of CypD as a potential target in NAFLD.

Mice fed the HFD for 4 weeks were given successive intraperitoneal injections of CsA for another 6 weeks (Fig. 7A). CsA could suppress HFD-induced mitochondrial ROS generation in a dose-dependent manner (Fig. 7B), and malondialdehyde levels were also decreased (Fig. 7C). Furthermore, CsA down-regulated SREBP-1c promoter activity (Fig. 7D) and its protein expression (Fig. 7E), which was followed by decreased lipid accumulation (Fig. 7F,G), especially TG levels (Fig. 7H).

We also checked the effect of CsA in a long-term HFD-fed mouse model. Mice were fed the HFD for 30 weeks, followed by intraperitoneal injections of CsA for another 6 weeks (Fig. 7I). Similar to the 4-week model, CsA also suppressed HFD-induced mitochondrial ROS generation (Fig. 7J), down-regulated SREBP-1c activity (Fig. 7K), and inhibited hepatic steatosis (Fig. 7L,M). Overall, these results suggested that CypD might be a preventive and therapeutic target for hepatic TG accumulation and that CsA would be a new pharmacological approach for improving NAFLD.