Niemann-Pick type C2 deficiency impairs autophagy-lysosomal activity, mitochondrial function, and TLR signaling in adipocytes

In this study, we investigated the role and mechanism of Niemann-Pick type C (NPC)2 in regulating lysosomal activity, mitophagy, and mitochondrial function in adipocytes. We found that knocking down NPC2 impaired lysosomal activity, as evidenced by the reduced mature cathepsin B, the increased accumulation of light chain 3 (LC3) and p62, and the decreased autophagic flux. In NPC2-knockdown (kd) adipocytes, the starvation-induced conversion of LC3-I to LC3-II was abolished. More interestingly, the majority of NPC2 was found in the mitochondrial fraction, and NPC2 deficiency led to impaired autophagic flux and decreased induction of LC3-II in the mitochondrial fraction during mitochondrial stress. Moreover, cellular respiration profiling revealed that NPC2-kd adipocytes had significantly decreased basal/maximal respiration and mitochondrial gene expression compared with scrambled cells, suggesting mitochondrial dysfunction. Additionally, we found that the mitochondrial recruitment of LC3-II induced by lipopolysaccharide (LPS), but not TNFα, was blunted in NPC2-kd adipocytes. Most intriguingly, NPC2-kd selectively diminished LPS-induced NFκB and ERK1/2 phosphorylation and the expression of pro-inflammatory genes, indicating that toll-like receptor signaling activation is impaired in the absence of NPC2. Our results suggest that NPC2 is in a mitochondrially associated autophagosome and plays an important role in regulating mitophagy, mitochondrial quality control, and mitochondrial function.

In another study, increased accumulation of mitochondrial cholesterol and mitochondrial dysfunction was observed in NPC1-deficient Chinese hamster ovary cells (18). Although the accumulation of cholesterol in the late endosomes and lysosomes is believed to be the cause of NPC disease, its underlying mechanism for the NPC2 deficiency-caused pathologies, especially defective autophagy in adipocytes, has not been fully understood.
In our previous studies, we found that the mRNA and protein expression of NPC2 was upregulated in epididymal adipose tissue, but downregulated in inguinal adipose tissue in high-fat diet-induced obese mice (19). In this study, we investigated the role and underlying mechanisms of NPC2 in regulating autophagy-lysosomal activity and maintaining mitochondrial health and function in adipocytes. We found that NPC2 plays a critical role in regulating lysosomal function and mitophagy. NPC2 deficiency reduces autophagy-lysosomal activity and impairs mitochondrial function in adipocytes. Most importantly, we discovered that NPC2 is a critical regulator of TLR signaling pathway activation and its mediated autophagy activation.

Animals
C57BL/6J mice (Jackson Laboratory, Bar Harbor, ME) were housed in a specific pathogen-free facility and were given free access to water and food. All experimental procedures were approved by the University of Minnesota animal care and use committee. Animal handling was performed according to National Institutes of Health guidelines. The tissues were collected from 12-week-old mice fed a regular chow diet.
The differentiation of 3T3-L1 fibroblasts was induced by a differentiation cocktail (100 IU/ml penicillin-streptomycin, 10% FBS, 115 g/ml methylisobutylxanthine, 390 ng/ml dexamethasone, and 1 g/ml insulin) for 2 days. The cells were subsequently maintained in culture medium supplemented with 10% FBS and 1 g/ml insulin for 5 days. Fully differentiated 3T3-L1 adipocytes with massive accumulation of fat droplets were used for the experiments.

Relative quantitative real-time RT-PCR
Total RNA was extracted from the cells with TRIZOL reagent (Invitrogen). First-strand cDNA was synthesized from activate autophagy. In the absence of insulin signaling and amino acids, mammalian target of rapamycin complex 1 (mTORC1) is inactivated and releases inhibition on Unc-51-like kinase 1 (ULK1) (4). Additionally, 5′AMP-activated protein kinase (AMPK) inhibits mTORC1 and may also directly interact with ULK1 to activate autophagy (4). These signaling events promote the formation of the ULK1 complex and initiate formation of the autophagosome. Induction of autophagy also requires the activity of sirtuin-1, which is activated by elevated levels of nicotinamide adenine dinucleotide (NAD) and deacetylates several key autophagy-related proteins (5). Therefore, there are redundant systems in place and multiple signaling events may be required to fully activate autophagy in response to nutrient deprivation. In addition to starvation, autophagy can be activated in response to inflammation. Activation of toll-like receptors (TLRs) by lipopolysaccharide (LPS) is known to stimulate autophagy via both the MyD88-and TRIF-dependent pathways (6)(7)(8). NFB, a key downstream component of the TLR signaling pathway, has long been known to be essential for the activation of autophagy (9,10). While autophagy can be activated by inflammation, it has also been shown to attenuate the inflammatory response. Autophagy activation promotes p62-mediated inflammasome degradation and prevents the conversion of pro-IL-1 to mature IL-1 (11). Hence, autophagy may serve to mitigate cellular stress to maintain homeostasis.
Niemann-Pick type C (NPC) disease is a childhood onset neurodegenerative disorder caused by mutations in the NPC1 and NPC2 genes that encode endosomal/lysosomal proteins. The major NPC phenotype is the cholesterol accumulation in endosomes and lysosomes. Most of the previous studies were focused on the role of NPC in cholesterol and other lipid metabolism. It has been documented that membrane-bound NPC1 and soluble NPC2 function in a common lysosomal pathway (12) to facilitate the transfer of unesterified cholesterol from endosomes and lysosomes to endoplasmic reticulum and plasma membrane (13,14). However, the defective cellular functions that are attributed to NPC deficiency have not been completely characterized. In particular, it is unclear whether NPC2 deficiency contributes to the development of cellular aging. In keeping with the common features of degenerative diseases, NPC1 deficiency impairs the clearance of autophagosomes leading to dysfunctional protein degradation, suggesting that it plays a role in the regulation of autophagy-lysosomal activity (15).
While most of previous studies on NPC proteins focus on their roles in neurodegenerative, liver and spleen disorders, few studies have investigated the association of NPC proteins with obesity, diabetes, and insulin resistance. In a recent study, a SNP in NPC1 has been associated with obesity and type 2 diabetes, and NPC1-deficient mice developed insulin resistance (16). Interestingly, recent studies unveiled that NPC2 has some NPC1-independent effects on metabolism. For instance, NPC2 has been demonstrated to transport endosomal cholesterol to mitochondria, which is important for maintaining mitochondrial membrane structure and function (17).

Mitochondria isolation
The differentiated adipocytes with different treatments were scraped and collected with ice-cold isolation buffer [20 mmol/l Tris and 1 mmol/l EDTA (pH 7.4)] supplemented with protease inhibitors. The cells were then lysed with 20 strokes of a glass-Teflon homogenizer at a speed of 1,600 rpm. After centrifugation at 700 g for 10 min, the supernatant (cytosolic fraction) was collected and lipids were removed. The pellet (mitochondrial fraction) was washed and recovered with isolation buffer by centrifugation at 9,000 g at 4°C for 10 min.

Filipin staining
Filipin staining was performed in adipocytes treated with U18666a, as described previously (23,24). U18666a is a widely used inhibitor of cholesterol trafficking and synthesis. Briefly, scrambled and NPC2-kd cells were seeded into 6-well plates and induced to differentiate into adipocytes. Prior to full differentiation, cells were treated with 10 g/ml U18666a (Sigma-Aldrich) for 48 h (from day 5 to day 7 of differentiation) and 72 h (from day 4 to day 7 of differentiation), respectively; PBS was used as a vehicle control. Cells were then fixed with 4% paraformaldehyde for 1 h at room temperature. After rinsing with PBS and incubation with 1.5 mg/ml glycine in PBS for 10 min to quench paraformaldehyde, cells were stained with 50 g/ml Filipin III (Sigma-Aldrich) in PBS for 2 h. The filipin signal was imaged using an Olympus IX70 inverted fluorescence microscope with UV filter package.

Statistical analysis
Results of gene expression are expressed as mean ± SEM. Results of protein expression by Western blot analysis are expressed as mean ± SD. Differences in the parameters between two groups were evaluated using Student's t-test with a 0.05 two-sided significance level. P < 0.05 was considered significant.

Establishment of NPC2-kd cell line
We have previously reported that NPC2 expression in adipose tissue is dysregulated in mice when challenged with high-fat diet or in macrophages in response to inflammatory stimulation (19). In this study, we showed that NPC2 was also expressed in brown adipose tissue and mesentery white adipose tissue (Fig. 1A). The gene expression of Npc2 was increased during adipocyte differentiation in 3T3-L1 cells (Fig. 1B).
To investigate the role and underlying mechanism of NPC2 in the pathogenesis of obesity and adipocyte dysfunction, we established NPC2-kd and scrambled cell lines using lentivirus delivery of shRNA directed against Npc2 mRNA or nonspecific scrambled control sequence. The selected oligomers targeting Npc2 sequence were 5′-GCTCTCGTTCTTTGGTAGTTT-3′ and 5′-CGGTTGTA-AGAGTGGAATCAA-3′. As shown in Fig. 1C, lentiviral shRNA was able to knock down the gene expression of Npc2 by 60% in 3T3-L1 adipocytes. Moreover, the protein level of NPC2 was markedly reduced by more than 90% in NPC2-kd adipocytes compared with scrambled controls (Fig. 1D). Oil Red O staining of adipocytes on day 8 of differentiation showed that both scrambled and NPC2-kd DNase-treated total RNA using a Superscript II reverse transcriptase kit (Invitrogen). Quantitative amplification by PCR was carried out using SYBR Green qPCR Master Mix (SABiosciences, Frederick, MD) by a StepOne Real-Time PCR System (Applied Biosystem, Foster City, CA). The Ct method was used to calculate the results. For quantification, results were normalized to TATA box binding protein (TBP) and presented as the levels of expression relative to those of endogenous controls. The primer sequences for amplifying the target genes and the GenBank accession number are summarized in supplemental Table S1.

Western blot analysis
Proteins were extracted using RIPA buffer (Sigma-Aldrich) containing protease inhibitors and phosphatase inhibitor cocktail (Roche Diagnostics Corporation, Indianapolis, IN). Proteins were separated on 8-10% SDS-PAGE gel and then transferred to nitrocellulose membranes. The primary antibodies probed to the target proteins included mouse monoclonal antibodies to NFB and rabbit polyclonal antibodies to Tom20 (Santa Cruz Biotechnology, Dallas, TX), goat polyclonal antibodies to cathepsin B (CTSB) and cathepsin L (CTSL) (R&D Systems, Minneapolis, MN), mouse monoclonal antibody to Parkin and rabbit monoclonal antibodies to NPC2, autophagy marker light chain 3 (LC3), SQSTM1/p62, autophagy-related (Atg)5, Atg7, phospho-p70 S6 kinase (Thr 389), p70 S6 kinase, phospho-NFB p65, phospho-Akt (ser473), Akt, tubulin, and -actin (Cell Signaling Technology, Danvers, MA); and rabbit polyclonal antibody to lysosome-associated membrane protein 1 (LAMP1; Bioss Antibodies, Woburn, MA). ECL Western Blotting Detection System (GE Healthcare BioSciences, Piscataway, NJ) was used to detect antibody reactivity. Densitometric quantification was determined using an image analysis program (Alpha Innotech, San Leandro, CA) and reported as a ratio to total protein or -actin as indicated in the result sections.

Glucose uptake assay
Uptake of 2-deoxy-d-[ 3 H] glucose (Amersham Biosciences, Piscataway, NJ) in 3T3-L1 adipocytes was measured as previously described (21). Briefly, the adipocytes were starved in Krebs-Ringer HEPES (KRH) buffer containing 0.5% BSA (pH 7.4) for 2 h before incubation either with or without 200 nmol/l insulin for 30 min at 37°C. Glucose uptake was initiated by adding [ 3 H]2deoxy-d-glucose to a final assay concentration of 100 mol/l. After 5 min, the cells were washed with ice-cold PBS three times and then solubilized with KRH buffer containing 1% Triton X-100. Incorporated radioactivity was determined by scintillation counting. Nonspecific 2-deoxyglucose uptake was measured in the presence of 20 mol/l cytocholasin B and subtracted from the total glucose uptake assayed to obtain specific uptake.

Cellular respiratory assay
Cellular respiration was analyzed in situ with the XF24 Extracellular Flux Analyzer (Seahorse Biosciences, Billerica, MA). The 3T3-L1 fibroblasts were plated on V7 microplates coated with 0.01% collagen solution type I (Sigma-Aldrich) and differentiated into adipocytes. Eight days postdifferentiation, cells were washed and incubated at 37°C without CO 2 in bicarbonate-free DMEM containing 1.0 mM sodium pyruvate, 2.0 mM glutamax, and 25 mM d-glucose. During the assay, cell monolayers were exposed to 2.0 M oligomycin, 0.5 M carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), and 4 m antimycin A. Respiration rates were calculated as previously described (22). found that both LC3-I and LC3-II levels were significantly higher in NPC2-kd adipocytes than those in scrambled cells, indicating a reduction in lysosomal degradation of autophagosomes (Fig. 2D). Consistently, the level of p62, a specific autophagic cargo (28), was also elevated in NPC2kd adipocytes (Fig. 2D), suggesting the reduced clearance of autophagic proteins by NPC2 deficiency.
To further assess the impact of NPC2 kd on autophagy, we examined autophagic flux in scrambled and NPC2-kd adipocytes. Autophagic flux was measured by comparing the difference in LC3-II and p62 levels in the presence and absence of lysosomal degradation inhibitor, BafA1. Autophagic flux was determined by calculating the ratio of protein levels of LC3-II and p62 in BafA1-treated adipocytes to those in untreated cells, as previously described (29). This method determines how much LC3-II and p62 can be accumulated by blocking lysosomal degradation, which reflects the autophagic ability. As shown in Fig. 2E, both LC3-II and p62 levels were higher in NPC2-kd adipocytes compared with scrambled cells in the basal condition. BafA1 treatment for 4 or 24 h significantly increased LC3-II and p62 levels in scrambled adipocytes. However, this BafA1-induced increase in LC3-II and p62 levels was significantly reduced in NPC2-kd adipocytes (Fig. 2E). This data suggests that NPC2 kd impairs autophagic flux; the accumulation of LC3-II in NPC2-kd cells in the basal state could result from the impairment of autophagic flux.

NPC2 kd reduces starvation-induced, but not rapamycininduced, autophagy
Next, we determined the effect of NPC2 deficiency on the autophagic process in adipocytes. Because nutrient deprivation is a common physiological inducer of autophagy, we examined the induction of autophagy by starvation in NPC2-kd adipocytes. First, we determined how NPC2 expression is regulated by starvation. We found that the NPC2 protein expression was significantly induced after 4 h treatment with HBSS, a medium containing 5 mM glucose and devoid of amino acids and other growth factors (Fig. 3A). Next, scrambled and NPC2-kd adipocytes were treated for 2 h with HBSS for the examination of autophagy activation as indicated by changes in LC3-II levels. As shown in Fig.  3B, there was an elevation of LC3-II in scrambled adipocytes when exposed to either HBSS or KRH buffer (containing 2.5 mM glucose) compared with those in the fed condition (Fig. 3B). However, in NPC2-kd adipocytes, neither HBSS nor KRH induced the conversion of LC3-I to LC3-II (Fig. 3B).
To determine whether the reduced LC3-II by starvation in NPC2-kd adipocytes was due to rapid autophagic flux or impaired induction, we treated adipocytes with or without BafA1 in a starvation milieu. Because the cells died within 3T3-L1 cells were able to differentiate into adipocytes to a similar extent (Fig. 1E), which is confirmed by no significant difference in the expression of adipogenic genes between scrambled and NPC2-kd adipocytes (Fig. 1F).
To determine whether NPC2 kd has an impact on cholesterol metabolism, we preformed filipin staining. Adipocytes were treated without or with U18666a, an inhibitor of cholesterol transport and synthesis, for 24 and 72 h. U18666a treatment led to a time-dependent increase in intracellular accumulation of cholesterol in both scrambled and NPC2-kd adipocytes (Fig. 1G). Interestingly, we found that NPC2-kd adipocytes displayed increased levels of intracellular cholesterol accumulation compared with scrambled cells in the basal and U18666a-treated conditions (Fig.  1G). This result suggests that NPC2 is involved in cholesterol metabolism, confirming the phenotype of NPC2 kd.

NPC2 kd reduces lysosomal activity
NPC2 is an endosomal/lysosomal protein and has been identified as an intracellular cholesterol transporter that acts in concert with NPC1 to remove cholesterol from the endosomal/lysosomal compartment (13,14,25,26). Cathepsins, the lysosomal proteases, are the markers of lysosomal activity and have been connected to NPC2 in inflammation and obesity in our previous studies (19). To evaluate how lysosomal activity affects lysosomal cathepsins in adipocytes, we detected the regulation of protein expression and secretion of CTSB by bafilomycin A1 (BafA1), an inhibitor of lysosomal activity, in 3T3-L1 adipocytes. Treatment with BafA1 for 24 h blocked the maturation of pro-CTSB, as evidenced by inhibiting the pro-CTSB conversion to mature CTSB in the cells ( Fig. 2A). Interestingly, pro-CSTB, but not mature CSTB, was secreted into culture media; BafA1 significantly increased pro-CSTB secretion ( Fig.  2A). Thus, the reduced expression of mature CTSB represents the suppressed lysosomal activity in adipocytes.
To investigate how NPC2 deficiency affects lysosomal activity in adipocytes, we first evaluated the lysomosal activity in NPC2-kd 3T3-L1 adipocytes by examining the expression of CSTB. As shown in Fig. 2B, C, the gene expression of Ctsb was downregulated (Fig. 2B) and the mature form of CSTB was significantly reduced (Fig. 2C) in NPC2-kd adipocytes compared with scrambled cells. The secreted mature CTSB was slightly deceased in the culture media of NPC2-kd adipocytes (Fig. 2C). These results indicate that lysosomal activity is impaired in NPC2-kd adipocytes as reflected by reduced expression of Ctsb gene and mature CTSB. Because the reduced hydrolytic activity of lysosomes is expected to attenuate the degradation of autophagosomes, causing the accumulation of autophagic flux protein LC3 (27), we then determined the levels of LC3 in NPC2-kd adipocytes in the basal state. Interestingly, we in 3T3-L1 adipocytes by Western blot. E: Oil Red O staining of 3T3-L1 cells on day 8 of differentiation. F: Adipogenic gene expression in 3T3-L1 adipocytes on day 8 of differentiation (n = 4 in each group). G: Filipin III fluorescence image of scrambled and NPC2-kd adipocytes treated with 10 g/ml U18666a for 48 and 72 h, respectively. Images were acquired using the automatic fluorescence microscope at 20× magnification. Scale bar = 50 m. Data are expressed relative to the value for scrambled (Scr) cells. The values are mean ± SEM. *P < 0.001 versus scrambled cells. Ap2, adipocyte P2; PPARg, peroxisome proliferator-activated receptor gamma; Adipoq, adiponectin; Glut4, glucose transporter type 4. levels compared with untreated controls (Fig. 3E). In order to better assess the changes in p62 proteins, cells were treated with rapamycin for a longer time (4 h). After 4 h treatment with rapamycin, p62 protein levels were decreased in both scrambled and NPC2-kd adipocytes (Fig.  3F), indicating that NPC2 kd does not affect rapamycin-induced degradation of p62 protein.
To further test to determine whether NPC2 kd impacts the mTOR signaling pathway, which is involved in the regulation of autophagy, we examined the effect of rapamycin on the mTOR signaling pathway, lysosomal proteases (CTSB and CTSL), and autophagy-related protein (Atg5). Consistent with the increased autophagy by rapamycin, we found that 48 h treatment with rapamycin significantly inhibited the phosphorylation of p70S6K in both scrambled and NPC2-kd adipocytes (Fig. 3G). Although the basal levels of CTSB and CTSL were lower in NPC2-kd adipocytes than in the scrambled controls, the expression of CTSB and CTSL was similarly upregulated by rapamycin in both types of cells (Fig. 3G, H). The expression of autophagic protein, Atg5, remained unchanged regardless of rapamycin treatment in the scrambled and NPC2-kd adipocytes (Fig. 3G). The above data suggests that NPC2 kd a very short period of time when incubated with HBSS plus BafA1, we assessed autophagic flux under the serum starvation condition. Scrambled and NPC2-kd adipocytes were treated with or without BafA1 in serum-free culture media for 24 h. Consistently, the basal level of LC3-II was higher in NPC2-kd compared with scrambled adipocytes after 24 h incubation with serum-free culture media (Fig.  3C). BafA1 treatment for 24 h led to accumulation of LC3-II in scrambled adipocytes, but this BafA1-induced LC3-II accumulation was attenuated in NPC2-kd adipocytes, as demonstrated by the ratio of LC3-II with BafA1 treatment to that without BafA1 (Fig. 3C), suggesting that the autophagic flux is defective in NPC2-kd adipocytes.
Rapamycin, an inhibitor of mTOR kinase, is another common inducer of autophagy by promoting autophagosome formation and autophagosome-lysosome fusion (30,31). Similar to starvation, rapamycin treatment for 48 h also significantly induced NPC2 protein expression in 3T3-L1 adipocytes (Fig. 3D). Next, we assessed the effect of NPC2 kd on rapamycin-induced autophagy by detecting the protein levels of LC3 and p62. In response to rapamycin treatment for 2 h, both scrambled and NPC2-kd adipocytes had elevated levels of LC3-II, but no change in p62 number of cytokines, including TNF and IL-1, during infection and inflammation (32). Activation of TLRs by LPS has been reported to stimulate autophagy in bone marrow-derived macrophages (33). To understand the role of NPC2 in the regulation of autophagy in response to cellular stress, we then investigated the effect of NPC2 kd on inflammation-induced autophagy in adipocytes. First, we studied the impact of NPC2 deficiency on TNFinduced autophagy. Consistent with the results shown in Fig. 2C, the basal levels of LC3 in NPC2-kd adipocytes were higher than those in scrambled cells, indicating a reduction in lysosomal degradation (Fig. 4A). Despite an increase in basal levels of LC3 in NPC2-kd adipocytes, TNF does not seem to affect rapamycin-induced autophagy and lysosomal activity, suggesting that NPC2 is involved in mTOR-independent signaling transduction rather than as a part of the autophagy-lysosomal degradation machinery.

NPC2 kd blunts LPS-induced, but not TNF-induced, autophagy
In addition to starvation, autophagy can be activated under the pathological condition (for instance, in response to inflammation/stress) and plays a role as a defensive mechanism in protecting against sustained inflammation and restoring cellular homeostasis. Numerous studies have shown that autophagy can be activated by LPS and a determined the cellular location of NPC2 and the effect of NPC2 kd on mitophagy and mitochondrial function. We found that NPC2 was highly abundant in the crude mitochondrial fraction and the mitochondrial abundance of NPC2 was increased in response to TNF (Fig. 4C) and LPS treatment (Fig. 4D). While LC3-I was mostly present in the cytosolic fraction, LC3-II was more abundant in the mitochondrial fraction (Fig. 4C, D). In accordance with the results obtained from the whole cell lysate of NPC2-kd adipocytes, there was an increased accumulation of LC3-II in the mitochondrial fraction of NPC2-kd adipocytes when compared with that of scrambled cells in the basal condition (Fig. 4C, D). Both the scrambled and NPC2-kd adipocytes responded to TNF treatment, resulting in an increase in LC3-II in the mitochondrial fraction (Fig. 4C). However, LPS treatment led to an increase in LC3-II in the mitochondrial fraction only in scrambled adipocytes, but failed to stimulate LC3-II in NPC2-kd adipocytes (Fig. 4D). treatment for 4 h led to an increase in LC3-II and p62 protein in both scrambled and NPC2-kd adipocytes (Fig. 4A). This data suggests that NPC2-kd adipocytes have a normal response to TNF-induced autophagy. Second, we tested to determine whether NPC2 deficiency influences the effect of LPS on autophagy. Interestingly, after 4 h treatment with LPS, the levels of LC3-II and p62 protein were increased in scrambled adipocytes. However, this stimulatory effect of LPS was not observed in NPC2-kd adipocytes (Fig. 4B). These results suggest that NPC2 kd selectively blunts LPS effect on autophagy activation.

NPC2 kd impairs mitophagy, mitochondrial function, and anti-oxidant response
One of the important purposes of inflammation-induced autophagy is to degrade mitochondria damaged by inflammation in a process known as mitophagy. To better understand the role of NPC2 in the regulation of cellular homeostasis, specifically mitochondrial quality control, we scrambled cells in the basal condition (Fig. 4E). When treated with CCCP or CCCP plus BafA1, NPC2-kd adipocytes displayed blunted induction of LC3-II with increased accumulation of Parkin in the mitochondrial fraction compared with scrambled cells (Fig. 4E).
Next, cellular respiration was evaluated by the Seahorse XF24 extracellular flux analyzer to assess mitochondrial function. We observed that NPC2-kd adipocytes had significantly decreased basal respiration, reduced proton leak, and lowered maximal respiration when compared with the scrambled cells (Fig. 5A, B). This data strongly suggests that NPC2-kd adipocytes have decreased mitophagy activation and impaired mitochondrial function. To determine whether mitochondrial dysfunction results from a decreased number of mitochondria or reduced We performed an experiment to provide additional evidence for the role of NPC2 in mitophagy in adipocytes. Because mitophagy does not occur frequently in the steady state, we treated adipocytes with carbonyl cyanide m-chlorophenylhydrazone (CCCP), a mitochondrial membrane depolarizing agent, to induce mitochondrial stress and to trigger mitochondrial damage-induced mitophagy (34). The translocation of Parkin and recruitment of LC3-II to the mitochondria is one of the most important mechanisms for promoting mitophagy. Parkin is known to play a key role in initiating mitophagy by specifically targeting damaged mitochondria and promoting the removal of damaged mitochondria by autophagic clearance (35,36). We observed that NPC2-kd adipocytes had higher levels of LC3-II and Parkin in the mitochondrial fraction than Zn superoxide dismutase (SOD1), was significantly downregulated in NPC2-kd adipocytes compared with scrambled cells. These results support that NPC2-kd reduces the antioxidant response.

NPC2 kd affects insulin sensitivity in adipocytes
Because mitochondrial dysfunction and oxidative stress are closely associated with insulin resistance, we next evaluated the impact of NPC2 kd on insulin sensitivity by measuring insulin-stimulated glucose uptake and Akt phosphorylation. We observed that the basal glucose uptake was higher in NPC2-kd adipocytes than that in the scrambled cells (Fig. 6A). Consistently, the gene expression of Glut1, which is responsible for basal glucose uptake and highly sensitive to hypoxia, was significantly elevated in NPC2-kd adipocytes compared with scrambled cells (Fig. 6B), suggesting high activity of anaerobic metabolism of glucose in NPC2-kd adipocytes. In response to insulin stimulation, both scrambled and NPC2-kd adipocytes had a similar increase in glucose uptake (Fig. 6A). There was also no difference in the gene expression of Glut4 between the scrambled and NPC2-kd adipocytes (Fig. 6B). Additionally, insulin was able to stimulate Akt phosphorylation to a similar extent in both scrambled and NPC2-kd adipocytes (Fig. 6C). However, pretreatment with LPS for 24 h induced a greater reduction of insulin-stimulated Akt phosphorylation in NPC2-kd adipocytes than the scrambled cells (Fig. 6D), suggesting mitochondrial activity, we evaluated mitochondrial DNA (mtDNA) copy number and the expression of key genes regulating mitochondrial biogenesis and function. There was no difference in the ratio of mtDNA number cyclooxygenase (COX) II and nuclear DNA (RIP 140) between scrambled and NPC2-kd adipocytes (Fig. 5C). The gene expression of Nrf1 and Nrf2 was also not different between the scrambled and NPC2-kd adipocytes (data not shown). However, the expression of peroxisome proliferator-activated  coactivator-1 (PGC1) was reduced in NPC2-kd adipocytes, as was the expression of mitochondrial genes, including COX IV, cytochrome c, and cyclooxygenase 8b (COX 8b) (Fig. 5D, E). This data suggests that impaired mitochondrial function, rather than changes in mitochondrial number, is the major contributor to the reduced mitochondrial activity in NPC2-kd adipocytes.
Mitochondria are an important source of superoxide production, which can be degraded to reactive oxygen species (ROS) (37). Mitochondrial dysfunction can cause increased ROS production and oxidative stress. To determine whether mitochondrial dysfunction-associated oxidative stress occurs in NPC2-kd adipocytes, we evaluated the gene expression of ROS detoxification enzymes. As shown in Fig. 5F, the gene expression of antioxidant enzymes, such as glutathione peroxidase 1 (Gpx1), peroxiredoxin 3 (Prdx3), glutathione S-transferase A4 (GstaA4), and Cu/ that NPC2-kd adipocytes become more insulin resistant under the condition of LPS-induced metabolic stress.

NPC2 kd impairs TLR signaling activation
The above results that NPC2 kd selectively impairs the LPS effect on autophagy/mitophagy activation suggest that NPC2 may play a role in regulating TLR signaling pathway activation. To prove this hypothesis, we further examined how NPC2 kd impacts the effect of TNF and LPS on the activation of TLR4 (inflammatory) signaling pathways, as well as the expression inflammatory cytokine genes by TNF and LPS in NPC2-kd adipocytes. As shown in Fig. 7A, both scrambled and NPC2-kd adipocytes had a similar response to TNF stimulation in NFB and ERK1/2 phosphorylation. Consistently, the gene expression of MCP1 and IL-6 was upregulated by TNF in both NPC2-kd and scrambled adipocytes (Fig. 7B). Strikingly, LPS treatment significantly stimulated NFB and ERK1/2 phosphorylation in scrambled adipocytes, but this stimulatory effect was almost completely diminished in NPC2-kd adipocytes (Fig.  7C). The LPS-induced expression of inflammatory genes, including TNF, IL-6, IL-18, and MCP1, was also significantly attenuated in NPC2-kd adipocytes (Fig. 7D). All these results suggest that NPC2 is essential for TLR signaling activation.

DISCUSSION
NPC2 is a protein in the late endosome/lysosome that, in concert with NPC1, modulates intracellular lipid metabolism. While most of the previous studies on NPC2 focused on its role in regulating cholesterol trafficking, the pathophysiological function of this protein in cellular homeostasis remains elusive. Our previous studies have connected dysregulation of adipose NPC2 expression to diet-induced obesity (19). Studies from others have demonstrated that NPC2 has a role in adipocyte differentiation (38). In this study, we show that NPC2 plays a critical role in the regulation of mitophagy and mitochondrial function in adipocytes. We also demonstrate that the impaired mitophagy activation may be partly attributed to the failure of TLR signaling pathway activation in NPC2-kd adipocytes.
Autophagy plays an essential role in maintaining cellular homeostasis and function in many cell types, thereby regulating multiple physiological processes. In obesity, autophagy activation is necessary in order to combat metabolic stress and inflammation. Studies have indicated that this regulation of autophagy activation is defective in adipose tissue during obesity (39,40). However, the pathological mechanism underlying this defect remains largely unknown. Autophagy activation is initiated by the formation of the autophagosome, a double-membraned vesicle, which is controlled by multiple autophagy-related proteins (ATG proteins). The mammalian homolog of yeast ATG8, LC3, is commonly used as a marker of autophagosome formation (41). The following and also most important step is the fusion of autophagosomes with late endosomes and lysosomes for the degradation of autophagic cargoes by vari-ous hydrolytic enzymes in the lysosome (42). Among the lysosomal enzymes, CTSB is one of the most abundant proteases (43) and is required for the activation of the autophagy-lysosome process (44,45). In our study, we found that inhibiting lysosomal activity by BafA1 blocks the conversion of pro-CTSB to mature CTSB, leading to an accumulation of pro-CTSB but a decrease in mature CTSB in adipocytes. Thus, the reduction in the mature form of CTSB could indicate the impairment of lysosomal activity. In NPC2-kd adipocytes, we found that the mature form of CTSB was significantly reduced compared with scrambled cells. Consistently, the accumulation of p62 and LC3 proteins was increased in NPC2-kd adipocytes under the basal condition. During autophagy, LC3-II localizes to the inner autophagosome for degradation by lysosomal enzymes. Although there was an increase in LC3-II and p62 in the steady state, reduced induction of LC3-II and p62 was observed in NPC2-kd adipocytes treated with BafA1, indicating the impaired autophagic flux by NPC2 deficiency. All of these results indicate that NPC2 kd impairs lysosomal proteolytic activity and function, thereby reducing autophagy activity. Our results are in line with previous studies demonstrating that autophagic activity is decreased in NPC1-deficient cells (15). In the previous studies, both NPC1 deficiency and inhibition of cholesterol efflux from the lysosome could lead to the enlargement of autophagosomes, the impairment of autophagy flux, and the accumulation of ubiquitinated proteins in the endosomal/ lysosomal compartment (27). Therefore, our studies have provided a new line of evidence that NPC2 has an important role in regulating autophagy-lysosomal activity.
Autophagy is involved in a wide variety of physiological and pathological processes and can be activated by various stimuli (46). Nutrient deprivation-induced autophagy is a physiological response to maintain cell survival. We found that starvation is able to induce the expression of NPC2 in adipocytes, suggesting a possible role of NPC2 in physiological autophagy. Our results showed that starvation with nutrient-deprived culture media failed to induce the recruitment of LC3-II in NPC2-kd adipocytes. Interestingly, rapamycin, a direct inhibitor of the mTORC1 and a strong inducer of autophagy, was also able to induce NPC2 expression in adipocytes. However, NPC2 deficiency does not influence the effect of rapamycin on the recruitment of LC3-II and the degradation of p62 cargo, as well as the induction of CTSB and CTSL in NPC2-kd adipocytes. Because rapamycin and nutrient deprivation have been reported to induce autophagy through two different mTOR signaling pathways, i.e., rapamycin suppresses the mTORC1 signaling pathway, while starvation inactivates the phosphoinositide-3-kinase (PI3K)-Akt and mTORC2 signaling pathway (47,48), it is likely that NPC2 may regulate autophagy-lysosomal activation via the mTORC1-independent signaling pathway in adipocytes.
In addition to starvation, autophagy can be activated as a protective mechanism under pathological conditions, such as inflammation, oxidative and endoplasmic reticulum stress, and hypoxia (49). In obesity, autophagy is activated to alleviate stress and inflammation, thereby restoring fact that NPC2 has an essential role in the regulation of mitophagy and mitochondrial function. Because mitochondrial function is important for glucose metabolism and insulin sensitivity, we further investigated the impact of NPC2 kd on glucose uptake and insulin signaling activity. We observed that the basal level of glucose uptake was increased in NPC2-kd adipocytes compared with scrambled adipocytes. This was supported by the upregulation of Glut1, a primary transporter that is responsible for basal glucose uptake (55) and highly sensitive to hypoxia (56). The upregulation of Glut1 indicates a hypoxic condition reflecting mitochondrial dysfunction of NPC2-kd adipocytes. While NPC2 kd does not affect insulin-stimulated glucose uptake and Akt phosphorylation, it does worsen LPS-induced downregulation of insulin-stimulated Akt phosphorylation.
Inflammatory cytokines have been well-known to link inflammation to mitochondrial dysfunction and insulin resistance in obesity. TNF has been known to alter mitochondrial membrane permeability and inhibit electron chain respiration, thereby stimulating ROS production and causing mitochondrial damage (57)(58)(59). LPS, a major component of the outer membrane of Gram-negative bacteria, has been recently considered as a critical stimulus of adipose tissue inflammation and a regulator of autophagy. LPS has been known to bind TLR2 and TLR4 on the plasma membrane with the involvement of MD-2 to activate an immune response and induce inflammation (60). The data from human studies has shown that obese subjects have altered gut microflora, leading to the elevation of circulating LPS (61). Circulating LPS levels were also found to be increased in aged humans (62). In animal studies, high-fat diet feeding has been shown to increase the proportion of LPS-containing microbiota in the gut leading to the elevation of circulating LPS levels in mice (63). In this sense, understanding the regulation of TLR signaling pathway activation and its related autophagy and metabolism in adipocytes is of importance.
Most intriguingly, we discovered that NPC2 kd selectively diminishes LPS-induced autophagy/mitophagy. NPC2-kd adipocytes have a normal response to TNF stimulation in autophagy/mitophagy activation. LPS is known to trigger the inflammatory cascade (NFB and Erk1/2) via activating TLR2 and TLR4 signaling (64)(65)(66)(67), whereas TNF stimulates NFB and Erk1/2 activation through a TLR-independent signaling pathway. We therefore speculate that NPC2 may be critical for TLR signaling activation. Our results showed that while NPC2 kd does not affect TNF-induced NFB and Erk phosphorylation, LPS-stimulated phosphorylation of NFB and Erk was almost completely diminished in NPC2-kd adipocytes. Furthermore, LPS failed to induce the expression of pro-inflammatory cytokines in NPC2-kd adipocytes as efficiently as it did in scrambled cells. This data strongly suggests that TLR signaling activation is impaired in NPC2-kd adipocytes. NFB activation is known as an important regulator of the autophagy activation process (9,10). It is likely that lack of NPC2 impairs the activation of TLR4 by LPS in adipocytes, leading to the failure of the NFB signaling pathway and autophagy activation. Further intracellular homeostasis and limiting cellular dysfunction (50). Mitophagy, a type of selective autophagy, plays an important role in mitochondrial quality control by eliminating damaged mitochondria (51,52). In the present study, we found that the majority of NPC2 is located in the mitochondrial fraction in adipocytes, and both TNF and LPS increase the abundance of NPC2 protein in the mitochondrial fraction. Because the isolated crude mitochondrial fraction is potentially contaminated with other microsomes such as lysosome, we detected the lysosome marker, i.e., LAMP1, to indicate the degree of lysosomal contamination in the mitochondrial fraction. We found that LAMP1 levels are much higher in the cytosolic fraction than in the mitochondrial fraction, which is opposite to that of NPC2 levels in both LPS-and TNF-treated conditions. Based on the distribution of TOM20 and LAMP1 between the mitochondrial and cytosolic fraction, we believe that mitochondria still account for the majority of the content in the crude mitochondrial fraction. To determine whether NPC2 is a mitochondrial targeting protein, we performed a computational prediction of mitochondrial targeting peptide on NPC2 sequence using the TargetP server. The likehood of NPC2 having a mitochondrial targeting peptide is low, suggesting that NPC2 may not be a mitochondrial protein or a mitochondrial targeting protein, but associates with mitochondria. Autophagosomes are known to have a close contact with mitochondria during mitophagy. In a previous study with electron microscopy, autophagosomes have been shown to contain mitochondria in the autophagic vacuoles (53). It is likely that the isolated mitochondrial fraction contains intact mitochondria as well as impaired mitochondria that are associated with autophagosomes, and NPC2 is located in autophagosomes/lysosomes that have been targeted to the damaged mitochondria for mitophagy.
Consistent with the marked accumulation of autophagosomes in NPC1-deficient cells in the previous report (54), we found that NPC2-kd adipocytes had significantly increased accumulation of LC3-II in the mitochondrial fraction in the steady state, indicating decreased mitochondrial degradation. Furthermore, we specifically tested the impact of NPC2 deficiency on mitophagy. Our results demonstrate that NPC2-kd adipocytes exhibit decreased induction of LC3-II and increased accumulation of Parkin in mitochondria during mitochondrial stress with CCCP. This strongly supports that NPC2 deficiency impairs mitophagy and reduces the removal of damaged mitochondria, leading to mitochondrial dysfunction. Moreover, the direct assessment of cellular respiration profiling demonstrated that NPC2-kd adipocytes have significantly impaired mitochondrial function, as evidenced by decreased basal and maximal respiration as well as reduced proton leak. This conclusion is further supported by our findings of downregulated expression of genes involved in the regulation of mitochondrial activity, including PGC1, cytochrome c, COX IV, and COX 8b, as well as the decreased expression of anti-oxidant genes such as GPX1, PRDX1, GSTA4, and SOD1 in NPC2-kd adipocytes. In accordance with the results from a recent study (18), our data clearly support the studies are needed to investigate the precise molecular mechanism for the role of NPC2 in the regulation of TLR signaling pathway activation.
In summary, we have demonstrated that NPC2 kd impairs autophagy-lysosomal activity in adipocytes. We found that NPC2 kd selectively blunts the effect of LPS on TLR signaling activation, thereby leading to the defect in mitophagy activation and mitochondrial dysfunction in adipocytes. Our results support that NPC2 is an important regulator of mitophagy, mitochondrial quality control, and mitochondrial function via controlling TLR-NFB signaling pathway activation. Our studies also suggest that NPC2 could serve as a potential therapeutic target for treating adipose tissue dysfunction in obesity and diabetes.