Extrusion of mitochondrial contents from lipopolysaccharide-stimulated cells: Involvement of autophagy

Abstract

Sepsis/endotoxemia is elicited by the circulatory distribution of pathogens/endotoxins into whole bodies, and causes profound effects on human health by causing inflammation in multiple organs. Mitochondrial damage is one of the characteristics of the cellular degeneration observed during sepsis/endotoxemia. Elimination of damaged mitochondria through the autophagy-lysosome system has been reported in the liver, indicating that autophagy should play an important role in liver homeostasis during sepsis/endotoxemia. An increased appearance of mitochondrial DNA and proteins in the plasma is another feature of sepsis/endotoxemia, suggesting that damaged mitochondria are not only eliminated within the cells, but also extruded through currently unknown mechanisms. Here we provide evidence for the secretion of mitochondrial proteins and DNA from lipopolysaccharide (LPS)-stimulated rat hepatocytes as well as mouse embryonic fibroblasts (MEFs). The secretion of mitochondrial contents is accompanied by the secretion of proteins that reside in the lumenal space of autolysosomes (LC3-II and CTSD/cathepsin D), but not by a lysosomal membrane protein (LAMP1). The pharmacological inhibition of autophagy by 3MA blocks the secretion of mitochondrial constituents from LPS-stimulated hepatocytes. LPS also stimulates the secretion of mitochondrial as well as autolysosomal lumenal proteins from wild-type (Atg5^+/+) MEFs, but not from atg5^−/− MEFs. Furthermore, we show that direct exposure of purified mitochondria activates polymorphonuclear leukocytes (PMNs), as evident by the induction of IL1B/interlekin-1β, a pro-inflammatory cytokine. Taken together, the data suggest the active extrusion of mitochondrial contents, which provoke an inflammatory response of immune cells, through the exocytosis of autolysosomes by cells stimulated with LPS.

Abbreviations

3MA

3-methyladenine

CPS1

carbamoyl-phosphate synthase 1, mitochondrial

CASP3

caspase 3, apoptosis-related cysteine peptidase

CTSD

cathepsin D

COX4I1

cytochrome c oxidase subunit IV isoform 1

GalN

D-galactosamine

IL1B

interleukin1, β

LPS

lipopolysaccharide

LAMP

lysosomal-associated membrane protein

MT-mKO1

mitochondrially tagged monomeric Kusabira Orange1

mGFP

monomeric green fluorescent protein

mRFP

monomeric red fluorescent protein

MEF

mouse embryonic fibroblast

PMN

polymorphonuclear leukocyte

SQSTM1/p62

sequestosome 1

tfLC3

tandem fluorescent-tagged LC3

TFEB

transcription factor EB

TNF

tumor necrosis factor

VDAC1

voltage-dependent anion channel 1.

Introduction

Lipopolysaccharide (LPS), a cell wall component of gram-negative bacteria such as Escherichia coli (E. coli), acts as a pathogen-associated molecular pattern that elicits innate immune responses in mammalian cells through the activation of pattern recognition receptors (PRRs).¹ Among PRRs, TLR4 (toll-like receptor 4) is thought to be the major PRR responsible for LPS recognition in many cells.¹ Innate immune signaling elicits inflammation by promoting cellular damage such as mitochondrial degeneration, the generation of reactive oxygen species, and the production of proinflammatory cytokines as well as type I interferons.² Despite its detrimental effect on hepatocytes, LPS alone causes little hepatocyte death,³ in contrast to the well-established acute liver injury model of LPS-D-galactosamine (GalN) co-administration. This is in part due to the suppression of LPS-induced autophagy by GalN,⁴ suggesting that autophagy is essential for the protection against immune damage in hepatocytes. One of the roles of autophagy in protecting hepatocytes against septic hepatotoxicity is the elimination of damaged mitochondria within the cells.^5,6

The elimination of aberrant proteins as well as organelles through autophagy is essential for cellular homeostasis,^7,8 as is obvious in liver homeostasis.^9,10 The conjugation of phosphatidylethanolamine (PE) to the C terminus of MAP1LC3 (LC3), which results in the conversion of LC3 from its cytosolic form (LC3-I) to a phagophore-associated form (LC3-II),¹¹ is an essential process in the nucleation step during autophagosome formation. The formation of the ATG12–ATG5-ATG16L1 complex is a prerequisite for this process,¹² so that Atg5-deficient cells cannot make LC3-II and autophagy is not induced.¹³ Autophagosomes, double-membrane vesicles that include cytoplasmic contents within their structure, fuse with lysosomes to deliver their contents for degradation by lysosomal hydrolytic proteases such as cathepsins. LAMPs, lysosomal membrane-associated proteins, are also suggested to be essential for the formation of autolysosomes, vesicles that are formed upon the fusion of autophagosmes and lysosomes. Prior to or following the formation of autolysosomes, LC3-II on the outer membrane is segregated and recycled to the cytoplasm, whereas LC3-II on the inner membrane is destined for degradation.¹¹

Recent research has pointed out the roles of the autophagy machinery in the secretion of intracellular contents into the extracellular milieu.^14,15 For example, the unconventional secretion of acyl coenzyme A-binding protein in yeast is dependent on the autophagy machinery.¹⁶ IL1B/interleukin-1β secretion from macrophages is enhanced by autophagy induction.¹⁷ A crucial role of the autophagy machinery has also been reported in the degranulation process of mast cells and in the release of lysosomal contents from ATP-stimulated microglia.^18,19 TFEB (transcription factor EB), a master regulator of autophagy and lysosome biogenesis,^20,21 also plays roles in lysosomal exocytosis,²² indicating that the role of autophagy is not limited to lysosomal degradation.

Other than the well-established role of mitochondrial dysfunction in the pathogenesis of sepsis/endotoxemia,² increased levels of mitochondrial contents, such as mitochondrial DNA as well as the liver mitochondrial protein CPS1 (carbamoyl-phosphate synthase 1, mitochondrial),^23-25 have been reported in plasma from critically ill and/or septic patients. Increased levels of CPS1 have also been reported in animal models of experimental sepsis suggesting that some mitochondrial contents are probably extruded into the blood during illness.^26,27 We have also shown the induction of autophagy and extrusion of mitochondrial proteins from liver during systemic LPS administration in rats.^6,26 These results prompted us to evaluate the relationship between autophagy and the extrusion of mitochondrial constituents in LPS-stimulated cells in vitro.

Results

LPS induces mitochondrial degeneration and cytoplasmic vacuolization in rat hepatocytes

To examine the effect of LPS administration on hepatocytes, primary cultures of parenchymal hepatocytes from rats were stimulated with 1 µg/ml LPS for 9 h. LPS-stimulated and unstimulated hepatocytes were then examined under transmission electron microscopy. A greater abundance of large or small mitochondria was observed in LPS-stimulated cells (Fig. 1B) as compared with unstimulated cells (Fig. 1A). The caspase inhibitor zVAD had no effect on the occurrence of these mitochondria in LPS-stimulated hepatocytes (Fig. 1B and C). In contrast to mitochondrial degeneration, little nuclear denaturation was observed (Fig. 1A–C). Notably, structures resembling damaged mitochondria were seen in the extracellular milieu of LPS-stimulated cells (Fig. 1D). Quantification of the EM images indicated that LPS-stimulated cells were more prone to be surrounded with mitochondria-like structures compared to unstimulated cells (Fig. 1E). We examined numbers and areas of mitochondria in unstimulated and LPS-stimulated cells with or without zVAD pretreatment. About a 40% decrease in the number of mitochondria per area (10 µm²) was observed in LPS-stimulated cells compared to unstimulated cells (Fig. 1F). Areas of mitochondria in LPS-stimulated cells showed a wider distribution compared to unstimulated cells, suggesting the accumulation of aberrant mitochondria (Fig. 1G). Both of these mitochondrial abnormalities were also observed in LPS + zVAD-treated cells (Fig. 1F and G), suggesting that caspases may not be involved in the mitochondrial degeneration. Taken together, mitochondrial degeneration occurs in LPS-stimulated hepatocytes, as is observed in rat liver.²⁶

Figure 1.

Degeneration of mitochondria in LPS-stimulated hepatocytes. Parenchymal hepatocytes prepared from rats were stimulated by LPS (9 h) with or without zVAD, and subjected to electron micrographic analysis. (A) Transmission electron micrograph of unstimulated hepatocytes (Nu, nucleus). (B and C) Transmission electron micrographs of LPS-stimulated hepatocytes with (C) or without (B) zVAD pretreatment. (D) Mitochondria-like structures in extracellular milieu of LPS-stimulated cells (PM, plasma membrane). Black arrowheads indicate mitochondria-like structures. Black bars: 2 µm. (E) Percentages of the cells that were surrounded with mitochondria-like structures in unstimulated and LPS-stimulated cells. (F) Numbers of mitochondria per area (10 µm²) in LPS-stimulated hepatocytes with or without zVAD pretreatmnt. The data shown are expressed as mean and SE of each group (>10 fields). (G) The distribution of mitochondrial areas (>100 mitochondria) in each experimental group. Normality of distribution was evaluated by Kolmogorov-Smirnov Test. **, p<0.01 versus control.

LPS induces autophagy and extracellular secretion of CPS1 in rat hepatocytes

The induction of autophagy as well as the secretion of CPS1 after LPS stimulation was next examined in the cells and in the culture supernatants. An extensive conversion of the autophagy marker LC3 from its cytosolic form (LC3-I) to its phagophore- and autophagosome-associated lipidated form (LC3-II), as well as a decrease of autophagy substrate SQSTM1/p62 (sequestosome 1), was observed, confirming that the direct administration of LPS induces autophagy in hepatocytes (Fig. 2A and B). An increase in the occurrence of the liver mitochondrial protein CPS1 was also observed in culture supernatants from LPS-stimulated cells (Fig. 2A and B). Extracellular levels of LC3-II, but not LC3-I, were also increased after LPS stimulation (Fig. 2A and B). In contrast, extracellular levels of EEA1 (an early endosome marker) as well as GAPDH (a cytosolic marker) were not increased. To further examine whether increased level of extracellular CPS1 indicates release of specific proteins or bulk mitochondrial proteins, we purified mitochondrial fractions from LPS-stimulated and unstimulated cells. Mitochondrial levels of CPS1, as well as COX4I1 (cytochrome c oxidase subunit IV isoform 1) and VDAC1 (voltage-dependent anion channel 1), seemed to be constant between LPS-stimulated and unstimulated cells (Fig. 2C). Therefore, CPS1 seemed to be secreted from LPS-stimulated cells as a constituent of mitochondria. In addition, we found that PARK2 and PINK1, proteins that play essential role in the progression of mitophagy,^28,29 were also secreted into culture supernatants in LPS-stimulated cells (Fig. 2D). These results confirm the induction of autophagy and the secretion of CPS1 in LPS-stimulated hepatocytes, as has been observed in the liver and serum of rats,^6,26 and also indicate that autophagy might be involved in the secretion of mitochondrial proteins including CPS1.

Figure 2.

Induction of autophagy and release of CPS1 in LPS-stimulated hepatocytes. (A) Activation of the autophagy marker LC3 in hepatocytes stimulated with LPS (24 h). (B) Extracellular release of CPS1 from LPS-stimulated (24 h) hepatocytes. Cell lysates (Intracellular) as well as culture supernatants (Extracellular) were subjected to immunoblot analysis for LC3, SQSTM1, CPS1, EEA1, GAPDH, and ACTA1/actin. ACTA1 served as a control for the analysis of cell lysates. ALB (albumin) levels were visualized by Ponceau S staining to confirm the equal loading of culture supernatants. (C) Relative levels of mitochondrial proteins (CPS1, COX4I1, and VDAC1) in purified mitochondrial fractions from unstimulated and LPS-stimulated (24 h) hepatocytes. (D) Extracellular release of PARK2 and PINK1 from LPS-stimulated (24 h) hepatocytes. Culture supernatants were subjected to immunoblot analysis for PARK2 and PINK1. The data shown are expressed as mean and SE of each group (n = 4 ). #, Not detected; **, p < 0.01.

LPS induces nuclear translocation of TFEB and expression of autophagy-lysosome system genes

TFEB localization was examined in LPS-stimulated hepatocytes, as nuclear TFEB translocation has been observed in the LPS-administered rat heart.³⁰ After transfection of GFP-TFEB, cells with or without LPS stimulation were observed under a fluorescence microscope. GFP-TFEB was observed in the cytoplasm of the nonstimulated cells, while it was observed in the nuclei of LPS-stimulated hepatocytes (Fig. 3A), indicating the nuclear translocation of TFEB follows LPS stimulation. GFP-LC3 puncta formation, an indicator of the formation of autophagosomes, was also observed in LPS-stimulated hepatocytes (Fig. 3A). Fluorescence from LAMP1-monomeric GFP (mGFP) seemed to be enhanced by LPS stimulation (Fig. 3A). Quantitative PCR (qPCR) analysis of several autophagy-lysosome system genes (LC3, SQSTM1, LAMP1, and LAMP2) showed that expressions of these genes were induced by LPS (Fig. 3B). Among these genes, SQSTM1 and LAMP1 have been suggested as direct targets of TFEB,²⁰ indicating the possible contribution of TFEB activation to the induction of the autophagy-lysosome system in LPS-stimulated hepatocytes. To examine whether the increase of LC3-II in LPS-stimulated hepatocytes (Fig. 2 A and B) indicates an increase of autophagic flux or not, we measured LC3-II levels both in the presence and absence of chloroquine (CQ), a lysosome inhibitor. Treatment with CQ significantly augmented the LPS-induced increase of LC3-II levels: LPS induced an approximately 4-fold increase of LC3-II, which was further increased to 6-fold in the presence of CQ (Fig. 3C). Estimation of autophagy flux by the use of a tandem fluorescent-tagged LC3 (tfLC3) vector further indicated that the flux was upregulated by LPS: both autophagosomes (yellow dots in merged images) and autolysosomes (red dots in merged images) were increased in LPS-stimulated cells, while only autophagosomes were observed in the cells co-administered with LPS and bafilomycin A₁ (Baf), another lysosome inhibitor (Fig. 3D).³¹ Baf, as well as a lysosomal cysteine protease inhibitor, E64d, partially suppressed LC3-II and CPS1 release into the medium (Fig. 3E and F), suggesting the involvement of lysosome in the secretion of mitochondrial proteins. Increased LC3-II and CPS1 levels in LPS-stimulated cell culture supernatants should not be due to increased cell lysis, since LDH release was constant between control, LPS, LPS + Baf, and LPS + E64d groups (Fig. 3G). We next examined whether autophagy induction as well as secretion of LC3-II and CPS1 is also observed in hepatoytes stimulated by pro-apoptotic stimuli such as LPS + GalN co-treatment.³² In LPS + GalN-stimulated hepatocytes, neither induction of autophagy nor secretion of LC3-II, CPS1, and COX4I1 was observed whereas apoptosis was evident by the immunoblot of active-CASP3 (caspase 3, apoptosis-related cysteine peptidase) (Fig. 3H). Collectively, secretion of LC3 as well as mitochondrial proteins is associated with increased autophagy flux, requires functional lysosomes, and is not observed in response to pro-apoptotic stimuli.

Figure 3 .

Nuclear translocation of TFEB, induction of the autophagy-lysosome system gene expression, and lysosome-dependent secretion of mitochondrial proteins in LPS-stimulated hepatocytes. (A) Hepatocytes were transfected with GFP-TFEB, GFP-LC3, and LAMP1-mGFP, treated with LPS for 24 h, and observed under fluorescence microscopy. Nuclear translocation of TFEB-GFP, as well as punctate formation of GFP-LC3, was observed in LPS-stimulated cells. The dotted structure of LAMP1-mGFP did not change during LPS stimulation, but the fluorescence tended to increase. (B) Quantitative RT-PCR analysis of the genes involved in the autophagy-lysosome system. The expressions of LC3, SQSTM1, LAMP1, and LAMP2 were all upregulated in response to LPS stimulation (24 h). (C) Increased autophagic flux in LPS-stimulated hepatocytes. The cells were stimulated with LPS (24 h) in the presence or absence of chloroquine (CQ, 3 µM). Cell lysates were subjected to immunoblot analysis for LC3, and ACTA1/actin. (D) Increased number of autolysosomes in LPS-stimulated hepatocytes. The cells were transfected with tfLC3 vector, stimulated with LPS (24 h) in the presence or absence of bafilomycin A₁ (Baf, 100 nM), and observed under fluorescence microscopy. Merged images of GFP and mRFP fluorescences were shown (upper panel). Numbers of autophagosomes (AP) and autolysosomes (AL) per cell were also shown (lower panel). (E) Effect of Baf on extracellular secretion of the proteins. Extracellular levels of LC3-II, and CPS1 in LPS and LPS + Baf groups were examined. (F) Effect of E64d on extracellular secretion of the proteins. Extracellular levels of LC3-II, and CPS1 in LPS and LPS + E64d groups were examined. (G) LDH release in control, LPS, LPS + Baf and LPS + E64d groups. (H) Lack of autophagy induction, activation of CASP3, and extracellular secretion of the proteins in LPS + D-galactosamine (GalN)-stimulated hepatocytes. Intracellular levels of cleaved (activated) CASP3 and LC3-II (left panels) and extracellular levels of LC3-II, COX4I1, and CPS1 (right panels) were examined after stimulation with LPS + GalN for 24 h. The data shown are expressed as the mean and SE of each group (n = 3 − 4). *, p < 0.05; **, p < 0.01.

Increased occurrence of CPS1, COX4I1, LC3-II, and CTSD, but not LAMP1, in the culture supernatant of LPS-stimulated hepatocytes

We next examined the mechanism of the release of CPS1 into the culture medium in response to LPS stimulation. Extrusion of mitochondrial protein from the cells might be achieved through at least 2 mechanisms: endocytosis-to-exocytosis pathway and autophagy-to-lysosome pathway. These 2 pathways might be interconnected, as autophagosomes often fuse with endosomes such as multivesicular bodies, a type of endocytic vesicle involved in the degradation of plasma membrane proteins and the secretion of intracellular proteins.³³ Whether endocytosis and/or autophagy is involved in the secretion of CPS1 was examined using pharmacological inhibitors of these processes. The administration of dynasore, a dynamin inhibitor that blocks endocytosis by inhibiting GTP hydrolysis of DNM1 (dynamin 1)-type small GTPases,³⁴ had no effect on CPS1 secretion (Fig. 4A and B). In contrast, 3-methyladenine, a PtdIsn3K inhibitor that blocks autophagosome formation in starved rat hepatocytes,³⁵ completely inhibited CPS1 upregulation in the culture supernatants of LPS-stimulated cells (Fig. 4A and B). Like CPS1, the mitochondrial inner membrane associated protein COX4I1 was also upregulated in the culture medium from LPS-stimulated hepatocytes (Fig. 4A and B), and LC3 was also released into the medium. Only the level of the autophagosome-associated form of LC3 (LC3-II) was found to be increased in the medium (Fig. 4A); levels of LC3-I seemed to be constant in the culture supernatants of all experimental groups (Fig. 4A). The activated form of CTSD, which resides in the lumenal space of lysosomes, was also secreted into the medium after LPS stimulation (Fig. 4A and B). In sharp contrast to CTSD, LAMP1, which is associated with lysosomal membranes, was not upregulated but rather was downregulated in the culture supernatants of the LPS group (Fig. 4A and B). Like CPS1, the upregulation of COX4I1, LC3-II, and CTSD in culture supernatants was reversed by 3MA, but not by dynasore (Fig. 4A and B). Intracellular levels of CPS1, COXI14, and CTSD decreased in response to LPS, and the decreases were cancelled by 3MA (Fig. 4A and B), correlating well with the extracellular upregulation. Plasma membrane integrities, as assessed by LDH leakage, were not changed significantly in any of the control, LPS, LPS + dynasore, or LPS + 3MA groups (Fig. 4C), suggesting that the increased occurrence of these proteins in culture supernatants is not the result of increased cell death. Collectively, these results show that mitochondrial proteins, as well as proteins that reside in the intralumenal space of autolysosomes, are secreted into the culture medium from LPS-stimulated hepatocytes. It is, therefore, feasible that mitochondria-loaded autolysosomes are secreted into the culture medium of LPS-stimulated hepatocytes through lysosomal exocytosis.

Figure 4.

Secretion of CPS1, COX4I1, LC3-II, and CTSD from LPS-stimulated hepatocytes and its suppression by the autophagy inhibitor 3MA. (A) Immunoblot analysis of cell lysates (intracellular) and culture supernatants (extracellular) of unstimulated and LPS-stimulated (24 h) hepatocytes pretreated with or without 3MA and dynasore (Dyn). Levels of CPS1, COX4I1, LC3-II, CTSD (activated form) and LAMP1 were determined in both cell lysates and culture supernatants in each experimental group. ACTA1 was also detected as a loading control. ALB was visualized by Ponceau S staining in culture supernatant samples. (B) Ratios of the levels of CPS1, COX4I1, LC3-II, CTSD, LAMP1 and ACTA1 (intracellular) or ALB (extracellular) were determined using densitometry analysis. (C) LDH release in control, LPS, LPS + Dyn, and LPS + 3MA groups. Each bar represents the mean and SE of 4 samples. #, Not detected; *, p < 0.05; **, p < 0.01.

Colocalization of mitochondria and autophagosomes-lysosomes during LPS administration in hepatocytes and secretion of mitochondrial DNA from LPS-stimulated cells

Whether mitochondria are engulfed within autophagosomes and delivered to lysosomes was examined in LPS-stimulated hepatocytes under fluorescence microscopy. Hepatocytes were transfected with vectors expressing MT-mKO1 (an orange fluorescent protein targeted to mitochondria), GFP-LC3, and LAMP1-mGFP to visualize mitochondria, autophagosomes, and lysosomes, respectively. Without LPS stimulation, GFP-LC3 was mainly observed to be diffused throughout the cytoplasm (Fig. 5A, a), while LAMP1-mGFP was observed as dotted structures, corresponding to the typical structure of lysosomes (Fig. 5A, b). An increased occurrence of GFP-LC3 dots was observed after LPS stimulation (Fig. 5A, c), corresponding to the increased formation of autophagosomes. In unstimulated cells, the MT-mKO1 and GFP-LC3 fluorescence overlapped only slightly with each other (Fig. 5A, a). In some of the LPS-stimulated cells, MT-mKO1 and GFP-LC3 fluorescence also overlapped only slightly (Fig. 5A, c), while in other cells GFP-LC3 fluorescence was colocalized with MT-mKO1 (Fig. 5A, e). Like MT-mKO1 and GFP-LC3, MT-mKO1 and LAMP1-mGFP fluorescence overlapped only slightly in some of the LPS-stimulated cells (Fig. 5A, d), while LAMP1-mGFP was colocalized with MT-mKO1 in other LPS-stimulated cells (Fig. 5A, f). Fig. 5B shows the percentage of the cells in which MT-mKO1 was colocalized with GFP-LC3 or LAMP1-mGFP. Differential response of the cells to LPS stimulation might be derived from cell-to-cell variability in their dynamics of autophagy induction, which has been pointed out in a previous report.³⁶ Collectively, it has been shown that at least some mitochondria should be delivered to the autophagy-lysosome system in LPS-stimulated hepatocytes. qPCR analysis showed that the relative abundances of mitochondrial DNA increased in the culture supernatants of LPS-stimulated hepatocytes in a 3MA-sensitive manner (Fig. 5C), suggesting the extrusion of mitochondria from LPS-stimulated cells via the autophagy-lysosome pathway. We further examined the effect of small interfering RNA-mediated knockdown of TFEB as well as ATG5 on the secretion of mitochondria in LPS-stimulated hepatocytes. Successful knockdown was confirmed by decreased levels of TFEB and ATG12–ATG5 conjugate in hepatocytes (Fig. 5D) and we found that the knockdown of gene expression efficiently suppressed the secretion of CPS1 and LC3-II (Fig. 5D).

Figure 5.

Colocalization of mitochondria and autophagosomes-lysosomes, and secretion of mitochondrial DNA in LPS-stimulated hepatocytes. (A) Colocalization of autophagosome marker GFP-LC3 and lysosome marker LAMP1-monomeric GFP (LAMP1-mGFP) with mitochondrial marker MT-monomeric KO1 (MT-mKO1) in control and LPS-stimulated hepatocytes. Hepatocytes were transfected with the vectors, stimulated with LPS for 9 h, and observed under fluorescence microscopy. Arrows indicate the colocalization of GFP-LC3 and MT-mKO1, as well as LAMP1-mGFP and MT-mKO1. (B) Percentages of cells with colocalization of MT-mKO1 and GFP-LC3 or LAMP1-mGFP in control and LPS groups. At least 100 cells were examined by visual inspection under fluorescence microscopy. (C) Secretion of mitochondrial DNA from LPS-stimulated hepatocytes. Relative levels of mitochondrial DNA (mtDNA) to those of nuclear DNA (nuDNA) in the culture supernatants of hepatocytes (control, LPS, and LPS + 3MA groups) were determined by qPCR analysis. (D) Effects of siRNA-mediated knockdown of TFEB and ATG5 on extracellular release of CPS1 and LC3-II from LPS-stimulated (24 h) hepatocytes. The cells were treated with siRNAs for the indicated targets, and examined with intracellular levels of TFEB and ATG12–ATG5 conjugate (left panels) as well as extracellular levels of CPS1 and LC3-II (right panels). NC indicates nontargeted control siRNA. Results of 2 independent experiments (#1 and #2) are shown. Each bar represents the mean and SE of 4 samples. **, p < 0.01.

Secretion of COX4I1, LC3-II, and CTSD into the culture supernatant of Atg5^+/+ MEFs, but not atg5^−/− MEFs, after LPS stimulation

To further confirm the involvement of autophagy, atg5^−/− mouse embryonic fibroblasts (MEFs) were used. The secretion of COX4I1, LC3-II, and CTSD into the culture supernatant was observed for LPS-stimulated Atg5^+/+ MEFs but not LPS-stimulated atg5^−/− MEFs (Fig. 6A and B). LAMP1 was not secreted into the culture supernatant of either LPS-stimulated Atg5^+/+ or atg5^−/− MEFs (Fig. 6A and B). The integrities of the plasma membranes did not change significantly in any experimental group, as LDH leakage was barely detectable in any group (Fig. 6C). We also confirmed absence of the ATG12–ATG5 conjugate in atg5^−/− MEFs (Fig. 6D). These results are in good agreement with the results for hepatocytes (Fig. 4), and further confirm the involvement of the autophagy machinery.

Figure 6.

For figure legend, see page 1529.

Colocalization of mitochondria and autophagosomes-lysosomes during LPS administration in Atg5^+/+ MEFs and the secretion of mitochondrial DNA from LPS-stimulated cells

We next examined the colocalization of MT-mKO1, GFP-LC3, and LAMP1-mGFP in LPS-stimulated Atg5^+/+ MEFs (Fig. 7A). Although MT-mKO1, GFP-LC3, and LAMP1-mGFP were not colocalized in unstimulated cells (Fig. 7A, a and b) and some of the LPS-stimulated cells (Fig. 7A, c and d), the colocalization of MT-mKO1 and GFP-LC3 as well as LAMP1-mGFP was observed in some of the LPS-stimulated cells (Fig. 7A, e and f). Figure 7B shows the percentage of the cells in which MT-mKO1 was colocalized with GFP-LC3 or LAMP1-mGFP. The relative abundance of mitochondrial DNA increased in the culture supernatants of LPS-stimulated Atg5^+/+ MEFs, but not that of LPS-stimulated atg5^−/− MEFs (Fig. 7C). Although no significant differences in LC3-II levels were observed between unstimulated and LPS (24 h)-stimulated cells (Fig. 6A and B), we observed an approximately 2-fold increase of LC3-II levels after stimulation with LPS for 5 h (Fig. 7D). Like hepatocytes, treatment with CQ significantly augmented the LPS-induced increase of LC3-II levels (Fig. 7D). Estimation of autophagy flux by the use of tfLC3 further indicated that the flux was upregulated by LPS in MEFs (Fig. 7E). Levels of COX4I1 and VDAC1 in mitochondrial fractions seemed to be unchanged by LPS stimulation (Fig. 7F), as observed in hepatocytes (Fig. 2C). These results are consistent with the results obtained for LPS-stimulated hepatocytes, and confirm the exocytosis of mitochondria-loaded autolysosomes from LPS-stimulated MEFs. Finally, we investigated the effect of lysosome inhibition on accumulation of intracellular oxidative stress, which is elicited by LPS. Accumulation of 4-hydroxy-2-nonenal (4-HNE)-modified proteins was observed in LPS-stimulated cells and was further enhanced in LPS + Baf-stimulated cells (Fig. 7G). Thus, clearance of damaged mitochondria through lysosomal degradation as well as lysosomal exocytosis should be essential against oxidative stress during LPS stimulation.

Figure 7.

For figure legend, see page 1531.Colocalization of mitochondria and autophagosomes-lysosomes, and secretion of mitochondrial DNA in LPS-stimulated MEFs. (A) Colocalization of autophagosome marker GFP-LC3 and lysosome marker LAMP1-mGFP with mitochondrial marker MT-mKO1 in control and LPS-stimulated MEFs. Atg5^+/+ MEFs were transfected with vectors, stimulated with LPS for 9 h, and observed under fluorescence microscopy. Arrows indicate the colocalization of GFP-LC3 and MT-mKO1, as well as LAMP1-mGFP and MT-mKO1. (B) Percentages of cells with colocalization of MT-mKO1 and GFP-LC3 or LAMP1-mGFP in control and LPS groups. At least 100 cells were examined by visual inspection under fluorescence microscopy. (C) Caspase-independent secretion of mitochondrial DNA from LPS-stimulated Atg5^+/+ MEFs but not LPS-stimulated atg5^−/− MEFs. Relative levels of mitochondrial DNA (mtDNA) to those of nuclear DNA (nuDNA) in the culture supernatants of MEFs were assessed by qPCR analysis. (D) Increased autophagic flux in LPS-stimulated Atg5^+/+ MEFs. The cells were stimulated with LPS (5 h) in the presence or absence of chloroquine (CQ, 3 µM). Cell lysates were subjected to immunoblot analysis for LC3, and ACTA1. (E) Increased number of autolysosomes in LPS-stimulated Atg5^+/+ MEFs. The cells were transfected with tfLC3 vector, stimulated with LPS (5 h) in the presence or absence of Baf (100 nM), and observed under fluorescence microscopy. Merged images of GFP and mRFP fluorescences were shown (upper panel). Numbers of autophagosomes (AP) and autolysosomes (AL) per cell were also shown (lower panel). (F) Relative levels of mitochondrial proteins (COX4I1 and VDAC1) in purified mitochondrial fractions from unstimulated and LPS-stimulated Atg5^+/+ MEFs. (G) Accumulation of 4-HNE-adducted proteins in LPS-stimulated Atg5^+/+ MEFs. Intracellular levels of 4-HNE-adducted proteins were determined in LPS (3 h)-stimulated cells with and without Baf pretreatment. Each bar represents the mean and SE of 4 samples. *, p<0.05; **, p<0.01.

Purified mitochondria activate PMNs

Since the accumulation of 4-HNE-modified proteins in LPS-stimulated cells and its augmentation by lysosomal inhibition indicated an essential role of lysosomes against LPS-elicited oxidative stress (Fig. 7G), the biological significance of the lysosomal exocytosis of mitochondria remains to be elucidated. It has been demonstrated that mitochondrial damage-associated molecular patterns, such as formyl peptides and mitochondrial DNAs, could activate polymorphonuclear neutrophils.³⁷ Moreover, a recent report shows that intact mitochondria could activate monocyte-derived macrophages.³⁸ Therefore, we examined whether direct exposure of mitochondria to immune cells upregulates cytokine production in these cells. Polymorphonuclear leukocytes (PMNs), which are composed mainly from neutrophils, were obtained from healthy rats, and mitochondrial as well as cytoplasmic fraction was purified from healthy rat liver (Fig. 8A). Then PMNs were exposed to purified mitochondria at a final concentration of 100 µg/ml. IL1B expression was significantly induced by exposure to the mitochondrial fraction, but not by the cytoplasmic fraction (Fig. 8B), suggesting that PMNs were activated by mitochondria. Next we evaluated whether mitochondria secreted from LPS-stimulated hepatocytes could also stimulate PMNs. Mitochondrial fractions were collected from the culture supernatants of LPS-stimulated hepatocytes and administered to PMNs at a final concentration of 10 µg/ml (1/100 dilution with the culture medium, which resulted in less than 10 pg/ml endotoxin levels). A significant increase of TNF/TNFα expression, as well as a trend of increased IL1B expression, was observed in the cells administered with the secreted mitochondrial fraction (Fig. 8C). In contrast to PMNs, no increase in cytokine production was observed in hepatocytes (Fig. 8D). These results indicate that the mitochondria released from LPS-stimulated hepatocytes should activate immune cells, rather than hepatocytes.

Figure 8.

Mitochondria activate PMNs. (A) Relative levels of COX4I1 (mitochondrial marker) and GAPDH (cytoplasmic marker) in mitochondrial (Mito) and cytoplasmic (Cyto) fractions purified from rat liver. (B) Effects of the exposures to liver mitochondrial or cytoplasmic fractions on pro-inflammatory cytokine (TNF and IL1B) expression in PMNs. The cells were treated with mitochondrial fraction (Mito [L], 100 µg/ml) or cytoplasmic fraction (Cyto [L],100 µg/ml) for 5 h, and subjected to qPCR analysis. (C and D) Effects of the exposures to mitochondrial fraction from LPS-stimulated hepatocyte culture supernatant on TNF and IL1B expression in PMNs (C) and hepatocytes (D). These cells were treated with liver mitochondrial or cytoplasmic fractions, the fraction of mitochondria from hepatocyte culture supernatant (Mito [S],10 µg/ml), or LPS (1 µg/ml) for 5 h and subjected to qPCR analysis. Each bar represents the mean and SE of 3–4 samples. *, p < 0.05; **, p < 0.01.

Discussion

In this study, we demonstrate a novel role of autophagy other than for cellular cannibalism in LPS-stimulated cells. Our current results show that LPS stimulates the secretion of mitochondrial proteins and DNA in an autophagy-dependent manner. The secretion of mitochondrial constituents was observed not only for rat hepatocytes, but also for mouse embryonic fibroblasts, suggesting that this process might be a relatively general, rather than cell type-specific, phenomenon. The results obtained in the current study are schematically summarized in Fig. 9.

Figure 9.

Scheme for mitochondrial secretion through the autophagy-lysosome system in LPS-stimulated cells. In LPS-stimulated cells, damaged mitochondria are engulfed by autophagosomes, which fuse to lysosomes to form autolysosomes. The resultant autolysosomes include mitochondria. Exocytosis of the autolysosomes results in the secretion of their lumenal contents (CPS1, COX4I1, PARK2, PINK1, LC3-II, and CTSD) into the extracellular space, while the autolysosomal membrane protein (LAMP1) is retained on the plasma membrane. The lysosomal membrane is indicated as a blue line and the autophagosomal membrane as a red line.

COX4I1, a component of the mitochondrial respiratory chain, is associated with the mitochondrial inner membrane, and resides in the intermembrane space, while CPS1, a urea cycle protein, resides in the mitochondrial matrix.³⁹ It has been reported that the mitochondrial membrane is one source of autophagosomal membranes: the mitochondrial membrane is transiently incorporated into the phagophore membrane during starvation-induced autophagy.⁴⁰ It has also been reported that ER-mitochondria contact sites are the site from which autophagosomes arise.⁴¹ Therefore, there might be a possibility that COX4I1 secreted from LPS-stimulated hepatocytes (Fig. 4) and MEFs (Fig. 6) is associated with the membrane incorporated into phagophores. However, this scenario should be the least likely, since the mitochondrial matrix protein CPS1 is also extruded from hepatocytes (Figs. 2 and 4). Mitochondrial constituents, not mitochondrial membrane-associated proteins that might be incorporated into the phagophore membrane, should be secreted from LPS-stimulated hepatocytes and MEFs. The secretion of mitochondrial DNA (Figs. 5C and 7C) further confirms the secretion of mitochondrial constituents from LPS-stimulated cells.

Although the secretion of IL1B from macrophages is enhanced by starvation-induced autophagy,¹⁷ autophagy negatively regulates LPS-stimulated cytokine production.^42,43 The generation of reactive oxygen species from mitochondria,⁴⁴ as well as the cytoplasmic release of mitochondrial DNA due to mitochondrial dysfunction, are involved in the activation of the NLRP3 inflammasome.⁴⁵ Autophagy negatively regulates inflammasome activation by eliminating damaged mitochondria, thereby reducing the production of pro-inflammatory cytokines such as IL1B and IL18.^44,45 In contrast to the negative regulation of cytokine production, positive regulation by autophagy has been reported for the secretion of cellular organelles/granules. For example, the impaired release of secretory granules that include histamine as well as β-hexosaminidase has been observed in Atg7-deficient bone marrow-derived mast cells during degranulation by Fc receptor stimulation.¹⁸ Takenouchi, et al. have reported that the autolysosomal components LC3-II and CTSD are released from ATP-stimulated glial cells in a Ca²⁺-dependent manner.¹⁹ This is accompanied by increases in intracellular lysosomal pH, suggesting that impaired lysosomal function by ATP stimulation results in the inhibition of the digestion of autolysosomal contents; these undigested lysosomal contents are instead released into the extracellular space.¹⁹ Thus, the extracellular release of autolysosomal contents might be a cellular compensation mechanism for the impairment of lysosomal degradation. Collectively, autophagy should be involved in the regulation of the secretion of cellular proteins as well as vesicles/granules, but the role of autophagy appears to be highly context dependent.

Mitochondrial extrusion through exocytosis has also been reported elsewhere. Nakajima et al. have reported that fragmented mitochondria in MEFs deficient in CFLAR/c-flip (cflar^−/− MEFs) are released through a mechanism mediated by an endocytosis-exocytosis pathway during the course of apoptosis elicited by TNF.⁴⁶ However, the autophagy machinery has proved not to be involved in this case, as 3MA has no effect.⁴⁶ To the best of our knowledge, our report is the first to demonstrate mitochondrial extrusion from LPS-stimulated cells through the autophagy machinery. Recent research has indicated that the final point of autophagy is not always lysosomal degradation. TFEB, the master transcriptional regulator of the autophagy-lysosome system, also induces lysosomal exocytosis.²² This autophagy-lysosome-exocytosis axis of the cellular elimination system is called the CLEAR system.⁴⁷ Suppression of CPS1 and LC3 release in response to TFEB knockdown in hepatocytes (Fig. 5D) suggests that TFEB is involved in autolysosomal exocytosis in LPS-stimulated cells.

It is worth noting that the possible involvement of mitochondrial constituents released into the plasma has been reported in a systemic inflammatory disease model: mitochondrial DNA, with an unmethylated CpG motif that resembles bacterial DNA, acts in a damage-associated molecular pattern that elicits multi-organ injuries following its release into the circulation due to trauma.³⁷ Very recently, Maeda et al. have shown that intact mitochondria are also released from cells undergoing TNF-induced necroptosis, and the released mitochondria could activate monocyte-derived macrophages.³⁸ Hence, our current results that PMNs could be activated by mitochondria (Fig. 8) are in agreement with these previous reports. Nevertheless, further studies are obviously needed to reveal more precise roles of mitochondrial constituents secreted from LPS-stimulated cells into the extracellular milieu.

Materials and Methods

Preparation of parenchymal hepatocytes from rats and the culture of hepatocytes and MEFs

Parenchymal hepatocytes were prepared from 8-wk-old male Sprague-Dawley rats by a standard collagenase-perfusion method with slight modifications.⁴⁸ In brief, livers of anesthetized rats were briefly perfused with Hank's solution containing glucose and EGTA, followed by perfusion with digestion buffer containing collagenase (Wako, 034–22363) and dispase (Roche, 4942086). Parenchymal hepatocytes were precipitated by centrifugation 2 times at 50 g for 1 min. The cells obtained were seeded on collagen-coated 3-cm diameter dishes (Iwaki, 4000–010) at 5×10⁴ cells/ml in Williams medium E (Sigma-Aldrich, W4125) containing 10% fetal bovine serum (FBS). Hepatocyte viability was estimated by a dye exclusion assay and routinely reached ~80%. Two d after seeding, the cells were stimulated with LPS (Sigma-Aldrich, L-2630) at a concentration of 1 µg/ml for the indicated time periods. For inhibitor studies, 50 µM z-VAD-FMK (zVAD, Calbiochem, 627610), 3 µM chloroquine (CQ, Wako, 038–17971), 100 nM bafilomycin A₁ (Sigma-Aldrich, B1793), 1 µg/ml E64d (Sigma-Aldrich, E8640), 2 mM 3-methyladenine (3MA, Sigma-Aldrich, M9281) and 50 µM dynasore (Sigma-Aldrich, D7693) were included in the medium prior to the administration of LPS. D-galactosamine (Wako, 075–05013) was also used in some experiments at a final concentration of 40 mM. Atg5 deficient (atg5^−/−) (RIKEN, RCB2710) as well as wild-type (Atg5^+/+) (RIKEN, RCB2711) MEFs were obtained from Dr. Noboru Mizushima (University of Tokyo), and maintained in DMEM supplemented with 10% FBS.¹³

Isolation of polymorphonuclear leukocytes and mitochondria

Neutrophil-enriched polymorphonuclear leukocytes (PMNs) were obtained as follows: Whole blood from rats was mixed with an equal volume of 3% dextran (molecular mass 450~650 kDa; Sigma, 31392), 0.9% NaCl solution at room temperature for 20 min to facilitate the formation of erythrocyte rouleaux. After centrifugation at 1000g for 10 min, the resultant cell pellet fraction was treated with 0.2% NaCl for 30 sec to facilitate the lysis of residual erythrocytes. These steps were repeated 2 times. After washing with 0.9% NaCl, neutrophil-enriched PMNs were resuspended in DMEM containing 10% FBS. Rat liver mitochondria were isolated as described previously.⁶ In brief, liver was homogenized with a Potter-Elvehjem homogenizer (400 rpm, 2 strokes) in buffer A (10 mM Tris, pH 7.4, 0.1 mM EDTA, 250 mM sucrose, 0.5 mM DTT, 0.5 mM PMSF), centrifuged at 1,000g for 10 min, and the resultant supernatant fraction was further centrifuged at 10,000g for 10 min. The resulting pellet fraction was resuspended in buffer A and further homogenized with a Dounce homogenizer (loose pestle, 5 strokes) to disrupt unbroken cells, followed by centrifugation at 10,000xg for 10 min. The resultant pellet fraction was used as the mitochondrial fraction while the supernatant fraction was used as the cytoplasmic fraction. The mitochondrial fraction from the culture supernatants in LPS-stimulated hepatocytes was collected by use of the Mitochondria Isolation Kit for Cultured Cells (Thermo scientific, 89874). Endotoxin levels in the fraction were determined to be less than 1 ng/ml using the limulus amebocyte lysate assay (GeneScript, L00350).

Transmission electron microscopy

Transmission electron microscopy analysis was performed as described previously.³⁰ In brief, unstimulated and LPS-stimulated hepatocytes were washed with 0.1 M phosphate buffer (pH 7.4) and fixed with phosphate buffer containing 2% glutaraldehyde. After washing in 0.2 M phosphate buffer, the fixed cells were incubated with 2% osmium tetroxide for 2 h, dehydrated in ethanol and embedded in Epon epoxy resin (Epon 812; TAAB, R3243). Ultrathin sections of the embedded hepatocytes were stained with uranyl acetate and lead citrate, and examined under an electron microscope (Hitachi, H800). The percentage of the cells that were surrounded with mitochondria-like structures (more than 10 mitochondria-like structures outside the cells) was counted for at least 20 cells in each control and LPS group. The number of mitochondria per area was counted for at least 10 cells in each experimental group. Mitochondrial areas in trans-sections of the hepatocytes were calculated using ImageJ software (1.47 v, downloaded from the National Institutes of Health). At least 100 mitochondria were examined in each experimental group.

Immunoblotting

The cells were lysed with lysis buffer (10 mM Tris-HCl, pH 8.0, 320 mM sucrose [Wako, 196–00015], 1 mM EDTA, protease inhibitor cocktail [Complete; Roche, 11697498001], 50 mM NaF, and 10 mM Na₃VO₄), and the protein concentrations of the lysates were determined using a Coomassie Protein Assay Kit (Thermo Fisher Scientific, 23200). Culture supernatants were spun at 500g for 5 min. to remove dead cells, and the resultant supernatant fractions were resolved directly in Laemmli sample buffer.⁴⁹ Equal amounts of cellular proteins and equal volumes of culture supernatants per lane were subjected to SDS-PAGE in discontinuous (4–15%) gels (Mini-PROTEAN-TGX precast gel; Bio-Rad, 456–1086), and electrophoretically transferred to a PVDF membrane using the Trans-Blot Turbo transfer system (Bio-Rad, 170–4157). Immunoblot analysis was performed with anti-CPS1 (1:1000 dilution; Santa Cruz Biotechnology, sc-10516), anti-LC3 (1:5000 dilution; Cell Signaling Technology, 4445), anti-COX4I1 (1:5000 dilution; Molecular Probes, A-21348), anti-VDAC1 (1:1000 dilution; Cell Signaling Technology, 4866), anti-EEA1 (1:1000 dilution; Cell Signaling Technology, 2411), anti-GAPDH (1:1000 dilution; Chemicon, MAB374), anti-PARK2 (1:1000 dilution; Cell Signaling Technology, 4211), anti-PINK1 (1:1000 dilution; Santa Cruz Biotechnology, sc-133796), anti-cleaved CASP3/caspase3 (1:1000 dilution; Cell Signaling Technology, 9661), anti-CTSD/cathepsin D (1:1000 dilution; Santa Cruz Biotechnology, sc-6486), anti-LAMP1 (1:5000 dilution; Cell Signaling Technology, 3243), anti-TFEB (1:500 dilution; abcam, 56330), anti-Atg5 (1:1000 dilution; MBL, PM050), anti-4-HNE (1:100 dilution; JaICA, MHN-100P) or anti-ACTA1/actin (1:5000 dilution; Sigma-Aldrich, A2066) antibodies. Peroxidase-conjugated anti-rabbit, -mouse, and -goat IgG antibodies (Promega, W4011, W4021, and V8051) were used as secondary antibodies. Relative protein levels were determined from a standard curve constructed by plotting band densities, and were normalized to the ACTA1 level using software for densitometric analysis (CS Analyzer, ver. Three.0, ATTO). For the determination of relative mitochondrial levels of CPS1, and COX4I1, mitochondrial fractions were isolated from the hepatocytes as well as MEFs using the Mitochondria Isolation Kit for Cultured Cells. For the estimation of relative extracellular levels of COX4I1, LC3-II, CTSD, and LAMP1 in MEFs, cell lysates (lysed in 100 µl of lysis buffer) as well as culture supernatants (1 ml) were resolved directly with Laemmli sample buffer and loaded on the same gel (5 µl for the cell lysate samples and 10 µl for the culture supernatant samples).

Transfection of fluorescence marker vectors and observation under a florescence microscope

Plasmid vectors expressing fluorescently tagged-proteins were transfected transiently into hepatocytes and MEFs by use of Lipofectamine 2000 (Invitrogen, 11688–019). Vectors expressing GFP-TFEB (Addgene plasmid 38119, deposited by Dr. Ferguson), GFP-LC3, LAMP1-mGFP, or tfLC3 (Addgene plasmid 21074, deposited by Dr. Yoshimori) were obtained from Dr. Shawn M. Ferguson (Yale University), Dr. Takeshi Noda (Osaka University), Dr. Esteban C. Dell'Angelica (University of California, Los Angels), or Dr. Tamotsu Yoshimori (Osaka University) respectively.^11,31,50,51 For the staining of mitochondria, pMT-mKO1, which expresses mitochondrially-targeted monomeric orange fluorescent protein (monomeric Kusabira Orange1, mKO1) by use of the signal peptide of COX4I1 for mitochondrial localization, was used (MBL, AM-V0221).⁵² One d prior to experiments, cells were transfected with plasmid vectors. The transfected cells were then stimulated with LPS, and followed by observation under a fluorescence microscope (Keyence, BZ-8100).

siRNA knockdown

siRNA for rat Tfeb (QIAGEN, SI01757063) and Atg5 (QIAGEN, SI01871695) were mixed with Lipofectamine RNAiMAX (Invitrogen, 13778030) at a final concentration of 50 nM in serum-free Williams medium E (400 µl/3 cm dish) and exposed to hepatocytes for 4 h, followed by addition of serum-containing medium (2 ml) and further incubation for another 44 h. Then, the cells were stimulated with LPS and protein extracts prepared for immunoblot analysis.

LDH release assay

Plasma membrane integrity was assessed by measuring the leakage of LDH using an LDH release assay kit (Wako, 299–50601) according to the manufacturer's instructions. In brief, cells attached to the dish were collected using a cell scraper, and lysed in phosphate-buffered saline (Nissui, 05913) containing 0.2% Tween 20 (Wako, 166–21115), whereas culture supernatants were lysed directly by adding 0.2% Tween 20. LDH activities in the cell and in the culture supernatants were measured. LDH release (%) was expressed as the relative LDH activities in the culture supernatants.

Quantitative PCR

For quantification of relative mRNA abundances, total cellular RNA was extracted from hepatocytes using TRIzol reagent (Invitrogen, 15596–026), and cDNA was synthesized by reverse transcription using SuperScript II enzyme (Invitrogen, 18064–014) and oligo (dT)₁₅. For quantification of relative abundance of mitochondrial and nuclear DNA, culture supernatants of hepatocytes and MEFs were used as templates. The mt-co1/Cox1 gene was used for the estimation of mitochondrial DNA while the Rn18s rRNA gene was used for the estimation of nuclear DNA. Quantitative PCR (qPCR) was performed with the StepOnePlus Real-Time PCR System (Applied Biosystems) using a GoTaq qPCR master mixture including SYBR green (Promega, A6001). The PCR reaction conditions were: 95°C for 20 sec, followed by 40 cycle of 95°C for 1 sec and 60°C for 20 sec. The primers used were: 5′-CGGGTTGAGGAGACACACAA-3' and 5′-TCTTTGTTCGAAGCTCCGGC-3′ for LC3; 5′-CGGGT-TGAGGAGACACACAA -3′ and 5′-TCTTTGTTCGAAGCT-CCGGC-3′ for SQSTM1; 5′-GCAAGGCGCTCGCCCTCAAT-3′ and 5′-GCCCGCGTGACTCCTCTTCC-3′ for LAMP1; 5′-AGCAGGTGGTTTCCGTGTCTCG-3′ and 5′-AGGGCTGCTCCCACCGCTAT-3′ for LAMP2; and 5′-GGCTCTCTGCTCCTCCCTGTTCTA-3′ and 5′-TGCC-GTTGAACTTGCCGTGGG-3′ for GAPDH; 5′-TCGGAGC-CCCAGATATAGCA-3′ and 5′-TTTCCGGCTAGAGGTG-GG-TA' for mouse mt-co1/Cox1; 5′-AGAGCGGGTAAGAGAGGT GT-3′ and 5′-GTCGGGGTCCGACAAAACC-3′ for mouse Rn18s rRNA; 5′-TGGAGGCTTCGGAAACTGAC-3′ and 5′-GCTAGGTTTCCGGCTAAGGG-3′for rat mt-co1/Cox1; 5′-CGAACGTCTGCCCTATCAACTT-3′ and 5′-CTT-GGATGTGGTAGCCGTTTCT-3′ for rat Rn18s rRNA; 5′-CATCCGTTCTCTACCCAGCC-3′ and 5′-AATTCTGAGC-CCGGAGTTGG-3′ for TNF; 5′-GCAGCTTTCGACAGTGAGGA-3′ and 5′-TCATCTGGACAGCCCAAGTC-3′ for IL1B.

Statistical analysis

Data are expressed as means ± SE of 3–4 samples in each experimental group. The Student t test was used to assess the statistical significance between 2 experimental groups. Multiple group comparisons were analyzed by analysis of variance (ANOVA) in combination with Tukey or Dunnett post hoc test. Statistical differences were considered significant at p < 0.05.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Dr. Noboru Mizushima (University of Tokyo), Dr. Takeshi Noda (Osaka University), Dr. Esteban C. Dell'Angelica (University of California, Los Angeles), Dr. Shawn M. Ferguson (Yale University), and Dr. Tamotsu Yoshimori (Osaka University) for providing materials. The mouse embryonic fibroblast cell lines, Atg5^+/+ MEFs and atg5^−/− MEFs, were provided by RIKEN BRC through the National Bio-Resource Project of the MEXT, Japan.

Funding

This study was supported by JSPS KAKENHI Grant Number 24790639 and 30586425 to KaU, and MEXT KAKENHI Grant Number 25460862 to TA.