Suppression of autophagy during liver regeneration impairs energy charge and hepatocyte senescence in mice

Authors


  • Potential conflicts of interest: Nothing to report.

Abstract

Autophagy is a homeostatic mechanism that regulates protein and organelle turnover and uses the amino acids from degraded proteins to produce adenosine 5'-triphosphate (ATP). We investigated the activity of autophagy-associated pathways in liver regeneration after partial hepatectomy (PHx) in liver-specific autophagy-related gene 5 (Atg5) knockout (KO) mice. Liver regeneration was severely impaired by 70% PHx, with a reduction in postoperative mitosis, but a compensating increase in hepatocyte size. PHx induced intracellular adenosine triphosphate and β-oxidation reduction as well as injured cellular mitochondria. Furthermore, PHx in Atg5 KO mice enhanced hepatic accumulation of p62 and ubiquitinated proteins. These results indicated that reorganization of intracellular proteins and organelles during autophagy was impaired in the regenerating liver of these mice. Up-regulation of p21 was associated with hepatocyte senescence, senescence-associated β-galactosidase expression, irreversible growth arrest, and secretion of senescence-associated molecules, including interleukin (IL)-6 and IL-8. Conclusion: These findings indicate that autophagy plays a critical role in liver regeneration and in the preservation of cellular quality, preventing hepatocytes from becoming fully senescent and hypertrophic. (Hepatology 2014;60:290–300)

Abbreviations
Akt

protein kinase B

ALT

alanine aminotransferase

AMP

adenosine monophosphate

AMPK

AMP-activated protein kinase

Atg

autophagy-related gene

ATP

adenosine 5′-triphosphate

BrdU

bromodeoxyuridine

CQ

chloroquine

ERK

extracellular signal-regulated kinase

FFAs

free fatty acids

HGF

hepatocyte growth factor

IHC

immunohistochemistry

IL

interleukin

KO

knockout

LC3

microtubule-associated protein 1 light chain 3

LW/BW

liver weight/body weight ratio

mRNA

messenger RNA

mTOR

mammalian target of rapamycin

NF-κB

nuclear factor kappa B NPCs, nonparenchymal cells

Nrf2

nuclear factor (erythroid-derived 2)-like 2

OS

oxidative stress

PCR

polymerase chain reaction

PHx

partial hepatectomy

SA-β-gal

senescence-associated β-galactosidase

SASP

senescence-associated secretory phenotype

SIRT1

sirtuin-1

Liver regeneration is a well-orchestrated process, in which complex signaling pathways coordinate the progression of distinct stages, including withdrawal of hepatocytes from quiescence (“priming phase”), cell-cycle entry and progression, cessation of cell division, and return of hepatocytes to quiescence.[1] Although hepatocytes rarely divide under normal circumstances, the liver has a remarkable ability to regenerate after surgical removal or after viral or chemical injury.[1-3] For example, after partial hepatectomy (PHx) of two thirds of the liver, a rodent model of liver regeneration, the remaining third grows rapidly to restore the liver's original mass, structure, and function within a few days. Indeed, >95% of mature hepatocytes synchronously exit the G0 phase and re-enter the cell cycle.[2, 3] Hepatocytes are the first cells to replicate, followed sequentially by biliary epithelial cells, Kupffer cells, stellate cells, and sinusoidal endothelial cells.[1-3] Hepatocyte proliferation in response to cytokines and growth factors plays a central role in liver regeneration.[1-3]

Early stress signals occurring after PHx may be the result of an increase in energy demand per unit liver mass. Remnant tissue retains liver-specific functions, such as gluconeogenesis and ureagenesis, and continues to produce adenosine 5'-triphosphate (ATP) for synthesis of proteins, nucleic acids, and other cell constituents.[4, 5] There was a marked decline in ATP within 6 hours after PHx, which was maintained throughout the prereplicative period. Moreover, the change in hepatocytes from the quiescent to replicative mode during the early phase of recovery was accompanied by reorganization of intracellular proteins and organelles.[1-3] However, the mechanisms underlying restoration of normal liver function, including the role of systemic metabolism, and reorganization of intracellular content throughout the replicative period after PHx remain unknown.

Autophagy is a homeostatic mechanism that regulates turnover of long-lived or damaged proteins and organelles, buffers intracellular constituents, and supplies amino acids taken from degradation products of the autolysosome.[6] The first step of autophagy involves the formation of a lipid bilayer structure, which sequesters cytoplasmic materials to form autophagosomes. These autophagosomes engulf organelles and then fuse with lysosomes to form mature autolysosomes, in which the sequestered proteins are digested into amino acids by lysosomal enzymes.[7] Based on its ability to reorganize intracellular proteins and organelles, and to modulate intracellular energy,[6-8] autophagy may be essential for liver regeneration, and its activities may be critical for mitotic or hypertrophic hepatocytes.

In a mouse model with liver-specific knockout (KO) of autophagy-related gene 7, livers exposed to long-term surveillance exhibited significant hepatomegaly with aggregates of unfolded proteins,[9] which is apparently inconsistent with the role of autophagy in promoting cell proliferation. To investigate the mechanism by which autophagy regulates hepatocyte proliferation and liver growth, we developed a mouse model of liver regeneration with liver-specific KO of autophagy-related gene 5 (L-Atg5 KO mice). Unexpectedly, liver regeneration after PHx was significantly delayed and was accompanied by delays in DNA synthesis and cell-cycle arrest during liver regeneration. Furthermore, we found that impaired hepatocyte proliferation was biologically distinct from cell-cycle arrest because it represented cellular senescence accompanied by increased cell size (hypertrophy) and accumulation of senescence-associated enzyme β-galactosidase (SA-β-gal).[10] These results suggest that autophagy plays critical roles in liver regeneration after PHx by maintaining intracellular energy production, as well as constitutive proteins and organelles, to prevent hepatocyte senescence.

Materials and Methods

Generation of Liver-Specific Atg5-Deficient Mice

Atg5flox/flox; Mx1-Cre mice were generated by crossing Atg5flox/flox mice, in which exon 3 of the Atg5 gene is flanked by two loxP sequences,[11] with transgenic mice expressing Cre recombinase under control of an Mx promoter (Mx-Cre).[12] Both strains of mice were purchased from The Jackson Laboratory (Bar Harbor, ME).[9] Recombination was successful in >90% of all hepatocytes from Atg5flox/flox; Mx-Cre mice. Mice were genotyped by polymerase chain reaction (PCR) to detect wild-type Atg5 and Atg5flox alleles, as previously described.[11] Cre expression in livers of Atg5flox/flox; Mx1-Cre mice was induced by intraperitoneal injection of 300 μL polyinosinic acid/polycytidylic acid (Sigma-Aldrich, St. Louis, MO) at a concentration of 1 mg/mL in water, three times at 48-hour intervals. Mice were maintained in a room with alternating 12-hour light/dark cycles. Animals received humane care in compliance with the institutional guidelines of the Graduate School of Medical Sciences, Kyushu University (Fukuoka, Japan). The study protocol conformed to the ethical guidelines of the 1975 Helsinki Declaration.

Animal Studies

PHx was performed in male and female mice 6 weeks of age. Mice were anesthetized with ether and subjected to approximately 70% PHx by removing the left lateral and median lobes, after midventral laparotomy.[2, 13] The mortality rate after 70% PHx was <1%.

Plasmids

A plasmid containing an inactive mutant of Atg4B (Atg4BC74A), a protease that processes pro-LC3 (microtubule-associated protein 1 light chain 3) paralogs and hampers conversion of LC3-I to LC3-II, was a kind gift from Dr. Yoshimori and was prepared as previously described.[14]

For materials and methods in more detail, see the Supporting Information.

Results

Generation of L-Atg5 KO Mice

PCR analyses showed that the subset of autophagic genes was expressed in livers of L-Atg5 KO mice (Atg5flox/flox; Mx-Cre mice; Supporting Fig. 1A). Livers of L-Atg5 KO mice displayed appropriate responses to 24-hour starvation with increased expression of Beclin-1 and LC3, similar to control mice, but lacked Atg5 expression. Structure and morphology of the liver (Supporting Fig. 1B) and other organs, including the spleen, heart, lung, kidney, and brain (Supporting Fig. 2), were normal. In terms of liver injury, serum alanine aminotransferase (ALT) levels were higher in L-Atg5 KO mice at 21 days after injection than in control mice, consistent with a previous report[9] (Supporting Fig. 3). At 8 weeks of age, liver and body weights of L-Atg5 KO mice were slightly lower than those of control mice (Supporting Fig. 4), consistent with previous results using Atg5flox/flox; nestin-Cre mice.[11]

Figure 1.

Effects of PHx on autophagic activity in the regenerating liver and in primary cultured hepatocytes. (A) Autophagosomes in the regenerating liver. Arrows, autophagosomes. Scale bar, 1 μm. (B) Number of autophagosomes in hepatocytes at the indicated times after hepatectomy. (C) Expression of the autophagy-related genes, LC3-I, LC3-II, Beclin-1, Atg5, and p62, in control mice at 0-196 hours after 70% PHx (representative western blottings are shown). (D) Immunofluorescent analysis of LC3 in primary cultured hepatocytes in response to HGF stimulation. Arrows, autophagosomes. Scale bar, 10 μm. Number of autophagosomes in hepatocytes treated without or with HGF by immunofluorescent analysis of LC3. (E) Electron microscopic images of autophagosomes (arrows) in hepatocytes treated without or with HGF. Scale bar, 1 μm. (F) Effects of HGF on expression levels of LC3-I, LC3-II, Beclin-1, Atg5-12, and cyclin D in primary cultured hepatocytes (representative western blottings are shown). *P < 0.05.

Figure 2.

Liver regeneration and survival after PHx is impaired in L-Atg5 KO mice. (A) Survival rate of control and L-Atg5 KO mice after 90% PHx. (B) Expression of autophagy-related proteins and Keap1/Nrf2 proteins in control and L-Atg5 KO mice at 0-48 hours after 70% PHx (representative western blottings are shown). (C) Fold change in p62 mRNA expression. (D and E) Changes in serum ALT and albumin levels after 70% PHx. *P < 0.05.

Figure 3.

DNA synthesis during liver regeneration is impaired in L-Atg5 KO mice. (A) BrdU-positive cell count at 24-72 hours after PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (B) BrdU staining of hepatocytes at 36 hours after PHx after 70% PHx in control and L-Atg5 KO mice. Scale bar, 50 μm. (C) Cell-cycle analysis of hepatocytes from control and L-Atg5 KO mice. (D) Changes in cleaved caspase-3 and caspase-7 expression levels at 0-72 hours after 70% PHx (representative western blottings are shown). (E) Changes in caspase-3 activity at 0-72 hours after 70% PHx in control and L-Atg5 KO mice. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05.

Figure 4.

Expression of cell-cycle–associated molecules in the regenerating liver in control and L-Atg5 KO mice. (A-E) Fold changes in cyclins D, E, A, and B as well as p21 mRNA levels in control and L-Atg5 KO mice at 0-72 hours after 70% PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (F) Changes in cyclins D, E, A, and B as well as p21 protein expression levels in control and L-Atg5 KO mice at 0-72 hours after 70% PHx. β-actin was used as a loading control, and these expression levels at 0 hours in control mice were defined as 1.0.

Autophagic Activation in the Regenerating Liver After PHx

In control mice, 70% PHx caused an increase in the number of autophagosomes, peaking at 18 hours after PHx, which was maintained until 48 hours, but returned to baseline after 4 days (Fig. 1A,B). Expression of p62, which regulates ubiquitin-positive protein aggregates during autophagic deficiency, remained elevated for the entire study period (Fig. 1C). In vitro, isolated hepatocytes showed marked dose-dependent autophagy in response to hepatocyte growth factor (HGF; Fig. 1D-F). Knockdown of autophagy using a lentivirus vector encoding mutant Atg4B, which is indispensable for lipidation of LC3 proteins, reduced HGF-stimulated autophagy (Supporting Fig. 5). HGF treatment (5-20 ng/mL) increased levels of phosphorylated protein kinase B (Akt; Thr308 and Ser473) and phosphorylated mammalian target of rapamycin (mTOR)1 (Supporting Fig. 6). HGF also increased LC3-II levels (Fig. 1F), which are usually down-regulated by phosphorylated mTOR. This finding may be the result of increased phosphorylated adenosine-monophosphate (AMP)-activated protein kinase (AMPK) levels in HGF-treated proliferative hepatocytes (Supporting Fig. 6).

Figure 5.

Liver hypertrophy and hepatocyte senescence in the regenerating liver. (A) Macroscopic images of the liver after PHx. (B) LW/BW at 0-48 hours after 70% PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (C) Size of hepatocytes in control and L-Atg5 KO mice. Scale bar, 10 μm. (D) Fold change in hepatocyte size after 0-48 hours after PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (E) IHC staining of SA-β-gal in hepatocytes. Scale bar, 100 μm. (F) Expression of polyubiquitinated protein at 0 and 24 hours after PHx in control and L-Atg5 KO mice. (G and H) Fold changes in IL-8 (G) and IL-6 (H) mRNA expression at 0-48 hours after 70% PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05.

Figure 6.

Suppression of autophagy activity and mitochondrial β-oxidation. (A) Intrahepatic ATP concentrations at 0-48 hours after PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (B) Electron microscopic images of hepatocyte mitochondria at 24 hours after 70% PHx. Scale bar, 2.5 μm. (C) Proportion of hepatocytes with a low mitochondrial membrane potential (ΔΨm) at 24 hours after 70% PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (D) mRNA expression of mitochondrial β-oxidation-related genes at 0 and 24 hours after PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (E) β-hydroxybutyrate concentrations at 0 and 24 hours after PHx. Open bars, control mice; closed bars, L-Atg5 KO mice. *P < 0.05. (F) Changes in total and phosphorylated levels of AMPK, Erk1/2, and cyclin D1 at 0-48 hours after PHx in control and L-Atg5 KO mice. (G) Changes in total and phosphorylated levels of c-jun and NF-κB at 0 and 24 hours after PHx in control and L-Atg5 KO mice.

In the LC3 turnover assay, LC3-II levels were increased by treatment with chloroquine (CQ), even before 70% PHx. However, differences in LC3-II levels in the presence and absence of CQ were greater at 24 and 96 hours after 70% PHx, compared with before 70% PHX (Supporting Fig. 7A). Levels of p62, which is normally degraded during autophagy, were increased by CQ, especially after 70% PHx (Supporting Fig. 7A). These results indicate that autophagic flux is increased during liver regeneration. In vitro, the increase in LC3-II levels induced by HGF was also greater in the presence of CQ than in the absence of CQ (Supporting Fig. 7), indicating that HGF increases autophagic flux in hepatocytes.

Impaired Recovery of Liver Regeneration in L-Atg5 KO Mice After PHx

All mice survived after 70% PHx in control and L-Atg5 KO mice. Excessive parenchymal damage by extended PHx has been proposed as the principal cause of hepatic failure, but little is known regarding the contribution of autophagic activity for primary deficiency in liver regeneration.[5] We developed a mouse model of 90% PHx to assess the effect of autophagic impairment on hepatic regenerative capacity of a critically small liver remnant. After 90% PHx, approximately 50% of control mice survived, whereas all L-Atg5 KO mice died after 24 hours (Fig. 2A). p62 expression after 70% PHx was higher in the regenerating liver in L-Atg5 KO mice than in control mice (Fig. 2B,C), which indicates that autophagic activity is essential to reorganize and modulate hepatocellular protein and organelle synthesis to achieve adequate regeneration. Serum ALT concentrations were significantly higher and serum albumin concentrations were lower after 70% PHx in L-Atg5 KO than in control mice (Fig. 2D,E and Supporting Fig. 8). These findings indicated that activation of autophagy in proliferative hepatocytes may be involved in cell survival and hepatic function after hepatectomy in the regenerating liver. In addition, Keap1 levels in liver of L-Atg5 KO mice were decreased by 70% PHx, and nuclear factor (erythroid-derived 2)-like 2 (Nrf2) levels were higher in L-Atg5 KO mice than in control mice, as expected (Fig. 2B). These results indicated that the increase in p62 messenger RNA (mRNA) levels may have been mediated by compensatory Nrf2 activation in L-Atg5 KO mice.

DNA Synthesis During Liver Regeneration Is Attenuated and Delayed in L-Atg5 KO Mice

DNA synthesis at the indicated times after PHx was determined by measuring bromodeoxyuridine (BrdU) incorporation, and the cell cycle was analyzed by propidium iodide PI staining. We found that the percentage of BrdU-positive cells in proliferating hepatocytes of control mice peaked at 36 hours after PHx, with a mean of 30.2% (Fig. 3A,B), similar to previous reports.[5, 15, 16] Percentage of BrdU-positive cells was then decreased at 48-72 hours after PHx. Percentage of BrdU-positive cells was significantly lower in L-Atg5 KO than in control mice at all times after PHx. Percentages of hepatocytes in the S and G2 phases were significantly lower in L-Atg5 KO mice (Fig. 3C). Western blotting analysis showed that expression levels of cleaved caspase-3 and cleaved caspase-7 were similar in L-Atg5 KO and control mice before 70% PHx, but their levels were greatly elevated after 70% PHx in L-Atg5 KO mice, compared to control mice (Fig. 3D). Caspase-3 activity was also elevated by 70% PHx in L-Atg5 KO mice (Fig. 3E). Hepatic mRNA and protein expression levels of cyclins A, B, D, and E, which regulate cyclin-dependent kinases, were significantly lower in L-Atg5 KO than in control mice at 36 hours after PHx (Fig. 4A-D, F). However, hepatic expression levels of cyclin D, which is initiated during G1 and drives the G1/S phase transition under the control of p21, were significantly lower in L-Atg5 KO than in control mice at 24 hours after PHx (Fig. 4E, F). Furthermore, hepatic mRNA and protein expression levels of p21 were higher in L-Atg5 KO than in control mice at 12 hours after PHx (Fig. 4F,G). These findings indicate that autophagy is involved in the mitotic response of hepatocytes after PHx, and is driven by down-regulation of p21, resulting in up-regulation of cyclin D. The proportion of dividing cells containing multipolar and lagging chromosomes was significantly lower in L-Atg5 KO than in control mice at 48 and 72 hours after PHx (Supporting Fig. 9). These findings indicate that impaired autophagy in the regenerating liver is associated with a decrease in altered nuclear ploidy levels in hepatocytes.

Liver Hypertrophy and Hepatocyte Senescence in the Regenerating Liver After PHx in L-Atg5 KO Mice

Unlike the mitotic responses, the liver mass and liver weight/body weight ratio (LW/BW) at 24 hours after PHx was significantly higher in L-Atg5 KO than in control mice (Fig. 5A,B). The size of hepatocytes, an indicator of liver hypertrophy, was significantly greater at 24 hours after PHx in L-Atg5 KO than in control mice (Fig. 5C,D). Stress responses in senescent cells include up-regulation of p21 and cellular hypertrophy.[10, 15, 19] To confirm the involvement of p21 and cellular hypertrophy in hepatocyte senescence during liver regeneration in L-Atg5 KO mice, we analyzed the expression of SA-β-gal and specific senescence-associated secretory phenotype (SASP) components. We found that levels of SA-β-gal and polyubiquitinated proteins at 24 hours after PHx were higher in L-Atg5 KO than in control mice (Fig. 5E,F), as were concentrations of the SASP components, interleukin (IL)-8 and IL-6 (Fig. 5G,H). These findings indicate that proliferating hepatocytes require activation of autophagy to prevent senescence, along with hepatic hypertrophy.

Suppression of Autophagic Activity and Mitochondrial β-Oxidation

Hepatic ATP levels were significantly lower in L-Atg5 KO than in control mice (Fig. 6A). Fragmentation of hepatocyte mitochondria in L-Atg5 KO mice was demonstrated, but not in control mice (Fig. 6B). In addition, the proportion of hepatocytes with a low mitochondrial membrane potential (ΔΨm) was higher in L-Atg5 KO than in control mice (Fig. 6C). Relative to control mice, L-Atg5 KO mice had lower hepatocyte expression levels of the genes encoding medium chain acyl-coenzyme A dehydrogenase and carnitine palmitoyl transferase-1, which have a rate-controlling effect on β-oxidation, and liver-type fatty acid-binding protein, which plays a role in the transportation of free fatty acids (FFAs) to mitochondria. Moreover, hepatocyte expression levels of FAS, which is involved in the synthesis of FFAsa from acetyl-coenzyme A, were significantly higher in L-Atg5 KO mice (Fig. 6D). Hepatocyte levels of β-hydroxybutyrate, a final ketone body product, were significantly lower in L-Atg5 KO than in control mice (Fig. 6E). In addition, hepatocyte expression of extracellular signal-related kinase (Erk)1/2, an ATP-dependent mitogen-activated protein kinase, was significantly lower in L-Atg5 KO than in control mice, as was cyclin D expression (Fig. 6F). By contrast, expression levels of the upstream targets of AMPK were slightly increased in L-Atg5 KO mice as part of the compensation to KO of autophagic activation (Fig. 6F). Therefore, phosphorylation of the transcription factor, c-jun, at serine 63, another ATP-dependent protein, was significantly lower at 24 hours after PHx in L-Atg5 KO than in control mice (Fig. 6G), whereas there were no changes in expression of nuclear factor kappa B (NF-κB) or phosphorylation of c-jun at threonine 91 in the regenerating liver of L-Atg5 KO. These findings indicated that, after PHx, hepatocytes maintain intracellular ATP levels by activating mitochondrial β-oxidation, possibly because of prompt removal of damaged mitochondria by activation of autophagy as a selective degradation system. Hepatocytes that fail to maintain their level of energy charge may become senescent.

Discussion

We have used L-Atg5 KO mice to investigate the roles of autophagy-associated pathways in liver regeneration after PHx. We found that liver regeneration was severely impaired by 70% PHx in these mice. Their livers showed an impaired postoperative mitotic response, with quiescent hepatocytes becoming senescent and hypertrophic. Moreover, PHx was followed by considerable damage to mitochondria, reduced β-oxidation, and reduced intrahepatic ATP generation. Thus, autophagy during the early phase of liver regeneration is critical for maintaining healthy mitochondria capable of producing ATP and prevents hepatocytes from becoming senescent.

Growth factors, such as HGF and IL-6, activate phosphatidylinositol 3-phosphate, which, in turn, phosphorylates Akt at Thr308 and Ser473 in hepatocytes.[17, 18] Akt also plays a critical role in proliferation by phosphorylating mTOR1, which suppresses autophagy by inhibiting TOR-dependent phosphorylation of Atg13 in a rapamycin-sensitive complex containing raptor.[17, 18] During proliferation, Atg13 is phosphorylated by phosphorylated AMPK-α, which is activated by an increase in the AMP/ATP ratio caused by cellular/environmental stressors, such as energy deficiency, hypoxia, and ischemia, during proliferation.[6-8] To date, a few reports have shown a relationship between HGF and autophagic activity. A recent report showed that HGF has a pivotal role in directly promoting autophagic activity for clearance of advanced glycation endproducts, which are involved in the pathogenesis of diabetic vascular complications, in primary mouse nonparenchymal cells (NPCs).[19] This previous report showed that LC3-II levels in NPCs are increased by HGF, similar to our results, and that this effect is inhibited by cotreatment with an anti-HGF neutralizing antibody.[19] In some ongoing clinical trials, pharmacological inhibitors of autophagy, such as CQ, which is an autophagolysosomal inhibitor, have been used for treating solid cancers (e.g., pancreatic adenocarcinoma and breast cancer), based on the knowledge that growth-factor–induced autophagic activity promotes proliferation of cancer cells through important components of cellular metabolism.[20] In proliferating hepatocytes, the level of phosphorylated AMPK is increased, activating autophagy to overcome the burden of various stressors, especially those that cause endoplasmic reticulum stress or oxidative stress (OS).[4] Therefore, HGF indirectly increases LC3-II levels by enhancing phosphorylated AMPK levels.

Mitophagy, the autophagy-dependent degradation of mitochondria, is a defensive mechanism that involves selective sequestration and subsequent degradation of dysfunctional mitochondria.[21] Occurrence of mitophagy in livers of patients with Reye's syndrome[22] suggests its importance for normal mitochondrial turnover and function.[23] Accumulation of mitochondria caused by disruption of autophagy is thought to lead to increased levels of reactive oxygen species and DNA damage, resulting in mitochondrial depolarization and permeability transition.[23-25] In this study, we showed that selective degradation of dysfunctional mitochondria in proliferating hepatocytes was impaired after PHx in L-Atg5 KO mice, resulting in a disruption of mitochondrial functions, such as maintenance of intracellular ATP levels, by mitochondrial β-oxidation. However, with the exception of hepatocyte senescence, which was accompanied by increased p21 expression, we were unable to detect any phenotypes of mitophagy in proliferating hepatocytes. Further detailed research is required to clarify the pivotal role of mitophagy in proliferating hepatocytes.

Recent studies have elucidated the mechanisms involved in cellular senescence in hepatocytes adapted to stress stimuli.[10, 26-30] As a stress response, senescence is a dynamic process involving multiple effector mechanisms, whose combination ultimately determines cellular phenotype. This process is characterized by a number of biological, biochemical, and molecular changes, including cell hypertrophy, up-regulation of SA-β-gal, irreversible growth arrest, and expression of a specific SASP.[31, 32] SASP components, such as IL-6, IL-8, and matrix metalloproteinases, can promote tissue repair by preventing the generation of a persistent acute inflammatory response and by attracting immune cells that kill and clear senescent cells by adhesion molecules expressed on the latter. Stimuli, such as OS, can induce autophagy and senescence and may be part of the same physiological process, known as the autophagy-senescence transition.[33, 34] However, the pivotal relationship between these two cellular responses has not been elucidated. In particular, it is unknown whether induction of autophagy is involved in the induction of senescence or vice versa.[26, 27] The results of our study suggest that autophagy regulates cellular senescence during liver regeneration. A subset of autophagy-related genes was up-regulated during liver regeneration in association with negative feedback in p21-activated senescence. Furthermore, inhibition of autophagy augmented the senescence phenotype, including senescence-associated secretion of IL-6 and IL-8. We reasoned that rapid protein turnover, which involves autophagy coupled with active protein synthesis, may facilitate this process by allowing the remodeling of proteins needed for liver regeneration. Indeed, senescent cells are usually hypertrophic, and we confirmed that the fall into senescence was dependent on cellular quality maintained by autophagy. Thus, autophagy may be complementary to epigenetic regulation in preventing the specific biochemical alterations observed during acute induction of senescence in the regenerating liver. Our findings suggest that autophagy, and its maintenance of cellular quality by rapid protein turnover during hepatocyte regeneration, negatively regulates the acquisition of the senescence phenotype.

Several genetic and pharmacological interventions can suppress transient hepatic steatosis, a characteristic of the early regenerative response accompanied by hepatic energy deficiency, resulting in impaired liver regeneration.[35] Subsequent to liver injury, liver mass is maintained or recovers in proportion to body mass.[1, 21] Moreover, hypoglycemia after PHx induces systemic lipolysis, which is followed by accumulation of fat derived from peripheral stores, during early regeneration.[35] These observations, and the central role of the liver as the principal intermediary between fatty acid metabolism and intrahepatic energy maintenance,[35, 36] prompted us to investigate the regulation and functional role of autophagy, a systemic metabolic phenomenon, during liver regeneration after PHx. In our study, we found that hepatocytes with activated autophagy, without senescence, stimulated mitochondrial metabolism during liver regeneration by maintaining healthy mitochondria and the production of high-energy mitochondrial fuels to support hepatocyte proliferation. The remnant liver relies on mitochondrial oxidative phosphorylation to satisfy its energy demands, and defects in mitochondrial permeability transition have been reported in mitochondria isolated from remnant livers after PHx.[36] In addition, autophagy is critical for regulating hepatocellular lipid stores.[37] Future studies should address whether systemic adipose stores during normal liver regeneration are needed as fuel sources to support regeneration, as lipid precursors for membrane synthesis, and/or act as specific signals that initiate the regenerative response itself.[37, 38] Mitochondrial oxidative damage is a fundamental component of liver regeneration, and multiple studies have shown that it is attenuated by autophagic activity to regulate hepatocellular lipid stores.[39] A recent study showed that sirtuin-1 (SIRT1), a nicotinamide adenine dinucleotide–dependent deacetylase, was involved in the protective effects of calorie restriction against hypoxia in the aged kidney, which is linked to calorie restriction-related longevity by mitochondrial autophagy.[40] Adult-onset and long-term calorie restriction in mice promoted increased SIRT1 expression in the aged kidney and attenuated hypoxia-associated mitochondrial and renal damage by enhancing mitophagy. Here, we have shown that autophagic activity in proliferating hepatocytes markedly diminished mitochondrial dysfunction in terms of increased mitochondria permeability transition and decreased β-oxidation. These data highlight the role of autophagy-dependent systemic lipolysis, including β-oxidation, in hepatocellular adaptation to mitochondrial oxidative damage, delineate a molecular mechanism of the autophagy-mediated antiaging effect, and could potentially direct the design of new therapies to promote liver regeneration after damage associated with OS and aging.

In our study, the peak percentage of Ki67-positive cells, as determined by immunohistochemistry (IHC), was 7.4% at 12 hours after PHx, 11.2% at 24 hours, and 31.2% at 36 hours (Supporting Fig. 10). These values are higher than those reported in previous studies.[5, 15] All of the mice subjected to 70% PHx were 6 weeks old in our study, whereas other studies induced PHx at 10-12 weeks. Hepatocyte senescence strongly down-regulates cellular proliferation by decreasing farnesoid X receptor levels[41] or SIRT1 levels.[42] Therefore, the difference in the percentage of BrdU-positive cells at 12 hours after PHx between our study and previous studies may be the result of a difference in the effect of senescence on proliferating hepatocytes labeled with BrdU.[41, 42]

Autophagy during the early phase of liver regeneration is critical for maintaining healthy mitochondria, which can produce ATP through β-oxidation after hypoglycemia-induced hepatic steatosis. Therefore, autophagy may be essential for preserving cellular quality by preventing hepatocytes from becoming senescent and hypertrophic.

Acknowledgment

The authors are grateful to T. Yoshimori (Osaka University, Osaka, Japan) for kindly providing the inactive mutant of Atg4B (Atg4BC74A) and N. Mizushima (Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan) for the Atg5flox/flox mice. The authors also thank N. Yamashita (Kyushu University, Fukuoka, Japan) for her expert advice related to statistical analysis.

Ancillary

Advertisement