Autophagy protects kidney proximal tubule epithelial cells from mitochondrial metabolic stress

Abstract

Chronic metabolic stress is related to diseases, whereas autophagy supplies nutrients by recycling the degradative products. Cyclosporin A (CsA), a frequently used immunosuppressant, induces metabolic stress via effects on mitochondrial respiration, and thereby, its chronic usage is often limited. Here we show that autophagy plays a protective role against CsA-induced metabolic stress in kidney proximal tubule epithelial cells. Autophagy deficiency leads to decreased mitochondrial membrane potential, which coincides with metabolic abnormalities as characterized by decreased levels of amino acids, increased tricarboxylic acid (TCA) ratio (the levels of intermediates of the latter part of the TCA cycle, over levels of intermediates in the earlier part), and decreased products of oxidative phosphorylation (ATP). In addition to the altered profile of amino acids, CsA decreased the hyperpolarization of mitochondria with the disturbance of mitochondrial energy metabolism in autophagy-competent cells, i.e., increased TCA ratio and worsening of the NAD⁺/NADH ratio, coupled with decreased energy status, which suggests that adaptation to CsA employs autophagy to supply electron donors from amino acids via intermediates of the latter part of the TCA cycle. The TCA ratio of autophagy-deficient cells was further worsened with decreased levels of amino acids in response to CsA, and, as a result, the deficiency of autophagy failed to adapt to the CsA-induced metabolic stress. Deterioration of the TCA ratio further worsened energy status. The CsA-induced metabolic stress also activated regulatory genes of metabolism and apoptotic signals, whose expressions were accelerated in autophagy-deficient cells. These data provide new perspectives on autophagy in conditions of chronic metabolic stress in disease.

Keywords: :

• autophagy
• metabolic stress
• mitochondria
• cyclosporin A
• kidney proximal tubule
• chronic kidney disease

Introduction

Cellular metabolism plays a vital role in living creatures, and chronic metabolic stress is strongly related with pathological conditions, including diabetes mellitus, cancer and aging process,^1-3 and mitochondria play pivotal roles of intracellular energy production through oxidative phosphorylation coupled with the electron transport chain and tricarboxylic acid (TCA) cycle. Therefore, elucidating the mechanism of mitochondrial metabolic stress-related pathological conditions could open a novel therapeutic strategy. Cyclosporin A (CsA), a frequently used immunosuppressant in the treatment of transplantation and diseases such as nephrotic syndrome, is known to induce mitochondrial dysfunction by blocking the conductance of the permeability transition pore of the inner membrane of mitochondria,⁴ and long-term usage of CsA is often limited by its toxicity to the proximal tubule epithelial cells of the kidney.⁵

Macroautophagy (hereafter referred to as autophagy) is a highly conserved bulk protein degradation system in eukaryotes, which may control the intracellular environment and energy status, not only through the elimination of abnormal proteins and organelles including mitochondria, but also through recycling nutrients.⁶ After the engulfment of components of the cytoplasm and organelles, the double-membrane vesicle called an autophagosome fuses with a lysosome, which results in the degradation of the sequestered materials by lysosomal enzymes. It has been shown that autophagy plays a protective role in several organs, including kidney proximal tubules, from aging-related stress and acute kidney injury.^7,8 The process of autophagy is followed by the generation of amino acids, sugars, fatty acids, and nucleosides that are recycled for macromolecular synthesis and energy production,^6,9 and a few studies have started to elucidate the specific events through which autophagy is strongly related with metabolism.^10-15 Autophagy also controls the quality of mitochondria by the selective elimination of impaired mitochondria, termed mitophagy,¹⁶ which substantiates the protective role of autophagy under mitochondrial chronic metabolic stress. In this study, we proved our hypothesis that autophagy guards against mitochondrial metabolic stress in kidney proximal tubule epithelial cells.

Results

Autophagy protects proximal tubular cells from CsA-induced mitochondrial injury

We first investigated whether CsA induces autophagy in proximal tubular cells (PTCs) of the kidney. Autophagic activity was measured by the degradation of SQSTM1/p62 protein, and by the accumulation of SQSTM1 and MAP1LC3-II in the presence of bafilomycin A₁. SQSTM1 is degraded mainly through autophagic clearance, and the induction of autophagy leads to decreased SQSTM1 abundance, whereas downstream suppression of autophagic flux or increased expression of SQSTM1 causes the accumulation of this protein. Induction of autophagy also promotes conversion of MAP1LC3-I to MAP1LC3-II, and, in the presence of bafilomycin A₁, results in the accumulation of MAP1LC3-II and SQSTM1 as a result of interference with autolysosome formation. In the presence of bafilomycin A₁, vehicle-treated wild-type PTCs accumulated both SQSTM1 and MAP1LC3-II with initial increase of the MAP1LC3-II/I ratio, indicating the presence of basal level of autophagy (Fig. S1A). CsA treatment alone induced initial degradation with a rebound of SQSTM1 and increased MAP1LC3-II/I ratio. Additionally, in the presence of bafilomycin A₁, CsA induced accumulation of both SQSTM1 and MAP1LC3-II with a further increase of the MAP1LC3-II/I ratio (Fig. 1A). The CsA-induced rebound of SQSTM1 in the absence of bafilomycin A₁ was due to the induction of its expression (Fig.S1B). When compared with vehicle-treated wild-type PTCs, CsA-treated cells exhibited accumulation of SQSTM1 and increased MAP1LC3-II/I ratio in the presence of bafilomycin A₁ (Fig. 1B), indicating that CsA treatment causes a significant autophagic flux.

Figure 1. Cyclosporin A (CsA)-induced autophagy protects kidney proximal tubular cells. (A) Wild-type proximal tubular cells were treated with 25 μM CsA for the indicated hours in the presence or absence of bafilomycin A₁ (Baf A1). (B) Wild-type proximal tubular cells were treated with either 25 μM CsA or vehicle for 24 h in the presence or absence of bafilomycin A₁. (C) Cell survival of autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells treated with indicated concentration of CsA for 24 h. (D) Autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells treated with either 25 μM CsA or vehicle for 24 h. Data are means ± SE from three to five experiments. Values were normalized to the leftmost column. *P < 0.05 vs. vehicle-treated cells of the corresponding hour (A) or cells of corresponding treatment (B–D); ^#P < 0.05 vs. untreated (A) or vehicle-treated cells (B–D).

To determine the precise role of autophagy, we exposed to CsA, two cultured cell lines, autophagy-deficient PTCs and their genetically reverted autophagy-competent ones.⁸ Autophagy-deficiency and autophagy-competency were confirmed by the levels of ATG5 and the conversion of MAP1LC3-I to MAP1LC3-II (Fig. S2). CsA decreased the survival ratio of cultured autophagy-deficient cells compared with that of autophagy-competent cells (Fig. 1C). CsA significantly increased the accumulation of SQSTM1 in autophagy-deficient cells compared with that of autophagy-competent cells (Fig. 1D). CsA induced mRNA expression of Sqstm1 in autophagy-competent cells, while its expression level was not altered by CsA in autophagy-deficient cells (Fig. S3A), indicating that CsA-induced accumulation of SQSTM1 in autophagy-deficient cells was due to autophagic stimuli of CsA. These data suggest the protective role of autophagy against CsA.

Because CsA affects mitochondrial function, we investigated whether autophagy could protect kidney PTCs from CsA-induced mitochondrial injury. The mitochondrial membrane potential of autophagy-deficient cells was slightly lower than that of autophagy-competent cells at baseline (Fig. 2A). CsA treatment induced a transient increase, followed by a decrease of mitochondrial membrane potential in both cells. Furthermore, exposure to CsA significantly increased mitochondrial reactive oxygen species, assessed by MitoSOX Red, in autophagy-deficient cells compared with that of autophagy-competent cells (Fig. 2B). These results suggest that autophagy plays a quality control of mitochondria, and that autophagy protects proximal tubules from CsA-induced mitochondrial damage.

Figure 2. Autophagy protects proximal tubules from cyclosporin A (CsA)-induced mitochondrial damage. (A) Autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells after treatment with either 25 μM CsA or vehicle for the indicated hours and stained with MitoTracker Red. (B) MitoSox Red staining of autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells treated with either 25 μM CsA or vehicle for 24 h. Scale bars: 10 μm. The images are representative of multiple experiments (n = 3 to 8). Values are normalized to the signal intensity of untreated (A) or vehicle-treated (B) atg5 (−) cells. Data are means ± SE *P < 0.05 vs. atg5 (−) cells of the corresponding hour (A) or treatment (B); ^#P < 0.05 vs. untreated (A) or vehicle-treated cells (B).

Autophagy deficiency exaggerates CsA-induced abnormal amino acid and TCA cycle metabolism

To evaluate the role of autophagy against CsA-induced metabolic stress, we employed metabolome analyses using the CE-TOFMS systems and compared the effect of CsA on autophagy-deficient and autophagy-competent PTCs. CsA increased the protein/DNA ratio, an index of cell size, in both autophagy-competent and deficient cells (Fig. S3B), and we measured the levels of metabolites on a per-protein basis.

The levels of most amino acids in vehicle-treated autophagy-deficient cells were significantly lower than those in vehicle-treated autophagy-competent cells (Fig. 3). CsA-treatment significantly decreased the level of both essential and nonessential amino acids in autophagy-competent and autophagy-deficient cells (the reduction of essential and nonessential amino acids were 35% and 58% in autophagy-competent cells, and 44% and 61% in deficient cells, respectively, compared with those of vehicle-treated cells). As a result, the level of both essential and nonessential amino acids was lower in in CsA-treated autophagy-deficient cells than in CsA-treated autophagy-competent cells. When assessed by the essential/total amino acid ratio, CsA-treatment significantly increased the ratio in both autophagy-competent and autophagy-deficient cells, with the increase being significantly smaller in autophagy-deficient cells than in autophagy-competent cells (46% and 34%, respectively). Among nonessential amino acids, the levels of aspartate and glutamate presented a characteristic decrease in both cells in response to CsA-treatment (the decrease of aspartate and glutamate were 88% and 74% in autophagy-competent cells, and 89% and 75% in deficient cells, respectively).

Figure 3. Autophagy deficiency affects amino acid metabolism. Levels of amino acids in autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells treated with either 25 μM cyclosporin A (CsA) or vehicle for 24 h were superimposed on a metabolic pathway map. Columns, average concentration (nmoL/mg protein); scale bars: SE. N.D., the metabolite concentration was below the detection limit of the analysis (n = 8). *P < 0.05 vs. Atg5 (+) cells of corresponding treatment; ^#P < 0.05 vs. vehicle-treated controls.

Regarding TCA cycle intermediates, the levels of intermediates of the former part of the cycle (citrate, cis-aconitate, and 2-oxoglutarate) were significantly lower, while the level of an intermediate of the latter part (succinate) was significantly higher in vehicle-treated autophagy-deficient cells than in vehicle-treated, autophagy-competent cells (Fig. 4A; Fig. S5). Because the former three intermediates of the TCA cycle are produced solely in mitochondria, their decrease in autophagy-deficient cells may reflect the mitochondrial dysfunction of autophagy-deficient cells as seen in decreased membrane potential (Fig. 2A). The decreased level of 2-oxoglutarate may be associated with the decreased levels of both glutamine and former TCA cycle intermediates. On the other hand, CsA-treatment significantly decreased the levels of all TCA intermediates in both cells, with a dominant decrease in the intermediates of the earlier part (the decreases in levels of intermediates of the earlier and latter parts were 63% and 71% in autophagy-competent cells, and 40% and 61% in autophagy-deficient cells, respectively). Considering that the latter part of TCA cycle could utilize amino acids as their source, it is possible that cells compensate for the CsA-induced metabolic stress by increasing the flow from amino acids to the TCA cycle. As a result, the TCA ratio, which is calculated by the equation, (succinate + fumarate + malate)/(citrate + cis-aconitate + iso-citrate + 2-oxoglutarate), increased in autophagy-competent cells after CsA treatment. The TCA ratio was higher in vehicle-treated, autophagy-deficient cells when compared with that of autophagy-competent cells, and CsA treatment further deteriorated this ratio in autophagy-deficient cells in comparison with that of CsA-treated autophagy-competent cells. Sensitive analyses using other definitions of the TCA ratio did not alter the results (Fig. S4A). The combination of the TCA ratio and the essential amino acid/total amino acid ratio differentiated well the cellular status (i.e., autophagy-deficiency vs. autophagy-competency, and CsA vs. vehicle, Fig. 4B).

Figure 4. Autophagy deficiency affects tricarboxylic acid (TCA) cycle metabolism. (A) Metabolite concentrations of autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells treated with either 25 μM cyclosporin A (CsA) or vehicle for 24 h were superimposed on a metabolic pathway map that included glycolysis and the pentose phosphate and TCA pathways (n = 8). Columns, average concentration (nmoL/mg protein); scale bars: SE. N.D., the metabolite concentration was below the detection limit of the analysis. (B) Scatter plot of the TCA ratio and essential amino acid (EAA)/total amino acid (TAA) ratio. *P < 0.05 vs. Atg5 (+) cells of corresponding treatment; ^#P < 0.05 vs. vehicle-treated controls.

With respect to the intermediates of glycolytic pathway, vehicle-treated autophagy-deficient cells demonstrated significantly lower levels of glucose 6-phosphate, fructose 6-phosphate, 3-phosphoglycerate, and phosphoenolpyruvate than those in vehicle-treated autophagy-competent cells (Fig. S5). After CsA-treatment, the levels of these intermediates decreased in autophagy-competent cells, whereas the levels remained low in autophagy-deficient cells. Regarding the intermediates of the pentose phosphate pathway, the levels of 6-phosphogluconate and ribulose-5-phosphate were lower in vehicle-treated autophagy-deficient cells than in the vehicle-treated, autophagy-competent cells (Fig. S5). CsA-treatment resulted in a significant increase of 6-phosphogluconate in autophagy-competent cells, and the levels of the intermediates (6-phosphogluconate, ribulose-5-phosphate, and sedoheptulose 7-phosphate) were lower in CsA-treated autophagy-deficient cells than in CsA-treated autophagy-competent cells.

Autophagy deficiency exaggerates CsA-induced reduced energy status and NADPH metabolism

These metabolic hallmarks dramatically affect energy metabolism. The TCA cycle generates NADH, and, coupled with the conversion from NADH (reduced form) to NAD⁺ (oxidative form), the electron transport chain of mitochondria exports H⁺ ions across the inner membrane of mitochondria to generate a proton gradient, which is utilized for the oxidative phosphorylation to generate adenosine triphosphate (ATP, Fig. 5A). Therefore, mitochondrial stress may disturb the balance of these players. On the other hand, CsA, a blocker of the permeability transition pore, could interfere with the oxidation of NADH.

Figure 5. Autophagy deficiency affects energy status and NADPH metabolism. (A) Quantified levels of metabolites involved in metabolism of nucleotides and NADH in autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells treated with either 25 μM cyclosporin A (CsA) or vehicle for 24 h were superimposed on a mitochondrial metabolic pathway map. (B) Levels of metabolites involved in NADPH in Atg5 (+) and atg5 (−) kidney proximal tubular cells treated with either 25 μM CsA or vehicle for 24 h. Columns, average concentration (nmoL/mg protein); scale bars: SE (n = 8). *P < 0.05 vs. Atg5 (+) cells of corresponding treatment; ^#P < 0.05 vs. vehicle-treated controls.

Under physiological conditions, the level of NAD⁺ is higher than that of NADH, as seen in vehicle-treated autophagy-competent cells (Fig. 5A), to exert its activity through oxidative form.¹⁷ CsA-treated autophagy-competent cells showed a significant decrease of the NAD⁺/NADH ratio due to the significant decrease of the level of NAD⁺ (the decreased levels were 60% and 39% for NAD⁺ and NADH, respectively, Fig. 5A). The reduction of NADH, as assessed by the intensity of autofluorescence signals, was confirmed in CsA-treated autophagy-competent cells (Fig. S6). On the other hand, vehicle-treated autophagy-deficient cells exhibited significantly lower levels of NAD⁺ than in vehicle-treated, autophagy-competent cells. After CsA treatment, the level of NAD⁺ in autophagy-deficient cells was still lower than that of autophagy-competent cells (the decreased levels were 55% and 41% for NAD⁺ and NADH, respectively). There was a strong correlation between the NAD⁺/NADH and TCA ratios (Fig. S4B), and, together with the increased mitochondrial membrane potential, the CsA-induced disturbance of TCA cycle may have affected this ratio.

The CsA-induced abnormal metabolism of NAD⁺/NADH resulted in decreased level of energy status. CsA-treated autophagy-competent cells significantly decreased the level of ATP and GTP (Fig. 5A; Fig. S4C). As a result, CsA treatment lowered the ATP/AMP ratio and the adenylate energy charge, which is calculated by the equation, (ATP) + 1/2 (ADP)]/[(ATP) + (ADP) + (AMP)¹⁸ in autophagy-competent cells. Autophagy-deficient cells showed a significantly lower ATP level even at basal conditions than autophagy-competent cells, and CsA-treatment accelerated the decrease in energy charge in autophagy-deficient cells (Fig. 5A).

Under physiological conditions, the level of NADPH is higher than that of NADP⁺ to exert its activity through the reduced form.¹⁷ CsA-treated autophagy-competent cells showed a significant increase in the NADP⁺/NADPH ratio, and this ratio was further deteriorated in CsA-treated autophagy-deficient cells compared with that of autophagy-competent cells (Fig. 5B).

We also performed similar analyses using the original data (measured by per cells, Figs. S7–S9). Although the effects of cell size modified the results, most results, especially the ratios, were largely unchanged.

Autophagy-deficiency accelerates the response to CsA-induced reduction of energy status

We then analyzed the cellular response against decreased energy status. CsA-treatment induced increased uptake of glucose in autophagy-competent cells (Fig. 6A), as indicated by a similar consumptive amount of glucose (decreased glucose concentration of medium per initial cell number) to vehicle-treated cells (Fig. 10S) in conditions with half the cell number (Fig. 1C). Autophagy-deficiency further enhanced the uptake of glucose (Fig. 6A). It is known that Slc2a1/Glut1 gene transcription is upregulated, coupled with CsA-induced increments of glucose consumption in kidney epithelial cells in vitro.¹⁹ We found that CsA-treatment upregulated Slc2a1 mRNA expression, and that deficiency of autophagy further enhanced CsA-induced Slc2a1 transcription (Fig. 6B). In addition, CsA significantly increased the levels of endoplasmic reticulum stress markers, HSPA5/GRP78 and HSP90B1/GRP94, which increase their levels in response to glucose starvation,²⁰ in autophagy-deficient cells compared with autophagy-competent cells (Fig. 6C). These results reflected a decreased intracellular energy status with increased uptake of glucose in CsA-treated autophagy-deficient cells compared with that of autophagy-competent cells.

Figure 6. Autophagy deficiency worsens energy status in response to cyclosporin A (CsA)-treatment. Autophagy-competent [Atg5 (+)] and autophagy-deficient [atg5 (−)] kidney proximal tubular cells were treated with CsA. (A) Glucose uptake. (B) mRNA expression level of Slc2a1. (C) Level of endoplasmic reticulum stress markers. (D and E) Protein (D) and mRNA (E) levels of Atpif1. (F) mRNA expression levels of metabolism-related genes. (G) Schematic representation of metabolic stress and autophagy. Magenta lines indicate the metabolic adaptation pathway through autophagy. Data are means ± SE of three to six experiments. Fold expression normalized to that of vehicle-treated Atg5 (+) cells. *P < 0.05 vs. Atg5 (+) cells of corresponding treatment; ^#P < 0.05 vs. vehicle-treated controls.

When mitochondrial respiration is compromised, the F₁F_o-ATP synthase reverses and consumes ATP, serving to maintain mitochondrial membrane potential, while ATPIF1 is known to mitigate this effect.²¹ The mRNA expression level of Atpif1 was increased in vehicle-treated autophagy-deficient cells compared with that in autophagy-competent cells, reflecting the decreased mitochondrial membrane potential in autophagy-deficient cells (Fig. 6D and E). CsA-treatment decreased the mRNA levels of Atpif1 in both cells (Fig. 6D and E). The transient increment of membrane potential (Fig. 2A) may have decreased the mRNA expression of Atpif1 despite the final decrease of membrane potential. The mRNA level of Atp5f1/ATP synthase subunit b, a component of F₁F_o-ATP synthase to which ATPIF1 binds, and Slc25a4/Ant1, a component of the mitochondrial permeability transition pore, were nearly identical between autophagy-competent and deficient cells (Fig. 6D and E). These results indicated that ATPIF1 may function to conserve ATP at the expense of mitochondrial membrane potential in vehicle-treated autophagy-deficient cells, and that CsA reduces the levels of ATPIF1 despite the final decrease of mitochondrial membrane potential.

As the CsA-induced characteristic metabolic alterations are similar to those observed in cancer,^2,22 and the expression of Slc2a1, one of the metabolic responsive genes, was upregulated, we further examined the gene-expression profile of target genes against metabolic stress. Among the metabolic regulators, Bbc3/Puma, a proapoptotic gene, and Cdkn1a/p21, a cell cycle inhibitor, increased expression in vehicle-treated autophagy-deficient cells compared with those of autophagy-competent cells (Fig. 6F). CsA-treatment increased the expression levels of these genes in autophagy-competent cells, and the expression of Bbc3 was further enhanced in CsA-treated autophagy-deficient cells. In addition, in vitro analysis demonstrated that other metabolic regulators (Sco2, Sesn2, 9630033F20Rik/Tigar [mouse ortholog of human C12ORF5], and Dram1) showed similar trends to those seen in Bbc3 and Cdkn1a (Fig. 6F). These results indicate that metabolic stress induces targeted gene expressions against metabolic stress under CsA treatment, and that, compared with autophagy-competent cells, autophagy-deficient cells increased their levels due to the exaggerated CsA-induced metabolic stress.

Discussion

We demonstrated that autophagy plays a protective role against chronic metabolic stress. By interfering with the energy metabolism of mitochondria, CsA stimulates cells to utilize amino acids, especially glutamine, to maintain levels of TCA cycle intermediates and to synthesize nucleotides from their nitrogen. Eventually, the cells would die if these compensations are insufficient. On the other hand, we clearly demonstrated that autophagy deficiency leads to metabolic abnormalities as characterized by decreased levels of amino acids, glycolytic intermediates, intermediates from the earlier part of the TCA cycle, and the products of oxidative phosphorylation (ATP) with decreased mitochondrial membrane potential. During CsA-induced metabolic stress, autophagy-deficient cells are unable to utilize amino acids efficiently, and therefore, this results in further deterioration of energy status and cell death (Fig. 6G). The upregulation of autophagy by CsA in the kidney epithelial cells in vitro has been reported,²³ however, its role was not determined due to the lack of both genetic ablation of autophagy genes and metabolic studies. Here, we proved the protective role of autophagy against metabolic stress in CsA-induced kidney injury.

We showed that, under pathological conditions, autophagy plays a role in utilization of amino acids as a source of energy and nucleotides. CsA induced the consumption of glutamine and other amino acids, while these amino acids are the source of intermediates that enter the latter part of the TCA cycle (2-oxoglurate, succinyl-CoA, succinate, fumarate, and malate). In combination with the decreased levels of amino acids, CsA-induced increases of the TCA ratio, (indicated by the increase of intermediates of the latter part of the cycle, with a decrease in intermediates of the earlier part), may represent amino acid flux into the TCA cycle.

We further demonstrated that autophagy maintains the levels of nicotinamide derivatives and energy status against CsA-treatment. The reduced level of NAD⁺, in the presence of reduced level of the NAD⁺/NADH ratio and energy status, may reflect the CsA-induced mitochondrial injury. Nicotinamide derivatives are known as classical molecules involved in energy metabolism, reductive biosynthesis and antioxidation.¹⁷ Under physiological conditions, NAD⁺/NAPH and NADP⁺/NADPH ratios are strictly controlled to protect cells from death, whereas CsA disrupts these balances as shown in this study. CsA induces hyperpolarization of mitochondria, which results in inhibition of electron transport by blocking the pumping of protons into the intermembrane space.²⁴ As a result, mitochondrial hyperpolarization could block the transformation of NADH to NAD⁺. NADH is one of the electron donors for electron transport chain in mitochondria in the transformation of NADH to NAD⁺, whose blockade results in defects of ATP/ADP exchange. NADH is generated through the three oxidative steps of TCA cycle using iso-citrate, 2-oxogluratate, and malate as substrates, and, taking the close correlation between the ratios of TCA and NAD⁺/NADH into account, the disturbance of the TCA ratio may also indicate the reduced production of NADH. The CsA-induced reduction of aspartate and glutamate may indicate a disturbance of the malate-aspartate shuttle. In a strict sense, the increased consumption of NAD⁺ in response to CsA could not be discounted, however it is unlikely as a primary cause, because mitochondria would supply NADH and convert it to NAD⁺ without the defect of oxidative phosphorylation and the TCA cycle. NADPH, a major component of antioxidant system and the source of reductive biosynthesis, may be rapidly consumed by CsA treatment. Accumulation of SQSTM1 in autophagy-deficient mice, which is associated with glutathione synthesis, may further decrease the level of NADPH.²⁵ The upregulation of the pentose phosphate pathway seems to be involved in the adaptive response against CsA-induced imbalance of the NADP⁺/NADPH ratio because this pathway serves to generate NADPH and nucleotides. Thus, it is implied that glucogenic amino acids, such as glutamine, may also be utilized as substrates for the pentose phosphate pathway under CsA treatment, and autophagy deficiency also affects this compensation.

Surprisingly, CsA-treatment decreased the levels of ATPIF1 (Fig. 6D), possibly due to the transient increment of membrane potential (Fig. 2A). ATPIF1 is known to play a protective role against hypoxia by conserving ATP, whereas cells with decreased level of ATPIF1 showed disorganization of mitochondria, coupled with preference of glycolysis rather than oxidative phosphorylation.²¹ CsA-induced reduction of ATPIF1 in the presence of decreased mitochondrial membrane potential may have worsened CsA-induced metabolic stress through the modulation of mitochondrial function.

The characteristic metabolic stress induced by CsA is quite identical to that of cancer in several points. It has been widely known that cancer cells show the Warburg effect, that is, a distinctive demand for glucose and amino acids, including glutamine.^26,27 Glutamine, which is the most abundant amino acid in the plasma and in the culture medium, compensates glucose and is utilized for energy and biosynthesis. Glutamine’s carbon is also used to maintain pools of TCA cycle intermediates, while its nitrogen is applied to produce nonessential amino acids, hexosamine, nucleotides and other molecules in tumor cells.²⁷ In addition, cancer cells are distinctive in that (i) the levels of intermediates of the latter part of the TCA cycle increase, while there is a decrease in intermediates of the earlier part, (ii) the level of NADPH is decreased while that of the intermediates of the pentose phosphate pathway is increased,² (iii) intracellular energy status is decreased, and (iv) metabolic regulatory genes are upregulated.²² While our study was in revision, the important role of autophagy in the maintenance of glutamine metabolism was also reported.²⁸ It is also interesting to consider the relationship between cancer and autophagy. Recently, it has been suggested that cancer cells utilize autophagy for energy supply.²⁹ Inhibition of autophagy is considered as a therapeutic option these days,³⁰ and this strategy may exert its effect through the metabolism as observed in autophagy-deficient cells of this study.

The results obtained with the current experimental system may have limitations due to the possible implication of Atg5 in apoptosis. Our observation could not deny the possibility that the decreased degradation of more long-lived proteins in autophagy-deficient cells may result in the reduction of the levels of metabolites when measured on a per-protein basis; however, as mentioned above, the mitochondrial quality control by autophagy should also be more important. We also relied on our results concerning the ratios (such as the TCA ratio); as these ratios would yield the same values regardless of the measurement conditions (i.e., per cell number or per-protein amount). In order to provide more profound analyses, metabolomics of each cellular compartment (i.e., cytoplasmic and nuclear compartment, or even each organelle) might be necessary.

In conclusion, we showed that autophagy plays a protective role in CsA-induced metabolic stress in proximal tubules of the kidney. Autophagy deficiency exaggerates CsA-induced metabolic stress through the abnormal metabolism of amino acids and TCA cycle, which decreases the energy status and results in cell death. CsA-induced metabolic stress also activates metabolism-related genes, whose expressions are increased in autophagy-deficient cells, and thereby, result in apoptotic signals. Metabolism is strongly related to both physiological and pathological conditions, including diabetes mellitus, cancer and the aging process,^2,3,31 and further metabolic study of autophagy will yield a novel therapeutic approach.

Materials and Methods

Cell culture

Generation of autophagy-deficient and autophagy-competent proximal tubular cell lines was described previously.⁸ Proximal tubular cell lines from wild-type mice were also cultured. Cells were grown in Dulbecco’s modified Eagle’s medium (1000 mg/mL glucose; Sigma, D6046) containing 5% fetal bovine serum (MP Biomedics, 2917054) at 37 °C under a humidified atmosphere of 5% CO2 and 95% air. Twenty-four hours after the seeding, cells were treated with ethanol-diluted CsA or vehicle, and the effects were examined at 24 h after treatment. Cell viability was determined by trypan blue exclusion. Cell size was determined by quantification of genomic DNA and total protein. The level of glucose in the culture medium was measured using the Glucose C-II test (Wako, 439-90901) and the consumption ratio was calculated by dividing it according to the number of cells before treatment.

Antibodies

The following antibodies were used: antibodies for SQSTM1 (MBL, PM045), ATG5 (MBL, M153-3), MAP1LC3 (Cell Signaling, 2755), ACTA2 (Sigma, F3777), ATPIF1 (Abcam, ab110277), ATP5F1 (Abcam, ab117991), and ACTB (Sigma, A5316).

Confocal microscopy

The fluorescence images were collected using a scanning confocal microscope (Olympus, FV1000-D). Cells were cultured in 35-mm glass-bottom dishes and were subject to the microscopy. Mitochodrial membrane potential was determined by staining with 25 nM MitoTracker Red (Invitrogen, M7512) for 15 min. To determine mitochondrial ROS production, cells were incubated with 0.3 μM MitoSOX Red (Invitrogen, M36008) for 10 min. UV-excited blue autofluorescence was measured to detect NADH autofluorescence. Intensity was measured using at least three independent experiments (> 150 cells per experiment).

Metabolite extraction

Extraction of metabolites from attached cells was performed as previously described, with modifications.^2,32 In brief, after washing twice with 10 mL 5% mannitol, cells were added 1 mL of methanol with 25 μM of internal controls [methionine sulfone (Wako, 502-76641) and D-camphor-10-sulfonic acid (Dojindo, 349-01032)]. After leaving at rest for 10 min, sample solutions were collected and mixed with 500 μL of Milli-Q water. A 1200-μL portion of the sample was extracted using chloroform, and the aqueous layer (800 μL) was filtrated using 5 kDa cutoff filters (Millipore, UFC3LCCNB_HMT). A 600-μL portion of the filtrate was dried, and the sample was dissolved in 50 μL of Milli-Q water that contained reference compounds (200 μM of 3-aminopyrrolidine [Sigma, 404624] and trimesate [Wako, 206-03641]).

Metabolite measurement

Measurement of metabolites was measured using CE-TOFMS.^32,33 In brief, CE-TOFMS was performed using a capillary electrophoresis system equipped with a Time-of-Flight mass spectrometer (Agilent, 6210), an isocratic HPLC pump (Agilent, 1100), a CE-MS adaptor kit (Agilent, G1603A), and a CE-ESI-MS sprayer kit (Agilent, G1607A). The system was controlled by ChemStation software version B.03.01 for CE (Agilent, G2201AA). Data acquisition was performed by Analyst QS Build: 7222 software for Agilent TOF (Applied Biosystems). For cationic metabolites, capillary electrophoreses were performed using a fused silica capillary. The electrolyte was 1 M formic acid. For anionic metabolites, a polymer coated COSMO (+) capillary (Nacalai, 07584-44i) was used. The electrolyte was 50 mM ammonium acetate (pH 8.5). For all analytical modes, the inner diameter and total length of capillary are 50 μm and 100 cm, respectively. The applied voltage was set at +30 kV and −30 kV for cation and anion modes and nucleotide mode, respectively. Electrosplay ionisation-TOFMS was operated in the positive ion mode (4 kV), the negative ion mode (3.5 kV), and the negative ion mode (3.5 kV) for cationic metabolites, anionic metabolites, and nucleotides, respectively. Exact mass data were acquired over a 50- to 1000-m/z range. The measured metabolite included amino acids, glycolysis intermediates, nucleoties and their derivatives.

Data processing

Raw data were processed using software (MasterHands) developed in-house as previously described.² The data processing flow consisted of noise filtering, baseline correction, peak detection and integration of the peak areas from 0.02 m/z-wide sections of the electropherograms. The accuracy of m/z of each peak was calculated by Gaussian curve-fitting on the m/z domain, and the migration times were normalized to match the detected peaks among the multiple data sets. The peaks were identified by matching m/z values and normalized migration times of corresponding authentic standard compounds. Quantification was performed by comparing the peak areas against calibration curves generated using internal standardization-techniques to eliminate systematic bias, which was derived from injection volume variance and MS sensitivity. The intracellular concentrations of metabolites were calculated using the measured cell count and the estimated protein content of a single cell. Data are presented as the mean ± SE from three independent experiments (n = 8).

Quantitative RT-PCR and western blot analysis

Quantitative RT-PCR and western blot analyses were performed as previously described.³⁵ The sequences of the primers used were as follows: Dram1-F, 5′-tcgtagccaa cttccaggag-3′; Dram1-R, 5′-tgagtgaagc acaggcaatc -3′; Bbc3-F, 5′-gcccagcagc acttagagtc-3′; Bbc3-R, 5′-tgtcgatgct gctcttcttg-3′; Cdkn1a-F, 5′-gtacttcctc tgccctgctg-3′; Cdkn1a-R, 5′-tctgcgcttg gagtgataga-3′; Sco2-F, 5′-ctgtggaccc agaacgagat-3′; Sco2-R, 5′-taggcccagc gctgtagtat-3′; Sesn2-F, 5′-tagcctgcag cctcacctat-3′; Sesn2-R, 5′-gattttgagg ttccgttcca-3′; 9630033F20Rik-F, 5′-ctccatcact cccaacactg-3′; 9630033F20Rik-R, 5′-tgctcctgga ggttcataca-3′; Actb-F, 5′-tgacaggatg cagaaggaga-3′; Actb-R, 5′-acatctgctg gaaggtggac-3′; Slc2a1-F, 5′-ctatggagct ggccaagaag-3′; Slc2a1-R, 5′-agaggccaca agtctgcatt-3′; Sqstm1-F, 5′-ccccagagtc gaagtagctg-3′; Sqstm1-R, 5′-agtgagaaga ggctggtgga-3′; Atpif1-F, 5′-ggttcggtgt ctggggtatg-3′; Atpif1-R, 5′-tctcgttttc cgaaggctcc-3′; Atp5f1-F, 5′-atggcgcctc ttttggagaa-3′; Atp5f1-R, 5′-tgctgtgcct tctccatgtc-3′; Slc25a4-F, 5′-gagagggtca aactgctgct-3′; Slc25a4-R, 5′-ggcttgagtg gggaagtacc-3′.

Statistics

All results are presented as means ± SE. The difference between two experimental values was assessed by the Student t-test. Statistical significance among groups was evaluated using ANOVA with the False Discovery Rate (FDR) method. Multiple comparison of metabolome data was also adjusted with FDR. Correlations were examined using Spearman rank correlation coefficient. Statistical significance was defined as P < 0.05. Statistical analyses were performed using STATA statistical software version 11 (STATA Corporation, College Station, TX, USA) and R environment for statistical computing, version 2.13.1.

Abbreviations:
CsA	=	cyclosporin A
TCA	=	tricarboxylic acid
TCA cycle	=	tricarboxylic acid cycle or Kreb cycle
PTCs	=	proximal tubular cells
ATP	=	adenosine triphosphate
NAD+	=	the oxidized form of nicotinamide adenine dinucleotide
NADH	=	nicotinamide adenine dinucleotide
SQSTM1	=	sequestosome 1
MAP1LC3	=	microtubule-associated protein 1 light chain 3
NADP+	=	nicotinamide adenine dinucleotide phosphate
NADPH	=	the reduced form of nicotinamide adenine dinucleotide phosphate
ATPIF1	=	ATPase inhibitory factor 1
EAA	=	essential amino acids
NEAA	=	nonessential amino acids
TAA	=	total amino acids
ROS	=	reactive oxygen species
CE	=	capillary electrophoresis
MS	=	mass spectrometer
TOF-MS	=	time-of-flight mass spectrometer

Acknowledgments

We thank: M Ohishi, K Endoh, A Suzuki, Y Hatakeyama, S Ohta, N Horimoto, and K Shibayama for technical and secretarial assistance. This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology in Japan (22590890 [HK], 24591196 [YT], 24890108 [TK], and 24659416 [YI]), and Novartis CPCF Research Grant 2013 (TK).

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Supplemental Materials

Supplemental materials may be found here: www.landesbioscience.com/journals/autophagy/article/25418